Production of quillaja saponaria saponins

Abstract

Quillaja saponins are produced from leaves using two-stage extraction / purification to remove sequentially phenolics, and then chlorophyll, to produce light-colored final extracts with minimum saponin losses and minimum color degradation products.

Description

Production of Quillaja Saponaria Saponins

[0001] Government Support Clause

[0002] This invention was made with government support under grant number 70NANB22H016 awarded by the National Institutes of Standards and Technology. The government has certain rights in the invention.

[0001] Introduction

[0002] Quillaja saponaria Molina is an endemic Chilean evergreen tree that grows in the sclerophyll forests of Central Chile (Donoso et al., 2011). Q. saponaria belongs to the genus Quillajacea, which has only one other species, Q. brasiliensis (Magedans et al., 2019). Unless explicitly stated otherwise, the terms quillaja, quillaja leaves, or any related terms as used in this text refer to Quillaja saponaria and its derivatives. Despite the geographic disparity of their respective ranges, mature Quillajacea spp. resemble, in both appearance and adaptation strategies, the Quercus spp. that grow in California (Avila et al., 1975).

[0003] Q. saponaria bark and biomass of wood and bark (without leaves) have been used for over 100 years for the extraction of triterpenoid quillaja saponins, which are high molecular weight glycosides containing a hydrophobic triterpenic nucleus and two hydrophilic sugar chains (Nord & Kenne, 1999; San Martin & Briones, 2000). Aqueous extracts of Q. saponaria bark and biomass contain a complex mixture of > 100 saponins (Nord & Kenne, 1999), plus aqueous-soluble compounds like phenolics, sugars, salts, and proteins.

[0004] Quillaja extracts are classified as GRAS (Generally Recognized as Safe) and approved for human consumption by regulatory bodies, including the U. S. FDA, the European Community, and Japan. This approval permits their use in the U. S. as foaming agents in beverages like root beer and slush drinks, and additionally as emulsifiers in various food applications. Ongoing research by both industry and academia highlights an expanding range of applications for quillaja extracts, including uses in food, biopesticides, cosmetics, and more (Reichert et al., 2019).

[0005] However, the supply of these extracts faces significant challenges due to the declining population of wild quillaja trees in Chile. Overexploitation, coupled with prolonged drought in the tree’s native regions, suggests that the sustainable limits of natural forests for biomass production have been exceeded. Moreover, recent media coverage on the use of quillaja saponin for COVID-19 vaccines has also heightened environmental concerns within Chile, prompting stricter regulations on the exploitation of wild quillaja forests.1 B25-074-2WO

[0006] These challenges have compelled some companies to establish quillaja plantations within the past two years. However, the trees require over 10-12 years of growth before they can produce economically viable amounts of wood and bark. After harvesting, new trees must be planted, demanding substantial land area to provide enough raw material to reduce dependence on wild forests.

[0007] Current Production Process

[0008] Currently, two main parts of wild quillaja trees, typically over 30 years old, are used to produce quillaja saponins:

[0009] Bark: Bark is primarily used for producing vaccine adjuvants, such as the QS-21 adjuvant, due to its high concentration of saponins at commercially viable levels. This material is obtained by debarking wild quillaja trees in Chile. Debarked trees do not regenerate.

[0010] Quillaja Biomass: This includes the wood, with or without bark, but excludes the leaves, which are left in the field to decompose. Biomass is utilized in applications other than vaccine adjuvants, such as food emulsifiers, foaming agents, biopesticides, antiviral agents, and feed additives. It is sourced from wild trees in Chile, either through pruning existing forests or by felling mature trees.

[0011] Both raw materials undergo aqueous extraction of milled material, mainly wood chips or quillaja bark, followed by various purification methods involving additives and membrane technologies (San Martin and Briones, 1999; Padilla Iglesias & Valencia Michaud, 2022). The methods described in these references are not suitable to produce saponins from leaves, as they do not address how to remove unique compounds present in leaves, such as chlorophyll.

[0012] Summary of the Invention

[0013] This invention transforms saponin production by unlocking the untapped potential of quillaja leaves, as well as all aerial biomass obtained during leaf harvesting, a renewable, fastgrowing resource containing 5-7 times more saponins than wood. Quillaja leaves can be sustainably and repeatedly harvested, and the trees require minimal water to survive, supporting large-scale cultivation in diverse climates, such as California, and breaking the industry's reliance on Chile as the sole supplier. Previously, the lack of a scalable method to extract saponins from leaves while efficiently removing chlorophyll and phenolics posed a major barrier to industrial adoption.

[0014] The invention introduces different extraction and purification systems to produce saponin-rich extracts from quillaja leaves, via: 1) single-stage batch extraction, 2) continuous extraction and 3) two-stage batch extraction. All of these processes result in high-purity, light-colored saponins for food, therapeutic and immunological applications.2 B25-074-2WO

[0015] Practical applications range from food emulsifiers and foaming agents to vaccine adjuvants for diseases like shingles, malaria, and COVID-19.

[0016] This invention provides a scalable process for producing saponins from the leaves of quillaja plants, including mature trees as well as shrubs cultivated in plantations, enabling sustainably expanding the supply of quillaja saponins. A prior study reported saponin content in leaf extractions (Schlotterbeck et al., 2015); however, no industrial, nor scalable production method has been developed until now.

[0017] Compared to quillaja bark or biomass, leaves present unique purification challenges due to the presence of strongly colored pigments, including chlorophyll and phenolics, that must be partially or fully removed to create a commercially viable product. These compounds can introduce unwanted flavors and colors. Industrial applications call for ingredients that are light-colored, neutral in taste, and contain a well-defined concentration of active ingredients. The lack of a scalable production process for food-grade, pharmaceutical and immunological saponins from leaves has, until now, limited the use of this part of the plant for saponin production.

[0018] In aspects this invention provides three scalable extraction systems: 1) single-stage batch extraction, 2) continuous extraction and 3) two-stage batch extraction. These approaches minimize the formation of unwanted degradation products, resulting in cleaner, higher-purity extracts. They also require fewer steps compared to previously developed methods for extracting and purifying saponins from biomass and bark (Padilla Iglesias et al., 2021; Padilla et al., 2022), resulting in reduced saponin losses.

[0019] In aspects and embodiments the invention provides:

[0020] 1. A method for producing saponins, comprising

[0021] a) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaja leaves;

[0022] b) treating the quillaja leaves with an aqueous solution comprising an antioxidant and a chelating agent at a pH between 7.5 and 8.5 and a temperature between 45-65 °C;

[0023] c) separating the treated quillaja leaves after step b) to obtain an extract;

[0024] d) contacting the extract from step c) with a nutritionally-inert sorbant, such as calcium bentonite, sodium bentonite or activated charcoal;

[0025] e) filtering the extract from step d) in the presence of divalent cations or at a pH between 7.5 and 8 to yield a purified saponin composition.

[0026] 2. I’he method of claim 1, to produce semi refined, refined and highly refined saponins from leaves.

[0027] 3. The method of claim 1, comprising a) batch extraction, b) continuous extraction, or c) two-stage extraction / purification, with sequential removal of phenolics (stage 1) and chlorophyll 3 B25-074-2WO(stage 2), to produce light-colored final extracts with minimal saponin losses and minimal color degradation products.

[0028] 4. The method of claim 1, providing significant enhancement of saponin recovery yields operating at basic pH and high temperatures.

[0029] 5. The method of claim 1, comprising use of magnesium ions to minimize chlorophyll degradation when extracting at basic pH and high temperature.

[0030] 6. The method of claim 1, comprising use of pH shifts, activated charcoal and bentonites to reduce chlorophyll content in the extract.

[0031] 7. The method of claim 1, comprising use of diafiltration with ultrafiltration membranes in the presence of divalent ions or the use of basic pH to disrupt the bonds between saponins and phenolics, effectively removing phenolics via diafiltration with water.

[0032] 8. A method of determining the source of a saponin composition, the method comprising detecting phenolic markers in the composition, wherein quillaja leaves have unique phenolic compounds that are not found in quillaja bark, e.g., quercetin (and derivatives), and oleuropein, wherein these compounds serve as markers specific to leaf derived extracts.

[0033] 9. A method for producing quillaja saponins from leaves, comprising processing steps to produce semi-refined quillaja extracts from leaves, substantially as shown in Fig. 3.

[0034] 10. A method for producing quillaja saponins from leaves, comprising processing steps to produce refined quillaja extracts from leaves, substantially as shown in Fig. 4.

[0035] 11. A method for producing quillaja saponins from leaves, comprising an cxtraction / purification process flowsheet to produce refined quillaja extracts using buffers, substantially as shown in Fig. 5.

[0036] 12. A method for producing quillaja saponins from leaves, comprising processing steps to produce refined quillaja extracts from leaves using a continuous process, substantially as shown in Fig. 12.

