Vibrating fluidized bed oligonucleotide synthesizer

Abstract

To provide an efficient solid phase oligonucleotide synthesis (SPOS) method.SOLUTION: A method of adding an oligonucleotide to a solid support in a bed reactor, the method comprising: removing a protecting group from a 5' - position of an oligonucleotide bound to the solid support; and adding an activated amidite solution to the bed reactor, wherein: Wherein the activated amidite solution comprises an amidite and flows up and down or is fluidized by nitrogen bubbling or other agitation in the bed reactor and reacts at the 5 ' position of the oligonucleotide, wherein the phosphite linkage found in the amidite comprises a P atom in oxidation state III, and then converting the P atom from oxidation state III to oxidation state V.SELECTED DRAWING: None

Description

[Technical Field]

[0001] The present disclosure relates to new systems and methods for synthetically producing oligonucleotides. More specifically, the disclosure relates to devices and methods that use oscillatory flow or gas bubbling to create a fluidized bed as part of solid phase oligonucleotide synthesis (SPOS) and completely drain the liquid from the solid resin after every reaction and wash step. [Background technology]

[0002] Solid-phase oligonucleotide synthesis ("SPOS") is the most commonly used method and system for synthesizing oligonucleotides. SPOS is typically performed on a solid-phase medium, typically a solid support made of controlled pore glass (CPG) or macroporous polystyrene (MPPS) spheres. SPOS is the solid-phase synthesis of oligonucleotides using various nucleoside derivative building blocks, the most common of which are phosphoramidites. Specifically, the starting phosphoramidite building block is attached to the solid phase, and then each nucleoside (phosphoramidite) is added and sequentially coupled to the phosphoramidite building block until the desired molecule is obtained. In other words, one phosphoramidite is added and coupled (usually at the 5'-terminal OH position), followed by the next phosphoramidite, and so on, thereby growing the chain until the desired sequence is obtained. Protecting groups are used on each amine base and phosphorus of the oligonucleotide, allowing the functional groups to withstand the acidic and neutral conditions utilized in the SPOS cycle. Once the oligonucleotide sequence is obtained, the molecule is cleaved from the solid support and globally deprotected to yield the desired oligonucleotide.

[0003] In the SPOS process, four chemical reactions generally occur to add a single phosphoramidite to a chain. The first step is a "deblocking" step, typically a detritylation reaction. Specifically, the 5'-hydroxyl group of the nucleotide is protected with an acid-labile protecting group such as DMT (4,4'-dimethoxytrityl). This protecting group is removed during a continuous flow of acid solution or by adding an acid to the solvent. The acid may be, for example, trichloroacetic acid (TCA), dichloroacetic acid (DCA), or some other acid carried in an inert solvent such as toluene or dichloromethane, or other solvent. In some embodiments, 2% TCA, 3% DCA, or 10% DCA is used with toluene. In the case of the DMT protecting group, the orange DMT cation formed during this "deblocking" reaction is washed away by the addition of a wash solution. This step therefore yields a solid-support-bound oligonucleotide precursor with a free 5'-terminal hydroxyl group.

[0004] Once the deblocking step has occurred, the "coupling" step follows. This involves adding a solvent solution of activated phosphoramidite (e.g., a 0.02-0.2 M solution of phosphoramidite in acetonitrile (ACN) (or anhydrous ACN)). This activated phosphoramidite reacts with and couples to the previously deprotected free 5'-terminal hydroxyl group. As is generally known in the art, the phosphoramidite solution can be "activated" by the addition of a catalyst to facilitate the coupling reaction. Various catalysts, including various azole or imidazole compounds, are known to "activate" phosphoramidites. Often, more than one equivalent of catalyst is used because the acidic nature of the catalyst helps neutralize the diisopropylamine by-product generated during coupling. Once coupling is complete, unbound reagents and by-products are removed by washing.

[0005] After the coupling step, the next step in SPOS synthesis is either oxidation, thiolation (also called sulfurization), or "capping." Capping is performed because a small percentage (0.1-1% or more) of the 5'-OH groups bound to the solid support remain unreacted and need to be blocked from further chain elongation to prevent the formation of oligonucleotides with internal base deletions, commonly referred to as shortmers, such as (n-1), (n-2), and (n-3). The unreacted 5'-hydroxy groups are largely acetylated by the capping mixture. Capping these unreacted OH groups allows these impurities to be more easily separated from the desired product by chromatography. Similarly, if the coupling reaction produces other undesired products (such as reaction of O with guanosine bases or other chemicals), these undesired products can also be blocked (capped) from further reaction, making them more easily separated in subsequent purification steps. In some embodiments, the capping step involves treating the solid support-bound material with a mixture of acetic anhydride and 1-methylimidazole. Other capping reagents may also be used.

[0006] In the oxidation step, a phosphoramidite reacts with and couples to the 5'-terminal OH group, generating a phosphite triester bond (e.g., where the P atom is in the +3 oxidation state). This phosphite triester bond is unnatural and has limited stability under the conditions of oligonucleotide synthesis. Therefore, the P atom is oxidized to the more stable +5 oxidation state by the addition of an oxidizing agent such as iodine and water in the presence of a weak base (pyridine, lutidine, or collidine). This reaction oxidizes the phosphite triester to a tetracoordinate phosphate triester, which is a protected precursor of the naturally occurring phosphodiester internucleoside linkage. Oxidation can be performed under anhydrous conditions using tert-butyl hydroperoxide or (1S)-(+)-(10-camphorsulfonyl)-oxaziridine (CSO). In other embodiments, oxidation is replaced by sulfurization to a phosphothiolate linker. Those skilled in the art will understand that some embodiments of SPOS may be best designed so that the oxidation or sulfurization step is followed by a capping step, or vice versa. Those skilled in the art will also appreciate that some embodiments of the SPOS may best be designed to omit the capping step from some of the cycles if high conversion rates are expected.

[0007] Once these four steps (deblocking, coupling, either oxidation or sulfurization, and capping) are complete, a phosphoramidite building block is added to the growing chain. As will be appreciated, the coupled phosphoramidite building block has its own DMT protecting group protecting the 5'-terminal OH group. Thus, this process can be repeated to add another phosphoramidite moiety until the chain reaches the desired length.

[0008] Once the chain reaches the desired length, the oligonucleotide protecting groups can be removed, cleaving the oligonucleotide from the resin and releasing it into solution. In some cases, these protecting groups from the nucleoside amines and the 2-cyanoethyl phosphate protecting group are globally deprotected in the same base-catalyzed hydrolytic cleavage reaction. Aqueous ammonia or a mixture of ammonia and methylamine is commonly used for this cleavage / deprotection step. These conditions also efficiently hydrolyze the 3' linker and cleave the oligonucleotide from the resin.

[0009] However, the acrylonitrile by-product generated during ammonolysis of the 2-cyanoethyl protecting group can alkylate the amino base moiety and form potentially problematic adducts. For this reason, it may be desirable to selectively deprotect the phosphate by treating the oligonucleotide with anhydrous secondary amine (e.g., diethylamine) solution while it is still attached to the resin. Once the acrylonitrile by-product is washed away with solvent, the oligonucleotide can be cleaved and deprotected in aqueous ammonia without fear of acrylonitrile adduct formation.

[0010] While the SPOS process is commercially used and remains the standard for oligonucleotide synthesis, it has clear drawbacks, the most significant of which are its high cost, the generation of large amounts of waste, and limited scalability. The multiple steps required make the process very expensive, consume large amounts of solvent, and generate waste. Even more problematic, many of these solvents are environmentally unfriendly. Furthermore, many SPOS solid supports have a low material load capacity, necessitating excessive batches to produce commercial quantities. Furthermore, conventional packed-bed plug-flow SPOS reactors are limited in batch size because the resin bed height is limited by the pressure drop of the liquid flowing downward through the bed, and the diameter is limited by issues related to the radial distribution of reagents and maintaining a uniform bed height across the entire cross-section. Furthermore, each oligonucleotide requires a protecting group, which increases the overall manufacturing cost.

[0011] Perhaps the most obvious weakness of SPOS is its inefficiency. Because four reactions are required to add a single phosphoramidite, low conversion rates in each cycle of even one reaction type significantly impact the overall process yield. Furthermore, the solutions used in the four reactions are typically added to the resin by adding them to the top of the vessel and reacting them as they are pumped downward and discharged out the bottom. This process generally results in uneven contact between the liquid and solid phases, especially when channels form in the resin bed, reducing reaction efficiency. Therefore, a large excess of reagents is required to achieve complete reaction, thorough washing, and high yield. Including reagents and washes, the amount of material required to produce commercial quantities of oligonucleotides is very large. This uneven contact also leads to uneven purity throughout the reactor vessel, especially from top to bottom. Given the limited scale of packed-bed reactors, we estimate that some of our portfolio assets require hundreds of synthesis batches per year. In addition, SPOS using a conventional downflow packed-bed reactor cannot easily accommodate flexible batch sizes in the same reactor, because varying the resin bed height can result in different yields and impurity profiles due to uneven top-to-bottom contact. Furthermore, the maximum resin bed height is also limited by the pressure drop through the resin bed, especially as the polystyrene resin particles simultaneously expand and compress while transferring solvent during flow. The loading of polystyrene resin is limited to approximately 300 μmol / g due to the need to limit resin expansion and therefore the pressure drop through the resin bed.

[0012] Therefore, it would be an improvement to find new uses for SPOS that address one or more of these shortcomings. A particular improvement would be to find an SPOS system that could be used on a commercial scale to produce large quantities of oligonucleotides, e.g., many metric tons per year. A further advance would be if such a system could be more environmentally friendly, reduce manufacturing costs, and increase overall efficiency. The present embodiments address one or more of these shortcomings. Summary of the Invention

[0013] This embodiment includes a method of adding a phosphoramidite to a solid-phase resin in a bed reactor, where a protecting group is removed from the 5' position of an oligonucleotide, and an activated amidite solution is coupled to the unprotected group, the activated amidite solution containing the amidite, and fluidizing the resin in the reactor. Fluidization can occur by forcing a liquid to flow up and down through the bed reactor, bubbling an inert gas through it, or other types of agitation to create a slurry. The amidite reacts at the 5' position of the oligonucleotide.

[0014] In addition to the coupling reaction, reagent solutions for deblocking, oxidation, thiolation, and capping can each flow with the resin, providing complete liquid / solid contact and resetting the channel-free resin bed. Fluidization can be followed by a plug flow reaction in which reagents flow downward through the resin bed, typical of traditional SPOS. The same fluidization section followed by a plug flow section can be used for solvent washing after each reaction. In this way, the majority of resin expansion and contraction can occur during the fluidization section of solid / liquid contact, where it is advantageous to overcome pressure drops and eliminate channeling.

[0015] The features and advantages of the present disclosure will become apparent to those skilled in the art upon consideration of the following detailed description in conjunction with the accompanying drawings. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 1 is a schematic diagram of the reactions carried out within the oligonucleotide SPOS system. [Figure 2] FIG. 1 is a schematic diagram of an SPOS system. [Figure 3-1] FIG. 1 is a schematic diagram of a small-scale vibrating fluidized bed oligonucleotide synthesizer setup. [Figure 3-2] FIG. 1 is a schematic diagram of a small-scale vibrating fluidized bed oligonucleotide synthesizer setup. [Figure 4]1 is a graph showing the resin bed height of Example 2 for each phosphoramidite cycle. [Figure 5-1] FIG. 1 is a schematic diagram of a pilot-scale fluidized bed oligonucleotide synthesizer setup. [Figure 5-2] FIG. 1 is a schematic diagram of a pilot-scale fluidized bed oligonucleotide synthesizer setup. [Figure 6-1] 1 is a schematic representation of a molecule that can be made using the techniques outlined herein. [Figure 6-2] 1 is a schematic representation of a molecule that can be made using the techniques outlined herein. [Figure 6-3] 1 is a schematic representation of a molecule that can be made using the techniques outlined herein. [Figure 6-4] 1 is a schematic representation of a molecule that can be made using the techniques outlined herein. [Figure 7-1] FIG. 1 is a schematic diagram of an alternative research-scale fluidized bed oligonucleotide synthesizer setup. [Figure 7-2] FIG. 1 is a schematic diagram of an alternative research-scale fluidized bed oligonucleotide synthesizer setup. [Figure 8] FIG. 1 is a schematic diagram of an alternative research-scale fluidized bed oligonucleotide synthesizer setup. [Figure 9] FIG. 1 is a schematic diagram of an alternative pilot-scale fluidized bed oligonucleotide synthesizer setup. [Figure 10-1] 1 is a schematic representation of a molecule that can be made using the techniques outlined herein. [Figure 10-2] 1 is a schematic representation of a molecule that can be made using the techniques outlined herein. [Figure 10-3] 1 is a schematic representation of a molecule that can be made using the techniques outlined herein. [Figure 10-4] 1 is a schematic representation of a molecule that can be made using the techniques outlined herein. [Figure 11] FIG. 1 is a schematic diagram of an alternative pilot-scale fluidized bed oligonucleotide synthesizer setup. [Figure 12] FIG. 1 is a schematic diagram of an in-process integrated multi-pass cleaning system after deblocking. [Figure 13] FIG. 1 is a schematic diagram of an in-process integrated multi-pass cleaning system after oxidation / thiolation. [Figure 14] 1 shows various UPLC chromatograms of the examples. [Figure 15] 1 shows various UPLC chromatograms of the examples. [Figure 16] 1 shows various UPLC chromatograms of the examples. [Figure 17] 1 shows various UPLC chromatograms of the examples. DETAILED DESCRIPTION OF THE INVENTION

[0017] A method for loading an oligonucleotide onto a solid-phase resin in a bed reactor is disclosed, which includes removing a protecting group from the 5'-position of an oligonucleotide bound to a solid-phase resin, and adding an activated amidite solution to the bed reactor, the activated amidite solution containing an amidite, flowing up and down through the bed reactor or fluidizing it by nitrogen bubbling or other agitation, to react at the 5'-position of the oligonucleotide, where the phosphite bond found in the amidite contains a P atom in oxidation state III, and converting the P atom from oxidation state III to oxidation state V.

[0018] In some embodiments, the method further comprises adding a capping solution before or after converting the P atom from oxidation state III to oxidation state V, wherein the capping solution caps the coupling moieties if they do not react with the amidite solution, thereby preventing further amidites from coupling to the coupling moieties, and the capping solution flows up and down within the bed reactor or is fluidized or mixed by nitrogen bubbling or other agitation. In some embodiments, capping is performed only on selected phosphoramidite cycles.

[0019] In an additional embodiment, the method further comprises removing the activated amidite solution from the bed reactor by passing the amidite solution through a filter located at the bottom of the bed reactor.

[0020] Further embodiments may be made that include the additional step of adding a first wash solution to the bed reactor, where the addition of the first wash solution occurs after removing the protecting groups. In additional embodiments, the method further includes the step of adding a second wash solution to the bed reactor, where the addition of the second wash solution occurs after the activated amidite solution has been added to the bed reactor. The first and second wash solutions may flow up and down through the bed reactor or may be mixed by gas bubbling or other agitation, and the method further includes the step of individually removing the first and second wash solutions from the bed reactor by passing them through filters located at the bottom of the bed reactor. More wash sections may be used and may be performed in an integrated multi-pass fashion as described herein.

[0021] In further embodiments, the step of adding the second wash solution occurs before the step of converting the P atoms from oxidation state III to oxidation state V. In other embodiments, the step of adding a third wash solution to the bed reactor occurs after the step of converting the P atoms from oxidation state III to oxidation state V. In other embodiments, the third wash solution flows up and down within the bed reactor or is fluidized or mixed by nitrogen bubbling or other agitation, and the method further includes removing the third wash solution from the bed reactor by passing it through a filter located at the bottom of the bed reactor. More wash sections may be used and may be performed in an integrated multi-pass manner as described herein. In other embodiments, the protecting group is a DMT group, and removing the protecting group comprises reacting the 5' position of the nucleotide with an activation solution comprising an acid in a solvent. Additional embodiments may be made that further include removing the activation solution from the bed reactor by passing the activation solution through a filter located at the bottom of the bed reactor. In some embodiments, up-and-down flow within the bed reactor is achieved by applying pressure to the top of the reactor. In further embodiments, fluidized-bed mixing of solids and liquids within the bed reactor is achieved by adding nitrogen or another gas to the bottom of the reactor or by some other type of agitation. In some embodiments, no fluidization or mixing occurs during the deblocking step, only plug flow through the resin bed.

[0022] In additional embodiments, a cleaner fraction of the wash solvent is recycled and reused after each phosphoramidite cycle. Further embodiments are designed in which a cleaner portion of the reagent solution used in the deblocking reaction is recycled and reused after each phosphoramidite cycle. Additional embodiments are made that include in-process integrated multi-pass washing as described herein.

[0023] A system for adding oligonucleotides to a solid-phase resin is also disclosed. The system includes a bed reactor and an activated amidite solution, where the activated amidite solution contains the amidite and flows up and down within the bed reactor or is fluidized by nitrogen bubbling or other agitation. The system may have a bed reactor including an inlet that allows pressurized gas to enter the bed reactor, where the pressurized gas or some other type of agitation causes the amidite solution to mix with the solids within the bed reactor. In other embodiments, the inlet is located at the bottom of the bed reactor. The bed reactor may be pressurized from the top of the bed reactor, where the pressure causes the amidite to flow up and down within the bed reactor. In some embodiments, rather than liquid flowing up and down within the reactor, an inert gas is bubbled in from the bottom, mixing the liquid and solids within the reactor.

[0024] For the purposes of promoting an understanding of the principles of the present disclosure, reference will now be made to the embodiments illustrated in the drawings and specific language will be used to describe the same, it being understood, however, that no limitation on the scope of the invention is intended.

[0025] Referring now to FIG. 1, a schematic diagram illustrating the reactions occurring within the SPOS system is shown. Specifically, an oligonucleotide 102 is attached to a resin 104. As shown in FIG. 1, the oligonucleotide 102 may be covalently attached to the resin 104 via an oxygen (ether) bond. Of course, other types and methods of attaching the oligonucleotide to the resin 104 may also be used. (The region of the resin is also referred to herein as the resin bed.) The oligonucleotide 102 generally includes a base 108, such as a base that binds to DNA or RNA. The base 108 may be protected (e.g., one or more functional groups of the base are protected, as known in the art).

[0026] The oligonucleotide may include a protecting group 110 that protects an O atom group at the 5' position 116. As shown in reaction 115 (represented by an arrow), the O atom at the 5' position 116 may be deprotected so that an OH group 117 is located at the 5' position 116. In some embodiments, the protecting group 110 is a DMT group, and removal of the protecting group comprises reacting the 5' position of the oligonucleotide with an activating solution comprising an acid in a solvent.

[0027] Once the 5' position 116 is deprotected, the oligonucleotide 102 can react with the amidite 102a. This amidite 102a reacts with the deprotected OH group 117 via a P bond. More specifically, the P atom 120 reacts with the OH group 117 to form a bond between the oligonucleotide 120 and the phosphoramidite 120a. This reaction is known as coupling reaction 121 (indicated by an arrow). The P atom 120 is in oxidation state 3 (also designated "III"). As a result of coupling reaction 121, the oligonucleotide 120 and the phosphoramidite 120a are bonded to each other, leaving one of the oligonucleotides coupled to the resin 104.

[0028] An oxidation step 128 (indicated by the arrow) then occurs, which converts the P atoms 120 from oxidation state III to oxidation state "V" (5 (five) or 5). One skilled in the art will understand the conditions used to achieve this oxidation. Although not shown in FIG. 1, a capping step may occur either before or after the oxidation reaction 128.

[0029] After oxidation reaction 128, the added amidite 120a also bears protecting group 110. Thus, a new "cycle" or "series" of reactions can occur. This may involve simply repeating the above reactions to add the next amidite to the chain. Specifically, protecting group 110 of amidite 120a may be removed (deprotected), followed by coupling reaction 121 and oxidation reaction 128 (and / or capping reaction) as needed. This iterative process may be repeated as many times as necessary to create an oligonucleotide chain of the desired length. Alternatively, reaction 128 may be a sulfurization (also called thiolation) reaction, converting P atom 120 from oxidation state III to oxidation state "V" (5 (five) or 5), where the P atom is attached to a sulfur (S) atom via a P=S double bond.

[0030] Referring now to FIG. 2, a schematic diagram of an SPOS reactor system 200 is illustrated. The system 200 includes a reactor 202 (also known as a reactor bed) containing a resin 204. The resin 204 is the same as the resin 104 described above. Thus, as described above, the resin 204 contains oligonucleotide chains that can be grown to a desired length (as known in SPOS synthesis). The reactor 202 includes a filter 206, which may be disposed at the bottom of the reactor 202. In the embodiment of FIG. 2, a gas chamber 210 is disposed below the filter 206, as is an outlet port 212. The outlet port 212 allows liquid and / or gas to exit the reactor 202. In other embodiments, the outlet port 212 and the gas chamber 210 may be the same opening. In the embodiment of FIG. 2, the outlet port 212 is shown within the reactor. In a more preferred embodiment, the outlet port may originate from the filter 206. The gas chamber may simply be a process tube or process pipe exiting the bottom of the reactor below the filter.

[0031] Reactor 202 also includes another inlet port 220. In the specific embodiment of FIG. 2, there are multiple inlet ports 220a, 220b, 220c, 220d, and 220e. The top portion of the reactor, including ports 220a, 220b, 220c, 220d, and 220e, and the bottom portion of the reactor, including ports 206, 210, and 212, may be separate vessels with tubing and optional valves between the first feed zone vessel and the second filter reactor zone vessel. Of course, those skilled in the art will understand that more or fewer ports 220 may be used. Indeed, in some embodiments, a single port may be used. Port 220a may be used to introduce a wash solution 240 (illustrated as a box) into reactor 202. Port 220b may be used to introduce an activated amidite solution 242 (illustrated as a box) into reactor 202. Port 220c may be used to introduce capping solution 244 (illustrated as a box) into reactor 202. Port 220d may be used to introduce oxidizing solution 246 (illustrated as a box for oxidation or thiolation) into reactor 202. Port 220e may be used to introduce deprotecting solution 248 (illustrated as a box) into reactor 202. Other embodiments may be designed where there is only one port 220 and all solutions enter reactor 202 through the single inlet port 220.

[0032] 1 and 2 collectively, the use of the SPOS system 200 will be described to show how the reactor 200 (also known as a bed reactor) is assembled and operated. As mentioned above, the solid support 204 is a solid support for attaching the phosphoramidite 102 to the growing oligonucleotide. Using port 220e, a deprotection solution 248 is used to remove the protecting group 110 from the 5′ position 116 of the oligonucleotide 102 attached to the solid support 204. The deprotection solution 248 flows through the solid support 204 and then downward through the filter 206.

[0033] Initially, the pressure difference across filter 206 is low, e.g., approximately 0 psig (pounds per square inch gauge). Pressure is then applied to the top of reactor 200 (via pressure port 252). Typically, this pressure is about 15 psig, although other pressures may be used. Such pressure forces a portion of liquid deprotection solution 248 downward through solid support 204 and filter 206 (as indicated by arrow 265). After solution 248 has been forced downward through filter 206, the gas chamber or process pipe 210 below filter 206 approaches 15 psig. System 200 then vents the top of filter 206 (e.g., via pressure port 252 (or some other similar mechanism / port)), and the approximately 15 psig pressure trapped below filter 206 forces solution 248 back up into filter 206 and solid support 204 (as shown by arrow 270) until the top and bottom pressures are again approximately equal at approximately 0 psig.

[0034] Such pressurization can be used to cause solution 248 to flow up and down through reactor 200 as many times (and at the rate used for flow) as desired. (Such pressure differentials can also be used to cause all solutions added to reactor 202 to flow in the same manner.) In some embodiments, solution 248 may flow up and down once every 10-15 seconds. In other embodiments, system 200 is designed so that solution 248 flows up and down one or more times to fluidize reactor 202, and then flows slowly downward to continue the reaction in a traditional plug flow manner. Flowing solution 248 up and down allows solution 248 to come into contact with solid support 204 multiple times, thus promoting complete contact and thorough distribution of the solid and liquid phases during the reaction. This also reduces the pressure drop when liquid flows downward, since much of the expansion and contraction occurs during fluidization. In other embodiments, only a small amount of liquid is forced downward through the filter screen at the bottom of the reactor, while nitrogen is blown upward from the bottom through the filter screen, bubbling to fluidize the solid support bed within the reactor and mix with the liquid. After fluidization with the upward flow of liquid reagents or inert gas from the bottom of the reactor, the next part of the deprotection reaction can utilize a controlled-rate downward flow of the reagent solution through the solid support bed, similar to a conventional packed-bed SPOS. However, this embodiment does not cause any mixing of the solution 248 within the reactor, and it may simply flow through the solid support plug flow and exit the filter of the reactor 206.

[0035] Deprotection solution 248 may be removed from reactor 202 through port 212. The use of pressure through pressure port 252 can facilitate removal of deprotection solution 248, and the liquid may be pumped out the bottom of reactor 202 through 212 at a controlled rate. A first wash solution 240a may be added through port 220a. This wash solution 240a may flow up and down through reactor 202 using the pressure differential outlined above, or may be mixed with the solid support by bubbling gas through the bottom of the reactor or some other mixing method, or may simply be passed through a plug flow of the solid support without flowing up and down or mixing. By flowing the wash solution up and down or fluidizing it with gas bubbling or other agitation, the same solution contacts (and "washes") solid support 204 one or more times. The reagent solution is completely emptied from the reactor before the addition of the wash solvent, and the wash solvent is completely emptied from the reactor before the addition of the next liquid. This may reduce the amount of wash solution 240a required (which may reduce costs associated with obtaining, using, and disposing of wash solution) compared to traditional packed-bed SPOS processes, where backmixing can occur within the liquid layer on top of the solid support bed during transfer. A distributor can be used to evenly pack the wash solvent 240a across the entire solid support surface in a manner that does not disturb the flatness of the solid support bed. The number of iterations of flowing the wash solution up and down through the reactor 202 depends on the particular reaction and the particular cycle. Furthermore, a fluidized wash may be followed by a plug flow wash, which then serves to de-swell the solid support bed and reset it to a level, channeling-free solid support bed at the top. Alternatively, if there are no pressure drop or channeling issues at a particular step, all washes may be performed in plug flow without fluidization. Upon completion, the wash solution 240a may exit the reactor 202 via outlet 212. Those skilled in the art will appreciate that one or more additional cleaning "cycles" or "rounds" may be performed by introducing larger amounts of first cleaning solution 240a, if desired.Additionally, the wash solution may be an integrated multi-pass reuse of wash solution from a previous cycle, as described herein.

[0036] Once the first wash step (or steps) has occurred and first wash solution 240a has been removed from reactor 202 (using pressure, pumping, or other driving force to force liquid through a filter), activated amidite solution 242 may be added via inlet 220b. The activated amidite solution 242 contains amidite 120a and is mixed with the solid support by flowing up and down through reactor 202 a desired number of times, or by bubbling gas through the bottom of the reactor, or by some other mixing method. By flowing up and down or mixing, activated amidite solution 242 comes into contact with oligonucleotide 102 on solid support 204 multiple times, thereby increasing the likelihood and / or efficiency of the coupling reaction. As described in detail above, the coupling reaction involves the amidite reacting at the 5' position of the oligonucleotide to form a phosphorus bond to P atom 120. In other embodiments, fluidization is achieved by bubbling nitrogen gas through the bottom of the reactor to achieve mixing of the solid and liquid phases. The same statements regarding bubbling nitrogen from the bottom of the reactor to mix the liquid and solid phases apply to each of the following fluidization descriptions in this story.

[0037] After the coupling reaction is complete, activated amidite solution 242 may be removed from reactor 202 (with or without pressure) via outlet 212, and second wash solution 240b may be added (via port 220a or otherwise). Second wash solution 240b may be the same solution as first wash solution 240a, or in other embodiments, a different wash mixture. This second wash solution 240b may flow up or down through reactor 202 in the manner described herein. Alternatively, second wash solution 240b may be mixed with the solid support by bubbling gas through the bottom of the reactor or by some other mixing method, or may flow in a solid support bed plug flow manner without any mixing or fluidization. Similar to the first cleaning solution 240a, embodiments may be designed in which the second cleaning solution 240b may exit the reactor 202 via outlet 212, and one or more additional cleaning "cycles" or "rounds" or "portions" may be performed as desired by introducing a new (clean) batch of the second cleaning solution 240b. Other embodiments may be designed in which a single batch of the second cleaning solution 240b is used.

[0038] After removing the final wash solution 240b (with or without pressure), an oxidation reaction may occur by introducing an oxidation or thiolation solution 246 through inlet 220d. As described above, the oxidation reaction converts P atoms 120 from oxidation state III to oxidation state V. Again, the oxidation solution 246 may flow up and down through the reactor 202 in the manner outlined herein, potentially improving reaction efficiency and reducing the amount of oxidation solution 246 required, or it may be mixed with the solid support by bubbling gas through the bottom of the reactor or some other mixing method, or it may flow in a solid support bed plug flow mode without any mixing or fluidization. The number of up and down flow repetitions and the duration of each cycle, as with other solutions, will vary depending on the conditions and can be modified by those skilled in the art. After fluidization with an upward flow of liquid reagent or inert gas from the bottom of the reactor, the next part of the oxidation reaction may utilize a controlled-rate downward flow of the oxidation reagent solution through the solid support bed, similar to a conventional packed-bed SPOS. Once the oxidation reaction is complete, the oxidizing solution 246 may exit the reactor 202 via port 212 (with or without pressure assistance).

