Methods and compositions for improving recombinant protein expression
Co-expression of phospholipases with recombinant proteins in a reduced genome E. coli strain targets proteins to the periplasm for high-yield, continuous fermentation and simplified purification, addressing inefficiencies in current production methods.
Patent Information
- Application Number
- JP2023055602
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2017-09-05
- Filing Date
- 2023-03-30
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2038-09-04
AI Technical Summary
Current methods for producing recombinant proteins in E. coli, such as CRM197, are expensive, time-consuming, and inefficient due to low yields, insoluble accumulation, and complex purification processes, which are exacerbated by potential toxicity and stress in host cells.
Co-expression of recombinant proteins with E. coli phospholipases in a reduced genome strain, followed by extraction into the culture medium using a Tris-EDTA buffer, facilitates high-yield and continuous production of proteins like CRM197, PD, and EPA by targeting them to the periplasm.
This method enables high-yield, continuous fermentation of recombinant proteins, simplifying purification and reducing production costs, with yields exceeding 3g/L/day and maintaining production for at least 30 days.
Smart Images

Figure 0007590485000007 
Figure 0007590485000008 
Figure 0007590485000009
Abstract
Description
[Technical field]
[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Patent Application No. 62 / 554,443, filed Sep. 5, 2017, the entire disclosure of which is incorporated herein by reference.
[0002] Methods and compositions are provided for enhancing the expression and recovery of difficult to produce proteins, such as single chain antibodies, single chain fragments, and those suitable as carriers for polysaccharides and haptens. [Background technology]
[0003] Thirty percent of FDA-approved recombinant therapeutic proteins are made in E. coli, including key vaccine components such as carrier proteins, single-chain antibodies, and single-chain antibody fragments. However, the standard method for producing recombinant proteins from E. coli has remained unchanged for decades. Traditional fed-batch fermenters are used to cultivate larger and larger batches of bacteria containing the target, which are then subjected to large-scale homogenization before the often complex and laborious task of protein purification. This process is costly, time-consuming, and often requires difficult and extensive purification steps, resulting in yield losses.
[0004] Cross-reactive substance 197 (CRM197), a mutant non-toxic form of diphtheria toxin, is an approved carrier protein used in a variety of successful conjugate vaccines. However, due to issues with its production, the demand for CRM197 exceeds supply. The original host, Corynebacterium diphtheriae, produces low levels of CRM197 (0.1-0.2 g / L), requiring biosafety level 2 facilities. Potential reversion of CRM197 to wild-type toxin remains an issue in alternative hosts, such as Pseudomonas fluorescens and most E. coli strains, and residual CRM197 toxicity causes significant physiological stress in these hosts, limiting production. Furthermore, in most E. coli hosts, CRM197 accumulates in an insoluble form (aggregating into inclusion bodies), further reducing yields and complicating downstream purification steps. Many E. coli strains favored for fermentation can 1) affect the expression vector or 2) destroy the host, in both cases resulting in rapid and total production loss in batch mode and precluding the use of the E. coli host in extended continuous fermentation processes.
[0005] Due to the importance of difficult to produce carrier proteins such as nonacylated protein D (PD) CRM197 of Haemophilus influenzae and carrier protein exoprotein A (EPA) of Pseudomonas aeruginosa, as well as single chain antibodies, single chain antibody fragment scFab YMF10, single chain antibody variable fragment scFv 75127, there is a need for improved and cost-effective methods for producing and purifying such proteins from E. coli host cells. Summary of the Invention [Means for solving the problem]
[0006] In some embodiments, methods are provided for achieving periplasmic targeted release of a recombinant protein in E. coli, comprising co-expressing an endogenous E. coli phospholipase with the recombinant protein, hi some related embodiments, the E. coli host strain is contacted with an extraction buffer after induction of the recombinant protein and the phospholipase.
[0007] In some embodiments, a method for producing a recombinant protein of interest is provided that includes fermentation of a recombinant (e.g., reduced genome) E. coli strain carrying a heterologous nucleic acid comprising a nucleotide sequence encoding the recombinant protein of interest fused to a signal sequence that directs import of the recombinant protein of interest into the periplasm, the nucleotide sequence being operably linked to an expression control sequence, and the E. coli strain co-expressing the recombinant protein of interest and one or more E. coli phospholipases. Exemplary periplasmic targeting signals include those described in U.S. Patent Application Publication No. 2017 / 0073379, paragraphs
[0055] to
[0059] , the entire disclosure of which is incorporated herein by reference. Representative signal sequences capable of directing CRM197 to the periplasm are listed at the end of paragraph
[0057] of US2017 / 0073379, and are preferably selected from OmpA, MalE, HdeA, OppA, HdeB, GlnH, MglB, agp, OmpC, RbsB, FkpA, or YtfQ signal sequences, more preferably selected from OmpA, OmpF, or YtfQ signal sequences. In some preferred embodiments, the YtfQ signal sequence directs PD, EPA, or CRM197 to the periplasm. In some preferred embodiments, the OmpA signal sequence directs the single chain antibody fragment scFab YMF10, single chain antibody variable fragment scFv 75127 to the periplasm.
[0008] According to the methods described herein, any E. coli phospholipase can be co-expressed with a protein of interest in an E. coli host. In one embodiment, the E. coli strain comprises a nucleic acid segment comprising a nucleotide sequence encoding a glycerophosphoryl diester phosphodiesterase (glpQ; b2239; GenBank Ref. No. NP_416742.1; SEQ ID NO: 1). In another embodiment, the E. coli strain comprises a nucleic acid segment comprising a nucleotide sequence encoding an Aes (ybaC; b0476; GenBank Ref. No. NP_415009.1; SEQ ID NO: 2). In another embodiment, the E. coli strain comprises a nucleic acid segment comprising a nucleotide sequence encoding a lysophospholipase L2 (pldB; b3825; GenBank Ref. No. YP_026266.1; SEQ ID NO: 3). In a preferred embodiment, the E. coli strain comprises a nucleic acid segment comprising a nucleotide sequence encoding OmpLA (pldA; b3821; GenBank Ref. NO.NP_418265.1; SEQ ID NO:4).
[0009] By "co-expression" is meant that the expression of the recombinant protein of interest and the phospholipase are coordinated so that the expression of the protein of interest and the phospholipase at least overlaps. In some preferred embodiments, the expression of the recombinant protein of interest and the expression of the phospholipase are under the control of one or more inducible promoters (e.g., on the same or separate expression vectors). In some aspects, the expression of the recombinant protein of interest and the phospholipase are under the control of the same inducible promoter (e.g., on the same or separate expression vectors). The expression of the protein of interest and the expression of the phospholipase can be induced at about the same time, the protein of interest can be induced before the phospholipase, or the phospholipase can be induced before the protein of interest. In some preferred embodiments, the protein of interest is induced before the induction of the phospholipase (e.g., 1 hour to 6 hours before). Preferably, the E. coli phospholipase is overexpressed and recombinantly produced relative to endogenous levels of the phospholipase (e.g., by introducing into E. coli an expression vector comprising a nucleotide sequence encoding the phospholipase operably linked to an expression control sequence, or by manipulating the promoter of the gene encoding the endogenous phospholipase). Alternatively, the endogenous phospholipase gene may be manipulated to overexpress the phospholipase. According to the methods described herein, any E. coli phospholipase may be co-expressed with the recombinant protein of interest. In some preferred embodiments, the E. coli phospholipase is OmpLA.
[0010] In a preferred embodiment, the nucleotide sequence encoding the protein of interest and the nucleotide sequence encoding the phospholipase are present on the same construct (e.g., an expression vector) and are either under the control of the same inducible promoter or under the control of different inducible promoters. In other embodiments, the nucleic acid segment comprising the nucleotide sequence encoding the protein of interest and the nucleic acid segment comprising the nucleotide sequence encoding the E. coli phospholipase are on separate constructs and may or may not be under the control of the same inducible promoter.
[0011] In some aspects, methods are provided for producing a recombinant protein of interest, comprising: (i) fermenting a recombinant (e.g., reduced genome) E. coli strain harboring a heterologous nucleic acid comprising: (a) a nucleotide sequence encoding a recombinant protein of interest fused to a sequence that targets the protein of interest to the periplasm, operably linked to an expression control sequence; and (b) a nucleotide sequence encoding an E. coli phospholipase operably linked to an expression control sequence; (ii) inducing expression of the recombinant protein of interest and the phospholipase; and (iii) extracting the protein of interest from the periplasm into the medium by contacting the recombinant E. coli strain with an extraction buffer.
[0012] In some embodiments, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95% or at least 98% of the protein of interest is released from the periplasm according to these methods.
[0013] In a preferred embodiment, the extraction buffer comprises a buffering agent (e.g., Tris) at a pH value of about 7-9 and a chelating agent (e.g., EDTA). In some embodiments, the chelating agent is selected from the group consisting of ethylenediaminetetraacetic acid (EDTA), ethylenediamine,nitrilotriacetic acid (NTA), ethyleneglycoltetraacetic acid (EGTA), diethylenetriaminepentaacetic acid (DTPA), 1,2-bis(o-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA), ethylenediamine-N,N'-disuccinic acid (EDDS), ethylenediaminetetra(methylenephosphonic acid) (EDTMP) and diethylenetriaminepenta(methylenephosphonic acid) (DTPMP). In a preferred embodiment, the chelating agent is EDTA. In a particularly preferred embodiment, the extraction buffer comprises 50 mM to 200 mM Tris, preferably about 150 mM Tris, and 2 mM to 10 mM ethylenediaminetetraacetic acid (EDTA), preferably about 5 mM EDTA, at a pH of about 7 to about 9, preferably about pH 7.75. In some embodiments, extraction of the protein of interest from the periplasm into the medium is achieved by contacting the recombinant E. coli with the extraction buffer for at least 2 hours, at least 3 hours, at least 4 hours, at least 6 hours, at least 8 hours, at least 10 hours, at least 12 hours, at least 13 hours, at least 14 hours, at least 15 hours, at least 16 hours, at least 18 hours, at least 20 hours, at least 24 hours or more, or any range therebetween (e.g., between 10-20 hours or between 12-16 hours). Preferably, the recombinant E. coli is contacted with the extraction buffer for at least 2 hours and no more than 24 hours, more preferably at least 6 hours and no more than 24 hours. Preferably, the contacting step is carried out at 20-30°C, more preferably at about 25°C.
[0014] In some embodiments, the method comprises batch fermentation. In other embodiments, the method comprises fed-batch fermentation. In preferred embodiments, the method comprises continuous fermentation. In particularly preferred embodiments, the method comprises continuous fermentation of reduced genome strains of E. coli. Reduced genome E. coli, including between about 5% and about 30% genome deletions, can be continuously fermented for at least 7 days or more, at least 14 days or more, at least 21 days or more, at least 30 days or more, at least 45 days or more, at least 60 days or more, at least 180 days or more, at least 240 days or more, or even at least 360 days or more, or any range in between. In contrast, E. coli strains commonly used for fermentation are subjected to stress induction systems that preclude their use in continuous manufacturing. In some embodiments, the continuous fermentation method comprises a two-vessel chemostat strategy that separates the growth of the seed culture from the induction phase. Briefly, freshly grown cells are fed from an uninduced seed vessel (bioreactor 1) to a production vessel (bioreactor 2) where an inducer is added and product synthesis is initiated. The steady-state chemostat comprises an uninduced seed vessel that provides a supply of healthy cells to the production vessel for induction, ensuring that seed cell growth is stress-free until introduction into the production vessel. Exemplary continuous fermentation apparatus include the apparatus illustrated in FIG. 17 and the apparatus described in U.S. Patent Publication No. 2018 / 087018, the contents of which are incorporated herein by reference, although any suitable continuous fermentation method and apparatus may be used to carry out the methods described herein.
[0015] A reduced genome E. coli strain comprising a nucleic acid segment encoding a protein of interest fused to a periplasmic targeting signal and co-expressing a phospholipase can be continuously fermented for 7 days or more, 14 days or more, 30 days or more, 45 days or more, 60 days or more, 180 days or more, 240 days or more, or even 360 days or more. In a preferred embodiment, a reduced genome E. coli strain comprising a nucleic acid segment encoding a carrier protein selected from CRM197, PD and EPA and co-expressing a phospholipase is continuously fermented for at least 5-65 days (or any range therebetween), thereby resulting in a yield of at least 1-9 grams / liter / day of carrier protein in the medium during continuous fermentation. Because the E. coli host cells remain intact during the continuous fermentation process and the medium is substantially free of contaminating cellular proteins and nucleic acids resulting from cell lysis, the carrier protein can be purified directly from the medium.
[0016] In another embodiment, the single chain antibody fragment scFab YMF10, the single chain antibody variable fragment scFv Of particular importance for the enhancement of difficult to produce recombinant proteins such as 75127, conjugated carrier proteins CRM197, PD, and EPA, genetically modified E. coli strains are provided that have improved properties that facilitate the recovery of these proteins from fermentation.
[0017] In some embodiments, an E. coli K-12 or B strain is provided that comprises a deletion of one or more toxin-antitoxin (TA) genes. In some embodiments, the E. coli K-12 or B strain comprises a chromosomal deletion of one or more TA genes. In other embodiments, the E. coli K-12 or B strain comprises a deletion of one or more TA genes from a plasmid. In preferred embodiments, the E. coli K-12 or B strain does not comprise a chromosomal or plasmid TA gene. In some embodiments, a wild-type E. coli K-12 or B strain is modified to delete one or more TA genes. In other embodiments, a reduced genome E. coli K-12 or B strain (e.g., an E. coli K-12 or B strain that has been genetically modified to be about 5% to about 30% smaller than the genome of its native parent E. coli K-12 strain) is modified to delete one or more TA genes.
[0018] The TA genes deleted from E. coli K-12 or B strains were type 1 (e.g., ldrA, rdlA, ldrB, rdlB, ldrC, rdlC, ldrD, rdlD, hokB, sokB, sibA, ibsA, sibB, ibsB, ohsC, shoB, sibC, ibsC, sibD, ibsD, dinQ, agrA, agrB, yhjJ, yhjM, yhjN, istR-2, tisB), type 2 (YafQ, dinJ, hha, to The TA genes can be mB, gnsA, ymcE, yoeB, yefF, yefM, mazF, mazE, mazG, cptA / ygfX, cptB / sdhE, mqsR, mqsA, higB, higA, yhaV, prlF, ldrD, rdlD, istr-2, tisB, chpB, chpS, ratA, ratB, ecnA, ecnB), type 4 (cbtA, cbeA), and / or type 5 (egghoT, ghoS) TA genes.
[0019] In some embodiments, one or more, and preferably all, of the following TA genes are deleted from the E. coli strain: YafQ, dinJ, hha, tomB, gnsA, ymcE, yoeB, yefM, mazF, mazE, mazG, cptA / ygfX, cptB / sdhE, mqsR, mqsA, higB, higA, yhaV, prlF, ldrD, rdlD, istR-2, tisB, chpB, chpS, ra tA, ratB, ldrA, rdlA, ldrB, rdlB, ldrC, rdlC, hokB, sokB, sibA, ibsA, sibB, ibsB, ohsC, shoB, sibC, ibsC, sibD, sibE, ibsD, ibsE, dinQ, agrA, agrB, ghoT, ghoS, yfeC, yfeD, fic, yhfG, yhjJ, yhjM, yhjN, yjjJ, ecnA, and / or ecnB. In some embodiments, hipA and / or hipB are deleted.