[0037] 13. A method for producing quillaja saponins from leaves, comprising a two-stage extraction / purification process, substantially as shown in Fig. 14.

[0038] 14. A method for producing saponins, comprising:

[0039] a) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaja leaves;

[0040] b) treating the quillaja leaves with an aqueous solution comprising an antioxidant and a chelating agent at a pH between 7.5 and 8.5 and a temperature between 45-65 °C;

[0041] c) separating the treated quillaja leaves after step b) to obtain an extract;

[0042] d) contacting the extract from step c) with a nutritionally-inert sorbant, such as calcium bentonite, sodium bentonite or activated charcoal;4 B25-074-2WO

[0043] e) filtering the extract from step d) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a purified saponin composition.

[0044] 15. The method of claim 14, wherein step b) is performed for a duration between 1 hour and 3 hours.

[0045] 16. The method of claim 14 or 15, further comprising:

[0046] washing the separated leaves from step c) with water to obtain a wash extract;

[0047] combining the wash extract with the extract in step c); and

[0048] containing the combined extract with the bentonite in step d).

[0049] 17. A method for producing saponins, comprising:

[0050] a) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaj a leaves;

[0051] b) treating the quillaja leaves for 1-3 h with an aqueous solution comprising an antioxidant and a chelating agent at a pH between 4 and 5.5 and at room temperature;

[0052] c) separating the treated quillaja leaves after step b) to obtain a first extract;

[0053] d) filtering the first extract from step c) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a first purified saponin composition;

[0054] e) subjecting the separated, treated quillaja leaves from step c) to a hot alkaline extraction at a pH between 7.5 and 8.5 and at a temperature between 45 °C and 65 °C to obtain a second extract;

[0055] f) contacting the second extract from step e) with a nutritionally-inert sorbant, such as calcium bentonite, sodium bentonite, or activated charcoal;

[0056] g) filtering the second extract from step f) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a second purified saponin composition.

[0057] 18. The method of claim 17, wherein step e) is performed for a duration between 0.5 hours and 3 hours.

[0058] 19. The method of claim 17 or 18, further comprising:

[0059] combining the first purified saponin composition and the second purified saponin composition.

[0060] 20. A method for producing saponins, comprising:

[0061] a) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaja leaves;5 B25-074-2WO

[0062] b) treating the quillaja leaves with an aqueous solution comprising an antioxidant and a chelating agent and buffer at a pH between 6.5 and 8 and at a temperature between 15 °C and 30 °C;

[0063] c) separating the treated quillaja leaves after step b) to obtain an extract;

[0064] d) filtering the extract from step c) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a purified saponin composition.

[0065] 21. The method of claim 20, wherein the buffer comprises phosphate ions.

[0066] 22. The method of claim 20 or 21, wherein the buffer can be used alone or in combination with sodium chloride.

[0067] 23. The method of any one of claims 20 to 22, wherein step b) is performed for a duration between 0.5 hours and 16 hours.

[0068] 24. The method of any one of claims 20 to 23, further comprising:

[0069] drying the purified saponin composition to obtain saponin-enriched powder.

[0070] 25. The method of any one of claims 14 to 24, wherein the quillaja leaves are fresh, dried, partially dried, intact, chopped, shredded, ground, milled, or photobleached.

[0071] 26. The method of any one of claims 14 to 25, wherein the antioxidant is a natural antioxidant.

[0072] 27. The method of any one of claims 14 to 25, wherein the antioxidant comprises ascorbic acid, cysteine or citric acid, or any combination thereof.

[0073] 28. The method of any one of claims 14 to 27, wherein the chelating agent comprises ethylenediaminetetraacetic acid (EDTA), citric acid or phosphoric acid, or any combination thereof.

[0074] 29. The method of any one of claims 14 to 28, wherein the divalent cations are calcium, zinc or magnesium cations.

[0075] 30. The method of any one of claims 14 to 29, wherein the method is performed as a batch process.

[0076] 31. The method of any one of claims 14 to 29, wherein the method is performed as a continuous process.

[0077] 32. The method of any one of claims 14 to 31, wherein the chlorophyll index of each extract obtained in the method is below 0.25.

[0078] 33. A purified saponin composition obtained according to a method of any one of claims 1 to 32.

[0079] 34. A purified saponin composition, having a saponin content of at least 80% w / w, and comprising detectable amounts of phenolic markers present in quillaja leaves from which the 6 B25-074-2WOpurified saponin composition is obtained, wherein the total phenolic markers present in the composition is no more than about 5 % w / w.

[0080] 35. The composition of claim 34, wherein the ratio of phenolic markers to saponins is between 0.01 and 0.15 w / w.

[0081] 36. A purified saponin composition comprising saponins and phenolic markers, wherein the ratio of phenolic markers to saponins is between 0.01 and 0.15 w / w.

[0082] 37. The composition of any one of claims 34 to 36, wherein the phenolic markers comprise: quercetin pentoside; dihexosylquercetin; or oleuropein, or any combination thereof.

[0083] 38. The composition of any one of claims 34 to 36, comprising: between 0.5 and 2 % w / w quercetin pentoside; between 0.5 and 2 % w / w dihexosylquercetin; or between 0.1 and 1 % w / w oleuropein, or any combination thereof.

[0084] 39. The composition of any one of claims 34 to 38, wherein the composition is in powder form.

[0085] 40. The composition of claim 39, wherein the composition has a Hunter Lab Lightness (L) value of at least about 75.0.

[0086] The invention encompasses all combinations of the particular embodiments recited herein, as if each combination had been laboriously recited.

[0087] Brief Description of the Drawings

[0088] Fig. 1. RP-HPLC chromatograms of commercial purified quillaja extract derived from biomass, 6 g saponins / L (top), vs. purified extract derived from Q. saponaria leaves of 2-year-old shrubs, 8 g saponins / L (bottom).

[0089] Fig.2. Milled green leaves (left) and photobleached leaves (right).

[0090] Fig.3. Processing steps to produce semi-refined quillaja extracts from leaves.

[0091] Fig.4. Processing steps to produce refined quillaja extracts from leaves.

[0092] Fig.5. Simplified extraction / purification process flowsheet to produce refined quillaja extracts using buffers.

[0093] Fig.6. RP-HPLC chromatogram of a purified Quillaja saponaria leaf extract showing the three major saponin fractions — QH-A, QH-B, and QH-C — commonly associated with immunological applications.

[0094] Fig.7. UPLC chromatograms of leaves extracted at 60°C and pH 4.6 (top) and 8.5 (bottom), showing a larger proportion of hydrophilic saponins, e.g., fraction QH-A.

[0095] Fig.8. Zeta-potential of thylakoid membranes with pH change. Thylakoid membranes were extracted from sugar beet leaves (Tamayo Tenorio et al., 2017).

[0096] Fig.9. RP-HPLC chromatogram of a non- refined extract from quillaja leaves.7 B25-074-2WO

[0097] Fig. 10. Extract of quillaja leaves, post diafiltration in the presence of divalent ions.

[0098] Fig. 11. Extract of quillaja leaves, post diafiltration at pH 8.

[0099] Fig. 12. Processing steps to produce refined quillaja extracts from leaves using a continuous process.

[0100] Fig. 13. Equipment used for the continuous production of saponins from quillaja leaves at a pilot plant scale: sugar cane shredder (left) and commercial screw dewatering press (right).

[0101] Fig. 14. Flow diagram of the two-stage extraction / purification system.

[0102] Fig. 15. RP-HPLC chromatogram of stage 1 of the two-stage process, during the percolation phase (Stage 1).

[0103] Fig. 16. RP-HPLC chromatogram of stage 1 of the two-stage process, during the wash phase of Stage 1.

[0104] Fig. 17. RP-HPLC chromatograms of extracts treated with activated charcoal filters at pH 5. Top: initial sample showing strong initial signal of hydrophilic phenolics and saponins.Bottom: Sample post- activated charcoal treatment showing removal of hydrophilic phenolics and saponins.

[0105] Fig. 18. RP-HPLC chromatograms of extracts extracted at pH 8 and 52 °C, diafiltered at pH 8. Top: initial sample showing strong initial signal of hydrophilic phenolics and saponins. Bottom: Sample post- diafiltration showing removal of hydrophilic phenolics and saponins.

[0106] Fig. 19. Dynamic light scattering analysis of purified leaf extracts in aqueous solution at room temperature, buffered to pH 7.2 with Tris buffer.

[0107] Fig. 20. Extractions with phosphate buffer at 75, 100 and 150 mM, RT, 12 h. The extraction at 100 mM shows a lighter color and lower volume of precipitates.

[0108] Fig. 21. Kinetics of saponin extraction yields (% w / w based on dry leaves), with 100 mM phosphate buffer, pH 7 at RT.