[0039] After the oxidation reaction, a third wash solution 240c may be introduced via inlet 220a (through port 220a or otherwise). The third wash solution 240c may be the same solution as the first wash solution 240a or the second wash solution 240b, or in other embodiments, a different wash mixture. This third wash solution 240c may flow up and down through reactor 202 in the manner described herein, or may be mixed with the solid support by bubbling gas through the bottom of the reactor or by some other mixing method, or may flow in a plug flow manner through the solid support bed without any mixing or fluidization. Again, such up and down flow and complete emptying of each liquid portion followed by plug flow washes may allow for more efficient washing by eliminating channeling rings or may alleviate pressure drop issues by allowing the solid support to expand or de-expand during fluidization or suspension, reducing the total amount of wash solution required. Similar to the first and second cleaning solutions 240a and 240b, a third cleaning solution 240c may exit the reactor 202 via outlet 212, and embodiments may be designed in which one or more additional cleaning "cycles," "rounds," or "portions" may be performed by introducing additional portions of the third cleaning solution 240c, as desired. Additionally, the cleaning solution may be integrated multi-pass reuse of cleaning solution from a previous cycle, as described herein. Other embodiments may be designed in which a single batch of the third cleaning solution 240c is used. As with all cleaning or reagent loading into the reactor, a distributor may be used to evenly load the cleaning solvent across the solid support surface in a manner that does not disturb the flatness of the solid support bed.

[0040] A capping reaction may also occur within reactor 202. This capping reaction may occur either before or after the oxidation reaction (i.e., the step in which the P atom is converted from oxidation state III to oxidation state V). To facilitate this capping reaction, a capping solution 244 may be added via inlet 220c. This capping solution 244 may be flowed up and down through reactor 202 in the manner outlined herein, potentially improving reaction efficiency and reducing the amount of solution 244 required. The number of repetitions of the up and down flow and the time of each cycle, as with other solutions, vary depending on the conditions and can be modified by those skilled in the art. Alternatively, the capping reagent solution may be mixed with the solid support by bubbling gas through the bottom of the reactor or by some other mixing method, or may flow in a solid support bed plug flow mode without any mixing or fluidization. After fluidization by the upward flow of liquid reagent or inert gas from the bottom of the reactor, the next part of the capping reaction may utilize a controlled-rate downward flow of the reagent solution through the solid support bed, similar to a conventional packed-bed SPOS. Once the capping reaction is complete, capping solution 244 may exit reactor 202 (with or without pressure assistance) via port 212. After removing capping solution 244, a washing step may occur. If the capping reaction occurs before the oxidation reaction, this would be the third wash, but if the capping reaction occurs after the oxidation step, this would be the fourth wash step. This wash may occur in the same manner as outlined herein.

[0041] After the oxidation or capping reaction (and washing), the cycle may be "restarted" to add a new post-phosphoramidite to the growing chain, starting with a deprotection reaction (e.g., addition of a deprotection solution), followed by completing the cycle as many times as necessary to obtain the desired product.

[0042] In some embodiments, solutions (such as wash solutions, activated amidite solutions, capping solutions, oxidation solutions, and / or deprotection solutions) may exit the reactor by passing through a filter at the bottom of the reactor. Of course, other methods of removing these solutions can also be used.

[0043] In the embodiment shown in FIG. 2, "up and down" flow through the reactor bed is achieved by pressure, moving the fluid vertically. However, as used herein, "up and down" also includes moving the fluid horizontally (e.g., from one side of the reactor through the bed to the other) or diagonally through the reactor. Any type of "oscillation" of the fluid through the reactor is included within the meaning of "up and down" flow. Such movement can also be achieved by a pressure differential, which is within the knowledge of those skilled in the art. Mixing may be induced by bubbling an internal gas from the bottom of the reactor. The gas bubbling may be intermittent, so that the liquid alternates between being forced down through the solid support and fluidizing with the solid support, or it may be constant bubbling throughout the entire reaction time. Intermittent fluidization may be more important for narrow, tall reactors to quickly achieve complete liquid contact with the entire solid support, but may be less important for larger diameter reactors.

[0044] In Examples 1-4 and 6-10, the wash solvent is drained from the bottom of the filter reactor before the reagents are charged. Similarly, the reaction solution is drained from the bottom of the filter reactor before the next wash solvent is charged. This reduces backmixing and makes the process more efficient compared to a packed bed reactor that does not drain mid-cycle.

[0045] Example 1 - Preparation of HPRT Div22 antisense strand using liquid upflow fluidization The HPRT Div22 antisense strand has the following sequence: 5'[Phos]mA*fU*mA mA mA fA mU mC mU mA mC mA mG fU mC fA mU mA mG mG mA*mA*mU, where * represents a P=S bond, and all other amidites have P=O bonds and RNA1{pm(A)[sp].[fl2r](U)[sp].m(A)pm(A)pm(A)p.[fl2r](A)pm(U)pm(C)pm(U)pm(A)pm(C)pm(A)pm(G)p.[fl2r](U)pm(C)p.[fl2r](A)pm(U)pm(A)pm(G)pm(G)pm(A)[sp].m(A)[sp].m(U)] (structure shown in Figure 6).

[0046] The synthesis of this molecule using the fluidized-bed method of the present invention is described herein, including deblocking, coupling, oxidation (or sulfurization), and capping steps to sequentially introduce the remaining phosphoramidites 3' to 5' of the HPRT Div22 antisense strand. The goal of Example 1 was to demonstrate the chemistry for the first time in a research-scale fluidized-bed reactor with high purity and yield. The goal was not to minimize the ACN wash solvent, minimize the DCA reagent solution, minimize the amount of amidite equivalents, or demonstrate a high solid support bed height. For examples minimizing the use of ACN solvent, see Example 6 at research scale and Examples 8 and 9 at pilot scale. For examples minimizing the amount of DCA solution, see Example 7. Furthermore, in Example 1, four equivalents of amidite were used in each cycle. In contrast, in Examples 2-9, two equivalents of amidite were used in all or most cycles. The maximum solid support bed height in Example 1 was only 2 cm, while the solid support bed heights in Examples 2-4 and Examples 6-10 were higher. For a 30 cm high resin bed, see Example 2. A guide to all examples is listed in Table 31.

[0047] Start by coupling mU (herein referred to as "mU resin") to NittoPhase HL 2'OMeU(bz)300 resin (Lot # E05005, 299 umol / g) using known methods; see Figure 3 for synthesizer setup.

[0048] Prepare the reagent solutions shown in Table 1. [Table 1]

[0049] The 0.1 M amidite solution shown in Table 2 is prepared as follows: weigh the amidite solid into a bottle, insert a dry pad, then add ACN to achieve a concentration of 0.1 M. Of course, one skilled in the art could also weigh the solid, dissolve it in ACN, then add a sieve and allow it to dry. [Table 2]

[0050] See Figure 3 for the setup of the vibrating fluidized-bed oligonucleotide synthesizer. In Figure 3, "ACN" refers to acetonitrile. Prime all pumps and feed lines. Place dry packs into the ACN bottle and all syringes. Use syringe pumps for the amidite solution (amidite pumps 101-108 in Figure 3), phosphorylation solution (amidite pump 109 in Figure 3), and activator solution (activator pump 110), and peristaltic pumps and feed vessels for all other reagent and solvent feeds. Equip the bottom of a 1 cm diameter, 20 cm tall reactor with a filter and automatic block valve (valve 24 in Figure 3), followed by 1.59 mm inner diameter tubing sufficient to accommodate approximately 2.5 mL of effluent volume from the reactor to one or more outlet valves (valves 9 and 10 in Figure 3). Fill the reactor with 0.1040 g of mU resin. The starting bed height of the dry resin was approximately 0.3-0.4 cm. Each addition of a phosphoramidite during synthesis is followed sequentially by a deblocking step, a coupling step, an oxidation step (or a sulfurization step if there is a P=S bond in the sequence), and a capping step, as described below.

[0051] In each step, resin bed fluidization occurs at two different times: once when the reagent mixture is loaded into the reactor and the resin is exposed to it, and once when the wash solvent is loaded into the reactor. However, fluidization continues throughout the coupling reaction for the entire 10-minute coupling time. In this example, during both the reagent load and solvent wash, the reagent mixture or wash solvent (or a portion thereof) is added to the feed zone and nitrogen pressure is applied to force the liquid into the reactor. Referring to Figure 3, this is accomplished by closing valves 9 and 10, opening valve 24, and applying nitrogen pressure through the appropriate inlet (valve 41, 42, 43, 44, or 45). This equalizes the pressure gradient between the top of the reactor and the tubing between the bottom of the reactor and valves 9 and 10, forcing the liquid in the reactor through the resin bed and into the tubing between the reactor and valves 9 and 10. The pressure at the top of the reactor is then released by opening the appropriate vent (valve 51, valve 52, valve 53, valve 54, or valve 55), creating a pressure gradient that equalizes with atmospheric pressure as liquid flows upward from the bottom of the reactor, agitating and fluidizing the resin bed. The nitrogen pressurization and venting process is repeated a specified number of times, each time for a duration at least sufficient to fluidize the resin bed. If only a portion of the reagent mixture or wash solvent is used to fluidize the bed, the remaining reagent mixture or wash solvent passes through the resin in a "plug flow" manner, where valve 9 is closed, valve 10 is open, nitrogen pressure is applied to the top of the reactor, and pump 9 is activated to meter the liquid out the bottom of the reactor as the liquid (reagent mixture or wash solvent) is added to the top of the reactor.

[0052] Because this first example was an early demonstration of an early prototype, the amidite + activator equivalents, DCA equivalents, and solvent wash volume were significantly higher in this first example compared to all subsequent examples. The reader will see that the process improves and the wash solvent is reduced as the examples and embodiments progress. See Table 31 for a summary of the embodiments.

[0053] Deblocking: Turn valve 8 to A, turn valve 7 to B, and close valve 24. Fill the feed zone with 8 mL of deblocking solution (Table 1) and pump it through the reactor with nitrogen pressure for 8 seconds. Open valve 24. The outlet waste valves (valves 9 and 10 in Figure 3) are closed. Nitrogen pressure is applied to the reactor for 5 seconds, forcing approximately 1.5 mL of reagent solution downward through the resin bed and through the filter at the bottom of the reactor into the process tubing, compressing any gas pockets in the tubing. Vent the pressure from the top of the reactor for 5 seconds, allowing the reagent liquid to flow back into the bottom of the reactor and agitate and fluidize the resin bed. Repeat the fluidization process (pressurize with nitrogen for 5 seconds, vent for 5 seconds). Open the waste valve (valve 10) and pump the deblocking solution through the resin bed with pump 9 at a rate of 16 mL per 330 seconds for 330 seconds. In parallel with pumping with Pump 9, open valve 14 and begin Pump 1 feeding deblocking solution at 10 mL / min until 8 mL of deblocking solution has been pumped (Pump 1 finishes before Pump 9). The liquid pumped from Pump 1 into the acid feed zone simultaneously flows into the reactor to maintain a liquid level above the resin bed, and the flow is maintained for 330 seconds. The total time the resin is in contact with the deblocking solution before the next ACN wash step is 6.9 minutes. Close valve 10 and perform ACN Wash Procedure A once, followed by ACN Wash Procedure B twice.

[0054] ACN Wash Procedure A: Open waste valve 9 and charge ACN (4 mL) to the feed zone, then close valve 9 and push nitrogen pressure into the reactor for 8 seconds. Fluidize the resin bed 5 times as above, pressurize the reactor with nitrogen for 5 seconds, and vent for 5 seconds. Open waste valve 9 and push nitrogen pressure into waste for 8 seconds.

[0055] ACN Wash Procedure B: Open valve 9 (to waste) and charge ACN (12 mL) to the feed zone, then close valve 9 and force nitrogen pressure into the reactor for 8 seconds. Fluidize the resin bed three times as above, pressurizing the reactor with nitrogen for 5 seconds and venting for 5 seconds. Open valve 10 and pump with pump 9 at a rate of 20 mL per 110 seconds for 110 seconds. In parallel with pumping with pump 9, open valve 34 and begin pump 2 feed at 40 mL / min until 8 mL of ACN has been pumped (pump 2 finishes before pump 9). The liquid pumped from pump 2 into the feed zone simultaneously flows into the reactor, maintaining the liquid level above the resin bed and continuing the flow for 110 seconds. At the end, close valve 10.

[0056] Coupling Reaction: After deblocking washes, the next successive phosphoramidite is coupled, introduced in sequential steps from 3' to 5'. For each phosphoramidite coupled in the sequence, the coupling reaction procedure is carried out essentially as described below, using the amidite solution corresponding to the phosphoramidite in the sequence (listed in Table 2). Valve 8 is turned to position B. Pre-wash the amidite zone and the flow path to the reactor twice, each time by pumping 4 mL of ACN into the amidite feed zone with valve 9 closed. Then, open valve 9 and push with nitrogen pressure to waste for 8 seconds. Pump first the activator solution (1.2 mL, 20 equivalents, Table 1) into the feed zone, followed by the appropriate amidite solution (1.2 mL, 4.0 equivalents) from Table 2 into the feed zone. Close valves 9 and 24, and push the mixture in the feed zone into the reactor with nitrogen pressure for 5 seconds. Then, open valve 24 and continue the nitrogen pressure for 8 seconds.

[0057] The amidite and activator solutions are mixed with the resin, and with valve 24 open and valve 9 closed, the bed is repeatedly fluidized as follows: nitrogen pressure is applied to the top of the reactor for 5 seconds, then the pressure is vented from the top of the reactor for 5 seconds. This process is repeated for 10 minutes, after which valve 9 is opened, nitrogen pressure is applied to the top of the reactor for 8 seconds, and the liquid is drained from the bottom of the reactor to waste. ACN (10 mL) is pumped into the amidite feed zone and forced through the reactor with nitrogen pressure for 30 seconds, and this ACN wash is repeated one more time.

[0058] Oxidation Reaction (If Required Instead of Sulfidation): After the coupling reaction wash, the oxidation reaction is carried out essentially as described below. Valves 6, 7, and 8 are turned to A, and valve 9 is opened. The oxidation solution (Table 1, 4.5 mL) is pumped into the feed zone, valve 9 is closed, and nitrogen pressure is forced into the reactor for 8 seconds. The reactor bed is fluidized twice as follows: the top of the reactor is pressurized with nitrogen pressure for 5 seconds, and then the nitrogen pressure is released by venting for 5 seconds. Valve 10 is opened, and pump 9 is used to pump 4.5 mL of liquid over 40 seconds, and then valve 10 is closed. Valve 9 is opened, and ACN (4 mL) is pumped into the feed zone, and then valve 9 is closed, and ACN is forced into the reactor with nitrogen pressure for 8 seconds. The reactor bed is fluidized five times as follows: the top of the reactor is pressurized with nitrogen pressure for 5 seconds, and then the nitrogen pressure is released by venting for 5 seconds. Valve 9 is opened and nitrogen pressure is applied to the top of the reactor for 8 seconds to push the liquid out of the reactor and discard it.

[0059] After the oxidation reaction, two "plug flow" ACN washes are performed: Open valve 9 and pump ACN into the feed zone (8 mL). Close valve 9 and use nitrogen pressure to force the liquid into the reactor for 8 seconds. Fluidize the reactor bed twice as follows: Pressurize the top of the reactor with nitrogen for 5 seconds, then release the nitrogen pressure by venting for 5 seconds. Open valve 10 and pump 9 to pump 12 mL of liquid over 95 seconds. In parallel with pump 9, open valve 33 and begin pump 2 feed at 30 mL / min until 4 mL of ACN has been pumped (pump 2 finishes before pump 9). The liquid pumped from pump 2 into the feed zone simultaneously flows into the reactor, maintaining the liquid level above the resin bed and allowing the flow to continue for 95 seconds. At the end, close valve 10.

[0060] Sulfurization (thiolation) reaction (if required instead of oxidation): After the coupling reaction wash, perform the thiolation reaction essentially as described below. Turn valve 6 to B, turn valves 5, 7, and 8 to A, and open valve 9. Pump the sulfurization solution (Table 1, 4.5 mL) into the feed zone, close valve 9, and force nitrogen pressure into the reactor for 8 seconds. Fluidize the reactor bed twice as follows: pressurize the top of the reactor with nitrogen pressure for 5 seconds, then release the nitrogen pressure by venting for 5 seconds. Open valve 9 and force the liquid in the reactor to waste while applying nitrogen pressure to the top of the reactor for 8 seconds. Perform two "plug flow" ACN washes identical to those described in the oxidation reaction procedure, except that the washes pass through the "XH feed zone" (Figure 3).

[0061] Capping Reaction: After the oxidation (or sulfurization) reaction wash, the capping reaction is carried out essentially as described below. Valves 5 and 6 are turned to B, and valves 7 and 8 are turned to A. Valve 9 is opened. Capping solution A (2.1 mL, Table 1) and capping solution B (2.1 mL, Table 1) are simultaneously pumped into the feed zone, then valve 9 is closed. The liquid is forced into the reactor with nitrogen pressure for 8 seconds. The reactor bed is fluidized twice as follows: pressurize the top of the reactor with nitrogen pressure for 5 seconds, then release the nitrogen pressure by venting for 5 seconds. Valve 10 is opened, and a 4.2 mL liquid volume is pumped in over 100 seconds. Valve 10 is closed, and two "plug flow" ACN washes are performed, identical to the washes described in the oxidation reaction procedure, except that the washes pass through the "cap feed zone" (Figure 3).

[0062] After the final phosphoramidite cycle is complete, the cycle is repeated using the phosphorylation solution (Table 1) instead of the amidite. After the phosphorylation reagent is coupled and oxidized, the deblocking step, followed by the solvent wash, is repeated. The resin is washed with DEA ​​solution (Table 1) for 10 min. The resin is washed with ACN and dried by blowing nitrogen downward through the resin bed to yield 380 mg of dry resin. The starting resin mass was 104 mg. This corresponds to a weight gain of 276 mg, or 8.88 g / mmol; therefore, the crude mass yield of the protected oligonucleotide product is 96% based on the mass gain.

[0063] The cleavage and deprotection reaction is carried out using concentrated NHOH solution at 50°C for 4 hours. UPLC shows that the cleaved and deprotected oligonucleotide product is 82% pure by peak area percent, as shown in the table of UPLC results for Examples 1-5 (Table 13). LCMS analysis confirms that the major product peak represents the correct HPRT div22 AS chain.

[0064] Referring to FIG. 3 and considering Example 1, the detailed automation procedure of the pump and valve operation sequence is described as follows.

[0065] An example of a detailed automated procedure for pump and valve operation sequences is provided below (explained by considering Example 1 in conjunction with the embodiment of FIG. 3).

[0066] Detailed automation procedure of pump and valve operation sequence of Example 1. Key: "O" means "open", "C" means "close", and "P" means "pump", e.g., "P9" in Figure 3 refers to "pump 9".

[0067] Unblocking Turn valve 8 to A. Turn valve 7 to B.

[0068] The acid solution is forced into the acid feed zone. Open 9 (O9) Open 54 (O54) Open 14 (O14) Acid (8 mL) is pumped into the acid feed zone. Close 14 (C14) Close 54 (C54) Close 9 (C9)

[0069] The acid solution is forced into the reactor, fluidized twice to achieve complete liquid-solid contact, and reset to a flat bed without channels. Open 44 (O44) Wait for the "time to push into reactor" (8 seconds). Do the next six rows twice. Open 44 (O44) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 44 (C44) Open 54 (O54) Wait "Time to vent fluid bed" (5 seconds). Close 54 (C54)

[0070] The acid solution is pumped through the resin plug flow for reaction. Open 44 (O44) Open 10 (O10) Pump P9 is started at a rate of 16 mL per 330 seconds for the next 330 seconds. In parallel with the pumping at P9, open 14 and begin P1 feed at 10 mL / min until 8 mL has been pumped. P1 ends before P9. The liquid pumped from P1 into the acid feed zone flows simultaneously into the reactor to maintain the liquid level above the resin bed, and plug flow is allowed to continue for 330 seconds. Close 14 (C14) Close 44 (C44) Close 10 (C10)

[0071] ACN is pumped into the acid feed zone. Open 9 (O9) Open 34 (O34) Open 54 (O54) Pump in the "ACN volume for fluidized bed wash deblocking" (4 mL). Close 34 (C34) Close 54 (C54) Close 9 (C9)

[0072] Small-scale fluidized bed ACN wash after deblocking. Open 44 (O44) Wait for the "time to push into reactor" (8 seconds). Repeat the following six lines five times: Open 44 (O44) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 44 (C44) Open 54 (O54) Wait "Time to vent fluid bed" (5 seconds). Close 54 (C54) Open 44 (O44) Open 9 (O9) Wait for the "time it takes to push out and discard after fluidization" (8 seconds). Close 44 (C44)

[0073] Plug flow wash after deblocking (do this twice). Start plug flow wash and fluidize three times to set the bed evenly and eliminate channeling. Open 9 (O9) Open 34 (O34) Open 54 (O54) ACN (12 mL) is pumped into the feed zone. Close 34 (C34) Close 54 (C54) Close 9 (C9) Open 44 (O44) Wait for the "time to push into reactor" (8 seconds). Repeat the next six lines three times. Open 44 (O44) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 44 (C44) Open 54 (O54) Wait "Time to vent fluid bed" (5 seconds). Close 54 (C54) Open 44 (O44) Open 10 (O10) Pump P9 is started at a rate of 20 mL per 110 seconds for the next 110 seconds. In parallel with the pumping at P9, open 34 and begin P2 feed at 40 mL / min until 8 mL of ACN has been pumped. P2 ends before P9. The liquid pumped from P2 into the acid feed zone flows simultaneously into the reactor to maintain the liquid level above the resin bed, and plug flow continues for 110 seconds. Close 34 (C34) Close 44 (C44) Close 10 (C10)

[0074] Coupling Reaction Turn valve 8 to B.

[0075] The amidite zone and the flow path to the reactor are pre-washed (this is done twice) before coupling. Open 35 (O35) Open 55 (O55) Pump 4 mL of ACN into the amidite feeding zone at P2. Close 35 (C35) Close 55 (C55) Open 9 (O9) Open 45 (O45) Wait the time (8 seconds) before pushing it out and discarding it. Close 45 (C45)

[0076] The amidite and activator are weighed into the amidite feed zone. Open valve 110A. The specified amount of activator (1.2 mL) is pumped. Valve 110A is closed. Valve 110B is opened. Wait 5 seconds and push the activator solution into the amidite mixing zone. Valve 110B is closed. Open valve 101A. Note: Valve 101 was used for mA. Each amidite had its own valve and its own feed line to the activation zone. The specified amount of amidite (1.2 mL) is pumped. Valve 101A is closed. Open valve 101B. Wait 5 seconds and push the amidite solution into the amidite mixing zone. Valve 101B is closed.

[0077] The amidite reaction solution is forced into the reactor and mixed with the resin for 10 minutes. Close 55 (C55) Close 9 (C9) Close 24 (C24) Open 45 (O45) Wait 5 seconds. Open 24 (O24) Wait for the "time to push into reactor" (8 seconds). Repeat the next 7 lines for "Fluid Bed Coupling Time" (10 minutes). Open 45 (O45) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 45 (C45) Open 55 (O55) Wait "Time to vent fluid bed" (5 seconds). Close 55 (C55) Wait for the "time between fluidizations in coupling" (2 seconds). After the "fluid bed coupling time" has ended Open 45 (O45) Open 9 (O9) Wait for the "time it takes to push out and discard after fluidization" (8 seconds). Close 45 (C45)

[0078] After coupling, solvent wash with ACN (do this twice). Open 35 (O35) Open 55 (O55) Pump in the "ACN volume for single pass wash coupling" (10 mL). Close 35 (C35) Close 55 (C55) Open 9 (O9) Open 101B, 102B, 103B, 104B, 105B, 106B, 107B, 108B, 109B, and 110B simultaneously (O101B, 102B, 103B, 104B, 105B, 106B, 107B, 108B, 109B, and 110B (simultaneously)) Wait for "Time to push out and discard single-pass coupling wash" (30 seconds) Close 101B, 102B, 103B, 104B, 105B, 106B, 107B, 108B, 109B, and 110B simultaneously (C101B, 102B, 103B, 104B, 105B, 106B, 107B, 108B, 109B, and 110B (simultaneously)). Close 9 (C9)

[0079] Oxidation (instead of sulfurization, if desired) Open 9 (O9) Turn valve 8 to A. Turn valve 7 to A. Turn valve 6 to A.

[0080] The iodine solution is pumped into the oxidation feed zone. Open 13 (O13) Open 53 (O53) Pump in 4.5 mL of iodine. Close 13 (C13) Close 53 (C53) Close 9 (C9)

[0081] The iodine solution is forced into the reactor, fluidized twice, and reset to a flat bed without channels. Open 43 (O43) Wait for the "time to push into reactor" (8 seconds). Do the next six rows twice. Open 43 (O43) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 43 (C43) Open 53 (O53) Wait "Time to vent fluid bed" (5 seconds). Close 53 (C53)

[0082] The iodine solution is pumped through the resin plug flow for reaction. Open 43 (O43) Open 10 (O10) Start pump P9 at a rate of 4.5 mL per 40 seconds for the next 40 seconds. Close 43 (C43) Close 10 (C10)

[0083] Small-scale fluidized bed ACN wash after oxidation. Open 9 (O9) Open 33 (O33) Open 53 (O53) Pump 4 mL of ACN into the oxidation feed zone. Close 33 (C33) Close 53 (C53) Close 9 (C9) Open 43 (O43) Wait for the "time to push into reactor" (8 seconds). Do the following six rows five times: Open 43 (O43) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 43 (C43) Open 53 (O53) Wait "Time to vent fluid bed" (5 seconds). Close 53 (C53) Open 43 (O43) Open 9 (O9) Wait for the "time it takes to push out and discard after fluidization" (8 seconds). Close 43 (C43)

[0084] Plug flow cleaning after oxidation (this is done twice). Start plug flow cleaning and fluidize three times to set the bed evenly and eliminate channeling. Open 9 (O9) Open 33 (O33) Open 53 (O53) ACN (8 mL) is pumped into the feeding zone. Close 33 (C33) Close 53 (C53) Close 9 (C9) Open 43 (O43) Wait for the "time to push into reactor" (8 seconds). Do the following six rows three times: Open 43 (O43) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 43 (C43) Open 53 (O53) Wait "Time to vent fluid bed" (5 seconds). Close 53 (C53) Open 43 (O43) Open 10 (O10) Pump P9 is started at a rate of 12 mL per 95 seconds for the next 95 seconds. In parallel with pumping at P9, open 33 and begin P2 feed at 30 mL / min until 4 mL of ACN has been pumped. P2 ends before P9. Liquid pumped from P2 into the oxidation feed zone flows simultaneously into the reactor to maintain the liquid level above the resin bed, and plug flow is allowed to continue for 95 seconds. Close 33 (C33) Close 43 (C43) Close 10 (C10)

[0085] Sulfidation (if desired instead of oxidation) Open 9 (O9) Turn valve 8 to A. Turn valve 7 to A. Turn valve 6 to B. Turn valve 5 to A.

[0086] The sulfiding solution is pumped into the sulfiding feed zone. Open 12 (O12) Open 52 (O52) Pump in the sulfurization solution (4.5 mL). Close 12 (C12) Close 52 (C52) Close 9 (C9)

[0087] The sulfurization solution is forced into the reactor, fluidized twice, and reset to a flat bed without channels. Open 42 (O42) Wait for the "time to push into reactor" (8 seconds). Do the next six rows twice. Open 42 (O42) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 42 (C42) Open 52 (O52) Wait "Time to vent fluid bed" (5 seconds). Close 52 (C52)

[0088] The sulfurization solution is pumped through the resin plug flow for reaction. Open 42 (O42) Open 10 (O10) Pump P9 is started at a rate of 4.5 mL per 90 seconds for the next 90 seconds. Close 42 (C42) Close 10 (C10)

[0089] Small-scale fluidized bed ACN wash after sulfidation. Open 9 (O9) Open 32 (O32) Open 52 (O52) Pump 4 mL of ACN into the sulfurization feed zone. Close 32 (C32) Close 52 (C52) Close 9 (C9) Open 42 (O42) Wait for the "time to push into reactor" (8 seconds). Do the following six rows five times: Open 42 (O42) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 42 (C42) Open 52 (O52) Wait "Time to vent fluid bed" (5 seconds). Close 52 (C52) Open 42 (O42) Open 9 (O9) Wait for the "time it takes to push out and discard after fluidization" (8 seconds). Close 42 (C42)

[0090] Plug flow cleaning after sulfiding (do this twice). Start plug flow cleaning and fluidize twice to set the bed evenly and eliminate channeling. Open 9 (O9) Open 32 (O32) Open 52 (O52) ACN (8 mL) is pumped into the feeding zone. Close 32 (C32) Close 52 (C52) Close 9 (C9) Open 42 (O42) Wait for the "time to push into reactor" (8 seconds). Do the next six rows twice. Open 42 (O42) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 42 (C42) Open 52 (O52) Wait "Time to vent fluid bed" (5 seconds). Close 52 (C52) Open 42 (O42) Open 10 (O10) Pump P9 is started at a rate of 12 mL per 95 seconds for the next 95 seconds. In parallel with pumping at P9, open 32 and begin P2 feed at 30 mL / min until 4 mL of ACN has been pumped. P2 ends before P9. Liquid pumped from P2 into the sulfurization feed zone flows simultaneously into the reactor to maintain a liquid level above the resin bed, and plug flow is allowed to continue for 95 seconds. Close 32 (C32) Close 42 (C42) Close 10 (C10)

[0091] Capping Turn valve 8 to A. Turn valve 7 to A. Turn valve 6 to B. Turn valve 5 to B.