[0020] In another embodiment, one or more of the TA genes listed in Table 1 below are deleted from the E. coli strain. In a preferred embodiment, all type II TA genes listed in Table 1 are deleted. In a particularly preferred embodiment, all TA genes listed in Table 1 are deleted. [Table 1-1] [Table 1-2]
[0021] In other embodiments, E. coli K-12 or B strains are provided that have been genetically modified to reduce or eliminate expression of the rnc(b2567) gene (encoding ribonuclease III). Preferably, expression of the rnc gene product is reduced or eliminated by partial or complete deletion of the rnc gene. In some embodiments, wild-type E. coli K-12 or B strains are modified to reduce or eliminate expression of the rnc gene. In preferred embodiments, a reduced genome E. coli K-12 strain (e.g., an E. coli strain that has been genetically modified to be about 2% or about 5% to about 22% or about 30% smaller than the genome of its native parent E. coli K-12 strain) is modified to reduce or eliminate expression of the rnc gene.
[0022] In some embodiments, E. coli K-12 or B strains are provided that contain a partial or complete deletion of the rnc gene and that contain a deletion of one or more, preferably all, of the TA genes listed in Table 1. In a preferred embodiment, reduced genome E. coli K-12 strains are provided that contain a partial or complete deletion of the rnc gene and that contain a deletion of one or more, preferably all, of the TA genes listed in Table 1.
[0023] In other embodiments, E. coli K-12 strains are provided that contain a partial or complete deletion of the rnc gene and / or a deletion of one or more, and preferably all, of the TA genes listed in Table 1, and further contain the following modifications: (a) enhanced orotate phosphoribosyltransferase activity, (b) production of active acetohydroxyacid synthase II, and (c) reduced expression of the iclR and arpA genes.
[0024] In related embodiments, the present invention comprises a partial or complete deletion of the rnc gene, and / or a deletion of one or more pairs of TA genes listed in Table 1, and preferably all pairs of TA genes, and / or the following modifications: (a) enhanced orotate phosphoribosyltransferase activity, (b) production of active acetohydroxyacid synthase II, and (c) reduced expression of the iclR and arpA genes, and includes genes including a non-functional dinB (b0231, coordinates 250898-251953 on the MG1655 genome) and optionally a non-functional polB (b0060, E. coli K12 In one embodiment, an E. coli K-12 strain is provided that comprises the umuDC gene (b1183-b1184, coordinates 63429-65780 on the MG1655 genome) and / or the umuDC gene (b1183-b1184, coordinates 1229990-1231667 on the MG1655 genome). Preferably, the (these) genes are inactivated by complete or partial deletion.
[0025] In some embodiments, reduced genome E. coli K-12 or B strains are provided having a genome that is genetically engineered to be about 2% or about 5% to about 22% or about 30% smaller than the genome of its native parent strain, where the reduced genome E. coli includes one or more deletions or modifications described herein and does not include an insertion sequence.
[0026] Reduced genome bacteria can be produced by deleting selected genes from a natural parent strain of the bacteria, or can be completely synthetic, for example, as an assembly of preselected genes. As will be readily apparent from the discussion herein, reduced genome bacteria contain less than the full complement of genes found in the compared natural parent strain and share certain essential genes.
[0027] Also provided are methods of producing a polypeptide of interest utilizing the modified E. coli K-12 or B strain of the invention, wherein the soluble polypeptide of interest is targeted to the periplasm and / or secreted into the fermentation medium. In one embodiment, the polypeptide is produced by culturing a modified E. coli K-12 or B strain comprising a nucleic acid encoding the polypeptide operably linked to an expression control sequence under conditions suitable for expressing the polypeptide. In one aspect, the nucleic acid encodes a heterologous protein. In a related aspect, the protein is produced in the modified E. coli strain of the invention at a higher level than in an E. coli K-12 or B strain not comprising the modifications described herein. In yet another aspect, protein production is increased to a higher level than in an E. coli K-12 or B strain not comprising the modifications described herein due to the higher cell numbers that such modified cells can achieve under the same growth conditions. In some embodiments, an engineered E. coli K-12 or B strain harboring a heterologous nucleic acid comprising a nucleotide sequence encoding a protein of interest fused to a periplasmic signal peptide operably linked to an expression control sequence, as described herein, and optionally co-expressing an E. coli phospholipase, is subjected to continuous fermentation, whereby a yield of at least 1, at least 2, at least 5, at least 7, or even at least 10 grams / liter / day is obtained directly from the periplasm (e.g., if a phospholipase is not co-expressed) and / or from the medium (if a phospholipase is co-expressed) for a period of at least 1 day, at least 2 days, at least 3 days, at least 4 days, at least 7 days, at least 14 days, at least 30 days, at least 60 days, at least 90 days, at least 120 days, or even at least 365 days.
[0028] Methods for amplifying heterologous nucleic acids (e.g., vectors such as plasmids) utilizing the modified E. coli K-12 or B strains of the invention are also provided. In one embodiment, nucleic acids are produced by culturing modified E. coli K-12 or B strains containing heterologous nucleic acids under suitable nutrient conditions, thereby amplifying the nucleic acids. In yet another aspect, nucleic acid production is increased to a higher level than E. coli K-12 or B strains not containing the modifications described herein due to the greater cell numbers that such modified cells can achieve under the same growth conditions. The present invention provides, for example, the following items. (Item 1) 1. A method for extracting a recombinant polypeptide from the periplasm of an E. coli host cell, the host cell comprising a first nucleotide sequence encoding the recombinant polypeptide operably linked to an expression control sequence, and a second nucleotide sequence encoding an E. coli phospholipase operably linked to an expression control sequence, the method comprising: (a) culturing the E. coli host cell in a medium under conditions that allow for co-expression of the recombinant polypeptide and the phospholipase in the E. coli host cell; (b) optionally contacting said E. coli host cells containing the expressed recombinant polypeptide and the expressed phospholipase with an extraction buffer containing a chelating agent; (b) optionally, collecting said recombinant polypeptide from said culture medium. (Item 2) (a) culturing the E. coli host cell in a medium under conditions that allow for co-expression of the recombinant polypeptide and the phospholipase in the E. coli host; (b) contacting the E. coli host cells containing the expressed recombinant polypeptide and the expressed phospholipase with an extraction buffer containing a chelating agent. (Item 3) 3. The method according to item 1 or 2, wherein the recombinant polypeptide is a carrier protein selected from CRM197, protein D and exoprotein A. (Item 4) The method according to any one of items 1 to 3, wherein the phospholipase is selected from GlpQ, YbaC, OmpLA, PldB, and TesA. (Item 5) 5. The method according to item 4, wherein the phospholipase is at least 80%, at least 85%, at least 90%, at least 95%, at least 98%, at least 99% identical or identical to any one of the amino acid sequences set forth as SEQ ID NOs: 1 to 5. (Item 6) 5. The method according to claim 4, wherein the phospholipase is pldA and is at least 80%, at least 85%, at least 90%, at least 95%, at least 98%, at least 99% identical or identical to the amino acid sequence set forth as SEQ ID NO:4. (Item 7) The method according to any one of items 1 to 6, wherein the E. coli host cell comprises an expression vector comprising the first nucleotide sequence and / or the second nucleotide sequence. 8. The method of claim 7, wherein one or both of the expression control sequences comprises an inducible promoter. (Item 9) 9. The method of claim 8, wherein both of the expression control sequences comprise an inducible promoter, and the inducible promoters are different. (Item 10) 9. The method of claim 8, wherein both of the expression control sequences comprise an inducible promoter, and the inducible promoters are the same. (Item 11) 11. The method according to any one of items 1 to 10, wherein the first nucleotide sequence encodes a recombinant polypeptide fused to a signal sequence that directs translocation of the recombinant polypeptide to the periplasm. (Item 12) 12. The method of claim 11, wherein the sequence directing the recombinant polypeptide to the periplasm is selected from the OmpA, MalE, HdeA, OppA, HdeB, GlnH, MglB, agp, OmpC, RbsB, FkpA or YtfQ signal sequences, more preferably the OmpA, OmpF or YtfQ signal sequence. (Item 12) 12. The method according to any one of items 1 to 11, wherein the extraction buffer in step (b) comprises 1 to 10 mM ethylenediaminetetraacetic acid (EDTA), preferably 3 mM to 7 mM EDTA, more preferably about 5 mM EDTA. (Item 13) The extraction buffer is 50 to 200 mM Tris, preferably 100 mM to 200 mM 13. The method of claim 12, comprising Tris, more preferably about 150 mM Tris. (Item 14) 14. The method according to item 12 or 13, wherein the extraction buffer has a pH of 6.5 to 9, preferably 7 to 8, more preferably about 7.75 to 8. (Item 15) 15. The method according to any one of items 1 to 14, wherein step (b) comprises incubating the E. coli host cells in an extraction buffer for at least 1 hour and less than 24 hours, preferably at least 6 hours and less than 24 hours, at a temperature between 20°C and 30°C, preferably at about 25°C. (Item 16) 16. The method of claim 15, wherein step (b) comprises shaking. (Item 17) 17. The method according to any one of items 1 to 16, wherein the E. coli host cell is a wild-type strain, preferably a wild-type K-12 or B strain. (Item 18) 17. The method of any one of items 1 to 16, wherein the E. coli host cell is a reduced genome strain having a genome that is genetically engineered to be about 2% to about 30% smaller than the genome of its native parent strain. (Item 19) 19. The method of claim 18, wherein the native parent strain of the E. coli host cell is a K-12 or B strain. (Item 20) The E. coli host cell is a recombinant host cell that is a recombinant host cell of the E. coli K-12 strain MG1655. The recombinant host cell is a recombinant host cell of the E. coli K-12 strain MG1655. b2968~b2972, b2975~b2977, b2979~b2987, b4466~4468, b1137~b1172, b0537~b0565, b0016~b0022, b4412~b 4413, b0577~b0582, b4415, b2389~b2390, b2392~b2395, b0358~b0368, b0370~b0380, b2856~b2863, b3042~b 3048, b0656, b1325~b1333, b2030~b2062, b2190~b2192, b3215~b3219, b3504~b3505, b1070~b1083, b1878~b 1894, b1917~b1950, b4324~b4342, b4345~b4358, b4486, b0497~b0502, b0700~b0706, b1456~b1462, b3481~b 3484, b3592 to b3596, b0981 to b0988, b1021 to b1029, b2080 to b2096, b4438, b3440 to b3445, b4451, b3556 to b3558, b4455, b1786, b0150 to b0153 and b2945, or the corresponding genes in a different K-12 or B strain. (Item 21) The E. coli host cell is a strain of E. coli K-12 strain MG1655 that is at least one of the following genes: b0245 to b0301, b0303 to b0310, b1336 to b1411, b4426 to b4427, b2441 to b2450, b2622 to b2654, b2657 to b2660, b4462, b1994 to b2008, b4435, b3322 to b3338, b2349 to b2363, b1539 to b1579, b4269 to b4320, b2968 to b2972, b2975 to b2977, b2979 to b2987, b4466 to 4468, b11 37~b1172, b0537~b0565, b0016~b0022, b4412~b4413, b0577~b0582, b4415, b2389~b2390, b2392~b2395, b0358~b0368, b0370~b0380, b285 6~b2863, b3042~b3048, b0656, b1325~b1333, b2030~b2062, b2190~b2 192, b3215~b3219, b3504~b3505, b1070~b1083, b1878~b1894, b1917~ b1950, b4324~b4342, b4345~b4358, b4486, b0497~b0502, b0700~b0706, b1456~b1462, b3481~b3484, b3592~b3596, b0981~b0988, b1021~b 1029, b2080~b2096, b4438, b3440~b3445, b4451, b3556~b3558, b4455, b1786, b0150~b0153, b2945, b2481~b2492, b2219~b2230, b4500, b3 20. The method of claim 19, wherein the strain is a reduced genome strain lacking the corresponding genes of any one of the following strains: b0611, b0488, b0500, and b0502, or a different K-12 or B strain. (Item 22) 22. The method of any one of items 19 to 21, wherein the E. coli host cell lacks a functional recA gene. (Item 23) 23. The method of any one of items 19 to 22, wherein the E. coli host cell lacks a functional dinB gene. (Item 24) 24. The method according to any one of items 19 to 23, wherein the natural parent of the E. coli host cell is a K-12 strain, and the E. coli host cell comprises the following modifications: (i) deletion of the rph gene to enhance orotate phosphoribosyltransferase activity, (ii) correction of a native-2 frameshift mutation in the ilvG gene to restore active acetohydroxyacid synthase II production, and (iii) deletion of all or part of the iclR and arpA genes. (Item 25) The E. coli host cell is adapted to express the following toxin-antitoxin genes: yafQ, dinJ, hha, tomB, gnsA, ymcE, yoeB, yefM, mazF, mazE, mazG, cptA / ygfX, cptB / sdhE, mqsR, mqsA, higB, higA, yhaV, prlF, ldrD, rdlD, istR-2, tisB, chpB, chpS, ratA, ratB, ldrA, rdlA, ldrB, rdlB, ldrC, rdlC 25. The method of any one of items 1 to 24, wherein the vector is lacking hokB, sokB, sibA, ibsA, sibB, ibsB, ohsC, shoB, sibC, ibsC, sibD, sibE, ibsD, ibsE, dinQ, agrA, agrB, ghoT, ghoS, yfeC, yfeD, fic, yhfG, yhjJ, yhjM, yhjN, yjjJ, ecnA, and ecnB, and optionally lacking hipA and hipB. (Item 26) 26. The method according to any one of items 1 to 25, wherein the E. coli host cell lacks a functional rnc gene, preferably wherein the E. coli host comprises a complete or partial deletion of the rnc gene. (Item 27) 27. The method according to any one of items 1 to 26, wherein at least 70%, at least 80%, or at least 90% of the recombinant polypeptide is extracted from the periplasm of the E. coli host cell. (Item 28) 28. The method according to any one of items 1 to 27, wherein the E. coli host cells are cultured in a continuous fermentation mode, preferably for a period of at least 14 days, at least 30 days or more, preferably thereby producing at least 1 gram, at least 2 grams, at least 3 grams or more of the recombinant protein per liter per day. (Item 29) It has a genome ranging from 4.41 Mb to 2.78 Mb and contains the following toxin-antitoxin genes: yafQ, dinJ, hha, tomB, gnsA, ymcE, yoeB, yefM, mazF, mazE, mazG, cptA / ygfX, cptB / sdhE, mqsR, mqsA, higB, higA, yhaV, prlF, ldrD, rdlD, istR-2, tisB, chpB, chpS, ratA, ratB, ldrA, rdlA, ldrB, rdlB, ldrC, ldrD, ldrD, ldrD, ldrC ... A non-naturally occurring E. coli bacterium lacking rC, rdlC, hokB, sokB, sibA, ibsA, sibB, ibsB, ohsC, shoB, sibC, ibsC, sibD, sibE, ibsD, ibsE, dinQ, agrA, agrB, ghoT, ghoS, yfeC, yfeD, fic, yhfG, yhjJ, yhjM, yhjN, yjjJ, ecnA, and ecnB, and / or lacking a functional rnc gene. (Item 30) 30. The bacterium according to item 29, comprising a complete or partial deletion of the rnc gene. (Item 31) 31. The bacterium of item 29 or 30, wherein the bacterium further lacks all of the IS1, IS2, IS3, IS5, IS150 and IS186 insertion sequences. (Item 32) 32. The bacterium according to any one of items 29 to 31, wherein the bacterium further comprises a deletion of the dinB gene. (Item 33) 33. The bacterium according to any one of items 29 to 32, wherein the bacterium further comprises a deletion of the recA gene. (Item 34) 34. The bacterium according to any one of items 29 to 33, wherein a parent strain of the bacterium is a K-12 or B strain. (Item 35) 35. The bacterium according to item 34, wherein the parent strain of the bacterium is a K-12 strain. (Item 36) 36. The bacterium according to item 35, wherein the bacterium comprises one or more, and preferably all, of the following modifications: (i) deletion of the rph gene to enhance orotate phosphoribosyltransferase activity, (ii) correction of the native -2 frameshift mutation in the ilvG gene to restore active