[0109] Fig. 22. Kinetics of saponin extraction (% w / w yield based on dry leaves) at pH 7.0 adjusted with phosphate / NaCl buffer. Room temperature (RT) vs initial 30 minutes at 50 °C, followed by extraction at RT.

[0110] Fig. 23. Two-stage process at bench-scale — Stage 1: Percolation at room temperature, pH 4.6. Top: RP-HPLC chromatogram of the extract immediately after percolation. Bottom: RP-HPLC chromatogram following downstream processing (DSP), showing the near-complete removal of phenolic compound signals in the early elution region.

[0111] Fig. 24. Two-stage process at bench-scale — Stage 2: Extraction at pH 8.3, 50 °C, 1.5 h. Top: RP-HPLC chromatogram of the extract immediately after extraction. Bottom: RP-HPLC chromatogram following downstream processing (DSP), showing the complete removal of phenolic compound signals in the early elution region.8 B25-074-2WO

[0112] Fig. 25. Two-stage process- Stage 1 extract and freeze-dried product with > 90% w / w saponin content.

[0113] Fig. 26. Two-stage process Pilot Plant scale — Stage 1: Percolation at room temperature, pH 4.6. Top: RP-HPLC chromatogram of the extract immediately after percolation. Bottom: RP-HPLC chromatogram following downstream processing (DSP), showing the near-complete removal of phenolic compound signals in the early elution region.

[0114] Fig. 27. Two-stage process — Stage 2: Extraction at pH 8.3, 50 °C, 1.5 h. Top: RP-HPLC chromatogram of the extract immediately after extraction. Bottom: RP-HPLC chromatogram following downstream processing (DSP), showing the almost complete removal of phenolic compound signals in the early elution region.

[0115] Fig. 28. Dynamic light scattering analysis of a 10% w / w canola oil in water emulsion with purified leaf saponins. Saponin load of 50 mg saponin / g oil, or 0.5% saponin.

[0116] Description of Particular Embodiments of the Invention

[0117] Unless contraindicated or noted otherwise, in these descriptions and throughout this specification, the terms “a” and “an” mean one or more, the term “or” means and / or. It is understood that the examples and embodiments described herein are for illustrative purposes only and that various modifications or changes in light thereof will be suggested to persons skilled in the art and are to be included within the spirit and purview of this application and scope of the appended claims. All publications, patents, and patent applications cited herein, including citations therein, are hereby incorporated by reference in their entirety for all purposes.

[0118] Composition and Properties of Q. saponaria Leaves

[0119] Aqueous extracts of quillaja leaves contain saponins, salts (mainly calcium oxalate), proteins, sugars, chlorophyll, carbohydrates (starch, pectin) and phenolics. A comprehensive analysis was conducted to determine if the saponin and phenolic profile of the leaves matched that of existing commercial products. Saponins were identified through UPLC / MS analysis by their m / z values and compared to profiles from bark and biomass, as documented in previous studies (Kite et al., 2004; Wallace et al., 2017). The results showed that the saponin profile in the leaves closely resembles that of bark and biomass, with QS-18 and QS-21 as the predominant saponins, along with smaller amounts of QS-7. These identifiers (e.g., QS-18, QS-21) refer to approximate retention times in minutes, as established in the original RP-HPLC study by (Kensil et aL, 1991).

[0120] Figure 1 shows a RP-HPLC chromatogram of a commercial purified Q. saponaria product vs. a purified extract derived from Quillaja leaves. The chromatograms show the same saponins, though their distribution may vary slightly from batch to batch. These saponins are 9 B25-074-2WOprimarily present as micelles in the extracts, as the critical micelle concentrations are quite low, e.g. 300- 500 mg saponins / L.

[0121] However, the UPLC-MS analyses revealed that the predominant saponins in quillaja leaves are QS-7-Rhanmose (m / z 1876.8), QS-18-Rhamnose (m / z 2165), and QS-21 -Rhamnose (m / z 2002.9), corresponding to saponins 18a, B3 / B5, and S3 / S5, respectively, as described by Fleck et al. (2019). These saponins are also in the extracts derived from quillaja biomass (wood chips) approved for human consumption. In contrast, bark extracts from mature trees contain the same structural types of saponins but with a higher proportion of xylose substitutions instead of rhamnose, yielding the QS-7-Xylose, QS-18-Xylose, and QS-21 -Xylose variants. Additionally, bark extracts contain saponin QS-17, which is absent in the leaves.

[0122] The UPLC / MS analysis of the phenolics also revealed that the main compounds are consistent with those in bark and biomass extracts, specifically piscidic acid (comprising 75-87% w / w of all phenolics) and quillajaside A and B (phenylpropanoid sucrose esters (Maier, Conrad, Carle et al., 2015; Maier, Conrad, Steinglass et al., 2015. Unique to the leaves, however, are small amounts of quercetin and oleuropein.

[0123] Sourcing and Preparation of Biomass

[0124] Definitions

[0125] “Leaves.” Refers to the foliage of Q. saponaria, including intact leaves and leaf fragments, and leaf petioles and midribs.

[0126] “Aerial biomass” or “biomass.” Refers to the above-ground portions of Q. saponaria obtained during harvesting, pruning, milling, shredding, or processing. Includes fresh, dried, partially dried, intact, chopped, shredded, ground, or milled material, such as leaves, leaf fragments, petioles, fine twigs, and small branch debris.

[0127] “Fresh leaves.” Leaves collected and processed without drying, bleaching, or storage.

[0128] “Dry green leaves.” Leaves dried using commercial kilns, air-drying, or similar methods, retaining natural green coloration and chlorophyll.

[0129] “Photobleached leaves.” Leaves exposed to natural or artificial light to reduce chlorophyll content.

[0130] “Milled biomass.” Leaves or aerial biomass reduced in size via milling, grinding, shredding, or chopping.

[0131] Sourcing and Preparation of Biomass

[0132] Q. saponaria leaves can be sustainably sourced from shrubs over one year old, which can be cultivated on plantations. Growth can be achieved without use of chemical pesticides. Plants younger than one year may also be used, but the yields of leaves may be less. Leaves can also be obtained from plants grown in greenhouses, both in traditional pots, as well as10 B 25 -074-2 WOhydroponically. The present invention also provides the use of leaves from mature trees.Currently, when trees are harvested in Chile, the leaves are discarded and left to decompose in the field. By incorporating these leaves into the methods described herein, their value can be significantly enhanced, transforming a waste product into a useful resource.

[0133] Extensive testing was done to compare different types of leaves obtained from pruning > 1 year old plants grown in greenhouses and in the field, including fresh leaves, dry green leaves (offer several advantages, and are stored until milling is required. These leaves will have a high chlorophyll content, but the described invention will still result in final products with acceptable color and only slightly increased purification costs), photobleached leaves (light exposure degrades chlorophyll, resulting in extracts with minimal green tones and significantly reduced purification requirements. The dried leaves are easily stored and can be milled efficiently.) The use of photobleached leaves treated with natural sunlight, artificial light or chemical methods presents a valuable innovation, which has not previously been explored with Q.saponaria leaves.

[0134] In summary, the process can handle different types of leaves. Figure 2 shows dry ground green and photobleached leaves.

[0135] Extraction Methods

[0136] In one aspect, provided is a method for producing saponins, comprising:a) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaja leaves;b) treating the quillaja leaves with an aqueous solution comprising an antioxidant and a chelating agent at a pH between 7.5 and 8.5 and a temperature between 45-65 °C for a duration between 1 hour and 3 hours;c) separating the treated quillaja leaves after step b) to obtain an extract;d) contacting the extract from step c) with a nutritionally-inert sorbant, such as calcium bentonite, sodium bentonite or activated charcoal;e) filtering the extract from step d) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a purified saponin composition.

[0137] In another aspect, provided is a method for producing saponins, comprising:a) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaja leaves;b) treating the quillaja leaves for 1-3 h with an aqueous solution comprising an antioxidant and a chelating agent at a pH between 4 and 5.5 and at room temperature;c) separating the treated quillaja leaves after step b) to obtain a first extract;11 B 25 -074-2 WOd) filtering the first extract from step c) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a first purified saponin composition; e) subjecting the separated, treated quillaja leaves from step c) to a hot alkaline extraction at a pH between 7.5 and 8.5 and at a temperature between 45 °C and 65 °C for a duration between 0.5 hours and 3 hours to obtain a second extract;f) contacting the second extract from step e) with a nutritionally-inert sorbant, such as calcium bentonite, sodium bentonite, or activated charcoal;g) filtering the second extract from step f) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a second purified saponin composition.

[0138] In some embodiments of the foregoing, the method further comprises combining the first purified saponin composition and the second purified saponin composition.

[0139] In another aspect, provided is a method for producing saponins, comprising:a) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaja leaves;b) treating the quillaja leaves with an aqueous solution comprising an antioxidant and a chelating agent and buffer at a pH between 6.5 and 8 and at a temperature between 15 °C and 30 °C for a duration between 8 hours and 24 hours;c) separating the treated quillaja leaves after step b) to obtain an extract;d) filtering the extract from step c) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a purified saponin composition.