[0092] The capping solution is pumped into the capping supply zone. Open 11A (O11A) Open 11B (O11B) Open 51 (O51) Pump capping solution A (2.1 mL) and capping solution B (2.1 mL) simultaneously. Close 11A (C11A) Close 11B (C11B) Close 51 (C51) Close 9 (C9)

[0093] The capping solution is forced into the reactor and fluidized twice to reset it to a flat bed without channels. Open 41 (O41) Wait for the "time to push into reactor" (8 seconds). Do the next six rows twice. Open 41 (O41) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 41 (C41) Open 51 (O51) Wait "Time to vent fluid bed" (5 seconds). Close 51 (C51)

[0094] The capping solution is pumped through the resin plug flow for reaction. Open 41 (O41) Open 10 (O10) Pump P9 is started at a flow rate of 4.2 mL per 100 seconds for the next 100 seconds. Close 41 (C41) Close 10 (C10)

[0095] Plug flow cleaning after capping (do this twice). Start plug flow cleaning and fluidize twice to set the bed evenly and eliminate channeling. Open 9 (O9) Open 31 (O31) Open 51 (O51) ACN (8 mL) is pumped into the capping feed zone. Close 31 (C31) Close 51 (C51) Close 9 (C9) Open 41 (O41) Wait for the "time to push into reactor" (8 seconds). Do the next six rows twice. Open 41 (O41) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 41 (C41) Open 51 (O51) Wait "Time to vent fluid bed" (5 seconds). Close 51 (C51) Open 41 (O41) Open 10 (O10) Pump P9 is started at a rate of 12 mL per 95 seconds for the next 95 seconds. In parallel with pumping at P9, open 31 and begin P2 feed at 30 mL / min until 4 mL of ACN has been pumped. P2 ends before P9. Liquid pumped from P2 into the capping feed zone simultaneously flows into the reactor to maintain the liquid level above the resin bed, and plug flow is allowed to continue for 95 seconds. Close 31 (C31) Close 41 (C41) Close 10 (C10)

[0096] Example 2 - Preparation of HPRT Div22 Antisense Strand with Resin Bed Height up to 30 cm The same HPRT Div22 antisense strand is prepared as in Example 1. The synthesis of this molecule using the fluidized-bed method of the present invention is described herein, which includes deblocking, coupling, oxidation (or sulfurization), and capping steps to sequentially introduce the remaining phosphoramidites. The main difference is that the reactor geometry and fluidization method are modified to allow for much higher resin bed heights. The process in this example was carried out on a 180 μmol scale, with the resin bed reaching a height of 25 cm wet with ACN solvent by the end of the experiment. During the downflow portion of the final deblocking step, the maximum resin bed height reaches 30 cm. The maximum pressure drop across the resin bed during the experiment is 20 psig. The reactor has a bottom section 32 cm high and 0.63 cm internal diameter, and a conical-bottom top section 10.5 cm high and 4.7 cm diameter. The reactor is equipped with a stainless steel filter screen at the bottom of the 0.63 cm diameter section.

[0097] Each time the resin bed was fluidized, nitrogen forced the liquid and solids up through the 0.63 cm ID section into the upper section of the conical bottom, where nitrogen bubbling thoroughly mixed and fluidized the solids. After each fluidization, the fluidized slurry was then forced down through the 0.63 cm ID section to reform the resin bed, while a small amount of liquid exited the bottom of the reactor through a filter screen. Liquid entering from the feed zone was forced into the top of the 4.7 cm ID section through a 1 / 8 inch stainless steel tube angled toward the wall and then radially angled, which allowed the entering liquid to vortex around the inner wall and prevent splashing. Four equivalents of amidite were used in Example 1, while two equivalents of amidite were used for coupling in Example 2.

[0098] We began by coupling mU (referred to herein as "mU resin") to NittoPhase HL 2'OMeU(bz)250 resin using known methods; see Figure 3 for the synthesis apparatus setup. The resin batch was G07010 with a loading of 246 umol / g. The initial weight of dry resin placed in the reactor was 0.7322 g. The scale of the experiment was therefore 180.1 umol.

[0099] Use the reagent solutions listed in Table 3. [Table 3]

[0100] Prepare a 0.1 M amidite solution as shown in Table 4 below. Weigh the amidite solid into a bottle, insert a dry pad, then add ACN to achieve a concentration of 0.1 M. [Table 4]

[0101] Prime all pumps and supply lines. Place dry packs into the ACN bottle and all syringes. Use syringe pumps for the amidite and activator, and peristaltic pumps and supply vessels for all other reagents and solvents. Use one of the amidite syringe pumps (Amidite 9 Pump) for the Phos reagent. Equip the 0.63 cm ID reactor described above with a filter. Attach the top of an automatic block valve (valve 24 in Figure 3) to the bottom of the reactor, then attach enough tubing from the reactor to one or more outlet valves (valves 9 and 10 in Figure 3) to accommodate an effluent volume of approximately 3-4 mL.

[0102] The total synthesis conditions are shown in Table 5. [Table 5]

[0103] Each addition of a phosphoramidite during synthesis is followed sequentially by a deblocking step, a coupling step, an oxidation step (or a sulfurization step if there is a P=S bond in the sequence), and a capping step, as described below.

[0104] With reference to Figure 3 and considering Example 2, a detailed automated procedure for pump and valve operation sequences is described herein. When the procedure states that liquid is pumped downward through the resin bed, this means that a waste pump at the bottom outlet of the reactor operates at a target setpoint while nitrogen pressure pushes against the top of the resin bed, forcing the liquid downward. The purpose of the peristaltic pump (Pump 9) is to meter the liquid through the bed at a controlled rate.

[0105] Deblocking: Turn valve 8 to A, turn valve 7 to B, and close valve 24. 30 mL of deblocking solution (3% by volume dichloroacetic acid (DCA) in toluene) is charged to the feed zone and forced into the reactor with nitrogen pressure for 8 seconds. Open valve 24. The waste outlet valves (valves 9 and 10 in Figure 3) are closed. Nitrogen pressure is applied to the reactor for 3 seconds. Pressure is vented from the top of the reactor for 10 seconds while valve 38 is opened, bubbling nitrogen through the resin bed to agitate and fluidize it with the reagent solution. The metering valve in series with valve 38 is adjusted to a height sufficient to raise and mix the solids and liquids in the upper zone, but not too high to prevent solids from splashing to the top of the upper section, and to minimize the amount of solvent removed. The fluidization process is repeated four more times. Most of the resin expansion occurs during fluidization. Open the waste valve (valve 10) and pump deblocking solution through the resin bed with pump 9 at a rate of 12 mL / min for the user-specified time (330 s for cycles 7–12 and 410 s for cycles 13–24). The total deblocking solution contact time was as follows: cycles 1 and 2 were 14 min, cycle 3 was 12 min, cycles 4 and 5 were 9 min, cycle 6 was 7 min, and cycles 7–24 were 5.5 min. In parallel with pump 9 injection, open valve 14 and begin pump 1 feeding deblocking solution at 15 mL / min until the user-defined amount has been pumped (40 mL for cycles 1–12 and 57 mL for cycles 13–24). Pump 1 finishes before pump 9. The liquid pumped from pump 1 into the acid feed zone simultaneously flows into the reactor, maintaining the liquid level above the resin bed and maintaining the flow for the specified duration. ACN Wash Procedure A was performed twice with 10 mL of solvent, with fluidization for each wash four times, followed by ACN Wash Procedure B once with 10 mL of solvent, followed by ACN Wash Procedure A once with 40 mL of solvent, with fluidization for each wash two times, followed by ACN Wash Procedure B once with 10 mL of solvent. All wash solvent enters the reactor through the acid feed zone (Figure 3). Most of the resin shrinkage occurs during the first two fluidization washes, which alleviates pressure drop issues within the high bed.For the first six phosphoramidites, the deblocking solution flow rate is slowed and the total contact time is increased due to resistance to flow and the pressure above the reactor bed being intentionally limited to 20 psig. For bases 1 through 7, the liquid flux gradually increases and the total contact time of the deblocking solution gradually decreases.

[0106] ACN Wash Procedure A (Fluidization Wash): Open waste valve 9 and charge ACN to the acid feed zone, then close valve 9 and force nitrogen into the reactor for 8 seconds. Fluidize the resin bed as above for the desired number of times, pressurize the reactor with nitrogen for 3 seconds, vent and blow nitrogen upward through the reactor for 10 seconds. Open valve 10 and start waste pump 9 to pump to waste at a rate of 30 mL / min for 20 seconds.

[0107] ACN Wash Procedure B (Plug Flow Wash, No Fluidization): Open valve 9 (to waste) and charge ACN to the acid feed zone, then close valve 9 and force ACN into the reactor with nitrogen pressure for 8 seconds. Open valve 10 and pump with pump 9 at a rate of 30 mL / min for 20 seconds.

[0108] Coupling Reaction: After deblocking and washing, the next successive phosphoramidite is coupled, introduced in sequential steps from 3' to 5'. For each phosphoramidite coupled in the sequence, the coupling reaction procedure is carried out essentially as described below, using the amidite solution corresponding to the phosphoramidite in the sequence (listed in Table 2). Valve 8 is turned to position B. Pre-wash the amidite zone and the flow path to the reactor twice, each time by pumping 8 mL of ACN into the amidite feed zone with valve 9 closed. Then, open valve 9 and push to waste with nitrogen pressure for 30 seconds. Pump first the activator solution (3.6 mL, 10 equivalents) into the feed zone, followed by the appropriate amidite solution (3.6 mL, 2.0 equivalents) from Table 2 into the feed zone. Close valves 9 and 24, and push the mixture in the feed zone into the reactor with nitrogen pressure for 5 seconds.

[0109] The amidite and activator solutions are mixed with the resin, and with valve 24 open and valve 9 closed, the bed is repeatedly fluidized as follows: nitrogen pressure is applied to the top of the reactor for 3 seconds, then the pressure is vented from the top of the reactor, valve 38 is opened, and nitrogen is blown upward through the reactor for 6 seconds. The resin is allowed to flow downward through the liquid for 8 seconds. This process is repeated for 10 minutes, after which valve 9 is opened and nitrogen pressure is applied to the top of the reactor for 30 seconds, and the liquid is drained from the bottom of the reactor to waste. ACN (10 mL) is pumped into the feed zone and forced through the reactor with nitrogen pressure for 30 seconds, and this ACN wash is repeated one more time.

[0110] Oxidation Reaction (If Required Instead of Sulfurization): After the coupling reaction wash, the oxidation reaction is carried out essentially as described below. Valves 6, 7, and 8 are turned to A, and valve 9 is opened. Oxidation solution (9 mL, 2.5 equivalents) is pumped into the feed zone, valve 9 is closed, and nitrogen pressure is forced into the reactor for 8 seconds. The reactor bed is fluidized five times as follows: pressurize the top of the reactor with nitrogen pressure for 3 seconds, then release the nitrogen pressure by venting, open valve 38, and blow nitrogen upward through the reactor for 10 seconds. Open valve 10, and pump 9 with pump 9 to pump a 9 mL volume over 60 seconds. ACN Wash Procedure A (Fluidization Wash) is performed twice with 10 mL of solvent, fluidizing four times in the first wash and twice in the second wash. This is followed by one ACN wash step B (plug flow wash) with 10 mL of solvent, followed by one ACN wash step A with 30 mL of solvent, followed by three fluidized washes, followed by one ACN wash step B with 10 mL of solvent. All wash solvent enters the reactor through the oxidation feed zone. Most of the resin shrinkage occurs during the first two fluidized washes, which alleviates pressure drop issues in the high bed.

[0111] Sulfurization (Thiolation) Reaction (If Required Instead of Oxidation): After the coupling reaction wash, the thiolation reaction is performed essentially as described below. Valve 6 is turned to B, valves 5, 7, and 8 are turned to A, and valve 9 is opened. The sulfurization solution (12 mL) is pumped into the feed zone, valve 9 is closed, and nitrogen pressure is forced into the reactor for 8 seconds. The reactor bed is fluidized 22 times as follows: the top of the reactor is pressurized with nitrogen pressure for 3 seconds, then the nitrogen pressure is released by venting, valve 38 is opened, and nitrogen is blown upward through the reactor for 10 seconds. Most of the resin expansion occurs during fluidization. The total time for the 22 fluidizations is approximately 8 minutes. The waste valve (valve 10) is opened, and the sulfurization solution is pumped through the resin bed with pump 9 at a rate of 12 mL per 30 seconds. ACN Wash Procedure A (Fluidization Wash) is performed twice with 10 mL of solvent, with four fluidizations in the first wash and two fluidizations in the second wash. This is followed by one ACN wash (plug flow wash) with 10 mL of solvent, followed by one ACN wash (plug flow wash) with 30 mL of solvent, followed by three fluidized washes with 10 mL of solvent, followed by one ACN wash (plug flow wash) with 10 mL of solvent. All wash solvent enters the reactor through the XH feed zone (Figure 3). Most of the resin shrinkage occurs during the first two fluidized washes, which alleviates the problem of high bed pressure drop.

[0112] Capping Reaction: After the oxidation (or sulfurization) reaction wash, the capping reaction is carried out essentially as described below. Valves 5 and 6 are turned to B, and valves 7 and 8 are turned to A. Valve 9 is opened. Capping solution A (6.3 mL) and capping solution B (6.3 mL) are simultaneously pumped into the feed zone, then valve 9 is closed. The liquid is forced into the reactor with nitrogen pressure for 8 seconds. The reactor bed is fluidized three times as follows: pressurize the top of the reactor with nitrogen pressure for 3 seconds, then release the nitrogen pressure by venting, open valve 38, and blow nitrogen upward through the reactor for 10 seconds. Most of the resin expansion occurs during fluidization. Valve 10 is opened, and a 12.6 mL liquid volume is pumped in over 70 seconds. ACN Wash Procedure A (fluidization wash) is performed twice with 10 mL of solvent, fluidizing three times in the first wash and twice in the second wash. This is followed by one ACN wash (plug flow wash) with 10 mL of solvent, followed by one ACN wash (plug flow wash) with 30 mL of solvent, followed by three fluidized washes with 10 mL of solvent, followed by one ACN wash (plug flow wash) with 10 mL of solvent. All wash solvent enters the reactor through the capping feed zone (Figure 3). Most of the resin shrinkage occurs during the first two fluidized washes, which alleviates pressure drop issues within the high bed.

[0113] After the final amidite coupling cycle is complete, the cycle is repeated using the phosphorylating solution instead of the amidite. After the phosphorylating reagent is coupled and oxidized, the deblocking step is repeated. The resin is washed with DEA ​​solution as follows: 9.3 mL of DEA solution is charged to the reactor and fluidized four times, then pumped out the bottom of the reactor at 8 mL / min. This 9.3 mL DEA wash is repeated three more times. Then, ACN washes are performed as follows: ACN Wash Procedure A (fluidized wash) is performed twice with 10 mL of solvent, the first wash is fluidized four times, and the second wash is fluidized twice. Then, ACN Wash Procedure B (plug flow wash) is performed once with 10 mL of solvent, followed by ACN Wash Procedure A once with 30 mL of solvent and fluidized twice, followed by ACN Wash Procedure B once with 10 mL of solvent.

[0114] Drying is accomplished by blowing nitrogen downward through the resin bed to yield 2.2319 grams of dry resin, which corresponds to a weight gain of 1.4997 grams, a weight gain of 8.83 g / mmol, and therefore a crude mass yield of the protected oligonucleotide product of 96% based on mass gain.

[0115] A small sample was subjected to cleavage and deprotection reactions using concentrated NH4OH solution at 50°C for 4 hours. UPLC analysis showed that the cleaved and deprotected oligonucleotide product was 80.85% pure by peak area percent, as shown in the table of UPLC results for Examples 1-5 (Table 13). LCMS analysis confirmed that the major product peak represented the correct HPRT div22 AS chain.

[0116] The resin bed expansion, contraction, and growth data throughout the 23-mer oligonucleotide build is shown in Figure 4. The maximum pressure drop across the resin bed during the experiment was 20 psig, as this was the pressure of the nitrogen feed used to force the liquid through the resin bed.

[0117] The trend labeled "detrit" in Figure 4 represents the resin bed height after all of the deblocking reaction solution before washing has been drained through the resin bed. The trend labeled "ACN_detrit" in Figure 4 represents the bed height after the final ACN solvent wash after deblocking. Similarly, the trend labeled "sulfurization / oxidation" represents the resin bed height after all of the sulfurization or oxidation reaction solution before washing has been drained through the resin bed, and so on. The resin bed height increased approximately linearly from amidite cycle 1 to cycle 24. The resin bed height varied approximately 5 cm from minimum to maximum within each cycle. For example, the resin beads expanded during the deblocking reaction, increasing the packed resin bed height by approximately 4 cm. Then, the resin beads contracted during the wash, reducing the resin bed height by approximately 5 cm. After coupling, the resin beads expanded during the oxidation or sulfurization reaction, increasing the bed height by approximately 4 cm. Then, the resin beads contracted during the subsequent wash, reducing the resin bed height by approximately 4 cm. Considering that such extreme expansion and contraction events occur three times per cycle over 24 cycles, it would be impossible to implement such high bed heights in a downflow-only packed-bed reactor because the pressure drop would be prohibitive. Pressure drop is reduced in a fluidized-bed reactor because the resin particle expansion and contraction occurs primarily during bed fluidization. Even though the resin bed is re-settled at each new bed height, the flow resistance through the cake remains low, with a maximum pressure drop of only 20 psig. Furthermore, channeling does not occur each time the resin bed is re-settled after each solvent exchange.

[0118] Referring to FIG. 3 and considering Example 2, the detailed automation procedure of the pump and valve operation sequence is described as follows.

[0119] The procedure was similar to that described in Example 1, except that fluidization was achieved by blowing nitrogen upward through the bottom of the reactor. Also, after each reaction, washing was performed in a different way. The first two washes were fluidized to aid in the subsequent liquid flux through the tall resin bed. Most of the resin expansion occurred during fluidization with reagents at the start of the reaction, and most of the resin deswelling occurred during fluidization with solvent at the start of the wash.

[0120] This sequence was repeated for each amidite. In this procedure, the pump speed and time during deblocking represent cycles 7–12. Because a large amount of deblocking solution was used after the first 12 cycles, the pump time for cycles 13–24 was 410 s instead of 330 s. The pump speed during deblocking was initially slower due to increased resistance to flow through the resin bed in the first six cycles. Over the first six cycles, the deblocking flow rate gradually increased and the time gradually decreased. For example, in the first two cycles, the deblocking plug flow reaction time was 840 s and the pump speed was set to 5 mL / min, but by the seventh cycle, the deblocking plug flow reaction time was 330 s and the pump speed was set to 12 mL / min. This was due to the decreased resistance to flow through the bed as the oligos grew longer on the resin.

[0121] Unblocking Turn valve 8 to A. Turn valve 7 to B.

[0122] The acid solution is pumped into the acid feed zone. O9 (This depressurizes the reactor through the resin bed and allows liquid to exit the bottom of the reactor while the acid is being measured). Open 54 (O54) Open 14 (O14) Acid (30 mL) is pumped into the acid feed zone. Close 14 (C14) Close 54 (C54) Close 9 (C9)

[0123] The acid solution is forced into the reactor, fluidized three times, allowed to react, and reset to a flat bed without channels. Open 44 (O44) Wait for the "time to push into reactor" (8 seconds). Do the following six rows three times: Open 44 (O44) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 44 (C44) Open 54 (O54) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 54 (C54), close 38 (C38)

[0124] The acid solution is pumped through the resin plug flow for reaction. Open 44 (O44) Open 10 (O10) Pump P9 is started at a rate of 12 mL / min for 330 seconds. In parallel with pumping at P9, open 14 and begin the acid feed at 15 mL / min until 40 mL has been pumped. The liquid pumped into the acid feed zone simultaneously flows into the reactor to maintain a liquid level above the resin bed, and plug flow is allowed to continue for 330 seconds. Close 14 (C14) Close 44 (C44) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user setpoint (from approximately 20 psig to 10 psig).

[0125] Fluidization washing. This is done twice, with the first time fluidizing four times and the second time fluidizing two times. Open 9 (O9) Open 34 (O34) Open 54 (O54) ACN (10 mL) is pumped into the feeding zone. Close 34 (C34) Close 54 (C54) Close 9 (C9) Open 44 (O44) Wait for the "time to push into reactor" (8 seconds). Repeat the next 6 rows 4 times in the first wash and 2 times in the second wash. Open 44 (O44) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 44 (C44) Open 54 (O54) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 54 (C54), close 58 (C58) Open 44 (O44) Open 10 (O10) Pump P9 is started at a rate of 20 mL / min for 30 seconds. Close 44 (C44) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user setpoint (from approximately 20 psig to 10 psig).

[0126] Plug flow cleaning. Open 9 (O9) Open 34 (O34) Open 54 (O54) ACN (10 mL) is pumped into the feeding zone. Close 34 (C34) Close 54 (C54) Close 9 (C9) Open 44 (O44) Wait for the "time to push into reactor" (8 seconds). Open 10 (O10) Pump P9 is started at a rate of 30 mL / min for 20 seconds. Close 44 (C44) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value.

[0127] A more extensive fluidized wash to wash down the top wall of the reactor from a high position (note that this was later determined to be unnecessary). Open 9 (O9) Open 34 (O34) Open 54 (O54) ACN (40 mL) is pumped into the feeding zone. Close 34 (C34) Close 54 (C54) Close 9 (C9) Open 44 (O44) Wait for the "time to push into reactor" (8 seconds). Do the following six rows three times: Open 44 (O44) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 44 (C44) Open 54 (O54) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 54 (C54), close 58 (C58) Open 44 (O44) Open 10 (O10) Pump P9 is started at a rate of 30 mL / min for 80 seconds. Close 44 (C44) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user setpoint (from approximately 20 psig to 10 psig).

[0128] Plug flow cleaning. Open 9 (O9) Open 34 (O34) Open 54 (O54) ACN (10 mL) is pumped into the feeding zone. Close 34 (C34) Close 54 (C54) Close 9 (C9) Open 44 (O44) Wait for the "time to push into reactor" (8 seconds). Open 10 (O10) Pump P9 is started at a rate of 30 mL / min for 20 seconds. Close 44 (C44) Close 10 (C10)

[0129] Open 9 (O9) and wait until the pressure drops to the user setpoint (from approximately 20 psig to 10 psig).

[0130] Coupling Reaction Turn valve 8 to B.

[0131] The amidite zone and the flow path to the reactor are pre-washed (this is done twice) before coupling. Open 35 (O35) Open 55 (O55) Pump 8 mL of ACN into the amidite zone at P2. Close 35 (C35) Close 55 (C55) Open 9 (O9) Open 45 (O45) Push down through the resin bed in the reactor and wait for the time (30 seconds) before discarding. Close 45 (C45)

[0132] The amidite and activator are weighed into the amidite+activator feed zone. Open valve 110A. The specified amount of activator (3.6 mL) is pumped. Valve 110A is closed. Valve 110B is opened. Wait 5 seconds to push the activator solution into the amidite + activator supply zone. Valve 110B is closed. Open valve 101A. Note: Valve 101 was used for mA. Each amidite had its own valve and its own feed line to the activation zone (Figure 3). The specified amount of amidite (3.6 mL) is pumped in. Valve 101A is closed. Open valve 101B. Wait 5 seconds and push the amidite solution into the amidite + activator delivery zone. Valve 101B is closed.

[0133] The amidite reaction solution is forced into the reactor and mixed with the resin for 10 minutes. Close 55 (C55) Close 9 (C9) Close 24 (C24) Open 45 (O45) Wait 5 seconds. Open 24 (O24) Wait for the "time to push into reactor" (8 seconds). Repeat the next 7 lines for "Fluid Bed Coupling Time" (10 minutes). Open 45 (O45) Wait for the "time it takes to push down the floor with N2" (3 seconds). Close 45 (C45) Open 55 (O55) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (6 seconds). Close 55 (C55) Wait for the "time between fluidizations during coupling" (8 seconds). After the "fluid bed coupling time" has ended Open 45 (O45) Open 9 (O9) Wait for the "time it takes to push out and discard after fluidization" (30 seconds). Close 45 (C45) Wait until the reactor pressure drops to the user set value (10 psig), indicating all of the coupling solution has been pushed out to waste.

[0134] After coupling, solvent wash with ACN (do this twice). Open 35 (O35) Open 55 (O55) Pump in the "ACN volume for single pass wash coupling" (10 mL). Close 35 (C35) Close 55 (C55) Open 9 (O9) Open 101B, 102B, 103B, 104B, 105B, 106B, 107B, 108B, 109B, and 110B simultaneously (O101B, 102B, 103B, 104B, 105B, 106B, 107B, 108B, 109B, and 110B (simultaneously)) Wait for "Time to push out and discard single-pass coupling wash" (30 seconds) Close 101B, 102B, 103B, 104B, 105B, 106B, 107B, 108B, 109B, and 110B simultaneously (C101B, 102B, 103B, 104B, 105B, 106B, 107B, 108B, 109B, and 110B (simultaneously)). Close 9 (C9)

[0135] Oxidation (instead of sulfurization, if desired) Open 9 (O9) Turn valve 8 to A. Turn valve 7 to A. Turn valve 6 to A.

[0136] The iodine solution is pumped into the oxidation feed zone. Open 13 (O13) Open 53 (O53) Pump in 9 mL of iodine feed solution. Close 13 (C13) Close 53 (C53) Close 9 (C9)

[0137] The iodine solution is forced into the reactor and fluidized 11 times over approximately 4 minutes. This is the batch portion of the reaction. Open 43 (O43) Wait for the "time to push into reactor" (8 seconds). Do the following six rows 11 times: Open 43 (O43) Wait for the "time it takes to push down the floor with N2" (3 seconds). Close 43 (C43) Open 53 (O53) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 53 (C53), close 38 (C38)

[0138] The iodine solution is pumped through the resin for the plug flow portion of the reaction. Open 43 (O43) Open 10 (O10) Pump P9 is started at a rate of 9 mL / min for 60 seconds. Close 43 (C43) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set point (from 20 psig to 10 psig).

[0139] Fluidization washing. This is done twice, with the first time fluidizing four times and the second time fluidizing two times. Open 9 (O9) Open 33 (O33) Open 53 (O53) ACN (10 mL) is pumped into the I2 feeding zone. Close 33 (C33) Close 53 (C53) Close 9 (C9) Open 43 (O43) Wait for the "time to push into reactor" (8 seconds). The next six rows are performed four times during the first fluidized wash and twice during the second fluidized wash. Open 43 (O43) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 43 (C43) Open 53 (O53) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 53 (C53), close 58 (C58) Open 43 (O43) Open 10 (O10) Pump P9 is started at a rate of 20 mL / min for 30 seconds. Close 43 (C43) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set point (from 20 psig to 10 psig).

[0140] Plug flow cleaning. Open 9 (O9) Open 33 (O33) Open 53 (O53) ACN (10 mL) is pumped into the I2 feeding zone. Close 33 (C33) Close 53 (C53) Close 9 (C9) Open 43 (O43) Wait for the "time to push into reactor" (8 seconds). Open 10 (O10) Pump P9 is started at a rate of 30 mL / min for 20 seconds. Close 43 (C43) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user setpoint (dropping from 20 psig to 10 psig).

[0141] A more extensive fluidized wash to wash down the top wall of the reactor from a high position (note that this was later determined to be unnecessary). Open 9 (O9) Open 33 (O33) Open 53 (O53) ACN (30 mL) is pumped into the I2 feeding zone. Close 33 (C33) Close 53 (C53) Close 9 (C9) Open 43 (O43) Wait for the "time to push into reactor" (8 seconds). Do the following six rows three times: Open 43 (O43) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 43 (C43) Open 53 (O53) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 53 (C53), close 58 (C58) Open 43 (O43) Open 10 (O10) Pump P9 is started at a rate of 30 mL / min for 60 seconds. Close 43 (C43) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set point (from 20 psig to 10 psig).

[0142] Plug flow cleaning. Open 9 (O9) Open 33 (O33) Open 53 (O53) ACN (10 mL) is pumped into the I2 feeding zone. Close 33 (C33) Close 53 (C53) Close 9 (C9) Open 43 (O43) Wait for the "time to push into reactor" (8 seconds). Open 10 (O10) Pump P9 is started at a rate of 30 mL / min for 20 seconds. Close 43 (C43) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value (10 psig).

[0143] Sulfidation (if desired instead of oxidation) Open 9 (O9) Turn valve 8 to A. Turn valve 7 to A. Turn valve 6 to B. Turn valve 5 to A.

[0144] The sulfurizing solution is pumped into the XH feed zone. Open 12 (O12) Open 52 (O52) Pump in the sulfurization solution (12 mL). Close 12 (C12) Close 52 (C52) Close 9 (C9)

[0145] The sulfurization solution is forced into the reactor and fluidized 22 times. Open 42 (O42) Wait for the "time to push into reactor" (8 seconds). Do the next six rows 22 times, which takes about eight minutes. Open 42 (O42) Wait for the "time it takes to push down the floor with N2" (3 seconds). Close 42 (C42) Open 52 (O52) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 52 (C52), close 38 (C38)

[0146] The sulfurization solution is pumped through the resin for the plug flow portion of the reaction. Open 42 (O42) Open 10 (O10) Pump P9 is started at a rate that will empty the reactor in about 30 seconds. Close 42 (C42) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set point (from 20 psig to 10 psig).

[0147] Fluidization washing. This is done twice, with the first time fluidizing four times and the second time fluidizing two times. Open 9 (O9) Open 32 (O32) Open 52 (O52) ACN (10 mL) is pumped into the XH feeding zone. Close 32 (C32) Close 52 (C52) Close 9 (C9) Open 42 (O42) Wait for the "time to push into reactor" (8 seconds). Repeat the next six rows four times in the first fluidized wash and twice in the second fluidized wash. Open 42 (O42) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 42 (C42) Open 52 (O52) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 52 (C52), close 58 (C58) Open 42 (O42) Open 10 (O10) Pump P9 is started at a rate of 30 mL / min for 20 seconds. Close 42 (C42) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value (10 psig).

[0148] Plug flow cleaning. Open 9 (O9) Open 32 (O32) Open 52 (O52) ACN (10 mL) is pumped into the XH feeding zone. Close 32 (C32) Close 52 (C52) Close 9 (C9) Open 42 (O42) Wait for the "time to push into reactor" (8 seconds). Open 10 (O10) Pump P9 is started at a rate of 40 mL / min for 15 seconds. Close 42 (C42) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value (10 psig).

[0149] A more extensive fluidized wash to wash down the top wall of the reactor from a high position (note that this was later determined to be unnecessary). Open 9 (O9) Open 32 (O32) Open 52 (O52) ACN (30 mL) is pumped into the feeding zone. Close 32 (C32) Close 52 (C52) Close 9 (C9) Open 42 (O42) Wait for the "time to push into reactor" (8 seconds). Do the following six rows three times: Open 42 (O42) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 42 (C42) Open 52 (O52) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 52 (C52), close 58 (C58) Open 42 (O42) Open 10 (O10) Pump P9 is started at a rate of 40 mL / min for 45 seconds. Close 42 (C42) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value.