acetohydroxyacid synthase II production, and (iii) deletion of all or part of the iclR and arpA genes. (Item 37) 37. The bacterium according to any one of items 29 to 36, wherein the parent strain of the bacterium is a K-12 or B strain. (Item 38) 38. The bacterium according to item 37, wherein the parent strain of the bacterium is K12 strain MG1655. (Item 39) The bacterium is selected from the group consisting of E. coli K-12 strain MG1655 and at least the following genes: b0245 to b0301, b0303 to b0310, b1336 to b1411, b4426 to b4427, b2441 to b2450, b2622 to b2654, b2657 to b2660, b4462, b1994 to b2008, b4435, b3322 to b3338, b2349 to b2363, b1539 to b1579, and b4269 to b4320; b2968~b2972, b2975~b2977, b2979~b2987, b4466~4468, b1137~b1172, b0537~b0565, b0016~b0022, b4412~b 4413, b0577~b0582, b4415, b2389~b2390, b2392~b2395, b0358~b0368, b0370~b0380, b2856~b2863, b3042~b 3048, b0656, b1325~b1333, b2030~b2062, b2190~b2192, b3215~b3219, b3504~b3505, b1070~b1083, b1878~b 1894, b1917~b1950, b4324~b4342, b4345~b4358, b4486, b0497~b0502, b0700~b0706, b1456~b1462, b3481~b 39. The bacterium of any one of items 29 to 38, which is lacking genes encoding the following genes: 3484, b3592 to b3596, b0981 to b0988, b1021 to b1029, b2080 to b2096, b4438, b3440 to b3445, b4451, b3556 to b3558, b4455, b1786, b0150 to b0153 and b2945, or the corresponding genes of a different K-12 or B strain. (Item 40) The bacterium is selected from the group consisting of E. coli K-12 strain MG1655 and E. coli K-12 strain MG1655, which is characterized in that at least the following genes are present: b0245 to b0301, b0303 to b0310, b1336 to b1411, b4426 to b4427, b2441 to b2450, b2622 to b2654, b2657 to b2660, b4462, b1994 to b2008, b4435, b3322 to b3338, b2349 to b2363, b1539 to b1579, b4269 to b4320, b2968 to b2972, b2975 to b2977, b2979 to b2987, b4466 to 4468, and b1137 ~b1172, b0537~b0565, b0016~b0022, b4412~b4413, b0577~b0582, b4415, b2389~b2390, b2392~b2395, b0358~b0368, b0370~b0380, b2856~b 2863, b3042~b3048, b0656, b1325~b1333, b2030~b2062, b2190~b2192 , b3215~b3219, b3504~b3505, b1070~b1083, b1878~b1894, b1917~b195 0, b4324~b4342, b4345~b4358, b4486, b0497~b0502, b0700~b0706, b1 456~b1462, b3481~b3484, b3592~b3596, b0981~b0988, b1021~b1029, b2080~b2096, b4438, b3440~b3445, b4451, b3556~b3558, b4455, b1786, b0150~b0153, b2945, b2481~b2492, b2219~b2230, b4500, b3707~b3 40. The bacterium of any one of items 29-39, wherein the bacterium is lacking any of the genes encoding the K-12 or B strains of the present invention, the K-12 or B strains being mutated to the K-12 or B strains of the present invention. [Brief description of the drawings]
[0029] [Figure 1] Extraction of carrier proteins in shake flask cultures. Periplasmic yields of CRM197, protein D and EPA (left panel) were compared to the yields obtained using extraction buffer (150 mM Tris pH 7.75, 5 mM EDTA) (right panel). Briefly, in each case, reduced genome E. coli strains harboring expression plasmids encoding the carrier protein fused to a secretion signal (e.g., YtfQ) were subjected to shake flask culture and carrier protein was induced once the host cells reached saturation. The percentage of periplasmic yield of each carrier protein after 6 h incubation with Tris / EDTA extraction buffer is shown. Protein yields from triplicate experiments in each case are shown. [Diagram 2] The SG69 meta lpp strain is able to release EPA into the medium, which facilitates purification. Sequential fermentations showing expression and release of EPA. Note that periplasmic EPA (dark) does not accumulate until 3 weeks after induction. The amount of EPA in the medium is shown in light. The lines of best fit in A (light light and medium) are of the fourth order. The inset is the EPA in the medium of 6 batches on one day (days 20-25, labeled D20-D25) analyzed by capillary western using a polyclonal antibody against EPA. Duplicate lanes in each example correspond to 1:10 and 1:100 dilutions. [Diagram 3]Protein D is released into the medium during continuous fermentation. Three separate continuous fermentations (Figures 3A, 3B, 3C) were performed using the SG69meta strain (also known as MDS69meta) that expressed periplasmically targeted PD. Induction (100 μM IPTG) was at day 1 in Figures 3A and 3B and day 0 in Figure 3C. Periplasmic PD (dark) and released PD (light) values were measured twice daily using Caliper Labchip capillary protein electrophoresis. Periplasmic PD values (g / L / day) are based on OD600 measurements (mean OD for AC was 224, 220, and 214, respectively). Released PD (light) was measured directly from the medium. The inset in Figure 3B shows the increase in released PD over the indicated days (caliper analysis, top panel). Capillary Western analysis (Wes Instrument, Protein Simpl, San Jose, CA) (bottom panel of Figure 3B) confirmed the presence of PD in the media samples. Note that the Western samples were diluted to obtain measurements that were within the linear range of the instrument. The best-fit lines in Figures 3B and 3C (very bright) are fourth-order polynomials. [Figure 4] Validation of PD sequences using mass spectrometry. Highly enriched bands from fermentation media corresponding to putative PDs were excised from 4-12% SDS-PAGE slab gels and subjected to mass spectrometry after trypsin or chymotrypsin digestion. Coverage was >99%. [Figure 5A]Extraction of carrier proteins from shake flask cultures following expression of pldA. Cultures of host strain SG69meta harboring plasmids encoding carrier proteins fused to the YtfQ secretion signal (under the control of an IPTG-inducible promoter) and pldA (under the control of a doxycycline (DOX)-inducible promoter) were induced with IPTG for carrier protein expression, and 4 hours later pldA expression was induced (DOX at the indicated concentrations) and cultures were incubated for 16 hours. 2 OD of culture was subjected to periplasmic isolation (Epicentre Technologies, an Illumina Company, Madison, WI) while 1.5 ml was brought to 150 mM Tris pH 7.75, 5 mM EDTA and shake extracted for 24 hours at 25°C. Extraction values are from intermediate supernatants after extraction and centrifugation. [Figure 5B] Extraction of carrier proteins from shake flask cultures following expression of pldA. Cultures of host strain SG69meta harboring plasmids encoding carrier proteins fused to the YtfQ secretion signal (under the control of an IPTG-inducible promoter) and pldA (under the control of a doxycycline (DOX)-inducible promoter) were induced with IPTG for carrier protein expression, and 4 hours later pldA expression was induced (DOX at the indicated concentrations) and cultures were incubated for 16 hours. 2 OD of culture was subjected to periplasmic isolation (Epicentre Technologies, an Illumina Company, Madison, WI) while 1.5 ml was brought to 150 mM Tris pH 7.75, 5 mM EDTA and shake extracted for 24 hours at 25°C. Extraction values are from intermediate supernatants after extraction and centrifugation. [Figure 5C]Extraction of carrier proteins from shake flask cultures following expression of pldA. Cultures of host strain SG69meta harboring plasmids encoding carrier proteins fused to the YtfQ secretion signal (under the control of an IPTG-inducible promoter) and pldA (under the control of a doxycycline (DOX)-inducible promoter) were induced with IPTG for carrier protein expression, and 4 hours later pldA expression was induced (DOX at the indicated concentrations) and cultures were incubated for 16 hours. 2 OD of culture was subjected to periplasmic isolation (Epicentre Technologies, an Illumina Company, Madison, WI) while 1.5 ml was brought to 150 mM Tris pH 7.75, 5 mM EDTA and shake extracted for 24 hours at 25°C. Extraction values are from intermediate supernatants after extraction and centrifugation. [Figure 6A]6A shows an embodiment of the present invention. FIG. 6A shows the sequence of the intervening junction between the C-terminus of CRM197 ORF and the pldA ORF. Part A: The stop codon of CRM197 ORF is immediately followed by a ribosome binding site (RBS) and the pldA ORF. Two representative ribosome binding sites of different strengths are shown. Based on the literature and previous experience with other bicistronic constructs, modification of the RBS can be used to increase or decrease the translation frequency of the downstream pldA ORF. Part B: The genome of MDS69meta (aka T69meta) was modified to allow DOX induction of endogenous genomic pldA. A tet repressor and T5DE20-dual tetO promoter / operator were inserted upstream of the native pldA gene and its native promoter. The pldA gene of this strain can be induced from its endogenous promoter and an introduced promoter. Figure 6B: Expression of PldA from the genome enhanced CRM197 extraction with Tris / EDTA but failed to result in 100% extraction of what was in the periplasm. We also achieved good results by (i) placing the pldA ORF under the control of a tet-operator / tet-repressor controlled promoter on a plasmid compatible with pSX2-ytfQ-CRM197, where OmpLA expression was controlled in a dose-dependent manner by the inducer doxycycline and CRM197 expression was controlled by the inducer IPTG, and (ii) placing CRM197 under a tet-Operator / tet-repressor controlled promoter and pldA expression under an ITPG inducible promoter. [Figure 6B]6A shows an embodiment of the present invention. FIG. 6A shows the sequence of the intervening junction between the C-terminus of CRM197 ORF and the pldA ORF. Part A: The stop codon of CRM197 ORF is immediately followed by a ribosome binding site (RBS) and the pldA ORF. Two representative ribosome binding sites of different strengths are shown. Based on the literature and previous experience with other bicistronic constructs, modification of the RBS can be used to increase or decrease the translation frequency of the downstream pldA ORF. Part B: The genome of MDS69meta (aka T69meta) was modified to allow DOX induction of endogenous genomic pldA. A tet repressor and T5DE20-dual tetO promoter / operator were inserted upstream of the native pldA gene and its native promoter. The pldA gene of this strain can be induced from its endogenous promoter and an introduced promoter. Figure 6B: Expression of PldA from the genome enhanced CRM197 extraction with Tris / EDTA but failed to result in 100% extraction of what was in the periplasm. We also achieved good results by (i) placing the pldA ORF under the control of a tet-operator / tet-repressor controlled promoter on a plasmid compatible with pSX2-ytfQ-CRM197, where OmpLA expression was controlled in a dose-dependent manner by the inducer doxycycline and CRM197 expression was controlled by the inducer IPTG, and (ii) placing CRM197 under a tet-Operator / tet-repressor controlled promoter and pldA expression under an ITPG inducible promoter. [Figure 7]We illustrate the effect of co-expression of a phospholipase (exemplified by PldA) on the release of recombinant CRM917 into the medium of the reduced genome E. coli strain T69 Meta recA (aka MDS69meta recA), harboring a single plasmid expression vector encoding CRM197 fused to the YtfQ secretion signal (under the control of an IPTG-inducible promoter) and OmpLA under the control of a doxycycline-inducible promoter. Once the host cells reached saturation in shake flask culture, CRM197 and OmpLA were expressed and the yield of soluble CRM197 in the medium and periplasm was assessed after 2 h or overnight incubation with Tris / EDTA. T69 (aka MDS69) is a multiple deletion strain containing all deletions of MDS42 and 27 additional deletions. T69 Meta recA (also known as MDS69meta recA) contains a deletion of MDS69 and also contains correction of the native rph and ilvG frameshift mutations, deletions of the iclR and arpA genes, and a deletion of the recA gene. [Figure 8] We illustrate the effect of co-expression of a phospholipase (exemplified by OmpLA) on the release of recombinant exoprotein A (rEPA) into the medium of reduced genome E. coli strain T69 Meta recA (also known as MDS69meta recA), harboring a single plasmid expression vector encoding rEPA fused to the YtfQ secretion signal (under the control of an IPTG-inducible promoter) and OmpLA under the control of a doxycycline-inducible promoter. Once host cells reached saturation in shake flask cultures, rEPA and OmpLA were induced and the yield of soluble rEPA in the medium and periplasm was assessed after 2 h or overnight incubation with Tris / EDTA. [Figure 9]We illustrate the effect of co-expression of a phospholipase (exemplified by OmpLA) on the release of recombinant protein D (PD) into the medium of reduced genome E. coli strain T69 Meta recA (also known as MDS69meta recA) harboring a single plasmid expression vector encoding PD fused to the YtfQ secretion signal (under the control of an IPTG-inducible promoter) and OmpLA under the control of a doxycycline-inducible promoter. Once host cells reached saturation in shake flask cultures, PD and OmpLA were induced and the yield of soluble PD in the medium and periplasm was assessed after incubation with Tris / EDTA for 2 h, 6 h or overnight. [Figure 10A] Basal endogenous expression in T69 Meta recA (aka MDS69meta recA) strains carrying single plasmid expression vectors encoding either CRM197 (Figure 10A), protein D (Figure 10B) or rEPA (Figure 10C) fused to the YtfQ secretion signal (under the control of an IPTG-inducible promoter), and OmpLA under the control of a single tetracycline operator (tetO) doxycycline-inducible promoter (SG7766 2nd 0-100 nm in 10A lanes 4-6). Illustrated are basal endogenous + uninduced (SG9040 lane 4) OmpLA levels (measured by release of carrier protein into the medium in the absence of DOX) compared to T69 meta recA (aka MDS69meta recA) strains harboring a single plasmid expression vector (i.e., no OmpLA-encoding sequence in this vector) encoding only CRM197 (Figure 10A), protein D (Figure 10B), or rEPA (Figure 10C) fused to the YtfQ secretion signal under the control of an IPTG-inducible promoter. Co-expression of phospholipase significantly increased the release of carrier protein into the medium, approximately 5-fold for rEPA and CRM197 at the highest inducer concentration, and approximately 3-fold for PD at the highest inducer concentration. [Figure 10B]Basal endogenous expression in T69 Meta recA (aka MDS69meta recA) strains carrying single plasmid expression vectors encoding either CRM197 (Figure 10A), protein D (Figure 10B) or rEPA (Figure 10C) fused to the YtfQ secretion signal (under the control of an IPTG-inducible promoter), and OmpLA under the control of a single tetracycline operator (tetO) doxycycline-inducible promoter (SG7766 2nd 0-100 nm in 10A lanes 4-6). Illustrated are basal endogenous + uninduced (SG9040 lane 4) OmpLA levels (measured by release of carrier protein into the medium in the absence of DOX) compared to T69 meta recA (aka MDS69meta recA) strains harboring a single plasmid expression vector (i.e., no OmpLA-encoding sequence in this vector) encoding only CRM197 (Figure 10A), protein D (Figure 10B), or rEPA (Figure 10C) fused to the YtfQ secretion signal under the control of an IPTG-inducible promoter. Co-expression of phospholipase significantly increased the release of carrier protein into the medium, approximately 5-fold for rEPA and CRM197 at the highest inducer concentration, and approximately 3-fold for PD at the highest inducer concentration. [Figure 10C]Basal endogenous expression in T69 Meta recA (aka MDS69meta recA) strains carrying single plasmid expression vectors encoding either CRM197 (Figure 10A), protein D (Figure 10B) or rEPA (Figure 10C) fused to the YtfQ secretion signal (under the control of an IPTG-inducible promoter), and OmpLA under the control of a single tetracycline operator (tetO) doxycycline-inducible promoter (SG7766 2nd 0-100 nm in 10A lanes 4-6). Illustrated are basal endogenous + uninduced (SG9040 lane 4) OmpLA levels (measured by release of carrier protein into the medium in the absence of DOX) compared to T69 meta recA (aka MDS69meta recA) strains harboring a single plasmid expression vector (i.e., no OmpLA-encoding sequence in this vector) encoding only CRM197 (Figure 10A), protein D (Figure 10B), or rEPA (Figure 10C) fused to the YtfQ secretion signal under the control of an IPTG-inducible promoter. Co-expression of phospholipase significantly increased the release of carrier protein into the medium, approximately 