[0140] In some embodiments of the foregoing, the buffer comprises phosphate ions. For example, in one variation, the buffer is a phosphate buffer.

[0141] In some variations of the foregoing methods, the quillaja leaves are fresh, dried, partially dried, intact, chopped, shredded, ground, milled, or photobleached.

[0142] In some variations of the foregoing methods, the antioxidant is a natural antioxidant. In certain variations, the antioxidant comprises ascorbic acid, cysteine or citric acid, or any combination thereof. In one variation, the antioxidant comprises ascorbic acid.

[0143] In some variations of the foregoing methods, the chelating agent comprises ethylenediaminetetraacetic acid (ED TA), citric acid or phosphoric acid, or any combination thereof. In one variation, the chelating agent comprises EDTA.

[0144] In some variations of the foregoing methods, the divalent cations are calcium, zinc or magnesium cations.12 B 25 -074-2 WO

[0145] In some variations of the foregoing methods, the method is performed as a batch process. In other variations, the method is performed as a continuous process.

[0146] In some variations of the foregoing methods, the chlorophyll index of each extract obtained in the methods (e.g., in the extraction step(s) of the method) is below 0.25.

[0147] Variations of the steps, conditions and parameters of the methods for producing saponins are described in further detail below.

[0148] Extraction Conditions

[0149] Extraction conditions: pH, temperature and time

[0150] Different temperatures, pH and extraction times were tested to maximize saponin yields, while minimizing downstream processing (DSP) requirements, as summarized in Table 1. Water was chosen as the preferred solvent due to its safety and cost-effectiveness. When using water, ascorbic acid (0.1-2 g / L) and EDTA (0.01-1 g / L) are added to the water to minimize oxidation of phenolics. However, mixtures of water with ethanol or methanol can also be used effectively (water to alcohol ratio of 2:8 to 8:2), provided the solvent is removed via evaporation. A solvent to dry leaves ratio of 5-15 is preferable.

[0151] Following extraction, the leaves are separated from the extract with an internal sieve (50-300 μm) or pumped from the extraction vessel to an external filtering sieve (50- 300 pm). About 60-85% of the initial solvent volume is recovered. To maximize saponin recovery, the leaves can optionally be washed with fresh water for 0.5-2 hours, after which the leaves are filtered out again. The extract and the wash can be mixed and sent to downstream process (DSP).

[0152] Table 1: Summary of extractions performed with water as a solvent in a batch mode. RT= Room temperature.pH Temperature, °C Saponin yield Color4.5 -5.2 20-25 (RT) + +4.5-5.2 40-70 +4- ++7-8.5 20-25 (RT) +++ +++7-8.5 40-70 ++++ +++-1-

[0153] Extraction at equilibrium pH (4.5-5.2, natural pH)

[0154] Background: Q. saponaria leaves and water naturally equilibrate at a pH of 4.3 -5.2, depending on the concentration of acid antioxidant used.

[0155] Extraction at natural pH and room temperature (RT) yields low colored extracts with saponin yields of 2-8% w / w based on dry leaves. Optimal extraction times for these conditions were determined to be 1-5 hours, recovering more than 90% w / w of saponins from the leaves.13 B 25 -074-2 WO

[0156] Extraction at natural pH and 55-65 °C leads to saponin yields of 4-10 % w / w. The extracts are darker due to oxidation of phenolics and chlorophyll degradation. Extraction time between 1-3 h was used to recover >90% w / w of the saponins in the leaves.

[0157] Extraction at pH 7.0 - 8.5

[0158] Background: Saponins are predominantly found intracellularly within the cell vacuoles (Fang & Xiao, 2021), so enhancing cell disruption was identified as a key area to improve saponin yields. To address this challenge, techniques used to disrupt leaf cells to extract proteins, e.g., rubisco, were tested, ultimately pinpointing basic pH and higher temperature as crucial factors (Zhang et al., 2015). To our knowledge, basic pH has never been tested to maximize saponin extraction from Q. saponaria leaves. Basic pH (pH 9.5) combined with elevated temperatures (68 °C) has been demonstrated to optimize saponin extractionfrom Camellia oleifera spent seed cakes, a byproduct of edible oil production in China (Zhang et al., 2015). In contrast with Q. saponaria leaves, these seed cakes comprise saponin-rich crushed seed hulls, rather than leaves.

[0159] Under basic conditions, chlorophyll is more stable, however under high temperature chlorophyll may undergo degradation, losing its magnesium ion and giving rise to undesirable grey-brown compounds, including pheophorbide and pheophytin (Indrasti et al., 2018).Phenolics, on the other hand, are rapidly oxidized to colored compounds (Maier, Conrad, Carle, et al., 2015), including reactions in the presence of iron (Bijlsma et al., 2022). This is valid for compounds found in quillaja leaves like quercetin, rutin (Buchner et al., 2006), and oleuropein. Saponins may also be degraded, especially if operating at high temperature, high pH and for long extraction times (Cleland et al., 1996).

[0160] Extensive experiments were performed at different temperatures (RT to 70 °C), pH (7-8.5) and reaction times (0.5-3 h) to determine conditions that maximized saponin yields, while minimizing phenolic and saponin degradation.

[0161] Neutral to basic pH (7.0-8.5) at room temperature (RT)

[0162] Extraction at neutral to basic pH (7.0- 8.5) and RT increase saponin yields to 8-12% w / w, but the extracts have a darker color when compared to natural pH extracts, due to the oxidation of phenolics and chlorophyll degradation.

[0163] Neutral to basic pH (7.0-8.5) at 40-70 °C

[0164] Extraction at neutral to basic pH (7.0- 8.5) and 40-70 °C boosts saponin yields to 9-17% w / w, due to cell disruption under alkaline conditions and higher temperature. However, these extracts exhibit a darker green color with a higher chlorophyll content, Also, due to oxidation of phenolics into quinones which are likely to polymerize to colored compounds, the extracts also have deeper reddish hues.14 B 25 -074-2 WO

[0165] Importantly, adding magnesium ions during extraction at neutral to basic pH (7.0- 8.5) and 40-70 °C is advantageous, as it reduces the loss of magnesium ions in the chlorophylls, thereby minimizing their degradation to colored compounds.

[0166] Kinetic experiments indicated that at basic pH and higher temperature, the optimum time to prevent saponin degradation and extensive phenolic oxidation was found to be 0.5-2 h. The higher degradation / oxidation of both chlorophyll and phenolics under basic pH and high temperature require higher DSP efforts to obtain light-colored products.

[0167] Buffer-Based Saponin Batch Extraction Process

[0168] An important finding of this research is that while extraction conditions employing higher pH (8-8.5) and elevated temperature (50-55 °C) yield the highest saponin recovery, they can still produce acceptable final products - specifically, light-colored powders containing more than 80% w / w saponins — provided that both chlorophyll and phenolic compounds are adequately removed, as demonstrated in this invention. However, these processes require additional separation steps, such as treatment with activated charcoal or bentonite, and subsequent disposal of these additives. A further and equally significant discovery is that the need for such color-reducing agents can be substantially reduced — or eliminated — by performing extractions in buffered solutions at pH 6.5-8 under room-temperature conditions with extended extraction times. Although saponin yields under these milder conditions are approximately 15% w / w lower than those achieved under high-pH, high- temperature extraction, the trade-off is highly favorable. The solubilization and degradation of phenolics and chlorophyll arc greatly minimized, resulting in extracts with exceptionally low color indices. These light-colored extracts are particularly advantageous for commercial applications requiring clear, refined products without additional purification steps.

[0169] To achieve this, the fundamental role of pH and temperature in saponin extraction was systematically investigated. pH functions as a primary driver of saponin solubilization through modulation of cell membrane integrity and saponin ionization state. However, elevated temperature presents a dual-edged effect: while it enhances saponin extraction kinetics, it simultaneously promotes undesirable secondary reactions. Temperature increases the rate of phenolic oxidation, chlorophyll degradation, and the formation of brown polymeric compounds that darken extracts. This temperature-induced degradation occurs even under alkaline conditions favorable for saponin extraction, creating a fundamental trade-off between yield and product quality. By conducting extraction at RT, this degradation pathway is substantially suppressed while maintaining adequate saponin solubilization through pH optimization alone.

[0170] Buffer systems proved essential for achieving superior extraction performance compared to simple pH adjustment with base alone. The choice of buffer ion is not merely incidental; the 15 B 25 -074-2 WOnature of the buffer anion fundamentally influences extraction outcome through effects on ionic strength, selective precipitation of co-extracted compounds, and protein-saponin-phenolic interactions. Phosphate buffers, particularly at neutral to slightly alkaline pH, provide a distinct advantage over alternative buffer systems. Phosphate is polyvalent, existing predominantly as the divalent HPO42-anion at pH 6.8-7.2, which enables optimal ionic strength for selective precipitation of colored protein-phenolic and protein-chlorophyll complexes. This selective precipitation removes visually undesirable compounds while preserving saponin solubility, a balance that cannot be achieved with monovalent buffer systems such as bicarbonate.