[0150] Plug flow cleaning. Open 9 (O9) Open 32 (O32) Open 52 (O52) ACN (10 mL) is pumped into the XH feeding zone. Close 32 (C32) Close 52 (C52) Close 9 (C9) Open 42 (O42) Wait for the "time to push into reactor" (8 seconds). Open 10 (O10) Pump P9 is started at a rate of 40 mL / min for 14 seconds. Close 42 (C42) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value (10 psig).

[0151] Capping Turn valve 8 to A. Turn valve 7 to A. Turn valve 6 to B. Turn valve 5 to B.

[0152] The capping solution is pumped into the capping supply zone. Open 11A (O11A) Open 11B (O11B) Open 51 (O51) Pump capping solution A (6.3 mL) and capping solution B (6.3 mL) simultaneously. Close 11A (C11A) Close 11B (C11B) Close 51 (C51) Close 9 (C9)

[0153] The capping solution is forced into the reactor, fluidized three times, allowed to react, and reset to a flat bed without channels. Open 41 (O41) Wait for the "time to push into reactor" (8 seconds). Do the following six rows three times: Open 41 (O41) Wait for the "time it takes to push down the floor with N2" (3 seconds). Close 41 (C41) Open 51 (O51) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 51 (C51), close 38 (C38)

[0154] The capping solution is pumped through the resin for the plug flow portion of the reaction. Open 41 (O41) Open 10 (O10) Pump P9 is started at a rate that will empty the reactor in about 70 seconds. Close 41 (C41) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value.

[0155] Fluidization washing. This is done twice, with the first time fluidizing three times and the second time fluidizing twice. Open 9 (O9) Open 31 (O31) Open 51 (O51) ACN (10 mL) is pumped into the capping feed zone. Close 31 (C31) Close 51 (C51) Close 9 (C9) Open 41 (O41) Wait for the "time to push into reactor" (8 seconds). Do the next 6 rows 3 times in the first fluidized wash and 2 times in the first fluidized wash. Open 41 (O41) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 41 (C41) Open 51 (O51) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 51 (C51), close 58 (C58) Open 41 (O41) Open 10 (O10) Pump P9 is started at a rate of 30 mL / min for 20 seconds. Close 41 (C41) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value (10 psig).

[0156] Plug flow cleaning. Open 9 (O9) Open 31 (O31) Open 51 (O51) ACN (10 mL) is pumped into the capping feed zone. Close 31 (C31) Close 51 (C51) Close 9 (C9) Open 41 (O41) Wait for the "time to push into reactor" (8 seconds). Open 10 (O10) Pump P9 is started at a rate of 40 mL / min for 15 seconds. Close 41 (C41) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value (10 psig).

[0157] A more extensive fluidized wash to wash down the top wall of the reactor from a high position (note that this was later determined to be unnecessary). Open 9 (O9) Open 31 (O31) Open 51 (O51) ACN (30 mL) is pumped into the capping feed zone. Close 31 (C31) Close 51 (C51) Close 9 (C9) Open 41 (O41) Wait for the "time to push into reactor" (8 seconds). Do the following six rows three times: Open 41 (O41) Wait for the "time it takes to push down the floor with N2" (5 seconds). Close 41 (C41) Open 51 (O51) and 38 (O38) Wait for "Time to vent fluidized bed by N2 bubbling" (10 seconds). Close 51 (C51), close 58 (C58) Open 41 (O41) Open 10 (O10) Pump P9 is started at a rate of 40 mL / min for 45 seconds. Close 41 (C41) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value (10 psig).

[0158] Plug flow cleaning. Open 9 (O9) Open 31 (O31) Open 51 (O51) ACN (10 mL) is pumped into the capping feed zone. Close 31 (C31) Close 51 (C51) Close 9 (C9) Open 41 (O41) Wait for the "time to push into reactor" (8 seconds). Open 10 (O10) Pump P9 is started at a rate of 40 mL / min for 15 seconds. Close 41 (C41) Close 10 (C10) Open 9 (O9) and wait until the pressure drops to the user set value.

[0159] Example 3 - Preparation of HPRT Div22 antisense strand The same HPRT Div22 antisense strand is prepared as in Examples 1 and 2. The synthesis of this molecule using the fluidized-bed method of the present invention is described herein, which includes deblocking, coupling, oxidation (or sulfurization), and capping steps to sequentially introduce the remaining phosphoramidites. The main differences are as follows: Example 3 is performed at a larger scale (1 mmol) in a larger fluidized-bed reactor with the same diameter from bottom to top, an internal diameter of 2.2 cm, and a height of 1 m. This larger-diameter reactor allows for sufficient fluidization without the need for a wide funnel zone at the top. The larger reactor diameter reduces wall effects, making it easier to fully fluidize and redistribute solids and liquids without a wide-diameter section at the top. Fluidization at the start of each reaction step typically reached a height of approximately 0.3 m within the reactor. Additionally, in Example 3, each reaction other than coupling (deblocking, oxidation, sulfurization, and capping) was performed by charging a first portion of the reagent to the reaction mass, fluidizing the first portion for the desired time, and then pumping the first portion in a plug-flow manner through the resin bed while simultaneously charging a second portion of the reagent to the top of the reactor, thereby ensuring that all reagents were pumped in a plug-flow manner. As in Examples 1 and 2, large excesses of wash solvent and DCA reagent solution were used in Example 3. See Example 7 for an example using a reduced amount of DCA reagent, and see Examples 6, 8, and 9 for examples using a reduced amount of ACN wash. Table 31 provides a guide to various embodiments of the fluidized-bed reactor examples.

[0160] The resin bed height reached 9 cm at the end of the ACN solvent wash after draining and was 7 cm dry by the end of the experiment. At the end of the final deblocking and oxidation steps after draining, the maximum resin bed height reached 11 cm. The maximum pressure drop across the resin bed at any time during the experiment was 20 psig, due to the pressure of the nitrogen feed used to force the liquid through the resin bed. As in Example 2, two equivalents of amidite were used in the coupling. The overall synthesis conditions are shown in Table 6. [Table 6]

[0161] Start by coupling mU (herein referred to as "mU resin") to NittoPhase HL 2'OMeU(bz)300 resin (299 μmol / g) using known methods; see Figure 5 for the synthesis apparatus setup. 3.344 g (1.00 mmol) of mU resin is placed in a 2.2 cm internal diameter reactor equipped with a filter frit at the bottom. The initial dry resin depth is approximately 3 cm in height.

[0162] Reagent and amidite solutions are prepared as in Examples 1 and 2. All pumps and feed lines are primed. ACN is passed through a bed of molecular sieves on its way to the inerted feed can. Peristaltic pumps and feed vessels are used for all feeds. The amidite solution is contained in a separate feed vessel labeled "AM.1 L" and connected to a peristaltic pump attached to valves V1A-V8A in Figure 5. The phosphorylation reagent is contained in a feed vessel labeled "AM.1 L" and connected to a peristaltic pump attached to valve V9A in Figure 5. The activator solution and DEA solution are contained in feed vessels labeled "Activator 5 gal" and "DEA" in Figure 5, respectively. As in Figure 3, acetonitrile is abbreviated as "ACN" in Figure 5.

[0163] Each addition of a phosphoramidite during synthesis is followed sequentially by a deblocking step, a coupling step, an oxidation step (or a sulfurization step if there is a P=S bond in the sequence), and a capping step, as described below.

[0164] In each step, fluidization of the resin bed was performed at two different times: once when the reactor was charged with the reagent mixture and the resin was exposed to it, and once when some of the wash solvent steps were charged to the reactor. However, fluidization continued throughout the coupling reaction for the entire 10-minute coupling time. As in Example 2, fluidization was performed by blowing nitrogen gas upward through the bottom filter screen by opening valves 58, 54, and 53 (V58, V54, and V53 in Figure 5) simultaneously with the opening of vent valve V52. When the procedure describes pumping liquid downward through the resin bed, this means that the waste pump at the bottom outlet of the reactor is operating at a target setpoint while nitrogen pressure pushes against the top of the resin bed, forcing the liquid downward. The purpose of the peristaltic pump is to meter the liquid through the bed at a controlled rate.

[0165] Deblocking Reaction: Charge the deblocking solution (100 mL) to the feed zone. Drive the deblocking solution into the feed zone with nitrogen and clear the feed tubing. Push the deblocking solution into the reactor. Fluidize the resin bed twice to ensure complete liquid-to-solid contact and reset the resin bed. The total time for both fluidizations is approximately 1 minute. Begin pumping the deblocking solution downward through the resin bed over 315 seconds at a pump setting of 110 mL / min. Simultaneously pump additional deblocking solution (428 mL) into the reactor, so that it enters the top of the reactor at approximately the same rate as the pump discharges. A total of 528 mL is pumped through the resin bed in 315 seconds. Drive the deblocking solution into the reactor with nitrogen and clear the feed tubing. Push any remaining deblocking solution out the bottom of the filter to waste.

[0166] Wash 1 (do this step twice): Charge ACN solvent (50 mL) to the reactor through the acid feed line. Drive the wash solvent into the reactor with nitrogen, sweeping the feed tube. Push the solvent through the resin bed to waste.

[0167] Wash 2: Charge ACN solvent (90 mL) to the reactor through the solvent feed line. Drive the wash solvent into the reactor with nitrogen, sweeping out the feed tube. Force nitrogen down through the reactor and into the bottom of the fluidization push zone (between V54 and V57). Fluidize the resin bed to ensure complete liquid-to-solid contact and reset the resin bed twice. Begin pumping ACN solvent through the resin bed at a pump setting of 110 mL / min for 120 seconds. Simultaneously pump additional ACN solvent (110 mL) into the reactor, so that it enters the top of the reactor at approximately the same rate as it is being pumped out. A total of 200 mL will be pumped through the resin bed. Push any remaining ACN solvent out the bottom of the filter to waste.

[0168] Wash 3 (do this step twice): Charge ACN solvent (40 mL) to the reactor through the solvent feed line. Drive the wash solvent into the reactor with nitrogen to purge the feed tube. Force the solvent through the resin bed via valve 57 to wash the fluidization push zone between valves 54 and 57 and the waste tube, then push it out of the reactor to waste.

[0169] Coupling Reactions: To minimize the possibility of cross-contamination, each of the six amidites has its own individual pump, valve, and supply line to the activation zone (activation zone labeled "2L"), as shown in Figure 5. Although this construct uses only six amidites, there are nine amidites total (mA, mC, mG, mU, fA, fC, fG, fU, phos), with 10 ports in the amidite zone (including the activator).

[0170] Pre-wash (do this step twice): ACN solvent is charged to the amidite activation zone (80 mL) and forced down through the reactor to waste, also washing the fluidization push zone between valves 54 and 57.

[0171] Reaction: The specified amount of amidite (20 mL) is pumped into the amidite activation zone and purged with nitrogen. The activator solution (20 mL) is pumped into the amidite activation zone and purged with nitrogen. This mixture is forced into the feed zone and then into the reactor to initiate the coupling reaction on the resin. During the 10-minute coupling time, the resin reactor is fluidized once every 30 seconds to mix the contents (in other words, fluidize for 15 seconds every 30 seconds). After the reaction time has elapsed, the coupling solution is forced out the bottom of the filter and discarded.

[0172] Wash 1 (do this step twice): ACN solvent (100 mL) is charged to the amidite activation zone and forced down through the reactor to waste, also washing the fluidization push zone between valves 54 and 57.

[0173] Wash 2 (do this step twice): Charge ACN (40 mL) to the reactor through the solvent feed line. Drive the wash solvent into the reactor with nitrogen, sweeping the feed tubing. Force the solvent through the resin bed and out of the reactor to waste, while simultaneously using the solvent to wash the fluidization push zone (between V54 and V57) and the waste pump tubing.

[0174] Oxidation Reaction (if desired instead of sulfurization): Charge ACN (100 mL) to the amidite activator mixing zone, allowing the resin to be washed immediately after the oxidation reaction is complete. Charge the oxidation solution (59 mL) to the feed zone and flush with nitrogen to clear the feed tube. Force the solution into the reactor and fluidize the resin bed twice to ensure complete contact between the liquid and solids and reset the resin bed. The total time for both fluidizations is approximately 1.2 minutes. Begin pumping the oxidation solution through the resin bed over 30 seconds at a pump setting of 130 mL / min. Simultaneously pump an additional iodine solution (1 mL) into the reactor, allowing it to enter the top of the reactor at approximately the same rate as the pump discharge. Force the iodine solution into the reactor with nitrogen to flush the feed tube. A total of 60 mL of oxidation solution is pumped through the resin bed over 30 seconds. Push any remaining oxidation solution out the bottom of the filter to waste. 100 mL of ACN wash solvent is forced through the reactor (from the amidite activator mixing zone) to wash the resin.

[0175] Wash 1 (do this step twice): Charge ACN (50 mL) to the reactor through the oxidizing solution feed line and sweep the feed tube with nitrogen. Push the solvent through the resin bed to waste.

[0176] Wash 2: ACN (40 mL) solvent is charged to the feed zone through the solvent feed line. Nitrogen is used to drive the wash solvent into the reactor and sweep the feed tubing. The solvent is forced through the resin bed and out of the reactor to waste, while simultaneously using the solvent to wash the fluidization push zone (between V54 and V57) and the waste pump tubing.

[0177] Wash 3: Charge ACN (90 mL) to the feed zone through the solvent feed line. Drive the wash solvent into the reactor with nitrogen, sweeping the feed tube. Force the solvent down through the reactor and into the bottom of the fluidization push zone (between V54 and V57). Fluidize the resin bed to ensure full liquid-to-solid contact and reset the resin bed twice. Begin pumping ACN solvent through the resin bed for 130 seconds at a pump setting of 110 mL / min. Simultaneously pump additional ACN (110 mL) into the reactor, so that it enters the top of the reactor at approximately the same rate as it is being pumped out. A total of 200 mL has been pumped through the resin bed. Push any remaining ACN solvent out the bottom of the filter to waste.

[0178] Wash 4 (do this step twice): Charge ACN (40 mL) into the feed zone through the solvent feed line. Drive the wash solvent into the feed zone with nitrogen, sweeping the feed tubing. Force the solvent through the resin bed and out of the reactor to waste, while simultaneously using the solvent to wash the fluidization push zone (between V54 and V57) and the waste pump tubing.

[0179] Sulfurization (thiolation) reaction (if required instead of oxidation): Charge the hydrogenated xanthan solution (90 mL) to the feed zone and reactor. Drive the solution into the reactor with nitrogen and purge the feed tube. Fluidize the resin bed twice to ensure complete contact between the liquid and solids and reset the resin bed. The total time for both fluidizations is approximately 1 minute. Begin pumping the hydrogenated xanthan solution through the resin bed over 80 seconds at a pump setting of 130 mL / min. Simultaneously pump additional hydrogenated xanthan solution (60 mL) into the reactor, so that it enters the top of the reactor at approximately the same rate as it is being pumped out. Drive the hydrogenated xanthan solution into the reactor with nitrogen and purge the feed tube. A total of 150 mL is pumped through the resin bed over 80 seconds. Push the remaining hydrogenated xanthan solution out the bottom of the filter to waste.

[0180] Wash 1 (do this step twice): Charge ACN (50 mL) to the reactor through the hydrogenated xanthan solution feed line. Drive the wash solvent into the reactor with nitrogen, sweeping the feed tube. Push the solvent through the resin bed to waste.

[0181] Wash 2: ACN (40 mL) is charged to the reactor through the solvent feed line. Nitrogen is used to drive the wash solvent into the reactor and sweep the feed tubing. The solvent is forced through the resin bed and out of the reactor to waste, while simultaneously using the solvent to wash the fluidization push zone (between V54 and V57) and the waste pump tubing.

[0182] Wash 3: Charge ACN (90 mL) to the reactor through the solvent feed line and sweep with nitrogen to clear the feed tube. Force the solvent down through the reactor and into the bottom of the fluidization push zone (between V54 and V57). Fluidize the resin bed to ensure complete liquid-to-solid contact and reset the resin bed twice. Begin pumping ACN solvent through the resin bed at a pump setting of 110 mL / min for 130 seconds. Simultaneously pump additional ACN (110 mL) into the reactor, allowing it to enter the top of the reactor at approximately the same rate as it is being pumped out. A total of 200 mL of ACN has been pumped through the resin bed. Push any remaining solvent out the bottom of the filter to waste.

[0183] Wash 4 (do this step twice): Charge ACN (40 mL) to the reactor through the solvent feed line. Drive the wash solvent into the reactor with nitrogen, sweeping the feed tubing. Force the solvent through the resin bed and out of the reactor to waste, while simultaneously using the solvent to wash the fluidization push zone (between V54 and V57) and the waste pump tubing.

[0184] Capping Reaction: Charge the reactor with Capping Solution A and Capping Solution B (45 mL each), driving each solution through the reactor with nitrogen and purging the feed tubing. Fluidize the resin bed twice to ensure complete liquid-to-solid contact and reset the bed. The total time for both fluidizations is approximately 1 minute. Begin pumping the Capping Solution A and Capping Solution B mixture through the resin bed over 70 seconds at a pump setting of 130 mL / min. Simultaneously pump additional Capping Solution A and Capping Solution B (25 mL each) into the reactor, allowing them to enter the top of the reactor at approximately the same rate as they are being pumped out. Drive Capping Solution A and Capping Solution B into the reactor with nitrogen and purging the feed tubing. A total of 140 mL is pumped through the resin bed in 70 seconds. Push any remaining Capping Solution A and Capping Solution B out the bottom of the filter and discard.

[0185] Wash 1 (do this step twice): Charge ACN (50 mL each) to the reactor through the Capping Solution A and Capping Solution B feed lines and sweep the feed tubes with nitrogen. Push the solvent through the resin bed to waste.

[0186] Wash 2 (do this step twice): Charge ACN (40 mL) to the reactor through the solvent feed line. Drive the wash solvent into the reactor with nitrogen, sweeping the feed tubing. Force the solvent through the resin bed and out of the reactor to waste, while simultaneously using the solvent to wash the fluidization push zone (between V54 and V57) and the waste pump tubing.

[0187] After the final coupling cycle is complete, the cycle is repeated using phosphorylation solution instead of amidite. After the phosphorylation reagent is coupled and oxidized, the deblocking step is repeated. The resin is reacted with 500 mL of DEA solution for 10 minutes. Wash with ACN and dry by blowing nitrogen downward through the resin bed for 30 minutes to yield 11.96 g of dry product on the resin. 20 mg of product plus resin is sampled, using 3.344 g of resin initially, leaving 8.636 g (94% crude mass yield) of protected oligonucleotide product.

[0188] A small sample was subjected to cleavage and deprotection reactions using concentrated NH4OH solution at 50°C for 4 hours. UPLC analysis showed that the cleaved and deprotected oligonucleotide product was 77.8% pure by peak area percent, as shown in the table of UPLC results for Examples 1-5 (Table 13). LCMS analysis confirmed that the major product peak represented the correct HPRT div22 AS chain.

[0189] Example 4 - Preparation of HPRT Div22 antisense strand As in Examples 1, 2, and 3, the same HPRT Div22 antisense strand was produced in this example. As in Examples 1, 2, and 3, the synthesis of this molecule using the fluidized-bed method of the present invention is described herein, which includes deblocking, coupling, oxidation (or sulfurization), and capping steps to sequentially introduce the remaining phosphoramidites. The same reactor and procedure were used in this experiment as in Example 2. The main differences were the smaller scale (0.1 mmol scale vs. 0.18 mmol scale) and higher resin loading (299 umol / g vs. 246 umol / g), and therefore the resin bed height was not as high as in Example 2, and the shorter resin bed allowed for shorter deblocking and post-deblocking wash times. In this experiment, 0.3346 g of Nittophase HL 2'OMeU300 resin (Lot # EO5005, 299 umol / g loading) was used. While 2.5 equivalents of amidite (cycles 17, 18, and 21) and 12.5 equivalents of activator were used for the mU17, fA18, and fU22 couplings, 2.0 equivalents of amidite and 10 equivalents of activator were used for all remaining amidite coupling steps, as in Example 2. Similar to Examples 1, 2, and 3, large excesses of wash solvent and DCA reagent solution were used in Example 4. See Example 7 for an example using a reduced amount of DCA reagent, and see Examples 6, 8, and 9 for an example using a reduced ACN wash. Table 31 provides a guide to various embodiments of the fluidized-bed reactor examples.

[0190] Use the reagent solutions listed in Table 7. [Table 7]

[0191] Prepare a 0.1 M amidite solution as shown in Table 8 below. Weigh the amidite solid into a bottle, insert a dry pad, then add ACN to achieve a concentration of 0.1 M. The synthesis conditions for Example 4 are listed in Table 9. The synthesis procedure is the same as that described in Example 2, see also Figure 3. [Table 8] [Table 9]

[0192] The resin bed height during the downflow portion of the final deblocking step was 12.7 cm, thus the resin bed height was less than half compared to Example 2. The maximum pressure drop across the resin bed during the experiment was 15 psig, as this was the pressure of the nitrogen feed used to force the liquid through the resin bed. The resin bed height from the start to the end of the synthesis is shown in Table 10. [Table 10]

[0193] After the final DEA treatment, the resin is washed with ACN and dried by blowing nitrogen downward through the resin bed to yield 1.1458 grams of dry resin. This corresponds to a weight gain of 0.811 grams, not including correction for the sample mass, which was approximately 6% of the total material. The crude mass gain is 8.60 g / mmol, representing a crude mass yield of approximately 93%. The cleavage and deprotection reaction is carried out at 55°C for 4 hours using 0.5 mL of concentrated NH4OH solution on 21.9 mg of sample. UPLC indicates that the cleaved and deprotected oligonucleotide product is 84.5% pure by peak area percent, as shown in the table of UPLC results for Examples 1-5 (Table 13). LCMS analysis confirms that the major product peak represents the correct HPRT div22 AS chain.

[0194] Example 5 - Comparison with Cytiva AKTA Oligonucleotide Synthesizer The same HPRT Div22 antisense strand from Examples 1-4 was also prepared on a Cytiva AKTA automated oligonucleotide synthesizer starting with NittoPhase HL 2'OMeU250 resin (246 μmol / g, 0.603 g, 148.4 μmol) and subjected to the same sequential deblocking, coupling, oxidation (or sulfurization if there is a P=S bond in the sequence), and capping steps. The synthesis conditions for Example 5 are listed in Table 11. [Table 11-1] [Table 11-2]

[0195] After the synthesis was completed and the resin was dried, the mass gain was measured at 8.64 g / mmol, representing a crude mass gain of 94%. The crude yield was also measured at 169 OD / umol. Upon cleavage and deprotection, the oligonucleotide product was 80.28% pure by UPLC. LCMS analysis confirmed that the major product peak represented the correct HPRT div22 AS chain. As shown in Table 13, the crude yield and purity of the oligonucleotide product were comparable to those obtained in Examples 1–4. However, in the Cytiva AKTA system, the resin bed is static, and all reagents and solvents pass through the resin bed in a “plug flow” fashion. A limitation of this system is that resin bed heights greater than 10 cm can cause adverse effects, such as increased pressure drop across the resin bed and channeling within the resin bed. In Example 5, the resin bed height was a maximum of 2 cm. In contrast, the present invention allows for higher resin bed heights (the maximum bed height during the experiment described in Example 2 was 30 cm during the downward push portion of the reaction), increasing the batch size capacity for a given reactor diameter and facilitating flexible batch sizes for a given reactor.

[0196] The ion-pairing UPLC method conditions for the purity analysis of HPRT div22 antisense strand are outlined in Table 12. [Table 12] [Table 13-1] [Table 13-2] [Table 13-3]

[0197] Example 6.5 1.5 umol ANGPTL3 antisense strand, integrating wash solvents per cycle The ANGPTL3 antisense strand was prepared using the fluidized-bed method of the present invention, which includes deblocking, coupling, oxidation (or sulfurization), and capping steps for sequentially introducing phosphoramidites. Cycle 21: Capping is not required after adding the MeMOP phosphoramidite. In this example, an alternative research-scale synthesizer design was used. This new design does not have feed zones for reagents other than the amidite and activator. The use of multiple heads in parallel reduces the number of pumps, and solvent washouts are reduced by integrating solvent recycling after each phosphoramidite cycle. This experiment represents a baseline synthesis of a 22-mer single-stranded RNA (ANGPTL3 antisense strand) at a 51.5 μmole scale. The sequence of this RNA strand is shown in Figure 10 and can be abbreviated as follows, where * indicates thiolation instead of oxidation: 5'MeMOP*fG*fU*fAfUmA fAmCmC fUmUmC mCfAmUmUmUmUmGmA*mG*mG3

[0198] PFA tubing with an inner diameter of 0.63 cm and a height of 10 cm was used as the reactor. The synthesis conditions are summarized in Table 14. 0.2087 g of NittoPhase HL 250 2'OMeG(IBU) resin was used, yielding a final resin mass of 0.6287 g. Therefore, the crude mass gain was 0.4200 g. Normalization to a 1 mmol scale resulted in 8.16 g / mmol. UPLC results indicated an FLP of 82.7%. UPLC and yield results are shown in Table 17. The overall OD normalized to the synthesis scale was 171 OD / umol. Table 15 lists the solvent usage compared to a typical synthesis on a Cytiva AKTA OP100 synthesizer. Compared to Examples 1-4, Example 6 required significantly less ACN wash solvent per mmol. Table 31 provides a guide to various fluidized-bed reactor examples. A comparison of purity, yield, and impurity profile between the Cart 314 fluidized bed reactor synthesis and the AKTA synthesis system is shown in Table 17. A schematic of the synthesis system is shown in Figure 7. The synthesis system automation procedure is listed in Table 18. The maximum pressure drop across the resin bed during the experiment is 15 psig, as this is the pressure of the nitrogen feed used to force the liquid through the resin bed. [Table 14] [Table 15]

[0199] Four experiments were performed on an AKTA OP100 synthesizer for comparison and are presented in Table 17.

[0200] The materials and synthesis conditions used for the four experiments carried out on the AKTA OP100 are listed in Table 16.

[0201] All four experiments in Table 16 used the following: ●Cytiva AKTA OP100 Kinnovate NittoPhase HL-2'-OMeG(iBu)250 resin, lot G08004, loading 247 umol / g Deblocking, dichloroacetic acid (DCA) in toluene Coupling, oxidation, thiolation, and capping followed by 2CV ACN injection Amidite equivalent = 2 equivalents Activator reagent, 0.5M 5-(ethylthio)-1H-tetrazole in ACN Coupling time = 10 minutes Oxidizing reagent: 0.05M iodine in pyridine / water solution (90 / 10 (v / v)) ●Oxidation equivalent = 4 equivalents Oxidation contact time = 3 minutes Sulfurization reagent, 0.2M hydrogenated xanthan in ACN / pyridine (70 / 30 (v / v)) ●Hydrogenated xanthan gum used, 2CV Hydrogenated xanthan gum = 13 equivalents ●Sulfurization contact time = 5.5 minutes Capping solution A: 1-methylimidazole / ACN (20 / 80 (v / v)) Capping solution B: 1:1 mixture of B1 and B2 Capping solution B1: 40% acetic anhydride in ACN Capping solution B2: 60% by volume of 2,6-lutidine in ACN ●Total volume of capping solution A and capping solution B in a 50 / 50 (v / v) mixture = 2 CV Capping contact time = 0.5 minutes DEA, 20% diethylamine / ACN solution (20 / 80 (v / v)) DEA contact time = 10 minutes ●DEA amount = 10mL [Table 16]

[0202] For each experiment listed in Table 16, after drying the resin-bound oligonucleotide, approximately 20 mg of resin was suspended in 0.50 mL of NH4OH and shaken at 55°C for 4 hours or at 38°C for 18 hours. The resin was filtered and the filtrate was analyzed by UPLC (50 μL of the filtrate was diluted with 1.5 mL of water). The UPLC purity is shown in Table 17. [Table 17-1] [Table 17-2] [Table 17-3] [Table 17-4] [Table 17-5] [Table 17-6]

[0203] While the area percent peaks identified in the table in the 0.979-0.984 RRT and 1.006-1.016 RRT regions might suggest differences between the synthesizers, a review of the chromatograms reveals that there are in fact no distinct peaks in these regions. For example, a peak not present in the fluidized-bed reactor example is identified at 0.979 RRT in the AKTA example. Similarly, a peak not present in the AKTA example is identified at 0.984 RRT in the fluidized-bed reactor example. However, examination of the chromatograms in Figure 17 reveals that these are similar left-hand shoulders on the main peak, and the identified peak times simply depend on where the auto-integration line is drawn. Similarly, there is an elevated region above the baseline between 1.006 RRT and 1.016 RRT, which is assigned differently depending on where the auto-integration line is drawn, but this is the same elevated baseline region in all samples. From this table, it may be thought that a peak absent from the fluidized bed reactor sample is present at 1.006 RRT in the AKTA sample, and a peak absent from the AKTA sample is present at 1.01 RRT in the fluidized bed reactor sample, however, inspection of the chromatograms in Figure 17 shows that there is in fact no significant difference between samples from different synthesis equipment in these RRT regions. [Table 18-1] [Table 18-2] [Table 18-3] [Table 18-4] [Table 18-5] [Table 18-6] [Table 18-7]

[0204] Example 7. 100 umol ANGPTL3 antisense strand, DCA reagent integrated per cycle The same ANGPTL3 antisense strand shown in Figure 10 was prepared using the fluidized-bed method of the present invention, which includes deblocking, coupling, oxidation (or sulfurization), and capping steps to sequentially introduce phosphoramidites. Example 7 demonstrated the lowest DCA reagent use of all examples. Capping was not necessary after adding the Cycle 21 MeMOP phosphoramidite. This example used fewer equivalents of DCA compared to Examples 1, 2, 3, 4, and 6. Table 31 provides a guide to various embodiments of the fluidized-bed reactor examples. One factor contributing to DCA reduction is the recycling of a clean portion of the acid effluent from the previous phosphoramidite cycle in each phosphoramidite cycle. The recycled acid accomplishes some of the deblocking. Additionally, and perhaps more importantly, the recycled acid washes away residual ACN in / on the wet beads from the end of the previous cycle, causing the resin beads to swell during fluidization. ACN is known to interfere with DCA deblocking. The previous embodiment achieved initial resin fluidization and ACN drainage with fresh DCA solution. The concept of this example is to avoid the need for fresh acid during this stage of the procedure by using recycled acid instead. Because the recycled acid is a cleaner portion of the acid effluent from the previous phosphoramidite cycle, it does not contain ACN. Expanding the resin beads during initial fluidization allows the resin bed to expand and expand during fluidization, thereby reducing the subsequent pressure drop during the downflow portion of the subsequent deblocking step. The maximum pressure drop across the resin bed during the experiment is 15 psig, since this pressure is that of the nitrogen feed used to force the liquid through the resin bed. Another difference compared to the previous example is the reduction in the capping solution volume: Capping Solution A and Capping Solution B were each reduced by 50% to 1.75 mL. A third difference is the elimination of the extensive fluidization wash between reactions. The process in this example uses 29.2% of the standard amount of acid (mL acid / mmol starting resin) typically used with the Cytiva AKTA. See Figure 8 for the synthesis apparatus setup for acid recycling.