5-fold for rEPA and CRM197 at the highest inducer concentration, and approximately 3-fold for PD at the highest inducer concentration. [Figure 11] Following induction of expression of RM197 and OmpLA in T69 Meta recA (also known as MDS69meta recA) host cells harboring a single expression plasmid containing sequences encoding CRM197 under the control of an IPTG-inducible promoter and pldA under the control of a doxycycline-inducible promoter containing dual tet operators, we illustrate the yield of CRM197 in periplasmic and Tris / EDTA extracts at 25°C. In the absence of DOX, CRM197 barely appears in Tris / EDTA extracts at 25°C, but this increases significantly when pldA is induced. The periplasmic yield of CRM197 is not affected by pldA gene expression. [Figure 12A]Figure 12A shows CRM197 yield in the periplasm (Epicentre periplasmic prep with lysozyme) of T69 Meta Tox / Atox LowMut recA strain (aka MDS69meta T / A LowMut recA) carrying (i) an expression plasmid encoding CRM197 fused to the YtfQ secretion signal under the control of a doxycycline inducible promoter or carrying a bicistronic expression plasmid encoding CRM197-ytfQ and pldA (codon optimized) with either two ribosome binding sites (see Figure 6) between CRM197 and pldA orfs. Figure 12B shows the same culture as in Figure 12A, but instead of Epicentre Peri prep with lysozyme, 1 ml of induced culture was subjected to overnight shaking at 25°C (no Tris / EDTA treatment). In the presence of OmpLA, approximately 5-fold more target protein enters the medium from the periplasm upon shaking at 25°C (still only 50% of the total amount of target protein extracted into the medium with OmpLA and Tris / EDTA treatment). Figure 12C illustrates induction of pldA expression and CRM197 yield in the medium after extraction with Tris / EDTA and shaking at 25°C. The leftmost panel shows that very little target is extracted from the periplasm into the medium. Meanwhile, the two right panels have on average 5-fold more produced into the medium compared to the left panel. The two right panels have on average 2-fold more CRM197 extracted than the same two panels in Figure 12B. Figure 12D shows expression of pldA and incubation of 1 ml samples of the induced culture for 2 days at 4°C. Incubation at 4°C (without Tris / EDTA) versus 25°C (without Tris / EDTA) does not ameliorate "leakage" of target from the periplasm into the medium. [Figure 12B]Figure 12A shows CRM197 yield in the periplasm (Epicentre periplasmic prep with lysozyme) of T69 Meta Tox / Atox LowMut recA strain (aka MDS69meta T / A LowMut recA) carrying (i) an expression plasmid encoding CRM197 fused to the YtfQ secretion signal under the control of a doxycycline inducible promoter or carrying a bicistronic expression plasmid encoding CRM197-ytfQ and pldA (codon optimized) with either two ribosome binding sites (see Figure 6) between CRM197 and pldA orfs. Figure 12B shows the same culture as in Figure 12A, but instead of Epicentre Peri prep with lysozyme, 1 ml of induced culture was subjected to overnight shaking at 25°C (no Tris / EDTA treatment). In the presence of OmpLA, approximately 5-fold more target protein enters the medium from the periplasm upon shaking at 25°C (still only 50% of the total amount of target protein extracted into the medium with OmpLA and Tris / EDTA treatment). Figure 12C illustrates induction of pldA expression and CRM197 yield in the medium after extraction with Tris / EDTA and shaking at 25°C. The leftmost panel shows that very little target is extracted from the periplasm into the medium. Meanwhile, the two right panels have on average 5-fold more produced into the medium compared to the left panel. The two right panels have on average 2-fold more CRM197 extracted than the same two panels in Figure 12B. Figure 12D shows expression of pldA and incubation of 1 ml samples of the induced culture for 2 days at 4°C. Incubation at 4°C (without Tris / EDTA) versus 25°C (without Tris / EDTA) does not ameliorate "leakage" of target from the periplasm into the medium. [Figure 12C]Figure 12A shows CRM197 yield in the periplasm (Epicentre periplasmic prep with lysozyme) of T69 Meta Tox / Atox LowMut recA strain (aka MDS69meta T / A LowMut recA) carrying (i) an expression plasmid encoding CRM197 fused to the YtfQ secretion signal under the control of a doxycycline inducible promoter or carrying a bicistronic expression plasmid encoding CRM197-ytfQ and pldA (codon optimized) with either two ribosome binding sites (see Figure 6) between CRM197 and pldA orfs. Figure 12B shows the same culture as in Figure 12A, but instead of Epicentre Peri prep with lysozyme, 1 ml of induced culture was subjected to overnight shaking at 25°C (no Tris / EDTA treatment). In the presence of OmpLA, approximately 5-fold more target protein enters the medium from the periplasm upon shaking at 25°C (still only 50% of the total amount of target protein extracted into the medium with OmpLA and Tris / EDTA treatment). Figure 12C illustrates induction of pldA expression and CRM197 yield in the medium after extraction with Tris / EDTA and shaking at 25°C. The leftmost panel shows that very little target is extracted from the periplasm into the medium. Meanwhile, the two right panels have on average 5-fold more produced into the medium compared to the left panel. The two right panels have on average 2-fold more CRM197 extracted than the same two panels in Figure 12B. Figure 12D shows expression of pldA and incubation of 1 ml samples of the induced culture for 2 days at 4°C. Incubation at 4°C (without Tris / EDTA) versus 25°C (without Tris / EDTA) does not ameliorate "leakage" of target from the periplasm into the medium. [Figure 12D]Figure 12A shows CRM197 yield in the periplasm (Epicentre periplasmic prep with lysozyme) of T69 Meta Tox / Atox LowMut recA strain (aka MDS69meta T / A LowMut recA) carrying (i) an expression plasmid encoding CRM197 fused to the YtfQ secretion signal under the control of a doxycycline inducible promoter or carrying a bicistronic expression plasmid encoding CRM197-ytfQ and pldA (codon optimized) with either two ribosome binding sites (see Figure 6) between CRM197 and pldA orfs. Figure 12B shows the same culture as in Figure 12A, but instead of Epicentre Peri prep with lysozyme, 1 ml of induced culture was subjected to overnight shaking at 25°C (no Tris / EDTA treatment). In the presence of OmpLA, approximately 5-fold more target protein enters the medium from the periplasm upon shaking at 25°C (still only 50% of the total amount of target protein extracted into the medium with OmpLA and Tris / EDTA treatment). Figure 12C illustrates induction of pldA expression and CRM197 yield in the medium after extraction with Tris / EDTA and shaking at 25°C. The leftmost panel shows that very little target is extracted from the periplasm into the medium. Meanwhile, the two right panels have on average 5-fold more produced into the medium compared to the left panel. The two right panels have on average 2-fold more CRM197 extracted than the same two panels in Figure 12B. Figure 12D shows expression of pldA and incubation of 1 ml samples of the induced culture for 2 days at 4°C. Incubation at 4°C (without Tris / EDTA) versus 25°C (without Tris / EDTA) does not ameliorate "leakage" of target from the periplasm into the medium. [Figure 13] Illustrates a comparison of periplasmic production of CRM197 from several E. coli strains. [Figure 14] Illustrates the periplasmic yield of CRM197 in strains deleted for RNase III and toxin / antitoxin genes compared to the parent strain containing these genes. [Figure 15]We illustrate the production of CRM197 during long-term continuous culture of a reduced genome E. coli strain (SG69meta alias MDS69meta) harboring an expression vector encoding CRM197 fused to a periplasmic expression signal sequence, compared to production of CRM197 in a commonly used fermentation host harboring the same expression plasmid. Fermentations producing between 4-6 grams of CRM197 per liter per day are shown. Data points are Wes capillary electrophoresis chemiluminescence immunoassay signals converted to CRM197 expression values using a CRM197 standard curve. Recognition of CRM197 was by anti-CRM197 polyclonal antiserum and HRP-luciferase chemiluminescence. [Figure 16] Illustrates an extended continuous fermentation (aka continuous flow fermentation or C-flow) of the SG69meta strain containing a single expression plasmid containing sequences encoding CRM197 under the control of an IPTG-inducible promoter and pldA under the control of a doxycycline-inducible promoter. C-flow produced an average of 1 g / L / day as determined by Caliper Labchip and Wes (capillary western) analysis. The left panel shows a caliper analysis of fermentation broth from a production tank 10 days after IPTG induction of CRM197 expression that was either 1) treated with Tris / EDTA and shaken at 25°C or 2) treated with DOX to induce expression of the pldA gene and then extracted with Tris / EDTA for 16 h. A graphical representation of the caliper results is shown in the right panel. [Figure 17] An example of an exemplary continuous fermentation apparatus is illustrated. Bacterial culture in two chemostat bioreactors is performed at high cell density (approximately 200 OD600). Following induction of target gene expression in bioreactor 2, product-containing cells are harvested at a rate of 1 liter / day (or a rate dependent on the size of the fermentation vessel used). Note that the uninduced "seed" culture from bioreactor 1 is transferred to bioreactor 2 to replenish cells removed during product harvest. Inducer concentration is maintained by adding inducer (IPTG is exemplified) to the feed of bioreactor 2. [Figure 18]Compare the periplasmic yield of CRM197 in T69 Meta recA (aka MDS69meta recA), T69 Meta recA T / AT (aka MDS69meta T / A recA), and T69 Meta recA T / AT LowMut (aka MDS69meta T / A LowMut recA) host cells harboring an expression plasmid encoding CRM197 fused to the YtfQ secretion signal under the control of a doxycycline-inducible promoter. Deletion of the toxin / antitoxin genes significantly enhances the yield of CRM197 in the periplasm. [Figure 19A] Illustrated are the extraction of single chain variable fragment scFv 75127 (provided by Fritz Schaumburg, Lytic Solutions, Madison WI) (Figure 19A) and single chain antibody scFab YMF10 (Figure 19B) from shake flask cultures following OmpLA expression. Figure 19A: Shown are periplasmic and extract yields of scFv 75127 following induction of scFv and pldA expression in reduced genome E. coli strain SG69meta (aka T69 Meta aka MDS69meta) containing a dual scFv, pldA bicistronic expression plasmid construct. Figure 19B: Shown are periplasmic and extract yields of scFab YMF10 following induction of scFv and OmpLA expression in reduced genome E. coli strain SG69meta (aka T69 Meta aka MDS69meta) containing a dual scFab, pldA bicistronic expression plasmid construct. [Figure 19B]Illustrated are the extraction of single chain variable fragment scFv 75127 (provided by Fritz Schaumburg, Lytic Solutions, Madison WI) (Figure 19A) and single chain antibody scFab YMF10 (Figure 19B) from shake flask cultures following OmpLA expression. Figure 19A: Shown are periplasmic and extract yields of scFv 75127 following induction of scFv and pldA expression in reduced genome E. coli strain SG69meta (aka T69 Meta aka MDS69meta) containing a dual scFv, pldA bicistronic expression plasmid construct. Figure 19B: Shown are periplasmic and extract yields of scFab YMF10 following induction of scFv and OmpLA expression in reduced genome E. coli strain SG69meta (aka T69 Meta aka MDS69meta) containing a dual scFab, pldA bicistronic expression plasmid construct. [Figure 20A] Illustrated are periplasmic (FIG. 20A) and extractable yields (FIG. 20B) of scFv 75127 in MDS69meta recA (aka T69 Meta recA) and MDS69meta T / AT LowMut RecA (aka T69 Meta T / AT LowMut recA) strains harboring expression plasmids encoding scFv fused to the OmpA secretion signal under the control of an IPTG-inducible promoter and pldA under the control of a DOX-inducible promoter; scFv expression was induced, pldA expression was induced 4 hours later, and induction was allowed to continue overnight, at which point cells were contacted with Tris / EDTA extraction buffer (above) overnight at 25° C. [Figure 20B]Illustrated are periplasmic (FIG. 20A) and extractable yields (FIG. 20B) of scFv 75127 in MDS69meta recA (aka T69 Meta recA) and MDS69meta T / AT LowMut RecA (aka T69 Meta T / AT LowMut recA) strains harboring expression plasmids encoding scFv fused to the OmpA secretion signal under the control of an IPTG-inducible promoter and pldA under the control of a DOX-inducible promoter; scFv expression was induced, pldA expression was induced 4 hours later, and induction was allowed to continue overnight, at which point cells were contacted with Tris / EDTA extraction buffer (above) overnight at 25° C. [Figure 21A] Illustrated are the periplasmic ( FIG. 21A ) and extraction yields ( FIG. 21B ) of scFab YMF10 in MDS69meta recA (aka T69 Meta recA) and MDS69meta T / AT LowMut RecA (aka T69 Meta T / AT LowMut recA) strains harboring expression plasmids encoding the scFab fused to the OmpA secretion signal under the control of an IPTG-inducible promoter and pldA under the control of a DOX-inducible promoter; scFab expression was induced, pldA expression was induced 4 hours later, and induction was continued overnight, at which point cells were contacted with Tris / EDTA extraction buffer (above) overnight at 25° C. [Figure 21B] Illustrated are the periplasmic ( FIG. 21A ) and extraction yields ( FIG. 21B ) of scFab YMF10 in MDS69meta recA (aka T69 Meta recA) and MDS69meta T / AT LowMut RecA (aka T69 Meta T / AT LowMut recA) strains harboring expression plasmids encoding the scFab fused to the OmpA secretion signal under the control of an IPTG-inducible promoter and pldA under the control of a DOX-inducible promoter; scFab expression was induced, pldA expression was induced 4 hours later, and induction was continued overnight, at which point cells were contacted with Tris / EDTA extraction buffer (above) overnight at 25° C. [Figure 22]Illustrates periplasmic yields of CRM197 and CRM197 after Tris / EDTA extraction in MDS69meta T / AT LowMut recA and MDS69meta strains. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0030] The present inventors have discovered that recombinant therapeutic proteins, including but not limited to, single chain antibody fragment scFab YMF10, single chain antibody variable fragment scFv 75127, CRM197, protein D and EPA, delivered to the periplasm of E. coli host cells and co-expressing endogenous E. coli phospholipases, can be directly extracted into the medium during fermentation, reducing production time and labor, greatly simplifying the purification of these proteins, and dramatically reducing production costs.
[0031] Use of the methods and compositions described herein can overcome several standard manufacturing challenges associated with CRM197 and other carrier proteins, including: (1) low levels of CRM197 (0.1-0.2 g / L) in the original Corynebacterium diphtheria host; (2) potential reversion of CRM197 to wild-type toxin in alternative hosts such as Pseudomonas fluorescens and many E. coli strains; (3) physiological stress in the fermentation host under high production rates of CRM197; (4) accumulation of CRM197 in an insoluble form in most E. coli hosts, reducing yields and significantly complicating downstream purification steps; and (5) activation of cryptic prophage or insertional elements, or disruption of toxin / antitoxin homeostasis and other mutagenic stress-inducible systems in E. coli host cells that could induce cell lysis and impair the use of these strains in continuous manufacturing.
[0032] Using the reduced genome E. coli strains and methods described herein, we have discovered and verified that periplasmically targeted proteins such as CRM197, PD, and EPA can be continuously produced and extracted from the cells directly into the medium for easy purification at high yields (greater than 3 g per liter per day) that can be maintained for periods of at least 30 days and even at least 65 days or more during continuous fermentation. In contrast, conventional E. coli strains cannot sustain production for more than a few days under the same conditions.