[0171] Optimization of phosphate buffer concentration revealed that an intermediate concentration point (80-120 mM) outperforms both lower and higher concentrations. At lower ionic strengths (75 mM), the buffer lacks sufficient ion concentration to effectively precipitate colored protein-phenolic complexes. Conversely, at higher ionic strengths (150 mM), excess ions shift the chemical equilibrium toward greater co-extraction of these undesired compounds. The 80-120 mM concentration thus represents an optimal balance — adequate ionic strength for selective complex precipitation while maintaining conditions favoring saponin solubilization.

[0172] Cost considerations were a major driver in buffer system development. An alternative formulation combining Na₂HPO₄·2H₂O with NaCl achieves substantially lower reagent costs, provided the ionic strength is equivalent to that of the 80-120 mM phosphate buffer (12-13 mS / cm). This low-cost alternative enables industrial-scale operation without sacrificing extraction performance, making the room-temperature extraction approach economically competitive with conventional high-temperature methods when total processing costs — including bentonite waste disposal — are considered.

[0173] To further improve saponin extraction efficiency while retaining the benefits of reduced chemical degradation, an intermediate extraction method was developed. This approach begins with a short heating period at 50 °C for 10-30 minutes to rapidly solubilize saponins and disrupt cellular structures, followed by cooling to room temperature and continuing extraction for 6-12 hours. The brief elevated-temperature step provides the kinetic advantages of higher temperature while minimizing exposure to conditions that promote degradation. The subsequent roomtemperature extraction allows saponin diffusion to continue without further degradation, effectively separating the kinetics of saponin release from undesired secondary reactions.

[0174] Agricultural pre -treatment represents an additional lever for improving extraction performance and further reducing chemical inputs. Quillaja leaves exposed to sunlight for 3-10 days post-harvest undergoes partial photobleaching that substantially reduces chlorophyll content while maintaining saponin integrity. This pre-treatment is both practical and economically favorable, as quillaja plants thrive in sunny locations and harvested material can 16 B 25 -074-2 WObe easily subjected to solar exposure. When the intermediate -temperature buffer-based extraction protocol is combined with photobleached leaves, chlorophyll levels are sufficiently reduced that a chlorophyll removal treatment becomes entirely unnecessary. This integrated approach — combining agricultural pre-treatment, optimized buffer chemistry, controlled extraction temperatures, and extended contact time — provides a comprehensive solution for producing high-quality saponin extracts with minimal chemical inputs and waste generation.

[0175] Types of products: semi-refined, refined products and highly refined products

[0176] The process can produce different kinds of extracts, such as semi-refined, refined and highly refined products, with saponin levels in the solubles of 20-30% w / w, 80-85% w / w and > 90% w / w, respectively.

[0177] Semi-refined products, which contain almost all the water-soluble compounds that are extracted from leaves. Downstream processing is kept to a minimum with an emphasis on removal of chlorophyll and its degradation products to prevent excessive darkening of the extracts. This is particularly relevant for extracts produced under conditions that maximize saponin extraction, e.g., basic pH and high temperature. To this end, activated charcoal or bentonite can be used as purifying compounds. The saponin content of these products determined by RP-HPLC or UPLC is 20-30% w / w of the total soluble compounds.

[0178] Figure 3 shows a flow diagram for producing semi-refined products. Because of the presence of colored phenolic compounds, as well as sugars, they can be used at ppm levels to produce foam in foods, e.g., foam of the root beer, or in industrial applications where purity is not so relevant, such as soil biopcsticidcs, wastewater treatment (Kaczorck ct al., 2016) or as a saponin-based natural surfactant for enhanced oil recovery (Imuetinyan et al., 2022).

[0179] Refined products, which result from a more thorough downstream process, during which not only chlorophyll, but also other non-saponin compounds, such as phenolic acids, carbohydrates and minerals, are removed. The saponin content of these products is 80-85% w / w of quillaja solubles.

[0180] Figure 4 illustrates a representative flow diagram for producingrefined Quillaja products using water as the extraction solvent, whereas Figure 5 depicts a simplified process employing buffered extraction. In the buffer-based, room temperature extraction process described above, the only solid waste streams are the extracted leaves and the spent cellulose filters. The extracted leaves can be dried and utilized in applications where purified extracts are not required, such as additives in animal feed formulations.

[0181] Highly Refined Products. Purified quillaja saponins are widely recognized as potent adjuvants in veterinary vaccines, such as those for foot-and-mouth disease. Commercially available extracts, including Quil-A® (Croda Pharma, USA) and Vet-Sap® (Desert King 17 B 25 -074-2 WOInternational, USA), typically contain around 90% w / w saponins. These products are exclusively derived from the bark of mature, over 30-year-old wild trees native to Chile, a practice that is both ecologically unsustainable and resource-limited.

[0182] The purification of bark-derived extracts to such high purity is complex and inefficient. The bark of aged trees contains highly polymerized and oxidized phenolics that form strong associations with saponin micelles, limiting membrane separation efficiency. As a result, ultrafiltration typically yields only 60-65% w / w saponins. Further refinement requires multiple additional steps, including adsorption on PVPP, selective precipitation (Padilla et al., 2022), or ion-exchange chromatography (Kamstrup et al., 2000).

[0183] Surprisingly, the present invention demonstrates that QuiJlaja leaves can yield extracts exceeding 90% w / w saponins through the combination of optimized extraction and controlled removal of chlorophyll and phenolic compounds. Once chlorophyll is minimized — particularly by employing photobleached leaves — the ultrafiltration step becomes markedly more effective, as leaf phenolics are simpler and less tightly bound to saponin micelles than those in bark extracts.

[0184] Accordingly, the two-stage extraction process disclosed herein — comprising an initial room-temperature extraction of photobleached leaves at acid pH, followed by extraction at pH 8.5 and 55°C — enables the production of highly purified saponin fractions (>90% w / w) without the need for complex downstream purification steps or chemical additives.

[0185] For immunological applications — particularly human vaccines — extracts should contain minimal amounts of the more toxic saponin fractions. RP-HPLC or UPLC chromatograms typically resolve three saponin groups with distinct immunological and toxicological profiles: QH-A, QH-B and QH-C (Johansson & Lbvgren-Bengtsson, 1999). Figure 6 shows a typical RP-HPLC chromatogram obtained from quillaja leaves, showing the three saponin fractions. QH-A (which contains QS-7) and QH-C (which contains QS-21) are associated with favorable toxicity and immunological profiles and are therefore preferred for human vaccine formulations. By contrast, QH-B, which includes the relatively abundant QS-18, exhibits higher relative toxicity and is unsuitable for use in human vaccine preparations. Commercial vaccine producers currently source crude extracts from Chile and apply internal purification steps to reduce QH-B (e.g., Novavax) or to isolate specific saponins such as QS-21 (e.g., GSK). Accordingly, processes that enrich QH-A and QH-C fractions are preferred for preparations intended for human use.

[0186] Table 2 summarizes how extraction conditions strongly influence the relative abundance of hydrophilic saponins, such as those in fraction QH-A, while having only minor effects on the major saponins in fractions QH-B and QH-C (QS-18 and QS-21, respectively). At pH 4.6 and 6018 B 25 -074-2 WO°C, QH-A represented 9.7% w / w of total saponins; when the pH was increased to 8.5, its proportion rose to 13.4% w / w at room temperature and 33.7% w / w at 60 °C. Figure 7 shows the corresponding UPLC chromatograms for extracts obtained at pH 4.6 and 8.5, both at 60 °C. The extract at pH 8.5 displays several additional peaks in the early retention region (preceding QS-7, m / z 1876), which were identified by MS as saponins 13a (m / z 1730.7), 15a (m / z 1640.7), and 16a (m / z 1802.8) (Fleck et al., 2019). These compounds share structural similarity with QS-7, differing mainly by the absence of the complex oligosaccharide moiety at position R4 (Fleck et al., 2019), characteristic of more hydrophobic saponins such as QS-18 and QS-21.

[0187] These results are significant for downstream processing (DSP), as the more hydrophilic saponins tend to remain in the aqueous phase and can permeate UF membranes during diafiltration (see discussion below).