[0205] Acid recycling is achieved using a separate acid recycling feed bottle and pump. Each synthesis cycle for introducing phosphoramidites has two acid deblocking steps. In the first acid step of each cycle, acid is charged to the reactor from the acid recycling feed bottle. This is primarily used to expand the bed during flow and subsequently drain the ACN from the bed, allowing for more rapid pumping through the resin bed. Upon exiting the reactor, the acid from the first acid step is pumped to waste. In the second acid step, fresh acid is charged from the fresh acid feed bottle. Upon exiting the reactor, the acid from the second acid step is pumped back into the acid recycling feed bottle for reuse in the next cycle. In the first cycle of this sequence, the recycling acid feed bottle is charged with enough fresh acid to be used in the first acid step. In all subsequent steps, the acid from the previous cycle in the recycling acid feed bottle is recycled. The pump parameters are set so that all recycled acid from the previous cycle is charged to the reactor. Emptying the reusable bottle each time limits carryover to only one cycle and prevents buildup in the reusable acid supply bottle.

[0206] The process for this example was run on a 100 μmol scale, with the resin bed reaching 11 cm in height with ACN solvent wet at the start of the final cycle. During the downflow portion of the final deblocking step, the maximum resin bed height reached 12 cm. The maximum pressure drop across the resin bed during the experiment was 15 psig. The reactor had a bottom section 23 cm high and 0.63 cm in diameter and a conical-bottom top section 25 cm high and 4.7 cm in diameter. Reagent solutions were prepared as in Example 6 (hydrogenated xanthan gum concentration was 0.2 M). All pumps and feed lines were primed. Dry packs were placed in the ACN bottles and all syringes. Syringe pumps were used for the amidite and activator, and peristaltic pumps and feed vessels were used for all other reagent and solvent feeds.

[0207] The synthesis begins with coupling mG (referred to herein as "mG resin") to NittoPhase HL 2'OMeG(ibu)250 resin (Lot No. H08023) using known methods. Overall synthesis conditions are shown in Table 19. [Table 19-1] [Table 19-2]

[0208] The final resin-bound oligonucleotide mass was 1.220 grams on a dry basis, corresponding to a weight gain of 0.817 grams or a mass gain of 8.13 g / mmol. The cleavage and deprotection reaction was carried out at 55°C for 5 hours using concentrated NHOH solution. UPLC indicated that the cleaved and deprotected oligonucleotide product was 79.92% pure by peak area percent, as shown in the UPLC results tables in Examples 6-9.

[0209] We now describe a process using the apparatus shown in Figure 8. Each addition of a phosphoramidite during synthesis is followed sequentially by a deblocking step, a coupling step, an oxidation step (or a sulfurization step if there is a P=S bond in the sequence), and a capping step, as described below. When the procedure describes pumping liquid downward through the resin bed, this means that a waste pump at the bottom outlet of the reactor operates at a target setpoint while nitrogen pressure pushes against the top of the resin bed, forcing the liquid downward. The purpose of the peristaltic pump is to meter the liquid through the bed at a controlled rate.

[0210] Deblocking Reaction: The deblocking process includes a recycled deblocking step and a fresh deblocking step. The increased volume and time as a function of oligonucleotide length are listed in Table 20. [Table 20]

[0211] Recycle Deblocking: Turn valve 808 to A, turn valve 807 to B, and close valve 824. An initial amount of recycled deblocking solution (5 mL for cycles 1-3, 6 mL for cycles 4-9, and 7 mL for cycles 10-21) is charged to the acid feed zone and then pushed into the reactor with nitrogen pressure for 6 seconds. The amount of recycled deblocking solution in each cycle matches the amount of fresh deblocking solution in the previous cycle. Open valve 824. The waste outlet valves (valves 809 and 810 in Figure 8) are closed. Vent pressure from the top of the reactor for 15 seconds while simultaneously opening valve 38 and bubbling nitrogen through to agitate and fluidize the resin bed with the reagent solution. Close valves 854 and 838, open valve 844, and push down with nitrogen for 3 seconds. Close valve 844, and repeat these fluidization steps four more times. Most of the resin expansion occurs during fluidization. Open the waste valve (valve 810) and pump the deblocking solution through the resin bed with pump 9 at a rate of 10 mL / min. While pump 9 serves as a metering device to set the exit rate, nitrogen pressure at the top of the reactor provides the driving force for the liquid to flow downward and out the bottom of the reactor. In parallel with pumping with pump 9, open valve 837 and start pump 10, feeding recycled deblocking solution at 30 mL / min until a second amount of 2–11 mL (depending on the cycle) is pumped (pump 10 finishes before pump 9). The liquid pumped from pump 10 into the acid feed zone simultaneously flows into the reactor, maintaining the liquid level above the resin bed and maintaining flow for the entire duration of pump 9's operation. Run pump 9 continuously for 110–160 seconds (increasing throughout the experiment) to ensure no recycled acid remains in the tubing between the reactor and valve 839. This allows fresh acid to flow through the column and clear the waste tube before being pumped back into the recycled acid bottle for the subsequent fresh acid deblocking step.

[0212] Fresh deblocking: Turn valve 808 to A, turn valve 807 to B, and close valve 824. Charge 5-7 mL of fresh deblocking solution into the acid feed zone and push it into the reactor with nitrogen pressure for 6 seconds. Open valve 824. The waste outlet valves (valves 809 and 810) are closed. Open the waste valve (valve 810) and pump deblocking solution through the resin bed at 4 mL / min with pump 9. In parallel with pump 9, open valve 814, turn on pump 1, and feed deblocking solution at 30 mL / min until 2-11 mL have been pumped (pump 1 finishes before pump 9). The liquid pumped from pump 1 into the acid feed zone simultaneously flows into the reactor, maintaining the liquid level above the resin bed and maintaining flow for the entire duration of pump 9's operation. Additional pumping time is used to purge all lines from the reactor to the recycle acid supply bottle, allowing the full amount of fresh acid to be used in the subsequent recycle acid step of the next cycle.

[0213] The following ACN wash procedure is performed four times: open waste valve 809 and charge ACN (4 mL) to the feed zone, then close valve 809 and allow nitrogen pressure to push into the reactor for 2 seconds. Open the waste valve (valve 810) and pump the ACN wash through the resin bed with pump 9 at a rate of 10 mL / min until the wash has passed through the resin. This is a significant reduction in ACN wash solvent compared to Examples 1, 2, and 4, which use the same synthesis equipment. During the first cycle only, an additional fluidized wash is incorporated between the plug flow wash steps.

[0214] Coupling Reaction: After deblocking washes, the next successive phosphoramidite is coupled, introduced in sequential steps from 3' to 5'. For each phosphoramidite coupled in the sequence, the coupling reaction procedure is performed essentially as described below, using the amidite solution corresponding to the nucleotide in the sequence. Valve 808 is turned to position B. Pre-wash the amidite zone and the flow path to the reactor twice, each time by pumping 4 mL of ACN into the amidite feed zone with valve 809 closed. Then, open valve 810 and pump to waste at a rate of 10 mL / min for 15 seconds. Pump first the activator solution (2 mL, 10 equivalents, Table 1), followed by the appropriate amidite solution (2 mL, 2.0 equivalents) into the feed zone. Close valve 809, and open valves 855, 824, and 838 for 3 seconds to mix the amidite and activator solutions with bubbling nitrogen. Valves 855, 824, and 838 are closed. Valve 845 is opened to push the mixture in the feed zone into the reactor with nitrogen pressure for 6 seconds, after which valve 824 is opened and nitrogen pressure is maintained for 8 seconds. The amidite solution and activator solution are mixed with the resin, and with valve 824 open and valve 809 closed, the bed is continuously fluidized as follows: nitrogen pressure is applied to the top of the reactor for 3 seconds. Pressure is vented from the top of the reactor for 15 seconds, while valve 838 is opened and nitrogen is bubbled through to agitate and fluidize the resin bed with the reagent solution. This process is repeated continuously for 10 minutes (15 minutes for the fluoroamidite and mA in cycle 16), after which valve 809 is opened and nitrogen pressure is applied to the top of the reactor for 8 seconds, and the liquid is drained from the bottom of the reactor to waste. ACN (4 mL) is pumped into the amidite feed zone and forced through the reactor with nitrogen pressure for 15 seconds. This wash is repeated once. This is a significant reduction in the ACN wash solvent compared to Examples 1, 2, and 4, which use the same synthesis equipment. Additionally, larger versions of the reactor (Examples 8, 9, and 10) use a spray ball and require even less ACN (mL / mmol).

[0215] Oxidation Reaction (If Required Instead of Sulfurization): After the coupling reaction wash, the oxidation reaction is performed essentially as described below. Valves 806, 807, and 808 are turned to A, and valves 824, 809, and 853 are opened. The oxidation solution (Table 19, 4.2 mL for cycles 3 and 4, and 5.3 mL for cycles 5-18) is pumped into the feed zone, valve 809 is closed, and the iodine solution is forced into the reactor with nitrogen pressure for 10 seconds. The reactor bed is fluidized 11 times as follows: The top of the reactor is pressurized with nitrogen pressure for 5 seconds. The pressure is vented from the top of the reactor for 15 seconds while valve 838 is opened and nitrogen is bubbled through to agitate and fluidize the resin bed with the reagent solution. Valve 810 is opened, and pump 9 pumps 5.3 mL of solution over 30 seconds, after which valve 810 is closed. Valves 843 and 809 are opened to force any remaining reagents out of the reactor. Valve 843 is closed and valve 853 is open.

[0216] The following ACN wash procedure is performed four times. This is a significant reduction in the ACN wash solvent compared to Examples 1, 2, and 4, which use the same synthesis equipment. Open waste valve 809 and charge ACN (4 mL) into the iodine feed zone, then close valve 809 and use nitrogen pressure to push into the reactor for 2 seconds. Open the waste valve (valve 810) and use pump 9 to pump the ACN wash through the resin bed at a rate of 10 mL / min until the wash has passed through the resin.

[0217] Sulfurization (thiolation) reaction (if required instead of oxidation): After the coupling reaction wash, the thiolation reaction is carried out essentially as described below. Valve 806 is turned to B, valves 805, 807, and 808 are turned to A, and valves 824 and 809 are opened. The sulfurization solution (Table 19, 6.5 mL) is pumped into the feed zone, valve 809 is closed, and nitrogen pressure is forced into the reactor for 6 seconds. The reactor bed is fluidized 22 times as follows: The top of the reactor is pressurized with nitrogen pressure for 3 seconds. Pressure is vented from the top of the reactor for 15 seconds while valve 838 is opened and nitrogen is bubbled through to agitate and fluidize the resin bed with the reagent solution. Valve 810 is opened, and pump 9 pumps 6.5 mL of liquid over 30 seconds, after which valve 810 is closed. Washing is carried out as described above for the oxidation reaction, except that ACN enters the reactor through the sulfurization feed zone.

[0218] Capping Reaction: Oxidation (or Sulfurization) Reaction After cleaning, the capping reaction is carried out essentially as described below. Valves 805 and 806 are turned to B, and valves 807 and 808 are turned to A. Valves 824 and 809 are opened. Capping solution A and capping solution B (1.75 mL each) are simultaneously pumped into the feed zone, after which valve 809 is closed. The liquid is forced into the reactor with nitrogen pressure for 6 seconds. The reactor bed is fluidized three times as follows: The top of the reactor is pressurized with nitrogen pressure for 5 seconds. The pressure is vented from the top of the reactor for 15 seconds while valve 838 is opened and nitrogen is bubbled through to agitate and fluidize the resin bed with the reagent solution. Valve 810 is opened and pump 9 pumps 3.5 mL of liquid over 35 seconds, after which valve 810 is closed. Cleaning is carried out as described above for the oxidation reaction, except that ACN enters the reactor through the capping feed zone.

[0219] Perform one ACN wash procedure as follows: Open waste valve 809 and fill the capping feed zone with ACN (4 mL), then close valve 809 and allow nitrogen pressure to push into the reactor for 2 seconds. Open the waste valve (valve 810) and use pump 9 to pump the ACN wash through the resin bed at a rate of 3 mL / min until the wash has passed through the resin.

[0220] The very last cycle uses a sulfurization step, followed by an ACN wash. The resin is then washed with a DEA solution (20% (V / V) ACN solution) for 10 minutes. Valve 808 is set to the open position (B), and valves 815 and 824 are opened. Pump 4 is used to load 9.3 mL of DEA into the feed zone. Fluidize four times, then valve 810 is opened, pump 9 is turned on for 60 seconds, and the DEA is pumped to waste at a rate of 8 mL / min. These steps are repeated three more times for a total of four DEA washes. Three 4 mL ACN washes are performed through the amidite feed zone, followed by two 12 mL ACN fluidization washes.

[0221] Resin bed heights throughout this experiment are shown in Table 21. [Table 21]

[0222] The resin bed was dried for 120 minutes by blowing nitrogen downward through the resin bed. The final mass was 1.220 grams of dried resin. This corresponds to a weight gain of 0.817 grams or 8.13 g / mmol. The OD / umol is listed in Table 17. Cleavage and deprotection were performed on 22.8 mg of dried oligonucleotide-bound resin samples. To do this, the resin was added to a UPLC vial along with 0.5 mL of ammonium hydroxide. The vial was placed on a shaker for cleavage and deprotection (55°C for 5 hours). After the allotted time, the vial was removed and allowed to cool to room temperature. The sample was placed in a 1.5 mL centrifuge tube containing a filter basket and centrifuged for 30 seconds to 1 minute. 1.5 mL of Milli-Q water was added to a separate UPLC vial, followed by 50 μL of sample liquid (containing the oligonucleotide). The spent resin was discarded. The UPLC tube was repeatedly inverted to mix the sample before running it on UPLC. UPLC analysis showed that the cleaved and deprotected oligonucleotide product was 79.92% pure by peak area percent, as shown in the UPLC results for Examples 6-10 and a comparison with Cytiva AKTA (Table 17). LCMS analysis confirmed that the major product peak represented the correct strand. Only one significant cleavage (greater than 1% area) was observed, with a 2.02% area peak corresponding to incomplete coupling of the final amidite MeMOP. Due to the high purity and yield, these results demonstrate that the ACN wash followed by acid recycling is an acceptable post-capping wash method, and that removal of the fluidized wash is not detrimental compared to Examples 2 and 4. The total fresh acid used was 263 mL, calculated from the before and after feed bottle masses, or approximately 2630 mL / mmol. This is 29.2% of the typical amount for Cytiva AKTA (typically approximately 9000 mL / mmol).

[0223] Example 8 - Preparation of AngPTL3 antisense strand at 10 mmol scale in a 4 inch internal diameter reactor The single antisense strand of AngPTL3 was synthesized in a fluidized bed reactor on a pilot scale (this is the same sequence as shown in Figure 10). 5'MeMOP*fG*fU*fAfUmA fAmCmC fUmUmC mCfAmUmUmUmUmGmA*mG*mG3'

[0224] The synthesis of this molecule using the fluidized-bed method of the present invention is described herein, and includes a deblocking step, a coupling step, an oxidation step (or sulfurization step), and a capping step to sequentially introduce the remaining phosphoramidites. One key difference in this example is that it is performed on a larger scale (10 mmol) in a larger fluidized-bed reactor with the same diameter from bottom to top, an internal diameter of 10.16 cm, and a height of 61 cm. The reactor has a flat bottom with a filter frit. This larger-diameter reactor allows for sufficient fluidization without the need for a wide funnel zone at the top. The larger reactor diameter reduces wall effects, making it easier to fully fluidize and redistribute solids and liquids without a wide-diameter section at the top. Fluidization at the start of each reaction step typically only extends the height of the slurry in the reactor by approximately 2–4 cm. The apparatus for this synthesis is shown in Figure 9. Another difference in this example is the use of toluene for washes prior to deblocking. Also, similar to Example 6, Example 8 integrates solvent recycle with each phosphoramidite cycle, reducing solvent wash volumes. The cleaner post-deblocking wash is pumped into the recycle ACN vessel and used for the first part of the post-deblocking wash in the next phosphoramidite cycle. Table 31 is a guide to various embodiments of the fluidized bed reactor examples.

[0225] By the end of the experiment, the resin bed height reaches 5 cm wet with ACN solvent and 4 cm dry. During the downflow portion of the final deblocking step, the maximum resin bed height reaches 6 cm. The maximum pressure drop across the resin bed during the experiment is 15 psig, since this is the pressure of the nitrogen feed used to force the liquid through the resin bed. As in Examples 6 and 7, two equivalents of amidite are used in the coupling. The overall synthesis conditions are shown in Table 22. [Table 22]

[0226] Begin by coupling mG (referred to herein as "mG resin") to NittoPhase HL 2'OMeG(iBu)250 resin (247 μmol / g) using known methods; see Figure 9 for the synthesis apparatus setup. 40.50 g (10.0 mmol) of mG resin is slurried using ACN and placed in a 10.16 cm internal diameter reactor fitted with a 40 micron sintered mesh filter frit at the bottom. The initial resin depth is approximately 1 cm in height.

[0227] Prepare the reagent solution as follows:

[0228] All amidite solutions were prepared in ACN, Fisher Lot No. 212215. Dissolve the amidite in the ACN solvent as follows: Mix until dissolved. Add the molecular sieve dry pack to the sealed bottle. [Table 23] Cap B1: Acetic Anhydride: Macron Fine Chemicals Lot Number 0000239131 Acetonitrile: Fisher Lot No. 206496 1657 mL of acetic anhydride and 2486 mL of ACN were charged to the supply vessel. Cap B2: 2,6-Lutidine: Acros Lot Number A0428332 Acetonitrile: Fisher Lot No. 206496 2486 mL of lutidine and 1657 mL of ACN were charged to the supply vessel. Cap A: 1-Methylimidazole: Acros Lot No. A0425789 ACN:Fisher Lot No. 206496 1657 mL of imidazole was charged. 6628 mL of ACN was charged. 0.2M Hydrogenated Xanthan Sulfide Solution: Hydrogenated Xanthan Gum: TCI Lot Number QLXKC-RI Pyridine: Fishers Lot No. 208059 Charge 3775 mL of pyridine. Charge 114 g of XH to the pyridine bottle. Mix until dissolved. Oxidizing solution: Iodine solution (0.05M) Honeywell Lot Number EA702-US 10 Kg of barrel stock was filled into supply cans. Activator solution: Honeywell Lot Number EA952-US 5 Kg of barrel stock solution was filled into a supply can. 3% DCA in toluene: DCA: Supelco Lot Number 61069116 Toluene: Superior Lot Number FH11313266 Lot Number 1 1.18,201 mL of toluene was charged to a carboy. 2.563 mL of DCA was charged to the carboy. 3. The carboy was filled into the supply can. Lot No. 2 1.18,428 mL of toluene was charged to a carboy. 2.570 mL of DCA was charged to the carboy. 3. The carboy was filled into the supply can. Lot No. 3 1.18,483 mL of toluene was charged to a carboy. 2.572 mL of DCA was charged to the carboy. 3. The carboy was filled into the supply can. Lot No. 4 1.18,575 mL of toluene was charged to a carboy. 2.575 mL of DCA was charged to the carboy. 3. The carboy was filled into the supply can.

[0229] All pumps and feed lines were primed. ACN was passed through a bed of molecular sieves on its way to the inactivation feed canister. ACN, toluene, and a toluene solution of DCA were fed from feed canisters by pressure and controlled by automatic flow control valves. Peristaltic pumps and feed vessels were used for all other feeds. The amidite solution was contained in a separate feed vessel labeled "AM.1 L" and connected to a peristaltic pump attached to valves V901A-V908A in Figure 9. The MeMOP phosphoramidite was used in one of the AM feed vessels. The activator solution and DEA solution were contained in feed vessels labeled "Activator 5 gal" and "DEA" in Figure 9, respectively.

[0230] Each time a phosphoramidite is added during synthesis, the steps are sequentially followed: deblocking, coupling, oxidation (or sulfurization, if there is a P=S bond in the sequence), and capping, as described below. Capping is not required after the addition of the Cycle 21 MeMOP phosphoramidite.

[0231] Fluidization continued on an on / off cycle throughout the coupling, oxidation, sulfidation, and capping reactions for the majority of the specified reaction time. Note that fluidization does not have to be on / off cycled. Fluidization can be constant bubbling without pushing up and down. This procedure is a carryover from research-scale experiments. At research scale, in small diameter reactors, there is some advantage to pushing up and down during fluidization because it helps wet and fluidize all the resin beads initially, but there is no further advantage to pushing up and down thereafter. In larger diameter reactors, such as this 4-inch diameter reactor, on / off pushing up and down is not necessary.

[0232] During the deprotection reaction, the resin bed is first fluidized with 200–250 mL of DCA solution, and then the deprotection reaction continues without fluidization until the end in plug flow mode. Fluidization is achieved by simultaneously opening either valve 956 or valve 958 (V956, V958) and valve 953, allowing nitrogen gas to flow upward through the bottom filter screen. When this procedure describes pumping liquid downward through the resin bed, this means that the waste pump at the bottom outlet of the reactor is operating at a target setpoint while nitrogen pressure pushes against the top of the resin bed, forcing the liquid downward. The purpose of the peristaltic pump is to meter the liquid through the bed at a controlled rate.

[0233] Toluene Wash: Charge toluene (300 mL) to the feed zone. Drive the toluene into the feed zone with nitrogen and purge the feed tube. Push the toluene into the reactor. Fluidize the resin bed four times for 2 seconds each to ensure complete liquid-to-solid contact, expand the resin beads, and reset the resin bed. Start the waste pump and apply nitrogen pressure to the top of the feed zone at valve 951, which causes the toluene to begin flowing downward through the resin bed and out the bottom of the reactor. Charge an additional 200 mL of toluene to the feed zone, allowing it to flow into the top of the reactor as it is pumped out the bottom. Set the outlet pump speed so that it takes 2 minutes to pump out 500 mL of toluene. Push any remaining wash solvent out the bottom of the filter to waste.

[0234] Deblocking Reaction: Charge the deblocking solution (Table 22, 200 mL) into the feed zone. Drive the deblocking solution into the feed zone with nitrogen to clear the feed tubing. Push the deblocking solution into the reactor. Fluidize the resin bed for 7 seconds to ensure complete liquid-to-solid contact and reset the bed. Start the waste pump and apply nitrogen pressure to the top of the feed zone through valve 951, causing the DCA solution to begin flowing downward through the resin bed and out the bottom of the reactor. Simultaneously, charge additional DCA solution into the feed zone, allowing it to enter the top of the reactor as it is pumped out the bottom. Pumping of the deblocking solution begins 5–30 seconds before the start of the second feed. This time is adjustable. The goal is to pump until the deblocking solution level is just above the top of the resin bed when fresh deblocking solution begins to flow into the reactor, minimizing backmixing above the resin bed. Set the outlet pump speed to achieve the total pumping out within the desired reaction time as specified in Table 22. Add the deblocking solution to the reactor at approximately the same rate as the pumping out by setting the percent the feed control valve is open. Push any remaining solution out the bottom of the filter to waste.

[0235] Wash 1: Charge ACN solvent (200 mL) to the feed zone through the acid feed line. Drive the wash solvent into the feed zone with nitrogen, sweeping the feed tube. Push the solvent into the resin bed reactor. As the solvent enters the reactor, it will spray onto the walls of the reactor. Fluidize once for 5 seconds. Push out of the reactor to waste.

[0236] Wash 2, Wash 3, Wash 4, Wash 5: Charge ACN solvent (200 mL) from the ACN recycle canister through the solvent feed line into the feed zone. Drive the wash solvent into the feed zone with nitrogen to purge the feed tube. Forcing the solvent into the reactor through a spray cone allows for uniform solvent distribution without disrupting the resin bed. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter to waste. Repeat this same wash three more times.

[0237] Wash 6, Wash 7, Wash 8: Fill the feed zone with fresh ACN solvent (200 mL). Drive the wash solvent into the feed zone with nitrogen and purge the feed tube. Forcing the solvent into the reactor through a spray cone allows for even distribution of the solvent without disrupting the resin bed. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter into the recycle ACN canister. Repeat this same wash two more times.

[0238] Wash 9: Charge ACN solvent (200 mL) into the feed zone. Drive the wash solvent into the feed zone with nitrogen and sweep the feed tube. Push the solvent into the resin bed reactor. As the solvent enters the reactor, it will spray onto the walls of the reactor. Fluidize once for 5 seconds. Push the ACN solvent out the bottom of the filter into the recycle ACN canister.

[0239] Coupling Reaction: The specified amount of amidite (200 mL) is pumped into the amidite activation zone and purged with nitrogen. The activator solution (200 mL) is pumped into the amidite activation zone and purged with nitrogen. The two are mixed together by bubbling nitrogen into the bottom of the amidite zone for approximately 2 seconds. This mixture is forced into the feed zone and then into the reactor to initiate the coupling reaction on the resin. The resin reactor is fluidized intermittently approximately once every 45 seconds over the coupling time (10 or 15 minutes), with nitrogen bubbled into the bottom of the resin reactor for 15 seconds each time. Continuous fluidization rather than intermittent fluidization is also acceptable throughout the entire reaction time. After the reaction time has elapsed, the coupling solution is pushed out the bottom of the filter and discarded.

[0240] Solvent wash: ACN solvent (200 mL) is charged into the amidite activation zone through the amidite feed tube to expel any remaining droplets from the inlet tube, then forced into the feed zone and then into the resin reactor. The solvent sprays onto the walls as it enters the reactor. It is forced through the reactor to waste without being fluidized.

[0241] Oxidation reaction (if required instead of sulfurization): Charge ACN (200 mL) to the amidite activator mixing zone, allowing the resin to be washed immediately after the oxidation reaction is complete. The solvent enters the amidite activator mixing zone through the spray ball, washing all walls. Charge 0.05 M iodine solution (530 mL) to the feed zone and sweep with nitrogen to clear the feed tube. Push the solution into the reactor to initiate the oxidation reaction on the resin. Fluidize the resin reactor intermittently approximately every 30 seconds over the oxidation period (approximately 4 minutes), bubbling nitrogen into the bottom of the resin reactor for 12 seconds each time. Continuous fluidization rather than intermittent fluidization is also acceptable throughout the entire reaction period. Begin pumping the oxidation solution through the resin bed at 540 mL / min for 65 seconds. Push the remaining oxidation solution out the bottom of the filter to waste.

[0242] Wash 1: Push 200 mL of ACN wash solvent (from the amidite activator mixing zone) into the feed zone and then into the reactor to wash the resin. The solvent enters the reactor through the cone spray and contacts the resin, spraying evenly onto the top of the resin bed and keeping the resin bed flat, which makes plug flow washing more efficient. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter to waste.

[0243] Wash 2: Charge ACN (200 mL) to the feed zone through the oxidizing solution feed line and sweep the feed tube with nitrogen. Push the solvent into the reactor. The solvent will spray on the walls as it enters the reactor. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter to waste.

[0244] Wash 3, Wash 4, Wash 5: Fill the feed zone with ACN solvent (200 mL). Drive the wash solvent into the feed zone with nitrogen and purge the feed tube. Forcing the solvent into the reactor through a spray cone allows for uniform distribution of the solvent without disrupting the resin bed. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter and discard. Repeat the same wash two more times.

[0245] Wash 6: Charge ACN solvent (200 mL) into the feed zone. Drive the wash solvent into the feed zone with nitrogen and sweep the feed tube. Push the solvent into the resin bed reactor. As the solvent enters the reactor, it will spray onto the walls of the reactor. Fluidize once for 5 seconds. Push the ACN solvent out the bottom of the filter to waste.

[0246] Sulfurization (thiolation) reaction (if required instead of oxidation): Charge ACN (200 mL) to the amidite activator mixing zone, allowing the resin to be washed immediately after the sulfurization reaction is complete. Charge 0.2 M hydrogenated xanthan solution (650 mL) to the feed zone and flush with nitrogen to clear the feed tube. Force the solution into the reactor to initiate the sulfurization reaction on the resin. Fluidize the resin reactor intermittently approximately every 30 seconds over the oxidation period (approximately 8 minutes), bubbling nitrogen into the bottom of the resin reactor for 12 seconds each time. Continuous fluidization rather than intermittent fluidization is acceptable throughout the entire reaction period. Begin pumping the hydrogenated xanthan solution through the resin bed for 60 seconds at a pump setting of 700 mL / min. Push any remaining hydrogenated xanthan solution out the bottom of the filter to waste.

[0247] Wash 1: Push 200 mL of ACN wash solvent into the reactor (from the amidite activator mixing zone) to wash the resin. The solvent enters the reactor through the cone spray and contacts the resin. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter to waste.

[0248] Wash 2: Charge ACN (200 mL) to the feed zone through the hydrogenated xanthan solution feed line and sweep the feed tube with nitrogen. Push the solvent into the reactor. The solvent will spray on the walls as it enters the reactor. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter to waste.

[0249] Wash 3, Wash 4, Wash 5: Fill the feed zone with ACN solvent (200 mL). Drive the wash solvent into the feed zone with nitrogen and purge the feed tube. Forcing the solvent into the reactor through a spray cone allows for uniform distribution of the solvent without disrupting the resin bed. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter and discard. Repeat the same wash two more times.