[0033] CRM197 is a genetically non-toxic form of diphtheria toxin (DT). A single base change in CRM197 (glutamic acid to glycine at position 52) abolishes the ADP-ribosylating activity of the A chain, reducing toxicity. Although CRM197 is non-toxic, it is immunologically indistinguishable from diphtheria toxin. CRM197 functions as a carrier for polysaccharides and haptens, making them immunogenic in some conjugate vaccines.
[0034] PD is an immunoglobulin D-associated membrane lipoprotein exposed on the surface of the gram-negative bacterium Haemophilus influenza. The protein is synthesized as an 18-residue-long signal sequence modified by the covalent attachment of both ester- and amide-linked palmitates to a cysteine residue, which becomes the amino terminus of the mature protein after cleavage of the signal sequence. Mutant protein D lacking the cysteine residue is not acylated. Nonacylated protein D molecules are localized in the periplasmic space of E. coli. The hydrophilic protein D molecule functions as a carrier for polysaccharides and haptens, making them immunogenic in conjugate vaccines.
[0035] The carrier protein EPA from Pseudomonas aeruginosa activates T cells in the same manner as CRM197. Because CRM197 is used in so many conjugate vaccines, including Prevnar-13, the world's best-selling vaccine, some vaccine manufacturers have chosen to use it to avoid epitope suppression. EPA is used in the nicotine vaccine NicVAX, which has been shown to reduce nicotine dependence. (Esterlis I, Hannestad JO, Perkins E, Bois F, D´Souza DC, Tyndale RF, Seibyl JP, Hatsukami DM, Cosgrove KP, O´Malley SS. Effect of a nicotine vaccine on nicotine binding to beta2*-nicotinic acetylcholine receptors in vivo in human tobacco smokers. Am J Psychiatry. 2013;170(4):399-407. doi:.1176 / appi.ajp.2012.12060793. PubMed PMID:23429725; PMCID:PMC3738000.) An antimalarial conjugate vaccine using EPA has been successfully developed and is in clinical trials.
[0036] The inventors have discovered that co-expressing an E. coli phospholipase with a periplasmic target protein of interest in an E. coli host cell provides a system in which nearly all of the protein of interest can be extracted into the medium using an appropriate buffer. In a preferred embodiment, the phospholipase is selected from the group consisting of GlpQ, YbaC, PldA, PldB and TesA. These exemplary phosphodiesterases are discussed below.
[0037] Glycerophosphoryl diester phosphodiesterase (GlpQ) is involved in the utilization of the glycerol moiety of phospholipids and triglycerides after it has been broken down into usable forms such as glycerophosphodiesters, sn-glycerol-3-phosphate (G3P) or glycerol. GlpQ catalyzes the hydrolysis of glycerophosphodiesters to sn-glycerol 3-phosphate (glycerol-P) and alcohol. Glycerol-P is then transported into the cell by GlpT, the glycerol-P transporter. Periplasmic GlpQ is specific for the glycerophosphomoiety of the substrate, but the alcohol can be any one of several long- and medium-chain alcohols. This gives the cell the ability to direct a wide variety of glycerophosphodiesters to the glp-encoded catabolic system. In some embodiments, an E. coli host cell, preferably a reduced genome E. coli host cell, for use in the methods described herein comprises a nucleic acid segment comprising a nucleotide sequence encoding a polypeptide of SEQ ID NO:1 or a sequence that is at least 80%, at least 85%, at least 90%, at least 95%, at least 98%, or at least 99% identical thereto.
[0038] Aes (encoded by the ybaC gene) catalyzes the hydrolysis of p-nitrophenyl esters of fatty acids, with a preference for substrates with short (less than 8 in length) acyl chains. Aes has structural similarity to lipases and esterases and belongs to the hormone-sensitive lipase (HSL) protein family. The enzyme may be either monomeric or homodimeric, and it has been crystallized and its crystal structure solved. Denaturant-induced unfolding of the enzyme has been studied, and mutants with higher thermostability have been isolated. The catalytic triad is predicted to include ser165, asp262, and his292, and site-directed mutagenesis confirmed that these residues are required for catalytic activity. Additionally, asp164 is thought to be structurally important. Point mutations at l97 and l209, respectively, reduced the k cat Increase k maes increases enzyme activity by decreasing p-nitrophenylacetate. Overexpression of aes results in increased hydrolysis of p-nitrophenylacetate compared to wild type. In some embodiments, an E. coli host cell, preferably a reduced genome E. coli host cell, for use in the methods described herein comprises a nucleic acid segment comprising a nucleotide sequence encoding a polypeptide of SEQ ID NO:2 or a sequence that is at least 80%, at least 85%, at least 90%, at least 95%, at least 98%, or at least 99% identical thereto.
[0039] The E. coli K-12 gene pldA encodes the outer membrane phospholipase A (OmpLA) protein. Purified OmpLA exhibits multiple lipolytic activities and is active against phospholipids, lysophosphophospholipids, diglycerides, and monoglycerides, although in vivo the predicted substrate is the phospholipid of the cell envelope. Dimerization of OmpLA is required for phospholipase activity. Interactions between the enzyme and the acyl side chain of the substrate are the main stabilizing factors for dimerization. Although pldA is constitutively expressed, OmpLA activity is not detectable under normal conditions and is induced by conditions that disturb the integrity of the outer membrane. OmpLA has been implicated in bacteriocin release. In E. coli strains containing colicin production plasmids, the activity of OmpLA was increased and phosphatidylethanolamine (the major bacterial outer membrane phospholipid) was degraded prior to lysis and bacteriocin release. In colicin producing E. coli pldA mutants, colicin is retained in the cytoplasm. In some preferred embodiments, E. coli host cells, preferably reduced genome E. coli host cells, for use in the methods described herein comprise a nucleic acid segment comprising a nucleotide sequence encoding the polypeptide of SEQ ID NO:4, or a sequence that is at least 80%, at least 85%, at least 90%, at least 95%, at least 98%, or at least 99% identical to that sequence.
[0040] Lysophospholipase L2 (PldB) hydrolyzes the bond to the remaining acyl group of lysophospholipids. The most effective substrates are 2-acylglycerophosphoethanolamine and 2-acylglycerophosphocholine, with the monoacyl versions of each compound being somewhat less efficient at hydrolysis. In addition, PldB can transfer an acyl group from certain lysophospholipids to phosphatidylglycerol to produce acylphosphatidylglycerol. PldB does not contain an obvious signal sequence or transmembrane domain and is believed to be associated with the cytoplasmic side of the inner membrane. In some embodiments, an E. coli host cell, preferably a reduced genome E. coli host cell, for use in the methods described herein comprises a nucleic acid segment comprising a nucleotide sequence encoding a polypeptide of SEQ ID NO:3 or a sequence that is at least 80%, at least 85%, at least 90%, at least 95%, at least 98%, or at least 99% identical thereto.
[0041] The TAP complex (TesA) is a broad specificity esterase that performs thioesterase, arylesterase, lysophospholipase, and protease activities. Its exact function in the cell is unknown and could potentially be involved in the generation of free fatty acids or salvage of compounds containing ester bonds. TesA (pldC; b0494; GenBank Ref. No. NP_416742.1; SEQ ID NO:5) is a multifunctional esterase that can act as a thioesterase, arylesterase, lysophospholipase, and has limited protease activity. It acts as a thioesterase for a very wide range of acyl-CoA molecules, but has been reported to be specific for fatty acyl thioesters with more than 10 carbons, with the highest activity for palmityl, palmitoleyl, and cis-vaccenyl esters. The actual in vivo role of TesA is unclear. Its localization in the periplasm makes it nearly inaccessible to fatty acyl-CoA molecules. In some embodiments, an E. coli host cell, preferably a reduced genome E. coli host cell, for use in the methods described herein comprises a nucleic acid segment comprising a polypeptide of SEQ ID NO:5 or a sequence that is at least 80%, at least 85%, at least 90%, at least 95%, at least 98%, or at least 99% identical to that sequence.
[0042] The methods described herein allow for the extraction of a protein of interest into any E. coli host cell. In some embodiments, the E. coli host cell according to the methods described herein is a wild-type E. coli B or K12 strain.
[0043] In some preferred embodiments of the invention, the recombinant E. coli host strain according to the methods and compositions described herein is a reduced genome bacterium. As used herein, a "reduced genome" bacterium means a bacterium having about 1% to about 75% of its genome (e.g., protein-coding genes, RNA-coding genes) deleted, such as about 2%, about 5%, about 10%, about 20%, about 30%, about 40%, about 50%, or about 60% of the genome deleted. In one embodiment, the reduced genome bacterium used in the practice of the invention has a genetically engineered genome that is preferably at least two percent (2%) to up to thirty percent (30%) (including any number therebetween) smaller than the genome of the native parent strain. Preferably, the genome is at least two percent (2%) to up to twenty-five percent (25%) smaller than the genome of the native parent strain. The genome is about two percent (2%), five percent (5%), eight percent (8%), fourteen percent (14%), twenty-two percent (22%), twenty-five percent (25%), thirty percent (30%) or more (including any number in between) smaller than the genome of the native parent strain. Alternatively, the genome may be engineered to be less than 10%, less than 15%, less than 20%, less than 25%, or less than 30% smaller than the genome of the native parent strain. Reduced genome bacteria may have a genome that is between 4.55 Mb (2%) and 3.25 Mb (30%), between 4.41 Mb (5%) and 3.62 Mb (22%), between 4.41 Mb and 3.25 Mb, or between 4.41 Mb and 2.78 Mb. The term "natural parent strain", as generally understood by the scientific community, refers to a bacterial strain found in nature or in a natural environment, which represents the establishment of a strain lineage, and on whose genome a series of deletions can be made to generate a bacterial strain with a smaller genome. A natural parent strain also refers to a strain against which an engineered strain is compared, which has less than the full complement of the natural parent strain. The percentage of genome reduction after a series of deletions is calculated by dividing the "total number of base pairs deleted after all deletions" by the "total number of base pairs in the genome before all deletions", and then multiplying by 100.Similarly, the percentage by which a genome is smaller than the native parent strain is calculated by dividing the total number of nucleotides in the strain with the smaller genome by the total number of nucleotides in the native parent strain (regardless of the process by which it was produced) and then multiplying by 100.
[0044] In one embodiment, a "reduced genome" bacterium refers to a bacterium in which removal of the amount of genome does not unacceptably affect the ability of the organism to grow in minimal medium. Whether removal of two or more genes "unacceptably affects" the ability of the organism to grow in minimal medium in this context depends on the particular application and is easily evaluated. For example, a 30% reduction in growth rate may be acceptable in one application but not another. Furthermore, the adverse effects of deleting DNA sequences from the genome may be mitigated by measures such as changing culture conditions. Such measures may turn an otherwise unacceptable adverse effect into one that can be tolerated. In one embodiment, the growth rate is approximately the same as the parent strain. However, growth rates that are about 5%, 10%, 15%, 20%, 30%, 40% to about 50% lower than the growth rate of the parent strain are within the scope of the present invention. More specifically, the doubling time of the bacteria of the present invention may range from about 15 minutes to about 3 hours. Non-limiting examples of suitable reduced genome bacteria and methods for cumulatively deleting genomic DNA from bacteria such as E. coli are disclosed in U.S. Pat. Nos. 6,989,265, 7,303,906, 8,119,365, 8,039,243, 8,178,339, 8,765,408, and 9,902,965, respectively, which are incorporated herein by reference.
[0045] A parent E. coli strain in the context of a reduced genome bacterium can be any E. coli strain. In some preferred embodiments, the parent E. coli strain is a K-12 strain (e.g., MG1655 (GenBank Accession No. U00096.3), W3110 (GenBank Accession No. AP009048.1), DH10B (GenBank Accession No. CP000948.1), DH1 (GenBank Accession No. CP001637.1), or BW2952 (GenBank Accession No. CP001396.1). In other embodiments, the parent E. coli strain is a B strain (e.g., BLR(DE3) (GenBank Accession No. CP020368.1), REL606 (GenBank Accession No. CP000819.1), BL21(DE3) (GenBank Accession No. CP001509.3).
[0046] In one embodiment, the parent E. coli strain is a K-12 or B strain lacking one or more of the genes listed in Tables 2-20 of US Pat. No. 8,178,339, the contents of which are incorporated herein by reference.
[0047] In some preferred embodiments, the reduced genome E. coli K12 or B strains described herein comprise at least the following genes of E. coli K-12 strain MG1655 (identified by a "b" number based on the designation set forth in Blattner et al., Science, 277:1453-74 and GenBank Accession No. U00096.3): b0245-b0301, b0303-b0310, b1336-b1411, b4426-b4427, b2441-b2450, b2622-b2654, b2657-b2660, b4462, b1994-b2008, b4435, b 3322-b3338, b2349-b2363, b1539-b1579, b4269-b4320, b2968-b2972, b2975-b2977, b2979-b2987, b4466-4468, b 1137-b1172, b0537-b0565, b0016-b0022, b4412-b4413, b0577-b0582, b4415, b2389-b2390, b2392-b2395, b0358-b 0368, b0370-b0380, b2856-b2863, b3042-b3048, b0656, b1325-b1333, b2030-b2062, b2190-b2192, b3215-b3219, b3504-b3505, b1070-b1083, b1878-b1894, b1917-b1950, b4324-b4342, b4345-b4358, b4486, b0497-b0502, b0700- b0706, b1456-b1462, b3481-b3484, b3592-b3596, b0981-b0988, b1021-b1029, b2080-b2096, b4438, b3440-b3445, b4451, b3556-b3558, b4455, b1786, b0150-b0153 and b2945, or lacking the genes corresponding to different K-12 or B strains. These are the genes that were deleted from the parent strain MG1655 to create the multiple deletion strain MDS42.
[0048] In other preferred embodiments, the reduced genome E. coli K12 or B strains described herein contain at least the following genes of E. coli K-12 strain MG1655: b0245-b0301, b0303-b0310, b1336-b1411, b4426-b4427, b2441-b2450, b2622-b2654, b2657-b2660, b4462, b1994-b2008, b4435, b3322-b3338, b2349-b2363, b1539-b1579, b4269-b4320, b2968-b2972, b2975-b2977, b2979-b2987, b4466-4468, b 1137-b1172, b0537-b0565, b0016-b0022, b4412-b4413, b0577-b0582, b4415, b2389-b2390, b2392-b2395, b0358-b0368, b0370-b0380, b 2856-b2863, b3042-b3048, b0656, b1325-b1333, b2030-b2062, b2190-b2192, b3215-b3219, b3504-b3505, b1070-b1083, b1878-b1894, b1 917-b1950, b4324-b4342, b4345-b4358, b4486, b0497-b0502, b0700-b0706, b1456-b1462, b3481-b3484, b3592-b3596, b0981-b0988, b1 021-b1029, b2080-b2096, b4438, b3440-b3445, b4451, b3556-b3558, b4455, b1786, b0150-b0153, b2945, b2481-b2492, b2219-b2230, b45 00, b3707-b3723, b0644-b0650, b4079-4090, b4487, b4092-b4106, b0730-b0732, b3572-b3587, b1653, b2735-b2740, b2405-b2407, b3896-b3900, b1202, b4263-b4268, b0611, b2364-b2366, b0839, b0488-b0500, and b0502, or are missing the corresponding genes in a different K-12 or B strain.These are the genes that were deleted from the parent strain MG1655 to create the multiple deletion strain MDS69 (also known as T69, also known as SG69).
[0049] Recombinant bacteria for use in the methods described herein may contain a functional recA gene (b2699) or may lack a functional recA gene (b2699). For example, reduced genome E. coli strains such as MDS42 or MDS69 (also known as T69 or SG69) can be modified by inactivation of b2699 by complete or partial deletion of the gene from a modified E. coli K-12 strain.