[0188] Table 2: Impact of pH and temperature on the distribution of hydrophilic (QH-A) and hydrophobic (QS-18, QS-21) saponins in quillaja leaf extractsExtraction conditions % QH-A % QS-18 % QS-21 pH 4.6, 60°C, 2h 9.7 35.0 2.8 pH 8.5, 60°C, 2h 33.7 21.2 1.8 pH 8.5 2h RT 13.4 32.2 2.7

[0189] Downstream processing (DSP)

[0190] Background: To obtain light-colored refined and highly refined products, the extracts produced under single step batch or continuous extraction require varying purification strategies, as described in Figures 4 and 5. The primary-colored compounds that need to be removed or minimized are phenolics, which impart a red hue, and chlorophyll and its degradation products, which impart a dark color.

[0191] Removal of chlorophyll and its degradation products

[0192] In all extracts, chlorophyll and its degradation products are present in varying amounts, requiring the use of effective DSP strategies to minimize or fully remove these molecules.Depending on the operational conditions, chlorophyll may undergo degradation, losing its magnesium ion and giving rise to undesirable grey-brown compounds, including pheophorbide and pheophytin (Indrasti et al., 2018). Although chlorophyll and its degradation products are poorly soluble in water, the presence of saponins probably enhances their solubility (Yasuda & Tabata, 2021). Moreover, while chlorophylls are not water-soluble, they may be present as water-soluble protein chlorophyll complexes (WSPCs), which are proteins that attach to four chlorophyll molecules. WSPCs are found in the chloroplasts and do not participate in photosynthesis (Girr & Paulsen, 2021). An additional source of color may be the presence of19 B 25 -074-2 WOthylakoid membrane fragments, which are membrane structures found in plant leaves responsible for photosynthesis, rich in chlorophylls and other pigments (Girr & Paulsen, 2021).

[0193] The downstream processing (DSP) challenge stems from the detrimental impact of even trace amounts of chlorophyll and its degradation products in the extracts. This issue is particularly pronounced when using membranes to separate non-saponin compounds and / or concentrate the extracts. The molecular cut-off size of these membranes permits the passage of lower molecular weight compounds, such as phenolic acids, which are natural antioxidants, while retaining larger molecules like saponin micelles, chlorophylls, and their degradation products. This retention of chlorophyll and its byproducts, and the concomitant loss of natural antioxidants, results in darker extracts that fail to meet the color and clarity specifications required for commercial application. This is more critical when using dry ground green leaves, whereas photobleached leaves allow for simpler DSP strategies.

[0194] The following DSP strategies may be used to overcome these challenges:

[0195] Precipitation of leaf proteins and WSPCs via pH change: to maximize the removal of WSPCs, the extracts are acidified to pH 3.0- 4.5 using hydrochloric acid, citric acid, phosphoric acid, or a comparable reagent. As the isoelectric point (pl) of the leaf proteins is 3.5 -5.5 (Tamayo Tenorio et al., 2017), reducing the pH of the extract causes significant precipitation of free proteins and WSPCs, and the extracts lose part, but not all, of their green color. Because of this, concentration of these extracts with ultrafiltration membranes may still cause darkening of the extracts.

[0196] Use of activated charcoal (AC): to eliminate residual chlorophyll following protein precipitation, the extracts can be treated with activated charcoal (AC) using two approaches: (1) contacting the extracts with powdered AC, followed by removal through filtration with diatomaceous earth, or (2) using cellulose filters impregnated with AC. The AC-impregnated cellulose filters primarily adsorb phenolic compounds, and lesser amounts of chlorophyll. Furthermore, their use may result in significant saponin losses (10-40% w / w).

[0197] Use of bentonites: to remove residual WSPCs and chlorophyll degradation products, following protein precipitation, food-grade bentonites (calcium or sodium) were tested due to their negative charge and high surface area (Mokaya et al., 1994). This process also benefits from the use of an acidic pH. For example, WSPCs, which carry a positive charge below their pl, can be effectively adsorbed by food-grade sodium and calcium bentonites, which have a negative charge. Additionally, as shown in Figure 8 ( Tamayo Tenorio et al., 2017), the zeta potential of thylakoid membranes indicates that at pH values below 4.7, they cany a positive charge. At a pH of 3-4, the positive charge of thylakoids allows them to be adsorbed by20 B 25 -074-2 WObentonite, facilitating their removal from the extract. This strategy yields good results, since saponins are minimally adsorbed by the bentonites.

[0198] Bentonites can be added as powder or as cellulose filters impregnated with bentonites, as are used in the production of botanical extracts, such as the production of THC from cannabis. When added as powder, a concentration of 0.5 to 10 g / L of bentonite is used. Following agitation for 0.1-6 h, diatomaceous earth is added, and the extract is filtered through plant-and-frame filters to produce a clear extract with minimal chlorophyll content. Alternatively, the solution can be left to settle during 3-24 h at 3-15 °C and the supernatant collected and filtered through 0.4-10 pm filters. The bottoms are filtered separately with diatomaceous earth and plate-and-frame filters, or with microfiltration tangential flow systems. Depending on the amount of bentonite added, the volume of the bottoms is 10-40% of the total volume. This process variation generates fewer filter cakes as waste.

[0199] Quantification of chlorophyll in the extracts

[0200] For the development of an industrial-scale process, it was essential to establish a rapid method for estimating chlorophyll content in the extracts to design the appropriate DSP strategy. To achieve this, initially a UPLC / MS method was developed to detect chlorophyll and its degradation products. However, because the aqueous extracts were relatively dilute and chlorophyll concentrations were low (typically 0-500 pg / mL), this analytical approach did not provide sufficient sensitivity or resolution to clearly differentiate chlorophyll levels among samples.

[0201] To address this limitation, chlorophyll was estimated from the spectral absorbance profile recorded with a TECAN spectrophotometer in the range of 650-700 nm, corresponding to the absorption region of chlorophyll a, the predominant chlorophyll species in leaves. Since absorbance in this region is proportional to the total soluble content of the extract, the absorbance was normalized by the concentration of solubles (g / L), determined gravimetrically by drying the extract at 70 °C for 24 hours. The resulting ratio, A670 / concentration of solubles, was defined as the chlorophyll index (CI).

[0202] For routine operation and faster analysis, a simplified version of the CI determination was implemented. In this method, the CI is based on the absorbance of the extract at 670 nm — the principal absorption maximum of chlorophyll a — and the concentration of solubles extracted is estimated from the degrees Brix (°Brix) of the extract, measured using a handheld refractometer after filtration through a 0.45 pm membrane. Degrees Brix are widely used in the food industry to express the concentration of solubles in a liquid as a function of its specific gravity (SG). One degree Brix corresponds to 1 g of sucrose in 100 g of solution (1 °Brix = 1% sugar).21 B 25 -074-2 WO

[0203] For Quillaja leaf extracts, the concentration of solubles (g / L) was found to correlate linearly with °Brix, following the relationship:

[0204] Concentration of solubles (g / L) = 10 × °Brix × 0.75

[0205] For example, an extract reading 3 °Brix corresponds to approximately 22.5 g / L of solubles.

[0206] Based on this correlation, the chlorophyll index (CI) was estimated using the following expression:

[0207] CI = A670× 10 / °Brix

[0208] Multiplication of A670by 10 allows for clearer visualization of differences between extracts. The overall procedure is simple and involves taking a small sample of the extract (5-10 mL), filtering it through a 0.45 pm membrane, and measuring (i) the absorbance at 670 nm in a spectrophotometer and (ii) the °Brix value using a handheld refractometer.

[0209] As will be further discussed in detail, it was determined that to obtain light-colored final products, the CI of the extract prior to membrane purification must be below 0.25.

[0210] Removal of phenolics and non-saponin compounds

[0211] Following the removal of chlorophylls, the next step is to reduce the concentration of phenolics that impart a dark reddish color to the final products. These phenolics are mainly piscidic acid dominates (83% of total quantified polyphenols), followed by -coumaric acid derivatives, identified as Quillajasides A and B and isomers (4.8% to 5.3%) and quercetin derivatives (3.7% to 4.2%). The total phenolic content of the extracts is about 5-10% w / w of all the water-soluble molecules extracted. At conditions that maximize saponin recovery (basic pH, higher temperature) phenolics degrade into quinones, which are likely to polymerize to colored compounds.

[0212] In practice, a first step prior to phenolic removal with UF membranes is to concentrate the extracts with nanofiltration or reverse osmosis, to minimize the volume of extract that will be subject to the removal of phenolics. Since both nanofiltration and reverse osmosis membranes do not allow the passage of phenolics, all the components in the extracts are concentrated. This is advantageous, since phenolic acids such as those present in quillaja extracts are natural antioxidants or chelating agents, minimizing oxidation and color development during the concentration processes. In this concentration process, saponin content can be increased from 5-12 g / L, to 25-120 g / L.

[0213] A way to monitor phenolic removal is through RP-HPLC, using the same method developed to quantify saponins at 210-220 nm. Figure 9 shows a typical chromatogram highlighting the main saponins found in leaves. In these chromatograms, water soluble phenolics22 B 25 -074-2 WOare also detected due to the wavelength used (210-220 nm) and are the first set of molecules that elute generating a large initial peak.