[0250] Wash 6: Charge ACN solvent (200 mL) into the feed zone. Drive the wash solvent into the feed zone with nitrogen and sweep the feed tube. Push the solvent into the resin bed reactor. As the solvent enters the reactor, it will spray onto the walls of the reactor. Fluidize once for 5 seconds. Push the ACN solvent out the bottom of the filter to waste.

[0251] Capping reaction: Charge capping solution A and capping solution B (350 mL each) into the feed zone and sweep the feed tubes with nitrogen. Push the solutions into the reactor to start the capping reaction on the resin. Fluidize the resin reactor twice, bubbling nitrogen into the bottom of the resin reactor for 12 seconds each time. The total time for both fluidizations is approximately 1 minute. Begin pumping the reaction solution through the resin bed at 400 mL / min for 70 seconds. Push any remaining reaction solution out the bottom of the filter to waste.

[0252] Wash 1, Wash 2: Charge ACN (100 mL) to the feed zone through the Capping A solution feed line and flush with nitrogen to clear the feed tube. Charge ACN (100 mL) to the feed zone through the Capping B solution feed line and flush with nitrogen to clear the feed tube. Push the solvent into the reactor. The solvent will spray on the walls as it enters the reactor. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter to waste. Repeat this same wash one more time.

[0253] Wash 3, Wash 4: Fill the feed zone with ACN solvent (200 mL). Drive the wash solvent into the feed zone with nitrogen and purge the feed tube. Forcing the solvent into the reactor through a spray cone allows for uniform distribution of the solvent without disrupting the resin bed. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter and discard. Repeat this wash one more time.

[0254] Wash 5: Charge ACN solvent (200 mL) into the feed zone. Drive the wash solvent into the feed zone with nitrogen, sweeping the feed tube. Push the solvent into the resin bed reactor. As the solvent enters the reactor, it will spray onto the walls of the reactor. Fluidize once for 5 seconds. Push the ACN solvent out the bottom of the filter to waste.

[0255] Timing: The overall timing for a typical complete amidite cycle starting at 9:22 AM was as follows: 9:22 AM: Toluene washing and fluidization is performed four times. 9:28 AM: Inject the acid reagent solution and fluidize once. 9:35 AM Drain the acid reagent solution. 9:37 AM Fluidization wash is performed, spraying the solvent onto the walls as it enters the reactor. 9:39 AM: A plug flow wash is performed with recycled ACN, with the solvent passing through the cone spray into the reactor and contacting the resin. 9:41 AM: A plug flow wash is performed with recycled ACN, with the solvent passing through the cone spray into the reactor and contacting the resin. 9:43 AM: A plug flow wash is performed with recycled ACN, with the solvent passing through the cone spray into the reactor and contacting the resin. 9:44 AM: A plug flow wash is performed with recycled ACN, with the solvent passing through the cone spray into the reactor and contacting the resin. 9:46 AM Plug flow wash is performed, with solvent entering the reactor through the cone spray and contacting the resin. 9:48 AM Plug flow wash is performed, with solvent entering the reactor through the cone spray and contacting the resin. 9:50 AM Plug flow wash is performed, with solvent entering the reactor through the cone spray and contacting the resin. 9:52 AM Fluidization wash is performed, spraying the solvent onto the walls as it enters the reactor. 9:57 AM: Inject the coupling reagent solution. 10:14 AM Drain the coupling reagent solution. 10:16 AM: Plug flow washing is performed, spraying the solvent onto the walls as it enters the reactor. 10:21 AM: Inject XH reagent solution. 10:30 AM: Discharge the XH reagent solution. 10:32 AM Plug flow wash is performed, with solvent entering the reactor through the cone spray and contacting the resin. 10:35 AM: Plug flow washing is performed, spraying the solvent onto the walls as it enters the reactor. 10:37 AM Plug flow wash is performed, with solvent entering the reactor through the cone spray and contacting the resin. 10:39 AM Plug flow wash is performed, with solvent entering the reactor through the cone spray and contacting the resin. 10:41 AM Plug flow wash is performed, with solvent entering the reactor through the cone spray and contacting the resin. 10:43 AM Fluidization wash is performed, spraying the solvent onto the walls as it enters the reactor. 10:48 AM: Inject capping reagent solution. 10:52 AM Drain the capping reagent solution. 10:54 AM Fluidization wash is performed, spraying the solvent onto the walls as it enters the reactor. 10:56 AM Fluidization wash is performed, spraying solvent onto the walls as it enters the reactor. 10:59 AM Plug flow wash is performed, with solvent entering the reactor through the cone spray and contacting the resin. 11:01 AM Plug flow wash is performed, with solvent passing through the cone spray into the reactor and contacting the resin. 11:03 AM Fluidization wash is performed, spraying the solvent onto the walls as it enters the reactor.

[0256] This process was carried out for 4 consecutive days with 5, 5, 6, and 5 amidite cycles per day. The resin was kept in the reactor overnight, submerged in ACN, and placed under nitrogen each night.

[0257] Final cycle: Because the final amidite (MeMOP) does not have a DMT protecting group at the 5' position, a final deblocking step is not necessary. After the final MeMOP coupling, washing, sulfurization, and washing are complete, a wash with DEA ​​solution is performed. The DEA solution (500 mL) is charged to the feed zone. The DEA solution is driven into the feed zone with nitrogen, purging the feed tubing. The solution is then forced into the reactor. The resin bed is fluidized twice to ensure complete liquid-to-solid contact and reset the resin bed. The total time for both fluidizations is approximately 1 minute. Begin pumping the DEA solution through the resin bed at 100 mL / min for 600 seconds. Simultaneously, additional DEA solution (500 mL) is pumped in parallel into the feed zone, allowing it to enter the top of the reactor at approximately the same rate as it is being pumped out. The DEA solution is driven into the feed zone with nitrogen, purging the feed tubing. A total of 1 L is pumped through the resin bed over 600 seconds. The remaining DEA solution is then pushed out the bottom of the filter to waste. This DEA treatment is repeated once more.

[0258] Wash 1, Wash 2: Fill the feed zone with ACN solvent (200 mL). Drive the wash solvent into the feed zone with nitrogen and purge the feed tube. Forcing the solvent into the reactor through a spray cone allows for uniform distribution of the solvent without disrupting the resin bed. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter and discard. Repeat this wash one more time.

[0259] Wash 3: Charge ACN solvent (200 mL) into the feed zone. Drive the wash solvent into the feed zone with nitrogen, sweeping the feed tube. Push the solvent into the resin bed reactor. As the solvent enters the reactor, it will spray onto the walls of the reactor. Fluidize once for 5 seconds. Push the ACN solvent out the bottom of the filter to waste.

[0260] Wash 4, Wash 5: Fill the feed zone with ACN solvent (200 mL). Drive the wash solvent into the feed zone with nitrogen and purge the feed tube. Forcing the solvent into the reactor through a spray cone allows for uniform distribution of the solvent without disrupting the resin bed. Pump the ACN solvent through the resin bed at 300 mL / min. Push any remaining ACN solvent out the bottom of the filter and discard. Repeat this wash one more time.

[0261] Wash 6: Charge ACN solvent (200 mL) into the feed zone. Drive the wash solvent into the feed zone with nitrogen and sweep the feed tube. Push the solvent into the resin bed reactor. As the solvent enters the reactor, it will spray onto the walls of the reactor. Fluidize once for 5 seconds. Push the ACN solvent out the bottom of the filter to waste.

[0262] Drying: The resin is slurried out of the reactor and filtered through a laboratory filter. Nitrogen is blown downward through the resin bed and allowed to dry for 5-6 hours. 2.5 g of resin-bound material is sampled, including 2 g washed from the reactor and 0.5 g from the bulk after drying. The crude mass gain including the sample was 7.99 g / mmol.

[0263] Bulk cleavage and deprotection (C / D) was performed on approximately half of the resin-bound crude product at once. C / D was achieved by mixing the resin with 28% aqueous ammonium hydroxide (30 mL / g resin) and heating to 38 °C in a sealed vessel for 18–20 h. A 1.85 L Ace-thread pressure vessel equipped with a pressure gauge, a 25 psig pressure relief valve, a thermocouple, a heating mantle, and a magnetic stirrer was charged with protected ANGPTL3 AS (56.1 g, 6.798 mmol) and an aqueous solution of ammonium hydroxide (28 wt%) (1.68 L, 2000 g, 10,000 mmol). The thin slurry was sealed and stirred overnight while being heated to 38 °C.

[0264] After 18 hours at 38°C, the heat was turned off and an ice-water bath was added to cool the reactor to below room temperature. The resin was allowed to settle and a weighed aliquot of the supernatant was diluted with a weighed amount of Mill-Q water. Mass of aliquot = 0.1754 g Mass of milli-Q water = 20.2768g

[0265] Once C / D is complete, proceed to work up the bulk reaction mixture.

[0266] The bulk solution was filtered to remove spent resin. The spent resin was washed three times with 150 mL of EtOH:HO (1:1). The filtrate and washes were combined and concentrated on a rotary evaporator (40°C bath) to remove most of the ammonia. The same procedure was repeated for the remaining half of the resin-bound material. UPLC results showed 75.3% FLP for the first half of the sample and 78.9% FLP for the second half of the sample. Details can be seen in Table 17, the UPLC results for Examples 6-10, and a comparison with Cytiva AKTA. OD / umol was determined as recorded in Table 23. The purity-corrected crude yield was approximately 58% for the first half of the material and 62% for the second half. By comparison, a purity-corrected crude yield of 57% was obtained for the previous 1 kg cGMP campaign. [Table 24]

[0267] This material was pre-treated by chromatographic purification, which is beyond the scope of this document.

[0268] Example 9: Pilot-scale fluidized bed synthesizer with in-process integrated multi-pass cleaning. The same antisense strand of AngPTL3 (FIG. 10) was synthesized in a modified version of the fluidized bed reactor system of Example 8. 5'MeMOP*fG*fU*fAfUmA fAmCmC fUmUmC mCfAmUmUmUmUmGmA*mG*mG3'

[0269] The synthesis of this molecule using the fluidized-bed method of the present invention is described herein, which includes deblocking, coupling, oxidation (or sulfurization), and capping steps to sequentially introduce the remaining phosphoramidites. The main differences between Examples 8 and 9 are that the system was modified to include in-process integrated multi-pass washing, and there was no capping in cycles 2-9 (phosphoramidites 3-10). Capping is not required after the addition of the cycle 21 MeMOP phosphoramidite.

[0270] By the end of the experiment, the resin bed height reaches 5 cm wet with acetonitrile solvent and 4 cm dry. During the downflow portion of the final deblocking step, the maximum resin bed height reaches 6 cm. The maximum pressure drop across the resin bed during the experiment is 15 psig, since this is the pressure of the nitrogen feed used to force the liquid through the resin bed. As in Examples 6, 7, and 8, two equivalents of amidite are used in the coupling. The overall synthesis conditions are shown in Table 24. The deblocking times and the amount of fresh DCA solution from the start to the end of the synthesis are shown in Table 25. [Table 25]

[0271] KF of ACN used to prepare the amidite solution: 56 ppm water [Table 26]

[0272] The reagents and lot numbers used in Example 9 are listed in Table 26. [Table 27]

[0273] All amidite solutions were prepared in ACN, Fisher Lot No. 212215. Dissolve the amidite in the ACN solvent as follows: Mix until dissolved. Add the molecular sieve dry pack to the sealed bottle. ACN lots used: EMD lot number 52261, EMD lot number 52261, Fisher lot number 214141 [Table 28]

[0274] The amidite molecular weights were as follows: mA DMT-2'-O-MeA(bz) phosphoramidite, molecular weight 887.97 mC DMT-2'-O-MeC(Ac) phosphoramidite, molecular weight 801.87 mG DMT-2'-O-MeG(iBu) phosphoramidite, molecular weight 869.95 mU DMT-2'-O-MeU-CE phosphoramidite, molecular weight 760.82 fA DMT-2'-F-dA(bz) phosphoramidite, molecular weight 875.93 fC DMT-2'-F-dC(Ac) phosphoramidite, molecular weight 789.84 fG DMT-2'-F-dG(iBu) phosphoramidite, molecular weight 857.9 fU DMT-2'-F-dU-CE phosphoramidite, molecular weight 748.8 MeMOP, molecular weight 556.5

[0275] Prepare the reagent solution as follows: Cap B1: Acetic Anhydride: Macron Fine Chemicals Lot Number 0000239131 Acetonitrile: Fisher Lot No. 214141 Charge 481 mL of acetic anhydride and 722 mL of ACN into the bottle. Cap B2: 2,6-Lutidine: Acros Lot Number A0428332 Acetonitrile: EMD Lot No. 52261 Fill the bottle with 722 mL of lutidine and 481 mL of ACN. Cap A: 1-Methylimidazole: Alfa Aesar Lot No. 5009J24W Acetonitrile: Fisher Lot No. 214141 Charge 481 mL of imidazole. Charge 1924 mL of ACN. 0.2M Hydrogenated Xanthan Sulfide Solution: Hydrogenated Xanthan: TCI Lot Number QLXKC-LI Pyridine: Fishers Lot No. 208059 Charge 3775 mL of pyridine. Charge 114 g of XH to the pyridine bottle. Mix until dissolved. Oxidizing solution: Iodine solution (0.05M) Honeywell Lot Number EA702-US Approximately 9 kg of barrel stock was filled into a supply canister. Activator solution: Honeywell Lot Number EA713-US Approximately 5 kg of the Taru Soutoku solution is filled into a supply can. 20% DEA in acetonitrile: DEA: Sigma-Aldrich Lot No. STBJ5069 Acetonitrile: Fisher Lot No. 214141 Fill the bottle with 400 mL of DEA. Fill the bottle with 1600 mL of ACN. 3% DCA in toluene: DCA: Sigma Aldritch Lot Number MKCQ92 Toluene: Superior Lot Number HX11315122 Lot Number 1 1.19,240 mL of toluene was charged to a carboy. 2.595 mL of DCA was charged to the carboy. 3. The carboy was filled into the supply can. Lot No. 2 1.20,311 mL of toluene was charged to a carboy. 2.628 mL of DCA was charged to the carboy. 3. The carboy was filled into the supply can. Lot No. 3 1.12,272 mL of toluene was charged to a carboy. 2.380 mL of DCA was charged to the carboy. 3. The carboy was filled into the supply can.

[0276] Start by coupling mG (referred to herein as "mG resin") to NittoPhase HL 2'OMeG(iBu)250 resin (249 μmol / g) using known methods; see Figure 11 for the synthesis apparatus setup. 40.40 g (10.06 mmol) of mG resin is slurried using ACN and placed in a 10.16 cm internal diameter reactor fitted with a 40 micron sintered mesh filter frit at the bottom. The initial resin depth is approximately 1 cm in height.

[0277] All pumps and feed lines were primed. ACN was passed through a bed of molecular sieves on its way to the deactivation feed canister. The toluene solution of ACN and DCA was fed from the feed canister by pressure and controlled by an automatic flow control valve. Peristaltic pumps and feed vessels were used for all other feeds. The amidite solution was contained in a separate feed vessel labeled "AM. 1 L" and connected to a peristaltic pump attached to valves V1101A-V1108A in Figure 11. The MeMOP phosphoramidite was used in one of the AM feed vessels. The activator solution and DEA solution were contained in feed vessels labeled "Activator 5 gal" and "DEA 1 L" in Figure 11, respectively.

[0278] Each time a phosphoramidite is added during synthesis, the deblocking step, coupling step, oxidation step (or sulfurization step if there is a P=S bond in the sequence), and capping step are performed sequentially, as described below. Cycles 2-9 (nucleosides 3-10) are not capped. Cycle 21: Capping is not required after the addition of MeMOP.

[0279] Fluidization continued on an on / off cycle throughout the coupling, oxidation, sulfidation, and capping reactions for the majority of the specified reaction time. Note that fluidization does not have to be on / off cycled. Fluidization can be constant bubbling without pushing up and down. This procedure is a carryover from research-scale experiments. At research scale, in small diameter reactors, there is some advantage to pushing up and down during fluidization because it helps wet and fluidize all the resin beads initially, but there is no further advantage to pushing up and down thereafter. In larger diameter reactors, such as this 4-inch diameter reactor, on / off pushing up and down is not necessary.

[0280] As in the previous example, fluidization is achieved by blowing nitrogen gas upward through the bottom filter screen by opening valve 1156 or valve 1158 (V1156, V1158 in FIG. 11) simultaneously with the opening of the feed zone vent valve (V1152 in FIG. 11).

[0281] Throughout this procedure, please refer to Figures 11, 12, and 13. First, all six reusable bottles A-F and all three reusable bottles A2-C2 are manually filled with approximately 400 mL of ACN.

[0282] When the procedure states that the liquid is pumped downward through the resin bed, this means that the waste pump at the bottom outlet of the reactor is operating at a target setpoint while nitrogen pressure pushes against the top of the resin bed, forcing the liquid downward. The purpose of the peristaltic pump is to meter the liquid through the bed at a controlled rate.

[0283] Unblocking: First, recycled acid from the previous step was used to fluidize the resin, expanding the resin bed and flushing out the ACN solvent. This step was performed using fresh DCA / toluene solution for the first amidite cycle, then recycled DCA / toluene solution for the remaining amidites. Valve 1125A was set to recycled acid, valve 1125F was opened, vent valve 1152 was opened, and FCV1 was opened to charge the first charge (250 mL) of recycled acid. The controller measured the correct mass using the balance weight on the supply canister. FCV1 was closed, nitrogen valve 1125B was opened, and nitrogen was forced down the feed line to the feed zone. Nitrogen valve 1125B was closed. Vent valve 1152 was closed, nitrogen valve 1151 was opened, and the acid solution was forced through the spray cone into the reactor. Nitrogen valve 1151 was closed, vent valve 1152 was opened, valve 1153 was opened, and metering nitrogen valve 1156 was opened. This flushes nitrogen into the bottom of the reactor, allowing the acid solution to fluidize the resin for a user-set time (20 seconds). Metering nitrogen valve 1156, valve 1153, and vent valve 1152 are closed, and nitrogen valve 1151 is opened and pushed down. After fluidization is complete, valve 1159 is opened, pump 1159 is started, valve 1154 is directed to valve 1160, valve 1160 is directed to valve 1157, and valve 1157 is directed to waste. Valve 1125F is opened, and FCV1 is opened. This forces recycled acid through the feed zone and into the reactor as it is pumped out the bottom. The recycled acid canister is completely emptied. Volumes ranged from approximately 1150 mL in cycle 1 to 2600 mL in cycle 21. See Table 25. The amount of fresh acid in cycle 1 became the amount of recycled acid in cycle 2, and so on. Therefore, the volume of the second charge of recycled acid in cycle 2 was 1400 mL minus 250 mL, since 250 mL was used in the initial fluidization portion. This step completely flushes all ACN solvent from the resin to waste. Once all recycled acid has been emptied from the vessel, close FCV1. Open nitrogen valve 1125B to flush the feed line to the feed zone with nitrogen. Pump all recycled acid to waste. Total pumping time to waste ranged from approximately 3 minutes in cycle 1 to 4 minutes in cycle 21, gradually increasing as the volume gradually increased with each cycle.

[0284] Fill the first charge of fresh acid (150 mL) into the feed zone by opening valve 1125A to the fresh acid source, opening valve 1125F, opening vent valve 1152, and opening FCV1 to charge the specified mass of fresh acid. The controller uses the balance weight on the feed canister to deliver the correct mass. Close vent valve 1152, open nitrogen valve 1151, and force acid into the reactor through the spray cone, spraying it evenly onto the top of the resin bed and keeping it flat. Open valve 1159, open valve 1154 to valve 1160, open valve 1160 to valve 1157, open valve 1157 to the recycled acid canister, and start pump 1159. Pumping of the deblocking solution begins 5 to 30 seconds before the second feed begins. This time is adjustable. The goal is to pump until the deblocking solution level is just above the top of the resin bed when fresh deblocking solution begins to flow into the reactor, minimizing backmixing above the resin bed. The outlet pump speed is set to achieve the total pumping within the desired reaction time, as specified in Table 25. The deblocking solution is added to the reactor at approximately the same rate as it is being pumped out by setting the percentage at which Feed Control Valve 1 is open. By opening valve 1125F, opening nitrogen valve 1125B, and opening FCV1 to a value balanced with the flow of Pump 1159, acid solution is pumped out the bottom of the reactor while simultaneously feeding acid solution to the top of the reactor. This maintains the acid level above the resin bed while the acid flows through the resin plug flow. FCV1 closes when the user-specified total acid mass is reached. The total acid charge volume, including the 150 mL used in the initial charge, is listed by cycle in Table 25. For example, the total acid volume in cycle 1 was 1400 mL, consisting of 150 mL in the first charge and 1250 mL in the second charge. The volume increased linearly with each cycle, reaching 2830 mL by cycle 21. At the end of the pump-out period, valves 1153 and 1155 were opened, and the nitrogen supply to the feed zone was closed.This forces any remaining acid into the recycle acid canister until the pressure in the feed zone drops below a user-set value (e.g., the pressure drops from 15 psig to 9 psig), ensuring the reactor is empty before the automation proceeds to the next step in the sequence.

[0285] In-process integrated multi-pass cleaning after deblocking The first step of the in-process integrated multi-pass wash after the acid feed is to use solvent from Bottle A to wash the resin and push it out to waste. The next step is to use solvent from Bottle B to pump through the resin and back to refill Bottle A. Then, solvent from Bottle C is used to wash the resin in the reactor and pump it out to refill Bottle B, and so on. See Figures 11 and 12. Overall, the wash schedule for the experiment is detailed in Table 27. "w2" in the table is the second wash portion after deblocking. It washes through the resin bed and is pumped out the bottom of the reactor into Bottle A. As shown in the table, "w2" is the first wash portion after deblocking in Cycle 2. As listed in the table, "w3" is the third wash portion after deblocking in Cycle 1, which becomes the second wash portion after deblocking in Cycle 2, which then becomes the first wash portion after deblocking in Cycle 3, and so on. The seventh wash portion is split into three portions, e.g., w7a, w7b, and w7c in cycle 1 of the table. All three are pumped through the reactor and returned individually to bottle F, whereby the combined solvent pooled together becomes the sixth wash portion of cycle 2, and so on. [Table 29]

[0286] Valve 12201A (to the feed zone) and nitrogen feed valve 12201B and nitrogen feed valve 12201C (FIG. 12) share the same actuator air line, so that when valve 12201 opens, it is forced from the bottle into the feed zone. Similarly, valve 12200A (return from reactor) and vent valve number 12200B share the same actuator air line, so that when valve 12200 opens, the designated bottle receives the used wash solvent from the reactor.

[0287] First, open vent valve 1152, then open valve 11201, and open valve 11202. Nitrogen pushes the contents of Bottle A into the feed zone. Bottle A becomes completely empty. The automated system recognizes that Bottle A is completely empty by closing vent valve 1152 and waiting until the pressure in the feed zone rises to a user-specified value, meaning all the solvent has been transferred from Bottle A to the feed zone and is forced into the feed zone with nitrogen. Open nitrogen valve 1151 and turn valve 1145 to the spray cone to push the solvent from the feed zone through the spray cone and into the reactor, spraying it evenly onto the top of the resin bed and keeping it flat, making the cleaning more efficient. Open valve 1159, open valve 1154 to 1160, open valve 1160 to 1157, and open valve 1157 to waste. Turn on pump 1159 and pump the cleaning solution through the resin bed to waste. At the end of the pump-out period, close valve 1151, open valve 1153, and open valve 1155 to push any remaining wash solution to waste until the pressure in the zone is below a user-specified value (e.g., the pressure drops from 15 psig to 9 psig). This ensures that all liquid is pushed out of the reactor and to waste. Similarly, simultaneously activate pump 1159, which ensures that all liquid is removed from the pump and to waste. Only the wash from Bottles A through F is discarded. This removes the toluene and most of the acid from the resin and pushes it out to waste. The remaining wash is returned to Bottles A through E. This described procedure describes the procedure for removing solvent from Bottle B, pushing it through the reactor, and then back to Bottle A; the rest is similar. Open vent valve 1152, valve 11201, and valve 11203 to push the wash solvent from Bottle B into the feed zone until the bottle is empty. This is confirmed by the automated system by closing vent valve 1152 and waiting until the pressure in the feed zone rises above a user set value (9 psig), indicating that all liquid has been transferred and is being driven in with nitrogen.Close valves 11201 and 11203, open nitrogen valve 1151, and direct valve 1145 to the spray cone, forcing wash solvent from the feed zone through the spray cone into the reactor and onto the top of the resin, spraying it evenly on top and keeping the resin bed flat. Open valve 1159, direct valve 1154 to valve 11200, open valve 11200, and open valve 11202. Turn on pump 1159, allowing the solvent to wash the resin bed and return to Bottle A. At the end of the user-specified pump time, open valves 1153 and 1155, close nitrogen valve 1151, and wait until the pressure in the feed zone falls below the user-set value (dropping from 15 psig to 9 psig). This allows all the solvent to be transferred through the reactor to Bottle A. Repeat this process, using the solvent in Bottle C to wash the resin bed and return it to Bottle B, then Bottle D and back to Bottle C, and so on. The user has the option to specify whether any of the washes are fluidized. If the user chooses to fluidize one of the washes, after the wash solution is transferred to the reactor and before pump 1159 is started, nitrogen is blown upward from the bottom of the reactor with valve 1152 open. None of the in-process integrated multi-pass washes were fluidized in this experiment. The user has the option to specify which of these washes the system will automatically drive into the acid feed line and which of these washes the system will automatically wash the feed zone walls and reactor walls. For example, suppose the user chooses to perform a feed line drive wash during the second wash. In this case, the wash solvent from bottle B is forced from the feed zone into the reactor and then remains there waiting before being pumped into the reactor so the system can drive into the feed line. This is done by opening valve 1125C and using pump number 1130 to pump the specified amount of ACN (50 mL) solvent through the acid feed line and into the feed zone. Solvent is forced forward by closing valve 1125C and opening nitrogen valve 1125B, and then nitrogen valve 1151 is opened to force the forced solvent from the feed zone into the reactor.The mixed solvents in the reactor are then pumped through the resin and into destination bottle A, as described above. Also, for example, suppose the user specifies that a reactor wall wash be performed during the ACN solvent wash from bottle F. In this case, after the solvent from bottle F is forced into the reactor, it remains there before being forced through the resin, waiting for the reactor wall wash to occur. The reactor wall wash is performed as follows: Vent valve 1152 is opened, valve 1130B is opened, and FCV2 is opened until the specified mass (50 mL) is forced through the wall-washing spray ball and into the feed zone. Valve 1130B is opened during the fill because it helps the spray ball function more effectively at this scale, and the nitrogen through valve 1130B also drives solvent into the feed zone. Then, FCV2 is closed, valve 1130B is closed, vent valve 1152 is closed, nitrogen valve 1151 is opened, and valve 1145 is directed to the wall spray device into the reactor. This procedure is repeated one more time, spraying the feed zone walls one more time and spraying the reactor walls one more time, where the combined wash solvent from Bottle F and both reactor wall washes is pumped back through the resin in the reactor via pump 1159 to Bottle E, as described above. At the end of these countercurrent washes, Bottle F is emptied.

[0288] Plug flow wash after deblocking: The next step is to wash the resin in the reactor with fresh ACN solvent and pump it from the reactor to Bottle F. This is done by using the Plug Flow Wash program and specifying the destination as Bottle F. The Plug Flow Wash program has three user-selectable destinations: Bottle F, Bottle C2, or waste. Plug flow washing is accomplished as follows: Open vent valve 1152, open valve 1130B, and open FCV2 until the specified ACN solvent mass (150 mL) is forced through the spray ball, which washes the wall, and into the feed zone. Valve 1130B is open during filling because it helps the spray ball function more effectively at this scale, and the nitrogen through valve 1130B drives the solvent into the feed zone. Then, close FCV2, close valve 1130B, close vent valve 1152, open nitrogen valve 1151, and direct valve 1145 toward the spray cone into the reactor to spray evenly over the top of the resin bed and keep it flat. The wash solvent is then pumped through the resin bed by opening valve 1159, starting pump 1159, and setting downstream valves V1154, V1160, V1157, V11200, V11300 in the appropriate position depending on the destination (Bottle F, Bottle C2, or waste). In this case, the destination is Bottle F. This step is performed two more times, for a total of three 150 mL plug flow washes from the reactor to Bottle F.

[0289] By cycle number 7, the countercurrent wash from Bottle A contained approximately 600 mL (450 mL fresh ACN for the three plug-flow wash steps, 100 mL for the reactor wall wash, and 50 mL for the feed line run-off wash). The second wash and 50 mL run-off wash from Bottle B contained 600 mL. The third through sixth washes from Bottles C, D, E, and F totaled 550 mL (the 100 mL reactor wall wash plus 450 mL from Bottle F). The total amount of wash solvent flowing through the resin for all post-deblocking washes was approximately 4 L. However, only 600 mL of fresh ACN was loaded into the system. The remainder was recycled ACN from Bottles A through F. This in-process integrated multi-pass wash strategy (Table 27) resulted in more efficient washes. Samples were taken from the final wash throughout this run, from cycle 1 to cycle 21, and all samples were >99.9% ACN as measured by NMR. The Cytiva AKTA synthesizer also achieves 99.9% ACN solvent at the end of the wash, but requires seven times more wash solvent per mmol to achieve the same wash endpoint. Flushing the reactor with the fluid improves efficiency and is therefore superior to solvent washing in packed-bed reactors for several reasons: (1) Because the reagents are drained before the wash begins, significant backmixing with the previous liquid is eliminated, in addition to the liquid retained by the resin after draining. (2) The resin bed is fluidized during the reaction, which results in a flat, channel-free bed at the start of the wash. (3) Because the reactor is not completely filled with liquid, gravity allows the solvent to distribute radially throughout the top of the resin bed. (4) The wash is split into multiple smaller wash loads, which allows for more plug flow with less backmixing. (5) In-process integrated multi-pass washing allows for much more efficient use of wash solvent. After each reaction, only the "dirtiest" wash solvent leaves the system and is discarded; only a new, clean solvent supply is required for the final wash section.