[0050] In some embodiments, the reduced genome E. coli for use in the methods described herein comprises one or more of the following deletions / modifications: (i) a deletion of one or more of the toxin-antitoxin genes in Table 1, (ii) a partial or complete deletion of the rnc gene encoding RNAse III, (iii) a deletion of dinB, (iv) the following modifications: (a) enhanced orotate phosphoribosyltransferase activity, (b) production of active acetohydroxyacid synthase II, and (c) reduced expression of the iclR and arpA genes and (v) a deletion of the recA gene. In related embodiments, the reduced genome E. coli for use in the methods described herein is based on the parent E. coli K12 strain MG1655 and comprises a deletion of MDS42 or MDS69 (also known as T69 or SG69) and one or more of the deletions / modifications listed above. EXAMPLES
[0051] The following examples are illustrative of the scope of the present invention. The specific elements of the examples are for illustrative purposes only and are not intended to limit the scope of the present invention. Those skilled in the art will be able to develop equivalent methods and utilize comparable materials that fall within the scope of the present invention.
[0052] The strains used in the following examples are as follows:
[0053] The "MDS69" strain (also known as SG69 or T69) is a reduced genome E. coli K12 strain whose native parent is MG1655 and which lacks the following genes of MG1655: b0245-b0301, b0303-b0310, b1336-b1411, b4426-b4427, b2441-b2450, b2622-b2654, b2657-b2660, b4462, b1994-b2008, b4435, b3322-b3338, b2349-b2363, b1539-b1579, b4269-b4320, b2968-b2972, b2975-b2 977, b2979-b2987, b4466-4468, b1137-b1172, b0537-b0565, b0016-b0022, b4412-b4413, b0577-b0582, b4415, b2389-b2390, b2392-b 2395, b0358-b0368, b0370-b0380, b2856-b2863, b3042-b3048, b0656, b1325-b1333, b2030-b2062, b2190-b2192, b3215-b3219, b3504 -b3505, b1070-b1083, b1878-b1894, b1917-b1950, b4324-b4342, b4345-b4358, b4486, b0497-b0502, b0700-b0706, b1456-b1462, b3 481-b3484, b3592-b3596, b0981-b0988, b1021-b1029, b2080-b2096, b4438, b3440-b3445, b4451, b3556-b3558, b4455, b1786, b0150- b0153, b2945, b2481-b2492, b2219-b2230, b4500, b3707-b3723, b0644-b0650, b4079-4090, b4487, b4092-b4106, b0730-b0732, b3572-b3587, b1653, b2735-b2740, b2405-b2407, b3896-b3900, b1202, b4263-b4268, b0611, b2364-b2366, b0839, b0488-b0500, and b0502.
[0054] The "MDS69meta" strain (also known as SG69meta or T69meta) lacks the genes listed above for strain MDS69 and further contains the following (meta) modifications: correction of the native rph and ilvG frameshift mutations and deletion of the iclR and arpA genes.
[0055] The "MDS69meta recA" strain (also known as SG69meta recA or T69meta recA) lacks the gene and contains the modifications described above for the MDS69meta strain, and further contains a deletion of the recA gene.
[0056] The "MDS69meta T / A recA" (also known as SG69meta T / A recA or T69meta T / A recA) strain is missing genes and contains the modifications described above for the MDS69meta recA strain, and additionally lacks the following toxin / antitoxin genes: YafQ, dinJ, hha, tomB, gnsA, ymcE, yoeB, yefM, mazF, mazE, mazG, cptA / ygfX, cptB / sdhE, mqsR, mqsA, higB, higA, yhaV, prlF, ldrD, rdlD, istr-2, tisB, chpB, chpS, ratA. , ratB, ldrA, rdlA, ldrB, rdlB, ldrC, rdlC, hokB, sokB, sibA, ibsA, sibB, ibsB, ohsC, shoB, sibC, ibsC, sib D, sibE, ibsA, ibsD, ibsE, agrB, ghoT, ghoS, yfeC, yfeD, fic, yhfG, yhjJ, yhjM, yhjN, yjjJ, ecnA, and ecnB.
[0057] The "MDS69meta T / A LowMut recA" strain (also known as SG69meta T / A LowMut recA or T69meta T / A LowMut recA) lacks the gene and contains the modifications described for the MDS69meta T / A recA strain, and additionally contains a deletion of dinB.
[0058] All of the above MDS strains do not contain an insertion sequence.
[0059] Example 1 Buffer extraction of recombinant carrier proteins in reduced genome strains Methods for extracting recombinant proteins from the periplasm of E. coli are common on a small scale. However, laboratory-scale periplasmic isolation methods use osmotic pressure and expensive enzymes or organic solvents that are not easily or inexpensively used on an industrial scale. The inventors have spent considerable time and effort to develop an economical method for periplasmic extraction that is scalable.
[0060] After testing numerous combinations of buffer systems, detergents, and enzymes, we identified an extraction buffer containing Tris-HCl and EDTA (final conditions = 150 mM Tris pH 7.75, 5 mM EDTA) that was capable of extracting recombinant proteins from the periplasm of E. coli strains (a reduced genome E. coli strain is exemplified). Compared to the standard Epiccenter periplasmic isolation method, this buffer was able to extract 25%, 59%, and 49% of CRM197, Protein D, and EPA, respectively (see Figure 1).
[0061] Example 2 Murein lipoprotein gene deletion Although results with the extraction buffer were promising, genetic methods to enhance extraction yields were investigated given the discrepancies in extraction success between different classes of proteins.
[0062] First, the gene encoding the murein lipoprotein (lpp) was deleted from E. coli strain MDS69meta (alias SG69meta alias T69meta) to generate strain MDSmetaΔlpp (alias SG69metaΔlpp, alias T69metaΔlpp). The MDS69meta strain is a reduced genome E. coli strain based on the parental K-12 strain MG1655 and contains the following deletions compared to MG1655: b0245-b0301, b0303-b0310, b1336-b1411, b4426-b4427, b2441-b2450, b2622-b2654, b2657-b2660, b4462, b1994-b2008, b4435, b3322-b3338, b2349-b2363, b1539-b1579, b4269-b43 20, b2968-b2972, b2975-b2977, b2979-b2987, b4466-4468, b1137-b1172, b0537-b0565, b0016-b0022, b4412-b4413, b0577-b0582 , b4415, b2389-b2390, b2392-b2395, b0358-b0368, b0370-b0380, b2856-b2863, b3042-b3048, b0656, b1325-b1333, b2030-b2062, b2190-b2192, b3215-b3219, b3504-b3505, b1070-b1083, b1878-b1894, b1917-b1950, b4324-b4342, b4345-b4358, b4486, b0497-b 0502, b0700-b0706, b1456-b1462, b3481-b3484, b3592-b3596, b0981-b0988, b1021-b1029, b2080-b2096, b4438, b3440-b3445, b4 451, b3556-b3558, b4455, b1786, b0150-b0153, b2945, b0315-b0331, b0333-b0341, b0346-b0354, b2481-b2492, b2219-b2230, b45 00, b3707-b3723, b0644-b0650, b4079-4090, b4487, b4092-b4106, b0730-b0732, b3572-b3587, b1653, b2735-b2740, b2405-b2407,b3896-b3900, b1202, b4263-b4268, b0611, b2364-b2366, b0839, b0488-b0500, and b0502 (genes deleted from MG1655 to create MDS69 (also known as SG69 or T69)), and further includes the following (meta)modifications: correction of the native rph and ilvG frameshift mutations, deletion of the iclR and arpA genes, and deletion of the recA gene. Lpp (Braun's lipoprotein) is a major outer membrane protein that anchors the periplasmic cytoskeleton-like peptidoglycan layer to the cell envelope, and inactivation of LPP has been found to enhance release of proteins from the periplasm. However, MDS69metaΔlpp strains were found to have a limited ability to release proteins from the periplasm and were susceptible to lysis during long fermentations. Furthermore, the extraction buffer did not cause or enhance the release of recombinant protein from the MDS69metaΔlpp strain (also known as SG69metaΔlpp, or T69metaΔlpp) during continuous fermentation. Nevertheless, some release of recombinant protein was observed. Figure 2 shows a continuous fermentation experiment in which the carrier protein EPA (fused to a periplasmic expression sequence) was expressed in the MDS69metaΔlpp strain (also known as T69metaΔlpp, or SG69metaΔlpp in Figure 2). Once expression was initiated, EPA was released into the medium (Figure 2). However, expression was low during the first three weeks, and seed and production cultures were found to be susceptible to cell lysis. Once conditions were changed to promote more consistent growth, the levels of EPA stabilized somewhat, and culture medium was collected for EPA purification. Production continued for approximately 35 days, although we observed variation in expression during this period, primarily due to changes in the amount of cell lysis. For example, cell density was relatively low (OD, 600 The OD values for fermentations using MDS69meta, also known as T69meta and SG69meta, were 220-270. 600 (Compared to the values, the range was 90-150.)
[0063] The identity of EPA in the medium was initially determined using a Caliper Labchip (Figure 2). However, EPA migrates slower on capillary electrophoresis compared to standard SDS PAGE, and therefore identity was confirmed using a capillary western with an EPA-specific polyclonal antibody (Figure 2 inset). Finally, partially purified EPA subjected to MS unambiguously identified the predicted protein (99% coverage) with no modifications or amino acid position errors (Figure 4).
[0064] The inset in Figure 2 shows the concentration of EPA in the medium of six consecutive daily harvests of 1 liter. The entire purification process was greatly simplified and after a single chromatographic purification step, EPA was purified to >90%, again showing the advantage of using recombinant proteins released into the culture medium as the first step of protein purification and demonstrating the value of developing methods for extracting proteins into the culture medium.
[0065] Although MDS69metaΔlpp (also known as T69metaΔlpp and SG69metaΔlpp) consistently released more EPA compared to MDS69meta (also known as T69meta and SG69meta), its susceptibility to lysis due to extensive modifications of the outer membrane likely limits the application of the released carrier protein to commercial production. For example, MDS69metaΔlpp expressed the carrier protein CRM197 in continuous fermentation (C-Flow), but no release of CRM197 into the culture medium was observed using MDS69metaΔlpp. However, intermittent release of moderate amounts of CRM197 into the medium was observed in continuous fermentation experiments using MDS69meta at normal levels of Lpp.
[0066] Example 3 Co-expression with OmpLA treatment results in extraction of recombinant protein We observed that H. influenzae protein D (PD), the carrier protein for protected nontypeable Haemophilus influenza antigens and pneumococcal conjugate vaccines, was released in moderate amounts into the medium during continuous fermentation of E. coli strain MDS69meta (aka SG69meta aka T69meta) without Tris / EDTA extraction and was removed from the periplasm more effectively than CRM197 and EPA. Figures 3A-C illustrate three PDs and illustrate continuous fermentations in which high levels of PD were observed in the bacterial cell periplasm followed by a transient spike in PD release into the culture medium. Note that the levels of PD in the medium were as high as 10 grams per liter, reflecting the high levels of released recombinant protein. Intact PD in the medium was confirmed by Caliper LabChip using anti-PD antibodies, capillary western (inset in Figure 3B), and SDS-PAGE electrophoresis (not shown). The release did not appear to be the result of extensive cell lysis, as no increase in cellular protein was observed in the medium, and there was no evidence of an increase in materials resulting from cell lysis (i.e., nucleic acids). Furthermore, PDs were readily purified from the medium with little pretreatment prior to chromatography (see below).
[0067] The release of PD was not parallel to the periplasmic PD concentration. The increased expression of PD during the later stages of the fermentation shown in Figure 3C did not result in enhanced levels of PD in the medium. That is, the periplasmic levels of PD increased from days 25 to 30 in Figure 3C, but the PD levels in the medium did not increase significantly. The absence of PD in the medium during this period was not the result of PD degradation, since partially purified PD from a previous fermentation "spiked" into the medium sample from day 28 remained intact after 24 h incubation at 25 °C.
[0068] To confirm the identity of PDs, putative PDs in the medium of C-Flow Figure 1B were separated from contaminating proteins on a 4-12% acrylamide gel, excised from the gel and subjected to mass spectrometry (MS). As shown in Figure 4, MS identified the excised bands as PDs. No contaminating proteins were evident in the excised region illustrating the advantage of using culture medium for PD purification (see below). The complete N-terminus of the PDs (excluding the terminal cysteine) was identified as predicted, with an overall coverage of >99%. Post-translational modifications such as acylation were not evident as expected.
[0069] To identify the mechanism responsible for the release of PD, several culture conditions were modified. One parameter tested, an increase in inducer concentration, was hypothesized to enhance PD in the periplasm and promote increased translocation of PD into the medium. An increase in the amount of PD in the periplasm was observed up to the highest inducer concentration tested (300 μM IPTG), while little or no PD was evident in the culture medium. To determine whether a stressful event caused PD release, we examined both an increase in temperature and a decrease in pH during PD expression in C-Flow. Neither change in fermentation conditions resulted in the release of PD into the medium.
[0070] It was hypothesized that the presence of protein D in the periplasm would alter the outer cell membrane through its endogenous glycerophosphodiesterase activity, but co-induction of protein D with CRM197 or EPA did not affect either the release or extraction of CRM197 or EPA in shake flask cultures.
[0071] Although the trigger for the release of PDs into the culture medium has yet to be identified, a simple purification method for PDs was developed using continuous fermentation (aka C-Flow) samples containing released PDs. Using medium samples containing high levels of PDs (approximately 10 g / L on day 21 in Figure 3B), we developed a purification scheme requiring only two chromatography steps to achieve 95% purity. The ease of purification of PDs was directly attributed to the high percentage of PDs in the medium (58%). Given the simplicity of the purification protocol, the described method is highly suitable for the automation of continuous end-to-end manufacturing processes for recombinant proteins.
[0072] Next, we investigated a series of enzymes that we predicted would promote more complete extraction from the periplasm using shake flask culture. The E. coli outer membrane phospholipase A (OmpLA) enzyme (encoded by the pldA gene) was found to enhance extraction of all three carrier proteins in shake flask culture (Figure 5A-C). Several other E. coli phospholipases were also tested. GlpQ, PldB, and TesA were also found to enhance extraction of carrier proteins, although to a lesser extent than OmpLA.
[0073] In these shake flask culture experiments, each expression plasmid was modified so that the plda gene was placed downstream of each carrier protein (bicistronic construct) and under individual control (i.e., each carrier protein was induced by IPTG and pldA expression was induced by doxycycline (DOX)). The plasmids were then sequence verified and subcloned into MDS69meta (aka SG69meta, aka T69meta). To establish robust carrier protein expression levels, pldA expression was induced 4 hours after induction of the recombinant carrier protein. Results after 16 hours of extraction of each carrier protein are shown in Figure 5A-C. In each case, the amount of carrier protein extracted equaled or exceeded the amount present in the periplasm, based on the Epicentre protocol. Consistent with the experiments described above, higher levels of protein D were present in the medium after extraction compared to the periplasm, as moderate levels of protein D had "leaked" out of the cells during culture. Very little DOX is required to achieve OmpLA levels, resulting in complete extraction of each carrier protein, a key feature for commercial scale-up.
[0074] The production of three recombinant carrier proteins (exoprotein A (rEPA), protein D (PD) and CRM197) in the periplasm and secretion into the medium was examined in shake flask cultures using the reduced genome E. coli host strain MDS69meta recA (also known as T69meta recA or SG69meta recA), which contains the deletion made to create MDS69met (described above) and also contains a deletion of recA using several expression strategies.
[0075] Host cells in each case were transfected with a single expression plasmid construct containing sequences encoding EPA, PD, or CRM197 fused to the YtfQ secretion signal under the control of an IPTG-inducible promoter, and further containing sequences encoding OmplA under the control of a doxycycline (DOX)-inducible promoter.