[0214] Common methods to remove phenolics from botanical extracts include treatment with activated charcoal, use of proteins (gelatin) or use of polyvinylpolypyrrolidone (PVPP). The compounds used must subsequently be removed through filtration, generating process losses and adding waste treatment costs.

[0215] The use of diafiltration with ultrafiltration membranes has been reported for the removal small molecular weight compounds from Quillaja saponaria extracts (Kamstrup et al., 2000). However, this method is only partially effective, as a significant fraction of phenolics can associate with saponin micelles. These phenolics are present in the extract in two forms: 1) free phenolics, which can be readily absorbed or removed with diafiltration and 2) bound phenolics, which are associated with the outer layer of saponin micelles through chemical bonds and intermolecular forces. This binding occurs due to the negative charge on the glucuronic acid groups in saponins at the working pH of 4.5-5, where the pKa is >5.2. In quillaja saponin micelles, prior studies suggest that, beyond covalently bound phenolic compounds, the interactions are likely driven by hydrogen bonding between the sugar groups on saponin molecules and phenolic glycosides (Tippel et al., 2017). Moreover, some hydrophobic phenolic derivatives can be incorporated within the hydrophobic core of the micelles.

[0216] To address this limitation, it was found that adding a 100-1500 ppm concentration of divalent ions, such as magnesium (Mg2+), zinc (Zn2+), and calcium (Ca2+), to the extracts prior to diafiltration effectively disrupts these associations, thereby enhancing the removal of phenolics.Figure 10 is the chromatogram of the same sample shown in Figure 9 after diafiltration with UF, showing almost complete removal of the phenolics signal.

[0217] A second strategy explored to disrupt the association between the phenolics and the saponin micelles was to perform the diafiltration at pH 7.5-8.0. This is a different mechanism than adding calcium, which relies on competitive binding. At pH 7.5-8, the electrostatic repulsion between the highly deprotonated, negatively charged phenolic molecules and the similarly negatively charged saponin micelles increases significantly, effectively pushing the phenolics out of their association with the micelles. Since the large saponin micelles are still retained by the ultrafiltration membrane, the now-separated, smaller phenolic molecules can pass through with the permeate, resulting in a cleaner product. Figure 11 shows that this strategy was equally effective, probably because at this pH both the phenolic acids and the micelles have a negative charge that reduces strong associations.

[0218] Following diafiltration, the extract is concentrated with nanofiltration or reverse osmosis, sterilized either with 0.22-0.45 pm filters, or through continuous High-Temperature Short-Time 23 B 25 -074-2 WO(HTST) pasteurization, before undergoing spray drying. Food-grade carriers may be added prior to drying to achieve the desired saponin concentration in the final product. The resulting powder is stored in sealed, air-proof bags with food-grade desiccants to maintain a low moisture content (6-7% w / w).

[0219] Use of phenolics to verify origin of leaf-derived quillaja extracts (e.g., to detect patent infringement): Quillaja leaves contain unique phenolic compounds that are absent in quillaja bark, such as quercetin (and its derivatives) and oleuropein. These phenolics can serve as reliable “markers” for identifying leaf-derived extracts.

[0220] Even after extensive diafiltration, some of these compounds persist in the final products. For instance, Table 3 highlights the phenolic composition of a highly refined quillaja extract produced via a two-stage purification process. This extract contains detectable levels of quercetin derivatives, including dihexosylquercetin and quercetin pentoside. In less refined extracts, oleuropein is also detectable in higher concentrations. These phenolic markers are exclusive to leaf-derived extracts and are absent in bark or wood-derived extracts, providing a robust method for distinguishing and verifying the source of quillaja extracts

[0221] Table 3. Comparison of phenolics before and after ultrafiltration with CaCl2. Phenolics determined by UPLC / MS.Sample - mg / mlCompound IdentityExtract Post UF with Ca++ Shikimic acid 0.10 0.10Quinic acid 0.84 0.19 Piscidic acid 3.15 0.09Quercetin pentoside 0.40 0.35Coutaric acid 0.19 0.001-O-(4-Coumaroyl)-glucose 0.15 0.00Quercetin-pentoside-hexoside 0.00 0.00Dihexosylquercetin 0.37 0.35Caffeic acid 1,1-dimethylallyl ester 0.00 0.00Rutin 0.39 0.00 Quillajaside B 0.03 0.00 Quillajaside A 0.42 0.10Quillajaside B isomer 0.00 0.00 Quillajaside A isomer 0.17 0.03 Rhamnosylhexosyl methyl quercetin 0.00 0.0024 B 25 -074-2 WOQuillajaside B isomer 0.00 0.00 Quillajaside A isomer 0.01 0.00Quercetin 3-O-pentosylrutinoside 0.00 0.00Oleuropein 0.2 0.10Total phenolics 6.42 1.31Total Saponins 7.74 18.33Ratio phenolics / saponins 0.83 0.07

[0222] Extraction Configuration

[0223] Extraction of dry milled green or photobleached leaves was extensively tested using the following extraction configurations:

[0224] Single-stage batch system

[0225] Continuous extraction system

[0226] Two-stage batch extraction system

[0227] Single-stage Batch Extraction

[0228] The extractions can be conducted in stainless steel tanks with one, two, three, four or five of the following features:

[0229] Agitation system: Ensures uniform mixing and optimal extraction efficiency.

[0230] Double-walled or internal coil construction: Allows for precise heating and cooling control during the process.

[0231] Automatic pH control: Maintains the desired pH range for efficient extraction.

[0232] Internal sieve: Includes a mesh with openings ranging from 50-300 pm to separate solid particles from the extract.

[0233] Bottom discharge: Facilitates easy removal of processed materials.

[0234] An alternative configuration involves the use of an external sieve. In this setup, the tank contents are pumped through an external filtration system to separate the extracted leaves, offering flexibility and ease of maintenance.

[0235] Continuous Process

[0236] Continuous processing offers several advantages, including reduced labor requirements and faster processing times. To explore this approach, experiments were conducted using freshly harvested leaves, following the process diagram illustrated in Figure 12.

[0237] The primary objective was to achieve high microstructure disruption through the application of high shear, while avoiding the use of soda I NaOH or elevated temperatures. Such conditions can lead to chlorophyll degradation and the oxidation of phenolics into colored compounds.25 B 25 -074-2 WO

[0238] Initially, the leaves are shredded using either a commercial branch-and-leaf chipper or a sugar cane press, as shown in Figure 13. The shredded material is then mixed with water at a temperature of 15-45°C, containing ascorbic acid (0.1-2 g / L) and EDTA (0.01-1 g / L). The optimal ratio of water to fresh leaves is between 1:1 and 10:1 (w / w). After agitation for 0.5-3 hours, the mixture is pumped through a continuous dewatering press. On a bench scale, screwtype juicer machines were tested, while pilot-scale experiments utilized a commercial dewatering press, illustrated in Figure 13. DSP for these extracts is similar to those developed for batch-extraction processes.

[0239] Two-stage Process

[0240] To achieve high saponin yields while minimizing phenolic and chlorophyll degradation, an innovative two-stage purification process was developed (see Figure 14).

[0241] Stage 1 involves a percolation process using water at room temperature and a pH of 4.5-5, yielding an extract rich in saponins and water-soluble phenolics with minimal chlorophyll solubilization. These mild conditions help prevent the degradation of phenolics and chlorophyll. In this process, water is slowly passed through the leaves, which are placed over a filter (70-300 pm) in a method similar to brewing coffee with ground beans.

[0242] It has been observed that saponins and phenolics readily dissolve in the initial volumes of water, and that chlorophyll is substantially reduced. Additional percolation with more water primarily extracts remaining soluble phenolics without further solubilizing intracellular saponins. This "wash step" is advantageous as it minimizes the amount of phenolics carried over to Stage 2. The water recovered in this wash step is cither discarded or purified through reverse osmosis and recycled.

[0243] In summary, Stage 1 comprises two steps: (1) percolation to extract saponins and phenolics, followed by (2) a wash step with water to continue removing phenolics without additional saponin extraction. Figure 15 shows the RP-HPLC chromatogram of the percolation Stage, showing significant amounts of both phenolics and saponins. Figure 16, in contrast, is the chromatogram during the wash Stage, showing phenolics being removed with minimal saponin losses. Similar saponin yields were obtained when the extraction was earned out in batch mode instead of through percolation; however, the removal of phenolic compounds was less efficient under these conditions.

[0244] Stage 2 involves a batch extraction of the washed leaves from Stage 1 under conditions designed to release additional saponins by breaking down the leaves' intracellular structure. In this step, the leaves remain in the percolation setup with a filter, and water at 45-65 °C is added with the pH adjusted to 7.5-8.5. Extraction times are kept to a minimum, e.g. 0.5-1.5 h, to minimize degradation of saponins and chlorophyll. This process yields an extract rich in 26 B 25 -074-2 WOsaponins, protein, and chlorophyll, present both as free chlorophyll and as water-soluble proteinchlorophyll complexes (WSPC).