[0290] Coupling: In the coupling step, the phosphoramidite and activator are pumped into the amidite zone, mixed within the zone, and then the coupling solution is forced into the feed zone and then into the reactor. The coupling solution is allowed to fluidize within the reactor for a user-specified time, e.g., 10 minutes, after which the reaction solution is forced out of the reactor, allowing it to be completely drained and discarded. More specifically, the specified amidite (200 mL) is pumped into the amidite activation zone and purged with nitrogen. The activator solution (200 mL) is pumped into the amidite activation zone and purged with nitrogen. After mixing, the mixture is forced into the feed zone and then into the reactor to initiate the coupling reaction on the resin. The resin reactor is intermittently fluidized approximately every 45 seconds over the coupling time (10 or 15 minutes), with nitrogen bubbling through the bottom of the resin reactor for 15 seconds each time. For example, if amidite number 4 is being used, the automation performs the following: Open vent valve 1142, open valve 1104A, and turn on pump number 1104. Pump a user-specified mass (200 mL). The control system monitors the change in mass on the feed vessel scale to measure the correct amount. At the end of pumping, close valve 1104A and open valve 1104B to drive the amidite feed solution into the amidite zone with nitrogen. Do the same for the activator. Open vent valve 1142, open valve 1120A, and turn on pump 1120 to charge a user-specified mass (200 mL), then close valve 1120A and open valve 1120B to drive nitrogen forward into the amidite zone. Open valve 1143 to blow nitrogen backward from the feed zone into the amidite zone, mixing the activator into the amidite. The coupling solution is forced into the feed zone by opening valve 1141, opening valve 1143, closing vent valve 1142, and opening vent valve 1152. The coupling solution is forced from the feed zone into the reactor by closing vent valve 1152, closing valve 1143, and opening nitrogen valve 1151. Valve 1145 is directed to the spray cone.The batch reaction is thoroughly mixed by opening vent valve 1152, opening valve 1153, and opening valve 1158, and bubbling nitrogen into the bottom of the reactor and out the feed zone vent. Alternately, press down to push the liquid into the resin, then blow nitrogen upward at a user-specified frequency to force the fluid out of the resin. Continuous fluidization, rather than intermittent, is also acceptable throughout the reaction time. When pressed down, vent valve 1152 closes, nitrogen valve 1151 opens, and valve 1153 closes. When pressed up, the valves are in the opposite position, allowing nitrogen to flow to the bottom of the reactor and out the vent. At the end of the coupling, it is pushed out to waste. This means the system closes valve 1152, opens nitrogen valve 1151, opens valve 1153, and opens valve 1155. Valve 1154 points to valve 1160, valve 1160 points to valve 1157, and valve 1157 points to waste. Valve 1104A is then opened and peristaltic pump number 1104 is started in the reverse direction for approximately 1 mL to remove the reagent solution from the dead leg, minimizing the possibility of amidite 4 dripping into the amidite zone during different cycles.

[0291] Cleaning the drive-in supply line after coupling: This is a continuation of the example using amidite valve 1104. Valve 1104C is opened, vent valve 1142 is opened, and pump 1130 is used to pump a user-specified amount of ACN solvent (100 mL). Valve 1104C is then closed and valve 1104B is opened to drive the solvent into the amidite zone with nitrogen. Vent valve 1142 is closed, valve 1141 is opened, valve 1143 is opened, and vent valve 1152 is opened to push the driven wash solvent into the feed zone. The driven wash solution is then forced through the spray cone into the reactor, and the reactor is pressurized by closing valves 1143 and 1152, opening valve 1151, and pumping the wash solvent out the bottom of the reactor to waste.

[0292] Oxidation: (if required instead of sulfurization): Charge the feed zone with 0.05 M iodine solution (530 mL) and flush with nitrogen to purge the feed tube. The solution is forced into the reactor to initiate the oxidation reaction on the resin. Fluidize the resin reactor intermittently approximately every 30 seconds for the oxidation time (approximately 4 minutes), bubbling nitrogen into the bottom of the resin reactor for 12 seconds each time. Continuous fluidization is acceptable throughout the entire reaction time. Begin pumping the oxidation solution through the resin bed at 540 mL / min for 65 seconds. Force any remaining oxidation solution out the bottom of the filter to waste. More specifically, fill the feed zone with iodine solution by opening vent valve 1152, opening valve 1123A, and pumping with pump 1123 until the user-specified iodine solution mass is reached. The control system uses a balance weight on the iodine supply vessel to deliver the correct amount. After reaching the correct iodine delivery mass, close valve 1123A and open valve 1123B, allowing nitrogen to drive iodine from the feed line into the feed zone. Close vent valve 1152, open nitrogen valve 1151, aim valve 1145 at the spray cone, and wait a user-specified time (approximately 10 seconds) to force iodine from the feed zone into the reactor on top of the resin. Open valve 1159 and turn on pump 1159 long enough to pump out the approximately 20 mL of ACN that was forced out the bottom of the reactor as iodine was forced down through the resin. Proceed to run the oxidation reaction batch-wise by repeatedly fluidizing the resin bed in the iodine solution, as was done for the coupling reaction. However, use more vigorous nitrogen bubbling during fluidization by opening metered nitrogen valve 1156 in addition to valve 1158. Note that valve 1156 is at a higher flow rate of nitrogen and valve 1158 is at a lower flow rate of nitrogen due to the metering valve settings and CV. Forcing iodine down through the resin and bubbling nitrogen through the resin for fluidization at a specified frequency for a user-specified time alternate for the desired duration of the oxidation reaction (e.g., 4 minutes). Continuous fluidization rather than intermittent fluidization throughout the entire reaction time is also acceptable. At the end of the fluidization oxidation time, the iodine solution is pumped out to waste.To do this, open nitrogen valve 1151, open valve 1159, direct valve 1154 to valve 1160, direct valve 1160 to valve 1157, and direct valve 1157 to waste. Turn on pump 1159 to pump to waste. After the specified pump-to-waste time has elapsed, close nitrogen valve 1151, open valve 1153, open valve 1155, and wait until the pressure in the feed zone is below a user-specified value (dropping from 15 psig to 9 psig), ensuring all liquid has been pushed to waste and can be driven in with nitrogen. Open valve 1123A and run pump number 1123 in the reverse direction for approximately 1 mL. This clears the reagent from the dead leg and helps ensure a clean subsequent drive of the feed line, eliminating any iodine remaining in the feed line and potentially dripping into the feed zone during other steps.

[0293] In-process integrated multi-pass cleaning after oxidation: The in-process integrated multi-pass wash after oxidation is very similar to the in-process integrated multi-pass wash after acid deblocking, except that it uses only three bottles: A2, B2, and C2. The equipment details are shown in Figure 13. Bottle A2 is used first, and its wash solvent is forced through the reactor resin bed and discarded. Solvent from bottle B2 is then used to wash the reactor, and the solvent is forced through the resin bed and into bottle A2. At the end of the in-process integrated multi-pass wash, bottles A2 and B2 are full, but bottle C2 is empty. Bottle C2 is refilled with the reactor wall wash, drive wash, and amidite zone wash, as described next.

[0294] By cycle number 9, the countercurrent washes from bottles A2, B2, and C2 contained approximately 450 mL (450 mL of fresh ACN entering bottle C2 from 100 mL post-coupling drive, 100 mL reactor wall wash, 50 mL feed line drive wash, 100 mL amidite zone wash, and 100 mL plug flow wash). The total amount of wash solvent flowing through the resin for all post-oxidation washes was approximately 1800 mL. However, only 450 mL of fresh ACN was loaded into the system. The remainder was recycled ACN from bottles A2, B2, and C2. This in-process, integrated, multi-pass wash strategy makes washing more efficient. Samples were taken from the final wash throughout this run, from cycle 1 to cycle 21, and all samples measured by NMR were >99.9% ACN, which is roughly the same as what the Cytiva synthesizer gets at the end of the wash, but Cytiva uses 7 times more wash solvent per mmol compared to the total Cytiva wash used after coupling and oxidation. The final wash after sulfurization was also measured to be >99.9% ACN by NMR.

[0295] Post-oxidation reactor wall cleaning: Open vent valve 1152, open valve 1130B, and open FCV2 until the desired mass of solvent (50 mL) has entered the feed zone. Solvent enters the feed zone through the spray ball, which sprays the walls of the feed zone. Close valve 1130B, open nitrogen valve 1151, and direct valve 1145 to the wall spray, which sprays the reactor walls. Close nitrogen valve 1151, open vent valve 1152, and repeat these steps to fill with additional wash solvent (50 mL) while spraying the walls of the feed zone and the reactor walls. Fluidization of the solvent and resin in the reactor is optional in this step, as selected by the user; fluidization was not performed in this experiment. Open nitrogen valve 1151 and open valve 1159. Direct valve 1154 to valve 1160, which directs valve 1160 to valve 11300. Open valve 11304 and turn on pump 1159 to pump the wash solvent through the reactor and into bottle C2. At the end of the pumping period, open nitrogen valve 1153, open valve 1155, and close valve 1151. Push to a user-defined pressure in the feed zone (dropping from 15 psig to 9 psig) to ensure all solvent is removed from the reactor and into bottle C2.

[0296] Cleaning of feed lines after oxidation: Open vent valve 1152, open valve 1123C, and start pump 1130 to pump the specified mass of solvent (50 mL) into the feed zone. Close valve 1123C and open valve 1130B to blow solvent forward into the feed zone. Point valve 1145 at the spray cone to spray evenly onto the top of the resin bed and keep it level. Open nitrogen valve 1151 to force the wash solvent into the reactor. Open valves 1159, 11300, and 11304. Point valve 1154 to 1160 and valve 1160 to valve 11300. Pump the wash solvent through the reactor into bottle C2. At the end of the pump time, close nitrogen valve 1151, open valve 1153, and open valve 1155 to force any remaining solvent out of the reactor and into bottle C2 until the pressure in the feed zone is below the user set value (drops from 15 psig to 9 psig).

[0297] Amidite zone washing. The amidite zone wash is performed after oxidation (or sulfurization) to obtain double the value of the wash solvent, washing any remaining droplets from the walls of the amidite zone and storing more iodine-, pyridine-, and water-free wash solvent in bottle C2 for the next phosphoramidite cycle. Open valve 1130E and start pump 1130 to pump ACN (100 mL) into the wash bottle. Stop pump 1130 and close valve 1130E. Open vent valve 1142 and open valve 1130F to force solvent through the spray ball into the amidite zone, thoroughly spraying all surfaces within the zone and flushing out any previous amidite droplets. The air from the solenoid to valve 1130F also supplies the actuator for the nitrogen supply valve on top of the wash vessel, allowing nitrogen to pressurize the wash vessel as valve 1130F opens. Valve 1130F is a three-way valve that fails open to vent. Close vent valve 1142, close valve 1130F (closing valve 1130F also allows the N2 supply valve on top of the wash vessel to vent), open valve 1141, open valve 1143, and open vent valve 1152. This forces all of the wash solvent into the feed zone for a user-specified time (e.g., 5 seconds). Close valves 1141, 1143, and 1152, and open nitrogen valve 1151. Open valve 1159, set valve 1154 to valve 1160, set valve 1160 to valve 11300, open valve 11300, and open valve 11304. Start pump 1159 to pump the wash solvent through the resin bed into bottle C2 for a user-specified time. At the end of pumping, open valves 1153 and 1155, close nitrogen valve 1151, and allow nitrogen pressure to push any remaining solvent out of the reactor and into bottle C2. Wait until the pressure in the feed zone drops below a user-specified value, which confirms that all solvent has been pushed out of the reactor and into bottle C2.

[0298] Post-oxidation plug flow cleaning: Perform plug flow washing as before, but push the washing solvent (100 mL) into bottle C2.

[0299] Coupling reactions: Capping solution A and capping solution B (100 mL each) are charged into the feed zone and driven with nitrogen to clear the feed tubes. The solutions are forced into the reactor to initiate the capping reaction on the resin. The resin reactor is fluidized twice, bubbling nitrogen into the bottom of the resin reactor for 12 seconds each time. The total time for both fluidizations is approximately 1 minute. Continuous fluidization rather than intermittent fluidization over the course of 1 minute is acceptable. Begin pumping the reaction solution through the resin bed at 200 mL / min for approximately 1 minute. Any remaining reaction solution is forced out the bottom of the filter to waste. The specific automation sequence for the capping step is similar to that for the oxidation step, except that the capping reagent flows through valves 1121A and 1122A, using valves 1121B, 1122B, 1121C, and 1122C for nitrogen and solvent drives, as described for the oxidation step.

[0300] Another embodiment of the synthesis apparatus also uses three in-process integrated multi-pass washes for post-capping washes: Bottle A, Bottle B, and Bottle C. However, in this experiment, the post-capping washes were sent directly to waste.

[0301] Sulfurization (thiolation) reaction (if desired instead of oxidation): Charge 0.2 M hydrogenated xanthan solution (650 mL) into the feed zone and flush with nitrogen to purge the feed tube. The solution is forced into the reactor to initiate the sulfurization reaction on the resin. Fluidize the resin reactor intermittently approximately every 30 seconds over the sulfurization fluidization period (approximately 8 minutes), bubbling nitrogen into the bottom of the resin reactor for 12 seconds each time. Continuous fluidization is acceptable throughout the entire reaction period. Begin pumping the hydrogenated xanthan solution through the resin bed at 700 mL / min for 60 seconds. Force any remaining hydrogenated xanthan solution out the bottom of the filter to waste. The detailed automation sequence for the sulfurization step is similar to that for the oxidation step, except that the hydrogenated xanthan solution is pumped by pump 1124 using valves 1124A, 1124B, and 1124C. Sulfurization was used for the first two and last three cycles. After the first two cycles, bottles A2, B2, and C2 were removed and replaced with new bottles A2, B2, and C2, each filled with approximately 400 mL of fresh ACN. Then, before the final three cycles, bottles A2, B2, and C2 were removed and replaced with old bottles A2, B2, and C2, still filled with hydrogenated xanthan-containing ACN wash solvent from the first two cycles. This was because we chose not to use hydrogenated xanthan-containing ACN wash solvent for post-oxidation washes in this experiment, and vice versa.

[0302] Similar to the post-oxidation washes, the total amount of wash solvent flowing through the resin for all post-sulfurization washes was approximately 1800 mL. However, only 450 mL of fresh ACN was charged to the system; the remainder was recycled ACN from bottles A2, B2, and C2. Again, this in-process integrated multi-pass wash strategy results in more efficient washes.

[0303] Final cycle: Because the final amidite (MeMOP) does not have a DMT protecting group at the 5' position, final deblocking is not necessary. After the final MeMOP coupling, washing, sulfurization, and washing are complete, rinse with DEA ​​solution. Charge the DEA solution (500 mL) into the feed zone. Drive the DEA solution into the feed zone with nitrogen and purge the feed tubing. Push the solution into the reactor. Fluidize the resin bed twice to ensure complete liquid-to-solid contact and reset the resin bed. The total time for both fluidizations is approximately 1 minute. Continuous fluidization rather than intermittent over the 1 minute period is acceptable. Begin pumping the DEA solution through the resin bed at 100 mL / min for 600 seconds. Simultaneously, pump additional DEA solution (500 mL) into the feed zone in parallel, allowing it to enter the top of the reactor at approximately the same rate as it is being pumped out. Drive the DEA solution into the feed zone with nitrogen and purge the feed tubing. A total of 1 L is pumped through the resin bed over 600 seconds. The remaining DEA solution is pushed out the bottom of the filter and discarded. Repeat this DEA treatment one more time.

[0304] Wash thoroughly with ACN as follows: All of these ACN after DEA were pushed out of the reactor to waste using fresh ACN from the feed can. Similar to the other drive washes previously described in this procedure, perform a plug flow wash of the resin bed using 200 mL of ACN driven into the DEA feed line. Perform three plug flow washes with 150 mL of ACN each, using the same procedure as for the plug flow washes previously described. Wash the reactor walls with 50 mL of ACN as previously described ("reactor wall wash"). Perform two plug flow washes with 150 mL of ACN each, using the same procedure as for the plug flow washes previously described.

[0305] Drying: The resin was slurried out of the reactor and transferred to a single plate filter. Nitrogen was blown downward through the resin bed for 5-6 hours to dry. After removing approximately 3 g of sample, the total weight of the dried resin recovered was 115.7 g.

[0306] A small sample was taken for cleavage, deprotection, and UPLC. The results are included in Table 17, a comparison of the UPLC results for Examples 6-10 and the Cytiva AKTA. The purity was 77.89% FLP. This is slightly lower than the other fluidized-bed reactor examples in the table, but there is a specific reason for this. A small amount of acid was accidentally added to the coupling feed line in cycle 19. This resulted in a larger-than-normal cleavage of the 19-mer, as shown in the table. This error resulted in a loss of FLP of approximately 1.5%. Otherwise, we estimate that the FLP for the run would have been 79-80%. Yield and purity data for this experiment are listed in Table 17.

[0307] Bulk cleavage and deprotection (C / D) was performed on two lots, each containing approximately half of the resin-bound crude product. For each of the two lots, C / D was achieved by mixing the resin with 28% aqueous ammonium hydroxide (30 mL / g resin) and heating to 38 °C in a sealed vessel for 18-20 hours. An 1850 mL Ace-thread pressure vessel equipped with a pressure gauge, a 25 psig pressure relief valve, a thermocouple, a heating mantle, and a magnetic stirrer was charged with an aqueous solution (1.68 L, 2000 g, 10,000 mmol) of protected ANGPTL3 AS and ammonium hydroxide (28 wt%). The thin slurry was sealed and stirred overnight while heating to 38 °C. 58.43 g of resin-bound product was charged to the first lot, and 56.91 g of resin-bound product was charged to the second lot.

[0308] After 18 hours at 38 °C, the heat was turned off and an ice-water bath was added to cool the reactor to below room temperature. The resin was allowed to settle, and a weighed aliquot of the supernatant was diluted with a weighed amount of Mill-Q water. Each lot was analyzed by UPLC. The FLP of lot 1 was 77.6%, and that of lot 2 was 78.4%. As previously mentioned, the 19-mer cleavage was 2% in both lots. The bulk solution was filtered to remove the spent resin. The spent resin from each lot was washed three times with 150 mL of EtOH:HO (1:1). This time, the ammonia was not removed by rotary evaporation; instead, it was removed along with the C / D byproducts by TFF. The resulting crude mass was 27.52 g for lot 1 and 28.07 g for lot 2. This material was pre-treated by chromatographic purification, which is beyond the scope of this specification.

[0309] Example 10: Pilot-scale fluidized bed synthesizer with in-process integrated multi-pass cleaning. Example 10 was very similar to Example 9. However, in Example 10, an in-process integrated multi-pass wash was also performed after capping. Example 10 demonstrated the lowest ACN solvent wash volume (mL / mmol) of all the examples.

[0310] The same antisense strand of AngPTL3 (FIG. 10) was synthesized in a modified version of the fluidized bed reactor system of Example 10. 5'MeMOP*fG*fU*fAfUmA fAmCmC fUmUmC mCfAmUmUmUmUmGmA*mG*mG3'

[0311] The synthesis of this molecule using the fluidized bed method of the present invention is described herein, which includes deblocking, coupling, oxidation (or sulfurization), and capping steps to sequentially introduce the remaining phosphoramidites.

[0312] By the end of the experiment, the resin bed height reaches 5 cm wet with acetonitrile solvent and 4 cm dry. During the downflow portion of the final deblocking step, the maximum resin bed height reaches 6 cm. The maximum pressure drop across the resin bed during the experiment is 15 psig, as this is the pressure of the nitrogen feed used to force the liquid through the resin bed. Except for the final cycle, 2 equivalents of amidite are used in the coupling. 2.1 equivalents were used for the MeMOP amidite coupling. The overall synthesis conditions are shown in Table 28. The deblocking times and the amount of fresh DCA solution from the start to the end of the synthesis are shown in Table 29. [Table 30] [Table 31]

[0313] The reagents and lot numbers used in Example 9 are shown in Table 30. [Table 32]

[0314] All amidite solutions were prepared in ACN, Fisher Lot No. 212215. Dissolve the amidite in the ACN solvent as follows: Mix until dissolved. Add the molecular sieve dry pack to the sealed bottle.

[0315] ACN lots used: EMD lot number 52261, EMD lot number 52261, Fisher lot number 214141 [Table 33]

[0316] The amidite molecular weights were as follows: mA DMT-2'-O-MeA(bz) phosphoramidite, molecular weight 887.97 mC DMT-2'-O-MeC(Ac) phosphoramidite, molecular weight 801.87 mG DMT-2'-O-MeG(iBu) phosphoramidite, molecular weight 869.95 mU DMT-2'-O-MeU-CE phosphoramidite, molecular weight 760.82 fA DMT-2'-F-dA(bz) phosphoramidite, molecular weight 875.93 fC DMT-2'-F-dC(Ac) phosphoramidite, molecular weight 789.84 fG DMT-2'-F-dG(iBu) phosphoramidite, molecular weight 857.9 fU DMT-2'-F-dU-CE phosphoramidite, molecular weight 748.8 MeMOP, molecular weight 556.5

[0317] Prepare the reagent solution as follows: Cap B1: Acetic Anhydride: Macron Fine Chemicals Lot Number 0000239131 Acetonitrile: Fisher Lot No. 214141 Charge 481 mL of acetic anhydride and 722 mL of ACN into the bottle. Cap B2: 2,6-Lutidine: Acros Lot Number A0428332 Acetonitrile: EMD Lot No. 52261 Fill the bottle with 722 mL of lutidine and 481 mL of ACN. Cap A: 1-Methylimidazole: Alfa Aesar Lot No. 5009J24W Acetonitrile: Fisher Lot No. 214141 Charge 481 mL of imidazole. Charge 1924 mL of ACN. 0.2M Hydrogenated Xanthan Sulfide Solution: Hydrogenated Xanthan: TCI Lot Number QLXKC-LI Pyridine: Fishers Lot No. 212147 Charge 3775 mL of pyridine. Charge 114 g of XH to the pyridine bottle. Mix until dissolved. Oxidizing solution: Iodine solution (0.05M) Honeywell Lot Number EA702-US Approximately 8 kg of barrel stock was filled into a supply canister. Activator solution: Honeywell Lot Number EA713-US Approximately 4.3 kg of barrel stock solution was filled into a supply can. 20% DEA in acetonitrile: DEA: Sigma-Aldrich Lot No. STBJ5069 DEA: Sigma-Aldrich Lot No. SHBK7197 Acetonitrile: Fisher Lot No. 214141 Fill the bottle with 400 mL of DEA. Fill the bottle with 1600 mL of ACN. 3% DCA in toluene: DCA: Sigma Aldritch Lot Number MKCQ92 Toluene: Superior Lot Number HX11315122 Lot Number 1 1.18,664 mL of toluene was charged to a carboy. 2.577 mL of DCA was charged to the carboy. 3. The carboy was filled into the supply can. Lot No. 2 1.19,516 mL of toluene was charged to a carboy. 2.604 mL of DCA was charged to the carboy. 3. The carboy was filled into the supply can.

[0318] Start by coupling mG (referred to herein as "mG resin") to NittoPhase HL 2'OMeG(iBu)250 resin (249 μmol / g) using known methods; see Figure 11 for the synthesis apparatus setup. 40.18 g (10.00 mmol) of mG resin is slurried using ACN and placed in a 10.16 cm internal diameter reactor fitted with a 40 micron sintered mesh filter frit at the bottom. The initial resin depth is approximately 1 cm in height.

[0319] All pumps and feed lines were primed. ACN was passed through a bed of molecular sieves on its way to the deactivation feed canister. The toluene solution of ACN and DCA was fed from the feed canister by pressure and controlled by an automatic flow control valve. Peristaltic pumps and feed vessels were used for all other feeds. The amidite solution was contained in a separate feed vessel labeled "AM. 1 L" and connected to a peristaltic pump attached to valves V1101A-V1108A in Figure 11. The MeMOP phosphoramidite was used in one of the AM feed vessels. The activator solution and DEA solution were contained in feed vessels labeled "Activator 5 gal" and "DEA 1 L" in Figure 11, respectively.

[0320] Cycles 2-9 (nucleosides 3-10) are not capped. Cycle 21: Capping is not required after the addition of MeMOP. Each time a phosphoramidite is added during synthesis, the deblocking, coupling, oxidation (or sulfurization, if there is a P=S bond in the sequence), and capping steps are performed sequentially as described in Example 9, with the following differences:

[0321] Three additional 1 L bottles, bottle A3, bottle B3, and bottle C3, were also used with sequential automatic block valves to achieve in-process integrated multi-pass cleaning after capping. These are not shown in FIG. 11 but are similar to the sections of FIG. 11 including bottles A2, B2, and C2 and the descriptions in FIG. 13 for A2, B2, and C2. The valves in bottles A3, B3, and C3 are all the same as those shown in FIG. 13, but are labeled as 400 series: valve 400A, valve 400B, valve 400C, valve 401A, valve 401B, valve 401C, valve 402, valve 403, and valve 404. The multi-pass cleaning procedure was similar to that described for the post-oxidation / thiolation cleaning in Example 9. All solvent from the Bottle A3 wash was pumped through the resin bed to waste. Then, the Bottle B3 wash was pumped through the resin bed to Bottle A3, and the Bottle C3 wash was pumped through the resin bed to Bottle B3. Two 150 mL washes of fresh ACN were then pumped through the resin bed and back to Bottle C3. The first 150 mL was used to wash the reagent supply tubing to the feed zone and then the resin bed plug flow. The second 150 mL wash was sprayed on the walls of the feed zone and then the reactor walls and then the resin bed plug flow. Overall, a total of 1.3 L of fresh ACN was used for each wash cycle, after pre-filling the multi-pass wash bottles with 300 mL each. This included 550 mL after deblocking, 100 mL after coupling, 350 mL after oxidation, and 300 mL after capping. By comparison, at the 10 mmol scale, a packed bed synthesizer typically uses about 8-10 L of fresh ACN for solvent push-out and solvent washes during each cycle. Samples demonstrated that the wash endpoints were each the same as those typically achieved using a packed bed synthesizer, i.e., about 99.9% of the ACN in the final volume of wash solvent exits the reactor.The approximately 85% reduction in ACN wash solvent compared to a packed-bed reactor is due to the following reasons: the reagents are drained from the reactor before washing; the resin bed is flattened by fluidization during the reaction, eliminating channels; the wash solvent is effectively distributed over the resin by a cone spray distributor; the wash is divided into multiple plug-flow sections; and, most importantly, the in-process integrated multi-pass wash results in much more efficient use of wash solvent. For example, while only 550 mL of unused ACN is used for the post-deblocking wash, the total amount of wash solvent passing through the reactor during the post-deblocking wash is 3500 mL. Similarly, while only 350 mL of unused ACN is used for the post-oxidation / thiolation wash, assuming 100 mL of post-coupling wash is pumped into bottle C2, the total amount of wash solvent passing through the reactor during the post-oxidation / thiolation wash is 1700 mL. Similarly, while only 300 mL of unused ACN is used for the post-capping wash, the total amount of wash solvent passing through the reactor during the post-capping wash is 1200 mL. Most importantly, reducing the wash does not change the wash endpoint. The wash endpoint in packed-bed synthesizer experiments using 800-1000 mL / mmol of total ACN wash solvent per cycle is 99.9% ACN, while the wash endpoint in fluidized-bed synthesizer experiments using 130 mL / mmol of total ACN wash solvent per cycle is 99.9% ACN. The difference is that the wash solvent is used more efficiently in the fluidized-bed reactor with in-process integrated multi-pass washing.

[0322] Another difference compared to Example 9 is that the post-coupling wash solvent is reused as the post-oxidation / thiolation wash solvent in Example 10. The post-coupling wash procedure is carried out as follows:

[0323] Cleaning the drive-in supply line after coupling: This is a continuation of the example using amidite valve 1104 (Figure 11). Valve 1104C is opened, vent valve 1142 is opened, and pump 1130 is used to pump a user-specified amount of ACN solvent (100 mL). Valve 1104C is then closed and valve 1104B is opened to drive the solvent into the amidite zone with nitrogen. Vent valve 1142 is closed, valve 1141 is opened, valve 1143 is opened, and vent valve 1152 is opened to push the driven wash solvent into the feed zone. The solvent is then forced through the spray cone into the reactor on top of the resin, through the resin, and pumped out the bottom of the reactor and into bottle C2. To do this, vent valve 1152 is closed, valve 1143 is closed, nitrogen valve 1151 is opened, valve 1145 is directed to the spray cone, valve 1159 is opened, valve 1154 is directed to valve 1160, valve 1160 is directed to valve 11300, valve 11300 is opened, valve 11304 is opened, and pump 1159 is turned on. At the end of the user-specified pump time, valve 1159 is closed, valve 1153 is opened, and valve 1155 is opened, allowing nitrogen to push any remaining solvent into bottle C2 until the pressure in the feed zone drops below a user-specified value (dropping from 15 psig to 9 psig).

[0324] All other parts of this procedure are the same as those described in Example 9, except that multiple pass washes are also performed after capping.

[0325] Drying: After the final cycle, DEA treatment, and washing, the resin is slurried out of the reactor. Transferred to a single plate filter. Dry for 6 hours by blowing nitrogen downward through the resin bed. The total weight of the recovered dried resin, including a small sample, was 116.24 g.

[0326] A small sample was taken for cleavage, deprotection, and UPLC, and the results are included in Table 17 (UPLC results for Examples 6-10 and comparison with Cytiva AKTA).

[0327] Examples 6-9 all synthesized the same chain. As mentioned above, Example 6 demonstrates an alternative research-scale synthesizer design. This new design does not have any feed zones for reagents other than amidites and activators. The use of multiple heads in parallel reduces the number of pumps, and solvent wash volumes are reduced by integrating solvent recycling after each phosphoramidite cycle.

[0328] Example 7 demonstrates an alternative research-scale synthesizer design that integrates the recycling of excess deblocking reagent solution after each phosphoramidite cycle, which helps reduce the amount of acid required for the deblocking reaction. This example used 29% of the DCA solution per mmol typically used in a Cytiva packed-bed synthesizer.