[0076] Stock cultures of host cells stored at -80 °C were thawed and used to inoculate 10 mL of defined medium (Korz medium supplemented with 0.2% glucose and 50 µg / ml kanamycin). Precultures were grown overnight and used to inoculate shake flasks at the same initial cell density. Once cells reached saturation, the target genes (EPA, PD or CRM197) were induced with 50 µM IPTG (EPA and CRM197) or 100 µM IPTG (PD) (very slow induction), and after approximately 4 h pldA expression was induced with 0–100 nM DOX, induction was allowed to continue overnight at 25 °C. Growth rates and final cell densities were comparable between strains.
[0077] The following morning, 1335 μl of culture was collected and placed into several 2 ml tubes containing concentrated Tris and EDTA (final conditions = 1.5 ml, 150 mM Tris pH 7.75, 5 mM EDTA). The tubes were then incubated for 2 hours (shaking at 25°C) and then assessed for yield, or incubated for 2 hours (shaking at 25°C) and then overnight (25°C) and then assessed for yield. For PD, the tubes were again incubated for 2 hours (shaking at 25°C) and then incubated at 25°C for 6 hours.
[0078] After shaking for 2 h, 6 h, or overnight, the tubes were centrifuged at 4000xg for 15 min and 500 μl of the supernatant was removed and applied to a YM10 spin filter and centrifuged at 13000xg for 30 min. The retentate (representing the medium) was diluted to the equivalent of 0.02 OD / μl with Korz / 0.2% glucose Kan (25 μg / ml) / Tris-EDTA. To assess the periplasmic yield, a center-point peri-prep was performed on a separate aliquot of the culture, final concentration = 0.2 OD μl. All samples were analyzed on caliper tips.
[0079] As can be seen from Figures 7-9, co-expression of the pldA gene resulted in nearly 100% extraction of each carrier protein into the medium, confirming that the method is not limited to any particular recombinant protein. Overnight extraction with Tris / EDTA released significantly more target protein than the 2 hour extraction, and between the 6 hour and overnight extractions nearly all protein was extracted into the medium (compare periplasmic yields against extraction yields - if they are the same, all protein has been extracted).
[0080] One curious observation was the fact that even in cultures in which expression of pldA was not induced (0 DOX), some of the carrier protein was released at these extended extraction times. The promoter controlling pldA expression is known to have some uninduced basal activity based on previous experiments with GFP. It was therefore assumed that in strains carrying a pldA-based plasmid, expression levels of pldA might be higher than wild type. To test that assumption, Tris / EDTA extraction of the target from MDS69meta recA strains transformed with an expression vector encoding only the recombinant carrier protein was compared to an expression vector encoding the recombinant carrier protein and OmpLA. All strains were induced against the target with IPTG and treated with 0, 10, or 100 nM DOX, followed by extraction overnight in Tris / EDTA extraction buffer. The results demonstrate that even in the absence of induction of pldA expression, significantly more target is extracted from strains carrying a pldA-based plasmid, indicating leaky basal expression of OmpLA in the uninduced state. See Figure 10A-C. Using this procedure, yields of 2-3 g / L of CRM197 in medium were obtained.
[0081] Because leaky expression of OmpLA was observed with a single tet operator (as in the carrier protein experiments above), in the following experiments pldA was placed under the control of a dual tet operator. MDS69meta recA host cells were transfected with single expression plasmid constructs containing sequences encoding EPA, PD, or CRM197 fused to the YtfQ secretion signal under the control of an IPTG-inducible promoter and further containing sequences encoding OmpLA controlled by dual tet operators.
[0082] Stock cultures of host cells stored at -80°C were thawed and used to inoculate 10 mL of defined medium (Korz / 0.2% glucose / kanamycin (25 μg / ml)). Precultures were grown overnight and used to inoculate shake flasks at the same initial cell density. Once cells reached saturation, the target genes (EPA, PD or CRM197) were induced with 100 μM IPTG (EPA and CRM197) or 200 μM IPTG (PD) (very slow induction), and after approximately 4 hours, pldA expression was induced with 0, 12.5, 25, 50, 75, or 100 nM DOX, with induction continued overnight at 25°C. Growth rates and final cell densities were comparable between strains.
[0083] The following morning, 1335 μl of culture was collected and placed into several 1.5 ml tubes containing concentrated Tris and EDTA (final conditions = 1.5 ml, 150 mM Tris pH 7.75, 5 mM EDTA). The tubes were then incubated for 6 hours (shaking at 25°C) or overnight, after which the yield was assessed.
[0084] After shaking for 6 hours or overnight, the tubes were centrifuged at 4000xg for 15 minutes and 500 μl of the supernatant was removed and applied to a YM10 spin filter and centrifuged at 13000xg for 30 minutes. The retentate (representing the medium) was diluted to the equivalent of 0.02 OD / μl with Korz / 0.2% glucose / Kan (25 μg / ml) / Tris-EDTA. To assess the periplasmic yield, an Epicentre Peri-Prep was performed on a separate aliquot of the culture to a final concentration = 0.2 OD / μl. All samples were analyzed on caliper tips.
[0085] Data for CRM197 are shown in Figure 11. Unless DOX is added to induce expression of the pldA gene, very little CRM197 is released by Tris / EDTA extraction. This is in stark contrast to previous results obtained with single tetO constructs (Figure 10A above). In these cases, in the absence of induction of pldA gene expression by DOX due to uninduced basal expression of pldA, typically more than 50% of the carrier protein was released after overnight Tris / EDTA extraction. Similar results were observed for PD and EPA (data not shown). Thus, Tris / EDTA extraction of carrier protein in the absence of induced pldA expression is observed only when expression of pldA is controlled by a leaky (e.g., single tetO) promoter and is due to uninduced background OmpLA.
[0086] As seen previously, due to the intrinsic lipase activity of PD, more PD is released in the absence of DOX (which induces OmpLA) compared to CRM197 and EPA, which is also true for the dual tetO-pldA construct, however, when using the dual tetO-PLDA construct, only 1 / 6 of the PD is released in the absence of induction of PLDA expression, compared to more than 1 / 2 with the single tetO-PLDA construct (data not shown).
[0087] Minimizing the expression of the target carrier protein and OmpLA in the non-induced state is beneficial for continuous fermentation to avoid potential selection bias.
[0088] The single plasmid / two promoter constructs described above resulted in yields of 2-3 g / L of CRM197 in medium, however, it was found that these constructs were prone to forming plasmid dimers. Therefore, a bicistronic message approach was tested in which an RBS (ribosome binding site)-pldA insert was placed immediately downstream of CRM197 under the control of a DOX-inducible promoter. In this construct, expression of both CRM197 and OmpLA is induced by DOX. Utilizing this construct in the above procedure yielded 4-5 g / L based on periplasmic preparations and 4 g / L based on Tris / EDTA extracts (data not shown).
[0089] In some experiments, the bicistronic CRM197-RBS-pldA construct was modified to replace wild-type pldA with a codon-altered pldA to greatly reduce the chance of chromosomal integration. At the same time, three different RBS sequences were tested to determine whether co-expression of CRM197 and OmpLA could be optimized to improve the yield of CRM197.
[0090] Example 4 Optimization of CRM197 inducer concentration and extended analysis of CRM197 release into the medium of MDS69meta T / A LowMut recA cells In previous experiments, media samples to monitor CRM197 release were taken after an overnight induction period. A better control sample is generated by taking a media sample from an aliquot of the overnight induced culture and then incubating this in parallel with the Tris / EDTA extract sample (again incubated overnight at 25°C with shaking).
[0091] Previous experiments showed that optimal CRM197 expression was achieved in the doxycycline concentration range of 50-250 nM, but began to decline above 500 nM.
[0092] The next experiment tested induction with 100, 250, and 400 nM DOX and analyzed the release of CRM197 into the medium using improved controls.
[0093] Briefly, MDS69meta T / AT LowMut recA host cells harboring expression plasmids encoding (i) CRM197 fused to the YtfQ secretion signal under the control of the doxycycline promoter or (ii) CRM197 fused to the YtfQ secretion signal under the control of the same doxycycline promoter and a codon-optimized pldA sequence (two separate constructs were used that differ only with respect to the RBS between CRM197 and the pldA ORFS, see Figure 6A) were induced with 100, 250 or 400 nM DOX (medium = Korz / 0.2% glucose / Kan). To assess periplasmic yields, an Epicentre Peri-Prep was performed on an aliquot of the culture equivalent to 2 OD, bringing the final concentration of the preparation to equivalent to 2 OD / μl. To determine the yield of CRM197 in the medium, 1335 μl of the induced culture was placed in a 1.5 ml tube containing concentrated Tris and EDTA (final condition = 1.5 ml, 150 mM Tris, pH 7.75, 5 mM EDTA) and all samples were shaken overnight at 25°C. The supernatant was collected by cfg at 4000xg for 15 min. 500 μl of the supernatant was removed and applied to a YM10 spin filter and cfg at 13000xg for 30 min. The retentate was diluted to the equivalent of 0.02 OD / μl with Korz / 0.2% glucose / Kan (25 μg / ml) / Tris-EDTA.
[0094] Medium control 1: 1 ml of induced culture was placed in a 1.5 ml tube and shaken overnight at 25°C in parallel with the Tris / EDTA extraction.
[0095] Medium control 2: 1 ml of induced culture was placed in a 1.5 ml tube and stored at 4°C for 2 days.
[0096] Supernatants from both media samples were collected and processed following the same procedure as for the Tris / EDTA extracts. A final dilution to 0.02 OD / μl was performed with Korz / 0.2% glucose / Kan (25 μg / ml) / Tris-EDTA.
[0097] The maximum concentration of DOX used for optimal induction of YtfQ-CRM197 from these strains is 400 nM (see FIG. 12C).
[0098] Storage of induced cultures at 4° C. resulted in the release of significantly less CRM197 into the medium than incubation at 25° C. with shaking (compare FIGS. 12B and 12D). Approximately 50% of the CRM197 expressed in the OmpLA co-expressing strain can be released simply by incubation at 25° C. with shaking in the absence of extraction buffer (FIG. 12B). With extraction buffer, nearly 100% of the CRM197 can be released.
[0099] These data demonstrate that although a wide variety of constructs can be used to co-express a phospholipase and a protein of interest with very good results according to the methods described herein, optimal results in terms of both expression of the protein of interest and extraction of the protein of interest into the medium were achieved using a bicistronic construct encoding the protein of interest and a phospholipase separated by an RBS (and under the control of the same promoter, in this case the DOX-inducible promoter), extracting virtually 100% of the protein of interest into the medium. Furthermore, the strength of the ribosome binding site (RBS) was found to directly correlate with the amount of OmpLA expressed (see the right two panels in Figure 12A demonstrating that differences in the strength of the RBS result in different levels of expression of the pldA gene).
[0100] Example 5 CRM197 Periplasmic Yields in Wild-Type E. coli Strains and E. coli Strains with RNase III+ Toxin / Antitoxin Gene Deletions Periplasmic yields of CRM197 were measured in several E. coli strains, including strain MG1655 (the natural parent strain of the multiple deletion strains E. coli described herein), wild-type K12 strain W3110, and strain BL21DE3 (with and without phage deletions), using either the OmpF or YtfQ secretion signals to direct CRM197 to the periplasm.
[0101] Figure 13 demonstrates periplasmic production of CRM197 in the following strains: (1) BL21DE3 (2) BL21DE3 with phage deletion (3) MG1655 and (4) W3110. Briefly, each strain harboring an expression plasmid was grown at 25 °C until the OD600 was sufficient for late induction (0.3 OD600) or very late induction (1.5-2.0 OD600). For late induction, a 20 ml aliquot of the culture was removed into a 125 ml baffled Erlenmeyer flask and induced with 25 μM IPTG for the OmpF-CRM197 construct and 50 μM IPTG for the YtfQ-CRM197 construct. For very slow inductions, 20 ml aliquots of cultures were removed into 125 ml baffled Erlenmeyer flasks and induced with 25-200 μM IPTG for the ompF-CRM197 construct and 500-400 μM IPTG for the YtfQ-CRM197 construct. All inductions were grown for 20 h at 25 °C after which 2 OD of cultures were harvested and periplasmic and spheroplast fractions were prepared using the Epicentre Periplasting method.
[0102] Expression of CRM197 in E. coli B strain BL21DE3 was significantly improved by deleting the phage, and similar results were obtained when either the OmpF or YtfQ secretion signals were utilized.
[0103] The periplasmic yield of CRM197 was assessed in reduced genome E. coli strains engineered to increase the amount of mRNA transcript by (i) deleting the RNA-targeting toxin-antitoxin genes and (ii) deleting RNAse III. CRM197 production in a T69meta Tox / Antitox RNase III strain (containing a deletion of MDS69, a partial (inactivated) deletion of RNase III, and a deletion of the RNA-targeting toxin-antitoxin genes) was compared to CRM197 production in a T69meta strain.
[0104] As illustrated in Figure 14, the combined deletion of the RNA-targeting toxin / antitoxin genes and RNase III significantly increased the periplasmic yield of CRM197, whether or not OmpF or YtfQ signal sequences were used to target CRM197 to the periplasm (yields of 3.5 and 3.3 g / L in SG7128 and SG800 strains compared to 2.6 g / L in SG7755). Additive and even synergistic increases in recombinant protein yields could be obtained in E. coli strains containing deletions of the toxin-antitoxin genes and RNase III. Co-expression of the chaperone pspA did not increase the yield of CRM197 and even tended to decrease it, possibly due to a shift in the cellular machinery towards the production of the chaperone.
[0105] The results are summarized in Table 2 below. [Table 2]
[0106] Example 6 Continuous fermentation system for the production of carrier proteins and their extraction into culture medium Reduced genome E. coli strains can sustain continuous growth and protein production for over 60 days in continuous fermentation culture (continuous flow or "C-Flow"). See FIG. 15, which illustrates continuous growth and production of CRM197 at (periplasmic) yields of 4-5 g / liter / day over a period of more than 30 days (3 weeks shown), which is not possible with conventional E. coli production strains BL21, MG1655 and BLR, where yields of about 1 g / L are initially obtained (which gradually decrease) and continuous fermentation cannot be sustained for periods longer than a few days as non-producing mutants take over the culture.
[0107] To determine whether buffer extraction of recombinant proteins in shake flask cultures is scalable to the high bacterial cell densities achieved during continuous fermentation, an extended fermentation using SG69meta (aka MDS69meta aka T69meta), which contains a single expression plasmid containing sequences encoding CRM197 under the control of an IPTG-inducible promoter and pldA under the control of a doxycycline-inducible promoter, is used in the shake flask culture experiments described above. An extended fermentation showing expression of CRM197 over a two week period is shown in Figure 16. Expression of CRM197 averages 1 g / L / day and the fermentation reaches a relatively consistent, high density OD of 200. 600was maintained. Extraction experiments performed on high-density continuous fermentation fermentates resulted in 90% recovery of CRM197 (see panels A and B of FIG. 16). In the experiment shown, 100 ml of undiluted fermentation was taken from a 1-liter batch collection vessel. Doxycycline was then added to the 100 ml fermentation aliquots and the samples were incubated for 4 hours in a shaking chest incubator with agitation at 25° C. Extraction buffer was added (equal to 1 / 10 volume of fermentation) and the samples were incubated for an additional 16 hours. Tris / EDTA extraction in the absence of DOX induction resulted in extraction of approximately half of the CRM197 present in the periplasm (due to uninduced expression of ompLA), while DOX induction in combination with Tris / EDTA resulted in extraction of 90% of CRM197. These results demonstrate that the extraction method described herein is adaptable to high-density continuous fermentation.
[0108] As described herein, continuous fermentation of reduced genome E. coli strains (capable of growing to extremely high densities) fused to a periplasmic signal sequence and an E. coli phospholipase (with the carrier protein and phospholipase regulated by a single promoter or separate promoters) allows long-term, high-yield production of the encoded protein and complete extraction of the target protein into the medium using simple buffers, greatly simplifying downstream processing (e.g., purification) steps.