[0329] Example 8 demonstrates a new reactor design for scale-up. The new 10 mmol-scale reactor design uses a different wash strategy with more small-scale washes. The washes are a combination of plug flow and fluidization, designed for efficient reagent removal. A cone spray distributor is used to keep the resin bed flat, allowing for efficient plug flow washes. This example demonstrates 40% less wash solvent per mmol compared to the Cytiva AKTA synthesizer. Toluene is used as a wash solvent prior to the deblocking reaction to pre-swell the resin and remove ACN, resulting in a more efficient deblocking reaction.

[0330] In Example 9, no toluene wash was performed before deblocking. Toluene was replaced with recycled acid. The cleanest portion of the acid deblocking solution from one cycle was used to pre-swell the resin, which was then rinsed with ACN from the resin at the start of the next cycle. In Example 9, no capping was performed from cycles 2 through 9. In Example 9, an in-process integrated multi-pass wash was performed after deblocking, oxidation, and thiolation.

[0331] Example 10 is similar to Example 9, but also has an in-process integrated multi-pass wash after capping, with the post-coupling wash being reused for the post-oxidation / thiolation wash. Example 10 has the least ACN wash solvent compared to all other examples herein, which is 85% less wash solvent per mmol compared to the Cytiva AKTA synthesizer.

[0332] A guide to all fluidized bed reactor embodiments is provided in Table 31. [Table 34]

[0333] UPLC chromatogram overlays for Examples 1-5 are shown in Figures 14 and 15, and UPLC chromatogram overlays for Examples 6-10 and the "AKTA Comparative 1-4" Examples are shown in Figures 16 and 17. These show that the impurity profiles from the fluidized-bed synthesizer experiments are similar to those from the Cytiva AKTA OP100 experiments. New impurities not present in the Cytiva AKTA OP100 experiments were also absent in the fluidized-bed synthesizer experiments. However, in the high-purity experiments, the resin bed height for the AKTA synthesizer was up to 2 cm, while the resin bed height for the fluidized-bed reactor was up to 30 cm. In addition to purity considerations, the fluidized-bed synthesizer also achieved higher yields, approximately 85% fewer solvent washes, and lower DCA equivalents compared to the AKTA.

[0334] 1 shows an outline of ion-pairing UPLC conditions for analyzing the purity of the antisense strand in Examples 1 to 9. Instrument: Waters I-Class Acquity UPLC with binary pump Column: 50 x 2.1 mm Waters BEH C18, 1.7 mm, 130A (product number 186003949) Column temperature: 55℃ Mobile phase A: 10mM DIPEA, 100mM HFIP aqueous solution Mobile phase B: acetonitrile gradient ●Initial conditions: 99%A / 1%B ● Increase B from 1% to 24.3% over 25 minutes Increase B from 24.3 to 100% over 0.1 minutes ●Hold B at 100% for 1.9 minutes ● Decrease B from 100% to 1% over 0.1 minutes ●Hold B at 1% for 2.9 minutes Total runtime: 30 minutes Flow rate: 0.6mL / min Wavelength: 260nm

[0335] Illustrative Embodiments Embodiment 1. A method of adding oligonucleotides to a solid phase resin in a bed reactor, comprising: removing the protecting group from the 5' position of the solid phase resin-bound oligonucleotide; adding an activated amidite solution to the bed reactor, the activated amidite solution containing amidites, flowing up and down through the bed reactor or fluidizing by nitrogen bubbling or other agitation, to react at the 5' position of the oligonucleotide, and the phosphorous bond found in the amidites containing P atoms in oxidation state III; and converting the P atoms from oxidation state III to oxidation state V.

[0336] Embodiment 2. The method of embodiment 1, further comprising the step of adding a capping solution before or after converting the P atom from oxidation state III to oxidation state V, wherein if the coupling moiety has not reacted with the amidite solution, the capping solution caps the coupling moiety, thereby preventing further amidite from coupling to the coupling moiety, and wherein the capping solution flows up and down within the bed reactor, or fluidizes the resin using inert gas bubbling or other agitation, or flows downward through the resin bed without the fluidization portion of the fluidization / mixing or reaction followed by the plug flow portion.

[0337] Embodiment 3. The method of embodiment 1, further comprising removing the activated amidite solution from the bed reactor by passing the amidite solution through a filter located at the bottom of the bed reactor.

[0338] Embodiment 4. The method of embodiment 1, further comprising adding a first wash solution to the bed reactor, wherein the addition of the first wash solution occurs after removing the protecting group.

[0339] Embodiment 5. The method of embodiment 4, further comprising adding a second wash solution to the bed reactor, wherein the addition of the second wash solution occurs after the activated amidite solution is added to the bed reactor.

[0340] Embodiment 6. The method of embodiment 5, wherein the first and second wash solutions flow up and down within the bed reactor, and the method further comprises the step of individually removing the first and second wash solutions from the bed reactor by passing the first and second wash solutions through filters located at the bottom of the bed reactor.

[0341] Embodiment 7. The method of embodiment 5, wherein the addition of the second washing solution occurs before the step of converting P atoms from oxidation state III to oxidation state V.

[0342] Embodiment 8. The method of embodiment 5, further comprising adding a third wash solution to the bed reactor, wherein the addition of the third wash solution occurs after converting the P atoms from oxidation state III to oxidation state V.

[0343] Embodiment 9. The method of embodiment 8, wherein the third wash solution flows up and down within the bed reactor, and the method further comprises removing the third wash solution from the bed reactor by passing the third wash solution through a filter located at the bottom of the bed reactor.

[0344] Embodiment 10. The method of embodiment 1, wherein the protecting group is a DMT group, and removing the protecting group comprises reacting the 5' position of the oligonucleotide with an activating solution comprising an acid in a solvent.

[0345] Embodiment 11. The method of embodiment 10, wherein the method further comprises removing the activation solution from the bed reactor by passing the activation solution through a filter located at the bottom of the bed reactor.

[0346] Embodiment 12. The method of embodiment 1, wherein the up-and-down flow within the bed reactor is achieved by applying pressure to the top of the reactor during downward pressing, and then releasing pressure from the top of the reactor during upward pressing.

[0347] Embodiment 13. The method of embodiment 1, wherein the fluidized bed mixing of solids and liquids in the bed reactor is achieved by adding nitrogen or another gas to the bottom of the reactor or by some other type of agitation.

[0348] Embodiment 14. A system for adding oligonucleotides to a solid phase resin, comprising a bed reactor and an activated amidite solution, wherein the activated amidite solution contains amidite and flows up and down within the bed reactor or is fluidized by inert gas bubbling or other agitation.

[0349] Embodiment 15. The system of embodiment 14, wherein the bed reactor comprises an inlet that allows pressurized gas to enter the bed reactor, and the pressurized gas or some other type of agitation causes the amidite solution to mix with the solids in the bed reactor.

[0350] Embodiment 16. The system of embodiment 15, wherein the inlet is located at the bottom of the bed reactor.

[0351] Embodiment 17. The system of embodiment 14, wherein the bed reactor is pressurized from the top of the bed reactor, and the pressure causes the amidite to flow up and down within the bed reactor.

[0352] Embodiment 18. The method of embodiment 5, wherein the first and second wash solutions are mixed in the bed reactor, and the method further comprises the step of individually removing the first and second wash solutions from the bed reactor by passing the first and second wash solutions through filters located at the bottom of the bed reactor.

[0353] Embodiment 19. The method of embodiment 5, wherein the wash solvent is drained from the bottom of the filter reactor before loading the next reagent, the reagent is drained from the bottom of the filter reactor before loading the next wash solvent, and the resin bed is mixed to suspend the resin particles in the reagent and / or wash solvent by inert gas bubbling or up and down flow of liquid at selected times during selected reactions and / or washes in each cycle.

[0354] Embodiment 20. The method of embodiment 19, wherein a first portion of the reagents is charged into the reactor, the first portion is fluidized for a target time at the beginning of the reaction to achieve complete contact and achieve resin swelling, and then the first portion is pumped in a resin bed plug flow manner while a second portion of the reagents is simultaneously charged into the top of the reactor, thereby allowing the remaining reagents to be pumped in plug flow. This embodiment is demonstrated in Examples 1, 2, 3, 4, 7, 8, 9, and 10.

[0355] Embodiment 21 The method of embodiment 19, wherein the final portion of the deblocking reagent solution is recycled for each phosphoramidite cycle, thereby reducing the amount of acid required for the deblocking reaction, and the resin is swelled to reset the bed without channels at the start of deblocking, flushing out the ACN before the plug flow reaction with fresh deblocking reagent solution. This embodiment is demonstrated in Examples 7, 9, and 10.

[0356] Embodiment 22. The method of embodiment 19, wherein each wash is divided into a series of multiple smaller wash portions that are completely drained, thereby minimizing backmixing compared to one large continuous wash. This embodiment is demonstrated in Examples 1, 2, 3, 4, 6, 7, 8, 9, and 10.

[0357] Embodiment 23. The method of embodiment 19, wherein some or all of the solvent washes are not fluidized, the washes begin with a fluidized section followed by a plug flow section, or the washes have a fluidized section somewhere during or at the end of the plug flow wash, custom designed to take into account efficiency of reagent removal and when fluidization is needed to overcome pressure drops. This embodiment is demonstrated in Examples 1, 2, 3, 4, 6, 7, 8, 9, and 10.

[0358] Embodiment 24 The method of embodiment 19, wherein the incoming reagents and wash solvent are evenly distributed radially over the top of the resin bed with a spray cone or other distributor to keep the resin bed flat and allow for efficient plug flow reaction and washing. This embodiment is demonstrated in Examples 8, 9, and 10.

[0359] Embodiment 25 The method of embodiment 19, wherein the cleaner fraction of the wash solvent is recycled and reused for each phosphoramidite cycle. This embodiment is demonstrated in Examples 6, 8, 9, and 10.

[0360] Embodiment 26. The method of embodiment 19, wherein an in-process integrated multi-pass wash is used after the reaction, as described herein. Solvent portions are passed through the reactor multiple times. For example, the sixth solvent wash portion after deblocking in cycle 1 becomes the fifth wash portion after deblocking in cycle 2, which then becomes the fourth wash portion after deblocking in cycle 3, and so on. In-process integrated multi-pass washing allows for much more efficient use of the wash solvent, since only the "dirtiest" fraction of wash solvent leaves the system to waste after each reaction, and only unused solvent supply is needed for the final wash portion. This embodiment is demonstrated in Examples 9 and 10.

[0361] Embodiment 27. The method of embodiment 19, wherein the reactor has a smaller diameter lower section that expands to a larger diameter upper section when the reagents or wash solvents first enter the reactor to facilitate fluidization. Upflow inert gas pushes some or all of the resin beads up into the larger diameter section, where the liquid and solids can interact without significant wall effects. This embodiment is demonstrated in Examples 2, 4, 6, and 7.

[0362] Embodiment 28. The method of embodiment 19, wherein the resin bed is fluidized / mixed with reagent liquids during the deblocking, coupling, oxidation, sulfurization, and capping reaction steps in each cycle to achieve thorough contact and also mitigate the otherwise high pressure drop as it flows downward through the resin bed during the reaction. This embodiment is demonstrated in Examples 1, 2, 3, 4, 6, 7, 8, 9, and 10.

[0363] Embodiment 29 The method of embodiment 19, wherein an initial portion of the solvent wash is fluidized to alleviate the otherwise high pressure drop as it flows downward through the resin bed during the wash. This embodiment is demonstrated in Examples 2, 3, and 4.

[0364] Embodiment 30. The method of embodiment 19, wherein resin expansion is permitted to occur primarily during fluidization, thereby easing the pressure drop as the liquid subsequently flows downward through the bed and out the bottom of the reactor. This embodiment is demonstrated in Examples 1, 2, 3, 4, 6, 7, 8, 9, and 10.

[0365] Embodiment 31 The method of embodiment 19, wherein capping is omitted from some of the cycles. This embodiment is demonstrated in Examples 9 and 10.

[0366] Embodiment 32 The method of embodiment 19, wherein some of the reactions are not fluidized at any point in the reaction, and only plug flow contact is performed, for example, when fresh DCA solution is charged, deblocking is performed without fluidization. This embodiment is demonstrated in Examples 7, 9, and 10.

[0367] Embodiment 33 The method of embodiment 19, wherein an inert gas pushes the liquid down through the resin bed and a pump or other metering device at the outlet of the reactor controls the flow rate of the liquid through the bed. This embodiment is demonstrated in Examples 1, 2, 3, 4, 6, 7, 8, 9, and 10.

[0368] Embodiment 34 The method of embodiment 19, wherein the amidite and activator solutions are charged in separate zones, optionally mixed within the zones by inert gas bubbling, and then forced into the reactor. This embodiment is demonstrated in Examples 1, 2, 4, 6, and 7.

[0369] Embodiment 35 The method of embodiment 19, wherein the amidite and activator solutions are charged to separate zones, optionally mixed within the zones by inert gas bubbling, and then forced into the feed zone before being forced into the reactor. This embodiment is demonstrated in Examples 3, 8, 9, and 10.

[0370] Embodiment 36 The method of embodiment 19, wherein the reagents are loaded into individual feed zones before being forced into the reactor. This embodiment is demonstrated in Examples 1, 2, 4, and 7.

[0371] Embodiment 37 The method of embodiment 19, wherein the reagents are charged to a common feed zone before being forced into the reactors. This embodiment is demonstrated in Examples 3, 8, 9, and 10.

[0372] Embodiment 38 The method of embodiment 19, wherein the reagents are forced directly into the reactor rather than into a feed zone. This embodiment is demonstrated in Example 6.

[0373] Embodiment 39 The method of embodiment 19, wherein the post-coupling wash solvent is reused in the wash solvent after oxidation / thiolation. This embodiment is demonstrated in Example 10.

[0374] There are various embodiments in which products (including oligonucleotides) are made by any of the methods and / or processes and / or embodiments outlined herein. For example, the following methods can be used to make products:

[0375] 1. A method for adding oligonucleotides to a solid phase resin in a bed reactor, comprising: removing the protecting group from the 5' position of the solid phase resin-bound oligonucleotide; adding an activated amidite solution to the bed reactor, the activated amidite solution containing amidites, flowing up and down through the bed reactor or fluidizing by nitrogen bubbling or other agitation, to react at the 5' position of the oligonucleotide, and the phosphorous bond found in the amidites containing P atoms in oxidation state III; and converting the P atoms from oxidation state III to oxidation state V.

[0376] The products made by the above methods can be made using a process further comprising the step of adding a capping solution before or after converting the P atom from oxidation state III to oxidation state V, where if a coupling moiety does not react with the amidite solution, the capping solution caps the coupling moiety, thereby preventing further amidite from coupling to the coupling moiety, and where the capping solution flows up and down in a bed reactor, or uses inert gas bubbling or other agitation to fluidize the resin, or flows downward through the resin bed without fluidization / mixing or a fluidized portion of the reaction followed by a plug flow portion.

[0377] The products made by the above methods can be made using a process that further includes removing the activated amidite solution from the bed reactor by passing the amidite solution through a filter located at the bottom of the bed reactor.

[0378] The products made by the above methods can be made using a process further comprising the step of adding a first wash solution to the bed reactor, wherein the addition of the first wash solution occurs after removing the protecting groups.

[0379] The products made by the above methods can be made using a process further comprising the step of adding a second wash solution to the bed reactor, wherein the addition of the second wash solution occurs after the activated amidite solution has been added to the bed reactor.

[0380] The products made by the above methods can be made using a process, wherein the first and second wash solutions flow up and down through a bed reactor, and the process further includes the step of separately removing the first and second wash solutions from the bed reactor by passing the first and second wash solutions through filters located at the bottom of the bed reactor.

[0381] The products made by the above methods can be made using a process in which the addition of a second washing solution occurs before the step of converting the P atoms from oxidation state III to oxidation state V.

[0382] The products made by the above methods can be made using a process further comprising the step of adding a third wash solution to the bed reactor, wherein the addition of the third wash solution occurs after the conversion of P atoms from oxidation state III to oxidation state V.

[0383] The products made by the above methods can be made using a process, wherein a third wash solution flows up and down through a bed reactor, and the process further includes removing the third wash solution from the bed reactor by passing the third wash solution through a filter located at the bottom of the bed reactor.

[0384] The products produced by the above methods can be produced using a process wherein the protecting group is a DMT group and removal of the protecting group comprises reacting the 5' position of the oligonucleotide with an activating solution comprising an acid in a solvent.

[0385] The products made by the above methods can be made using a process, the process further comprising removing the activation solution from the bed reactor by passing the activation solution through a filter located at the bottom of the bed reactor.

[0386] The products made by the above methods can be made using a process in which up and down flow within a bed reactor is achieved by applying pressure to the top of the reactor during downward pressure and then releasing pressure from the top of the reactor during upward pressure.

[0387] The products made by the above methods can be made using a process in which fluidized bed mixing of solids and liquids in a bed reactor is achieved by adding nitrogen or another gas to the bottom of the reactor or some other type of agitation.

[0388] The products made by the above methods can be made using a process, wherein the first wash solution and the second wash solution are mixed in a bed reactor, the process further comprising the step of separately removing the first wash solution and the second wash solution from the bed reactor by passing the first wash solution and the second wash solution through filters located at the bottom of the bed reactor.

[0389] Products made by the above methods can be made using a process in which wash solvent is drained from the bottom of the filter reactor before loading the next reagent, reagents are drained from the bottom of the filter reactor before loading the next wash solvent, and the resin bed is mixed to suspend the resin particles in the reagent and / or wash solvent by inert gas bubbling or up and down flow of liquid at selected times during selected reactions and / or washes in each cycle.

[0390] The products made by the above methods can be made using a process in which a first quantity of reagents is charged to a reactor, the first quantity is fluidized at the start of the reaction for a target time to achieve complete contact and achieve resin swelling, and then a second quantity of reagents is simultaneously charged to the top of the reactor at the same time that the first quantity is pumped in a resin bed plug flow manner, allowing the remaining reagents to be pumped in a plug flow manner.

[0391] Products made by the above method can be made using a process in which the final portion of the deblocking reagent solution is recycled for each phosphoramidite cycle, thereby reducing the amount of acid required for the deblocking reaction, and the resin is swelled and the bed is reset without channels at the start of deblocking, and ACN is washed out before the plug flow reaction with fresh deblocking reagent solution.

[0392] Products made by the above methods can be made using a process in which each wash is divided into a series of multiple smaller wash portions that are completely drained, thereby minimizing backmixing compared to one large continuous wash.

[0393] Products made by the above methods can be made using processes where some or all of the solvent washes are not fluidized, where the washes begin with a fluidized section followed by a plug flow section, or where the washes have a fluidized section somewhere during or at the end of the plug flow wash, custom designed to take into account efficiency of reagent removal and when fluidization is needed to overcome pressure drops.

[0394] Products made by the above methods can be made using a process in which the incoming reagents and wash solvents are evenly distributed radially over the top of the resin bed with a spray cone or other distributor to keep the resin bed flat and allow for efficient plug flow reaction and washing.

[0395] The products made by the above methods can be made using a process in which the cleaner fraction of the wash solvent is recycled and reused for each phosphoramidite cycle.

[0396] The products made by the above methods can be made using a process in which an in-process integrated multi-pass wash is used after the reaction, as described herein. Solvent portions are passed through the reactor multiple times. For example, the sixth solvent wash portion after deblocking in cycle 1 becomes the fifth wash portion after deblocking in cycle 2, which then becomes the fourth wash portion after deblocking in cycle 3, and so on. In-process integrated multi-pass washing allows for much more efficient use of the wash solvent, as only the "dirtiest" fraction of wash solvent leaves the system to waste after each reaction, and only unused solvent supply is needed for the final wash portion.

[0397] Products made by the above methods can be made using a process in which the reactor has a smaller diameter lower section that expands to a larger diameter upper section when the reagents or wash solvents first enter the reactor to facilitate fluidization. Upflow inert gas pushes some or all of the resin beads up into the larger diameter section where the liquid and solids can interact without significant wall effects.

[0398] The products made by the above methods can be made using a process in which the resin bed is fluidized / mixed with reagent liquids during the deblocking, coupling, oxidation, sulfiding, and capping reaction steps in each cycle to achieve complete contact and also mitigate the otherwise high pressure drop as it flows downward through the resin bed during the reaction.

[0399] The products made by the above methods can be made using a process in which the initial portion of the solvent wash is fluidized to alleviate the otherwise high pressure drop as it flows downward through the resin bed during the wash.

[0400] Products made by the above methods can be made using a process in which resin expansion is allowed to occur primarily during fluidization, thereby easing the pressure drop as the liquid subsequently flows downward through the bed and out the bottom of the reactor.

[0401] Products made in the above manner can be made using a process in which capping is omitted from some of the cycles.

[0402] The products made by the above methods can be made using a process where some of the reactants are not fluidized at any point in the reaction and only plug flow contacting occurs, e.g., unblocking occurs without fluidization when fresh DCA solution is charged.

[0403] The products made by the above methods can be made using a process in which an inert gas forces a liquid down through a resin bed and a pump or other metering device at the outlet of the reactor controls the flow rate of the liquid through the bed.

[0404] The products made by the above methods can be made using a process in which the amidite and activator solutions are charged to separate zones, optionally mixed within the zones by inert gas bubbling, and then forced into the reactor.

[0405] The products made by the above methods can be made using a process in which the amidite and activator solutions are charged to separate zones, optionally mixed within the zones by inert gas bubbling, and then forced into a feed zone before being forced into the reactor.

[0406] Products made by the above methods can be made using a process in which the reagents are charged to individual feed zones before being forced into the reactor.

[0407] Products made by the above methods can be made using a process in which the reagents are charged to a common feed zone before being forced into the reactor.

[0408] The products made by the above methods can be made using a process in which the reagents are forced directly into the reactor rather than into a feed zone.

[0409] The products made by the above methods can be made using a process in which the post-coupling wash solvent is recycled in the wash solvent after oxidation / thiolation.

[0410] Alternatively, other products may be made using other methods.

Claims

1. 1. A method for adding oligonucleotides to a solid support in a bed reactor, comprising: removing a protecting group from the 5' position of the solid support-bound oligonucleotide; adding an activated amidite solution to the bed reactor, the activated amidite solution containing amidites, flowing up and down in the bed reactor or fluidized by nitrogen bubbling or other agitation, to react at the 5' position of the oligonucleotide, and the phosphorous bond found in the amidites containing P atoms in oxidation state III; and converting said P atoms from oxidation state III to oxidation state V.

2. 10. The method of claim 1, further comprising the step of adding a capping solution before or after converting the P atom from oxidation state III to oxidation state V, wherein if a coupling moiety does not react with the amidite solution, the capping solution caps the coupling moiety, thereby preventing further amidite from coupling to the coupling moiety, and wherein the capping solution flows up and down within the bed reactor, or is fluidized by nitrogen bubbling or other agitation, or flows downward through a resin bed without fluidization / mixing or the fluidized portion of the reaction followed by the plug flow portion.

3. 10. The method of claim 1, further comprising the step of removing the activated amidite solution from the bed reactor by passing the amidite solution through a filter located at the bottom of the bed reactor.

4. 10. The method of claim 1, further comprising the step of adding a first wash solution to said bed reactor, said adding of said first wash solution occurring after removing said protecting groups.

5. 5. The method of claim 4, further comprising the step of adding a second wash solution to the bed reactor, wherein the addition of the second wash solution occurs after the activated amidite solution is added to the bed reactor.

6. 6. The method of claim 5, wherein the first wash solution and the second wash solution flow up and down within the bed reactor, and the method further comprises the step of individually removing the first wash solution and the second wash solution from the bed reactor by passing the first wash solution and the second wash solution through a filter located at the bottom of the bed reactor.

7. 6. The method of claim 5, wherein said adding said second cleaning solution occurs before the step of converting said P atoms from oxidation state III to oxidation state V.

8. 6. The method of claim 5, further comprising the step of adding a third wash solution to said bed reactor, said adding of said third wash solution occurring after converting said P atoms from oxidation state III to oxidation state V.

9. 9. The method of claim 8, wherein the third wash solution flows up and down within the bed reactor, and the method further comprises the step of removing the third wash solution from the bed reactor by passing the third wash solution through a filter located at the bottom of the bed reactor.

10. 2. The method of claim 1, wherein the protecting group is a DMT group and the removal of the protecting group comprises reacting the 5' position of the oligonucleotide with an activating solution comprising an acid in a solvent.

11. 11. The method of claim 10, wherein the method further comprises purging the activation solution bed reactor by passing the activation solution through a filter located at the bottom of the bed reactor.

12. 10. The method of claim 1, wherein the up-and-down flow within the bed reactor is achieved by applying pressure to the top of the reactor during downward pushing and then releasing pressure from the top of the reactor during upward pushing.

13. 10. The method of claim 1, wherein fluidized bed mixing of solids and liquids in the bed reactor is achieved by adding nitrogen or another gas to the bottom of the reactor or by some other type of agitation.

14. A system for adding oligonucleotides to a solid support, comprising a bed reactor and an activated amidite solution, wherein the activated amidite solution contains an amidite and flows up and down within the bed reactor or is fluidized by nitrogen bubbling or other agitation.

15. 15. The system of claim 14, wherein the bed reactor comprises an inlet that allows pressurized gas to enter the bed reactor, and the pressurized gas or some other type of agitation causes the amidite solution to mix with the solids in the bed reactor.

16. 16. The system of claim 15, wherein the inlet is located at the bottom of the bed reactor.

17. 15. The system of claim 14, wherein the bed reactor is pressurized and depressurized from the top of the bed reactor, and the pressure fluctuations cause the amidite to flow up and down within the bed reactor.

18. 6. The method of claim 5, wherein the first wash solution and the second wash solution mix in the bed reactor, and the method further comprises the step of individually removing the first wash solution and the second wash solution from the bed reactor by passing the first wash solution and the second wash solution through filters located at the bottom of the bed reactor.

19. 6. The method of claim 5, wherein wash solvent is drained from the bottom of the filter reactor before loading the next reagent, the reagents are drained from the bottom of the filter reactor before loading the next wash solvent, and the resin bed is mixed to suspend the resin particles in the reagents and / or wash solvent by inert gas bubbling or up and down flow of the liquid at selected times during selected reactions and / or washes in each cycle.

20. 20. The method of claim 19, wherein a first portion of the reagents is charged to the reactor, the first portion is fluidized at the beginning of the reaction for a target time to achieve complete contact and achieve resin swelling, and then a second portion of the reagents is simultaneously charged to the top of the reactor at the same time that the first portion is pumped in a plug flow manner into the resin bed, thereby allowing the remaining reagents to be pumped in a plug flow manner.

21. 20. The method of claim 19, wherein a final portion of the deblocking reagent solution is recycled for each phosphoramidite cycle, thereby reducing the amount of acid required for the deblocking reaction, and the resin is swelled to reset the bed without channels at the start of deblocking, flushing out ACN before the plug flow reaction with fresh deblocking reagent solution.

22. 20. The method of claim 19, wherein each wash is divided into a series of smaller wash portions that are completely drained, thereby minimizing backmixing compared to one large continuous wash.

23. 20. The method of claim 19, wherein some or all of the solvent washes are not fluidized, the washes begin with a fluidized section followed by a plug flow section, or the washes have a fluidized section somewhere during or at the end of the plug flow wash, custom designed to take into account efficiency of reagent removal and when fluidization is needed to overcome pressure drops.

24. 20. The method of claim 19, wherein the incoming reagents and wash solvent are uniformly distributed radially over the top of the resin bed with a spray cone or other distributor to keep the resin bed flat and allow for efficient plug flow reaction and washing.

25. 20. The method of claim 19, wherein the cleaner fraction of the wash solvent is recycled and reused for each phosphoramidite cycle.

26. As described herein, in-process integrated multi-pass washing is used after the reaction, in the method of claim 19, where the solvent portion is passed through the reactor multiple times. For example, the sixth solvent wash portion after deblocking in cycle 1 becomes the fifth wash portion after deblocking in cycle 2, which then becomes the fourth wash portion after deblocking in cycle 3, etc. In-process integrated multi-pass washing allows for much more efficient use of the wash solvent, as only the "dirtiest" wash solvent leaves the system to waste after each reaction, and only a new clean solvent supply is needed for the final wash section.

27. 20. The method of claim 19, wherein the reactor has a smaller diameter lower section that expands to a larger diameter upper section when the reagents or the wash solvent initially enter the reactor to facilitate fluidization. The upflow inert gas pushes some or all of the resin beads up into the larger diameter section where the liquid and solids can interact without much wall effect.

28. 20. The method of claim 19, wherein the resin bed is fluidized / mixed with reagent liquids in each cycle during the other reaction steps to achieve thorough contact and also mitigate the otherwise high pressure drop as it flows downward through the resin bed during reaction.

29. 20. The method of claim 19, wherein an initial portion of the solvent wash is fluidized to alleviate an otherwise high pressure drop as it flows downward through the resin bed during the wash.

30. 20. The method of claim 19, wherein the resin expansion is permitted to occur primarily during fluidization, thereby mitigating the pressure drop as liquid subsequently flows downward through the bed and out the bottom of the reactor.

31. 20. The method of claim 19, wherein capping is omitted from some of the cycles.

32. 20. The method of claim 19, wherein some of the reactions are not fluidized at any point in the reaction and unblocking is performed by plug flow contact only, e.g., without fluidization, when the fresh DCA solution is charged.

33. 20. The method of claim 19, wherein an inert gas pushes liquid down through the resin bed and a pump or other metering device at the outlet of the reactor controls the flow rate of liquid through the bed.

34. 20. The method of claim 19, wherein the amidite and activator solutions are charged in separate zones, optionally mixed in said zones by inert gas bubbling, and then forced into the reactor.

35. 20. The method of claim 19, wherein the amidite and activator solutions are charged to separate zones, optionally mixed in said zones by inert gas bubbling, and then forced into a feed zone before being forced into the reactor.

36. 20. The method of claim 19, wherein reagents are loaded into individual feed zones before being forced into the reactor.

37. 20. The method of claim 19, wherein the reagents are charged to a common feed zone before being forced into the reactor.

38. 20. The method of claim 19, wherein reagents are forced directly into the reactor rather than into a feed zone.

39. 20. The method of claim 19, wherein the post-coupling wash solvent is recycled in the wash solvent after oxidation / thiolation.

Images