[0109] Importantly, production of CRM197 and other carrier proteins is not feasible in commercially available E. coli B strains such as BL21 and BLR, nor in E. coli K12 strain MG1655, the parent strain of the reduced genome E. coli strains described herein. Using these strains, continuous fermentation longer than 7 days is not possible, and production of CRM197 drops rapidly from about 1 g / L / day to about 0 g / L / day. Using the reduced genome E. coli strains described herein, continuous fermentation runs of up to 65 days have been achieved, with consistent production of PD at about 9 g / L / day and CRM197 at about 4-5 g / L / day for up to 30 days, with significant improvements in purification from the medium that are possible with co-expression of lipase. There appears to be no limit to extending continuous flow to continuous production for 365 days or more.
[0110] Selective downstream processing Continuous culture (also known as continuous flow or "C-Flow") may include a downstream culture clarification process for continuous automated clarification to remove intact cells or debris from the fermentation supernatant, and may include miniaturized, low-cost chromatography systems for easy purification of the desired product from the clarified supernatant by co-linking a chromatography process with the flow rate of the continuous culture.
[0111] CRM197 produced using the compositions and methods described herein is highly consistent, up to 98% pure, free of methylation, acetylation, gluconoylation or phosphorylation, and has extremely low endotoxin levels (<25EU / mg). Considering that the amount of CRM197 per dose in commercial vaccines ranges from 12-65ug, potential endotoxin contamination is less than 0.5EU per dose in standard vaccines, well below the industry standard of <10EU / dose for polysaccharide vaccines.
[0112] Example 7 Deletion of the toxin-antitoxin genes significantly improves carrier protein production The effect of deletion of toxin-antitoxin genes on carrier protein production in E. coli was evaluated. The following toxin-antitoxin genes were deleted from several MDS E. coli strains to generate the corresponding MDS T / A strains: YafQ, dinJ, hha, tomB, gnsA, ymcE, yoeB, yefM, mazF, mazE, mazG, cptA / ygfX, cptB / sdhE, mqsR, mqsA, higB, higA, yhaV, prlF, ldrD, rdlD, istr-2, tisB, chpB, chpS, ratA, ratB, ldrA, rdlA, tisB, tisC, tisD, tisD, tisE, tisF ... ldrB, rdlB, ldrC, rdlC, hokB, sokB, sibA, ibsA, sibB, ibsB, ohsC, shoB, sibC, ibsC, sibD, sibE, ibs D, ibsE, dinQ, agrA, agrB, ghoT, ghoS, yfeC, yfeD, fic, yhfG, yhjJ, yhjM, yhjN, yjjJ, ecnA, and ecnB.
[0113] Production of CRM197, EPA, and PD was evaluated in MDS69meta rec A (containing the toxin-antitoxin genes) and in the derivatives MDS69meta T / A recA and MDS69meta T / A LowMut recA, in which the toxin-antitoxin genes were deleted.
[0114] Periplasmic yields of CRM197, PD, and EPA were assessed in each host strain harboring expression plasmids encoding CRM197, EPA, or PD fused to the YtfQ secretion signal under the control of an IPTG- or doxycycline-inducible promoter.
[0115] For CRM197 with IPTG induction, three colonies from each strain were used to inoculate separate flasks with 25 ml Kor / 0.2% glucose / Kan (25 μg / ml) at 37°C. 25 ml of inoculum was used to inoculate subsequent 100 ml cultures in 500 ml baffled E-flasks to OD600=0.25. All cultures were grown at 25°C to the following approximate OD600 (corrected): 2.8 for MDS69meta recA (SG7766 in Table 3 below) and 3.0 for MDS69meta T / A LowMut recA (SG8081 in Table 3 below). 20 ml aliquots were then removed and placed in 125 ml baffled E-flasks and induced with 25 μM, 50 μM, 100 μM or 200 μM IPTG and the cultures were shaken overnight at 25°C. After 16-20 h of incubation at 25°C, 2 OD of cultures were harvested by cfg (10 min at 7500xg) and periplasm was isolated in triplicate from 2 OD by EpiCentre periplasm preparation (15 min at 4000xg). All samples were at a final concentration of 0.02 OD / μl and analyzed with the same caliper tip in the same assay.
[0116] As can be seen from Table 3 below, production of CRM197 was significantly increased in strains in which the toxin-antitoxin genes were deleted (approximately 38% increase in CRM197). [Table 3]
[0117] For EPA and PD with IPTG induction, three colonies from each strain were used to inoculate separate flasks with 25 ml Kor / 0.2% glucose / Kan at 37°C. The 25 ml inoculum was used to inoculate subsequent 100 ml cultures to OD600=0.25 in 500 ml baffled E-flasks. All cultures were grown to saturation at 25°C, OD600m (corrected): 2.9 for MDS69meta recA and 3.2 for MDS69meta T / A LowMut recA. 20 ml aliquots were then removed and placed in 125 ml baffled E-flasks and induced with 25 μM, 50 μM, 100 μM or 200 μM IPTG overnight at 25°C. After 16–20 h incubation at 25°C, 2 OD of cultures were harvested by centrifugation (10 min at 7500xg) and periplasmic contents were isolated in triplicate from 2 OD (15 min at 4000xg) by the EpiCentre periplasmic preparation method. All samples were at a final concentration of 0.02 OD / μl and analyzed with the same caliper tip in the same assay.
[0118] As can be seen from Table 4 below, production of PD was significantly increased in the strain in which the toxin-antitoxin genes were deleted (approximately 25% increase in PD). [Table 4]
[0119] We next assessed the periplasmic yield of CRM197 in three strains harboring an expression plasmid encoding CRM197 fused to the ytfQ signal sequence under the control of a dual tetO operator / tet repressor (two copies of the operator but only one copy of the tet repressor): (i) MDS69meta RecA (aka T69 Meta recA aka SG69 Meta recA), (ii) MDS69meta T / A recA (aka T69 Meta Tox / Atox recA aka SG69 Meta Tox / Atox recA), and (iii) MDS69meta T / A LowMut recA (aka T69 Meta Tox / Atox LowMut recA (aka SG69 Meta Tox / Atox LowMut recA).
[0120] Three colonies of each strain were used to inoculate separate flasks with 25 ml Kor / 0.2% glucose / Kan (25 μg / ml) at 37°C. The 25 ml inoculum was used to inoculate subsequent 100 ml cultures to OD600=0.25 in 500 ml baffled E-flasks. All cultures were grown to saturation at 25°C. Then, 20 ml aliquots were removed and placed in 125 ml baffled E-flasks and induced with 50 or 100 nM DOX overnight at 25°C. After 16–20 h of incubation at 25°C, cultures were harvested at 2 OD by centrifugation (10 min at 7500xg) and periplasmic proteins were isolated in triplicate from 2 OD (15 min at 4000xg) by the EpiCentre periplasmic preparation method. All samples were at a final concentration of 0.02 OD / μl and analyzed with the same caliper tip in the same assay.
[0121] Deletion of the toxin / antitoxin genes significantly enhanced the yield of CRM197 in the periplasm (from 3.9 g / L to 5.1 g / L, see also Table 5 below, Figure 18). [Table 5]
[0122] These results demonstrate the advantages of E. coli strains lacking toxin-antitoxin genes, and these advantages do not appear to be restricted to specific E. coli strains.
[0123] Example 8 OmpLA co-expression facilitates complete extraction of recombinant antibodies. Next, the extraction of relevant recombinant proteins was expanded to include a single chain antibody (scFab YMF10) and a single chain variable fragment (scFv 75127, kindly provided by Fritz Schaumburg, Lytic Solutions, Madison, WI). Both targets represent highly relevant recombinant therapeutic proteins. Briefly, MDS69metaRecA (aka T69 MetaRecA, aka SG69metaRecA) carrying an expression plasmid encoding either antibody fused to the OmpA periplasmic signal sequence under IPTG control and encoding OmpLA under the control of a dual tetO DOX inducible promoter is induced (very slow induction) with 100 μM IPTG and pldA induced 4 hours later with 0, 50 or 100 nM DOX. After overnight induction, 1335 ml of culture was placed into a 2 ml tube containing concentrated Tris and EDTA (final concentration = 1.5 ml, 150 mM Tris pH 7.75, 5 mM EDTA) and incubated overnight at 25°C with shaking. For the periplasm, an Epicentre Peri-Prep was performed on a separate aliquot of the culture, final concentration = 0.02 OD / μl. All samples were analyzed on caliper tips.
[0124] High levels of expression and subsequent periplasmic delivery of scFv were observed (scFv was co-expressed with the chaperone protein skp, which provides proper folding in the periplasm). Large amounts of scFv 75127 were present in the culture medium after extraction (Figure 19A), levels that were much higher than those observed in the periplasm. See also Figures 20A and 20B illustrating periplasmic and extraction yields of scFv 75127. These results confirm that the expression and extraction method to produce large amounts of carrier protein works for single chain antibodies as well.
[0125] Expression and extraction of the single chain antibody scFab YMF10 was also investigated. The scFab YMF10 was found to be highly sensitive to degradation in conventional E. coli expression strains and to undergo slow degradation in MDS69meta (aka SG69meta, aka T69meta). To eliminate the degradation issue, an SG69 expression strain was utilized in which additional proteases were removed compared to SG69meta. Stable expression of scFab YMF10 (and the chaperone, skp) was achieved in this protease-reduced strain (Figure 19B). See also Figures 21A and 21B illustrating the periplasmic and extraction yields of scFab YMF10. Similar to the scFvs, higher amounts of scFab YMF10 were evident in the extracted samples than in the Eccenter periplasmic isolation samples. Small amounts of scFab were present in the pre-extraction medium (green triangles in Figure 19B), which could explain this discrepancy. These results highlight the versatility of both the novel extraction method and the reduced genome E. coli described herein.
Claims
1. It has a genome of 4.41 Mb to 2.78 Mb and contains at least the following toxin-antitoxin genes: yafQ, dinJ, hha, tomB, gnsA, ymcE, yoeB, yefM, mazF, mazE, mazG, cptA / ygfX, cptB / sdhE, mqsR, mqsA, higB, higA, yhaV. , prlF, ldrD, rdlD, istR-2, tisB, chpB, chpS, ratA, ratB, ldrA, rdlA, ldrB, rdlB, l drC, rdlC, hokB, sokB, sibA, ibsA, sibB, ibsB, ohsC, shoB, sibC, ibsC, sibD, sibE, A non-naturally occurring E. coli bacterium lacking ibsD, ibsE, dinQ, agrA, agrB, ghoT, ghoS, yfeC, yfeD, fic, yhfG, yhjJ, yhjM, yhjN, yjjJ, ecnA, and ecnB, and lacking all of the IS1, IS2, IS3, IS5, IS150, and IS186 insertion sequences.
2. The bacterium of claim 1 , comprising a complete deletion of the rnc gene.
3. The bacterium of claim 1 , wherein the bacterium further comprises a deletion of the dinB gene.
4. The bacterium of claim 1 , wherein the bacterium further comprises a deletion of the recA gene.
5. The bacterium according to claim 1, wherein the parent strain of the bacterium is a K-12 or B strain.
6. The bacterium according to claim 5 , wherein the parent strain of the bacterium is a K-12 strain.
7. 7. The bacterium of claim 6, further comprising one or more of the following modifications: (i) a deletion of the rph gene to enhance orotate phosphoribosyltransferase activity, (ii) a correction of the native -2 frameshift mutation in the ilvG gene relative to the wild-type ilvG gene to restore active acetohydroxyacid synthase II production, and (iii) a deletion of all of the iclR and arpA genes.
8. The bacterium according to claim 6, wherein the parent strain of the bacterium is K-12 strain MG1655.
9. The bacterium comprises at least the following genes of E. coli K-12 strain MG1655: b0245-b0301, b0303-b0310, b1336-b1411, b4426-b4427, b2441-b2450, b2622-b2654, b2657-b2660, b4462, b1994-b2008, b4435, b3322-b3338, b2349-b2363, b1539-b15 79, b4269-b4320, b2968-b2972, b2975-b2977, b2979-b2987, b4466-4468, b1137-b1172, b0537-b0565 , b0016-b0022, b4412-b4413, b0577-b0582, b4415, b2389-b2390, b2392-b2395, b0358-b0368, b0370-b 0380, b2856-b2863, b3042-b3048, b0656, b1325-b1333, b2030-b2062, b2190-b2192, b3215-b3219, b3 504-b3505, b1070-b1083, b1878-b1894, b1917-b1950, b4324-b4342, b4345-b4358, b4486, b0497-b050 2, b0700 to b0706, b1456 to b1462, b3481 to b3484, b3592 to b3596, b0981 to b0988, b1021 to b1029, b2080 to b2096, b4438, b3440 to b3445, b4451, b3556 to b3558, b4455, b1786, b0150 to b0153 and b2945.
10. The bacterium comprises at least the following genes of E. coli K-12 strain MG1655: b0245-b0301, b0303-b0310, b1336-b1411, b4426-b4427, b2441-b2450, b2622-b2654, b2657-b2660, b4462, b1994-b2008, b4435, b3322-b3338, b2349-b2363, b1539-b1579, b4269-b4320, b2968-b2972, b2975-b2977, b2979-b2987, b4466-b4468, b4469-b4469 ... 4468, b1137-b1172, b0537-b0565, b0016-b0022, b4412-b4413, b0577-b0582, b4415, b2389-b2390, b2392-b2395, b0358-b0368, b037 0-b0380, b2856-b2863, b3042-b3048, b0656, b1325-b1333, b2030-b2062, b2190-b2192, b3215-b3219, b3504-b3505, b1070-b1083, b 1878-b1894, b1917-b1950, b4324-b4342, b4345-b4358, b4486, b0497-b0502, b0700-b0706, b1456-b1462, b3481-b3484, b3592-b359 6, b0981 to b0988, b1021 to b1029, b2080 to b2096, b4438, b3440 to b3445, b4451, b3556 to b3558, b4455, b1786, b0150 to b0153, b2945, b2481 to b 6. The bacterium of claim 5, which is lacking the following genes: b2492, b2219 to b2230, b4500, b3707 to b3723, b0644 to b0650, b4079 to 4090, b4487, b4092 to b4106, b0730 to b0732, b3572 to b3587, b1653, b2735 to b2740, b2405 to b2407, b3896 to b3900, b1202, b4263 to b4268, b0611, b2364 to b2366, b0839, b0488 to b0500, and b0502.
11. 2. The bacterium of claim 1, comprising a nucleic acid comprising a nucleotide sequence encoding a recombinant polypeptide, said nucleotide sequence being operably linked to an expression control sequence.
12. 11. A method for producing a recombinant polypeptide, the method comprising culturing the bacterium of claim 10 under conditions suitable for expressing the recombinant polypeptide.
13. The bacterium of claim 1, wherein the bacterium comprises a nucleic acid comprising a nucleotide sequence encoding a recombinant polypeptide fused to a signal sequence that directs import of the recombinant polypeptide into the periplasm, and a nucleic acid comprising a nucleotide sequence encoding an E. coli phospholipase.
14. The bacterium of claim 6, further comprising all of the following modifications: (i) deletion of the rph gene to enhance orotate phosphoribosyltransferase activity, (ii) correction of the native-2 frameshift mutation in the ilvG gene relative to the wild-type ilvG gene to restore active acetohydroxyacid synthase II production, and (iii) deletion of all of the iclR and arpA genes.
Citation Information
Patent Citations
Method for polypeptide production in the periplasm of prokaryotic cells
JP2015528477A
Enhanced production of recombinant crm197 in E. coli
JP2017508454A
Method for Enhancing Extracellular Secretion of Recombinant Proteins in Escherichia coli by Co-expressing Thermobifida Fusca Cutinase
US20140011239A1