[0245] Under these mildly basic conditions, chlorophyll degradation is minimized, preventing the loss of magnesium ions that would otherwise lead to brown degradation pigments. Phenolic oxidation is also kept low, as only residual phenolics remain from Stage 1.

[0246] Two-stage process, downstream processing

[0247] Following extraction, the extracts are purified in two separate process lines.

[0248] Purification of stage 1 extract. The main purification challenge for this extract is to effectively remove most of the water soluble phenolics that impart a reddish / yellow color to the extracts, with minimal saponin losses. These phenolics are present in the extract in two forms: 1) free phenolics, which can be readily absorbed or removed with diafiltration and 2) bound phenolics, which are associated with the outer layer of saponin micelles through chemical bonds and intermolecular forces. This binding occurs due to the negative charge on the glucuronic acid groups in saponins at the working pH of 4.5-5, where the pKa is >5.2. In quillaja saponin micelles, prior studies suggest that, beyond covalently bound phenolic compounds, the interactions are likely driven by hydrogen bonding between the sugar groups on saponin molecules and phenolic glycosides (Tippel et al., 2017).

[0249] Attempts to remove these compounds using activated charcoal may result in large saponin losses, as activated charcoal also adsorbs quillaja saponins (Mironenko et al., 2024). Adsorption with polyvinylpolypyrrolidone (PVPP), a fining agent commonly used in quillaja extracts from biomass and bark, yields low phenolic adsorption. Another common method, diafiltration with water using ultrafiltration membranes to wash away the bound phenolics, is also ineffective to remove these colored phenolics.

[0250] After extensive testing it was discovered that adding divalent ions, such as magnesium (Mg ), zinc (Zn ), and calcium (Ca ) at a concentration of 100-1500 ppm, to the extracts prior to diafiltration effectively disrupts these associations, thereby enhancing the removal of phenolics. Divalent ions effectively shield the negative charges of the glucuronic group and disrupt the ionic forces between phenolics and saponins, resulting in the release of phenolics into the bulk solution, subsequently facilitating the phenolic passage through the filtration system.Table 3 shows how this strategy allows the removal of most of the phenolic compounds via diafiltration in the presence of CaCl2. The ratio of phenolics to saponins drops dramatically from 0.8 to 0.07 post UF / diafiltration. Piscidic acid, the main phenolic acid present in quillaja extracts, is effectively removed, e.g. >95% w / w removal.

[0251] Following diafiltration, the extract is concentrated with nanofiltration or reverse osmosis, sterilized either with 0.22-0.45 pm filters, or through continuous High-Temperature Short-Time 27 B 25 -074-2 WO(HTST) pasteurization, before undergoing spray drying. Food-grade carriers may be added prior to drying to achieve the desired saponin concentration in the final product. The resulting powder is stored in sealed, air-proof bags with food-grade desiccants to maintain a low moisture content (6-7% w / w). The extract can also be commercialized as a concentrated liquid product.

[0252] Purification of Stage 2 extract: the main purification challenge is to remove protein and chlorophyll that are co-extracted with intracellular saponins, with minimal saponin losses. Traditional approaches to remove chlorophyll with activated charcoal are ineffective, as quillaja saponins are also significantly adsorbed (Mironenko et al., 2024). Following extensive testing, we discovered that bentonite can effectively remove protein, as well as adsorb chlorophyll. To this end the extract is rapidly cooled down to avoid chlorophyll degradation and its pH lowered to 3.5-4.5, so that both the protein and the WSCP complexes have a positive charge, thus forming bonds with the negatively charged bentonite. Contact with bentonite does not result in significant saponin losses. Both calcium and sodium bentonite can be used.

[0253] Bentonite is then removed using plate-and-frame filters and the extract is concentrated with nanofiltration or reverse osmosis, sterilized, either with 0.22-0.45 pm filters, or through continuous High-Temperature Short-Time (HTST) pasteurization, before undergoing spray drying.

[0254] The extracts from stages 1 and 2 are light colored, with a saponin concentration > 90% w / w. For food applications, foaming and emulsion properties are comparable to those of present commercial products. Taste and smell are much lower than commercial products derived from biomass and bark.

[0255] Importantly, the extracts produced by the two-stage process — particularly those from the second stage — exhibit minimal phenolic content as indicated by the absence or very low intensity of initial peaks in UPLC and RP-HPLC chromatograms, and have saponin concentrations that exceed 90% w / w. These materials are comparable in purity to extracts used in veterinary vaccines and are suitable either for direct use or for further fractionation to enrich QH-A and QH-C fractions or to isolate individual saponins (e.g., QS-21) for human vaccine formulations. A principal advantage of the two-stage process is that the downstream processing required to reach this level of purity is substantially simpler than the multi-step purification protocols typically applied to bark-derived extracts (Padilla et al., 2022). Moreover, extracts from both stages are enriched in the less-toxic fractions: QH-A and QH-C constitute >10% and >15% w / w of total saponins, respectively.

[0256] Saponin Compositions

[0257] In some aspects, provided is a purified saponin composition obtained according to any of the methods described herein.28 B 25 -074-2 WO

[0258] In certain aspects, provided is a purified saponin composition having a saponin content of at least 80% w / w. In some embodiments of the foregoing, the composition comprises detectable amounts of phenolic markers present in quillaja leaves from which the purified saponin composition is obtained. In certain embodiments, the total phenolic markers present in the composition is no more than about 5 % w / w. In certain embodiments, the ratio of phenolic markers to saponins is between 0.01 and 0.15 w / w.

[0259] In another aspect, provided is a purified saponin composition comprising saponins and phenolic markers, wherein the ratio of phenolic markers to saponins is between 0.01 and 0.15 w / w.

[0260] In some variations of the foregoing compositions, the phenolic markers comprise quercetin pentoside; dihexosylquercetin; or oleuropein, or any combination thereof. In certain variations, the phenolic markers are present in the following amounts: between 0.5 and 2 % w / w quercetin pentoside; between 0.5 and 2 % w / w dihexosylquercetin; or between 0.1 and 1 % w / w oleuropein, or any combination thereof.

[0261] In some variations of the foregoing compositions, the composition is in powder form. In one variation where the composition is in powder form, the composition has a Hunter Lab Lightness (L) value of at least about 75.0. It should be understood that this L value is a numerical scale from 0 to 100 that represents the brightness or darkness of a color, where 0 is black and 100 is white.

[0262] EXAMPLES

[0263] Maximizing saponin yields

[0264] Table 4 shows saponin yields based on dry leaves and the chlorophyll index (CI) for 2-year-old leaves harvested in the greenhouse, extracted at pH 4.6 and pH 8.5, at RT and 55 °C (average value from 3 experiments). The leaves were ground to 1 mm and extracted for 2 h at a water to dry leaves ratio of 10:1, with agitation in a 2-1 extraction vessel equipped with temperature and pH control. Table 4 shows that both temperature and pH have an impact on saponin yields. The lowest yield, e.g., 6.3% w / w saponins, occurs at RT and pH 4.6, where the CI is also the lowest, e.g., 0.16, indicating a lower solubilization and degradation of chlorophyll and phenolics. Increasing the temperature to 55 °C raises the yield to 7.7% w / w, but also the CI increases to 0.25. Interestingly, at pH 8.5, the saponin yields are significantly higher, with 9 % w / w vs 9.6% w / w at RT and 55 °C, respectively. However, at 55°C and pH 8.5, the CI increases significantly to 0.73. Also, the average percentage of saponins in the extracted solubles was 24% w / w for extractions at RT and pH 4.6, and 28% w / w at 55°C and pH 8.5. These results are similar to the percentage of saponins in bark and biomass aqueous extracts, which is about 20-25% w / w of the extracted solubles (San Martin & Briones, 2000). The higher yields obtained at 29 B 25 -074-2 WOpH 8.5 and 55 °C agree with research on the extraction of intracellular proteins from plant leaves, e.g. rubisco, that indicate that basic pH and higher temperature are crucial factors to increase protein yields due to disruption of the leaf microstructure (Zhang et al., 2015).Similarly, since saponins are predominantly found intracellularly within the cell vacuoles (Fang & Xiao, 2021), enhancing cell disruption was expected to improve saponin yields. To our knowledge, basic pH has never been tested to maximize saponin extraction from Q. saponaria leaves. However, basic pH (pH 9.5) combined with elevated temperatures (68°C) has been demonstrated to optimize saponin extraction from Camellia oleifera spent seed cakes, a byproduct of edible oil production in China (Zhang et al., 2015). In contrast with Q. saponaria leaves, these seed cakes comprise saponin-rich crushed seed hulls, rather than leaves.

[0265] Table 4: Saponin yields and chlorophyll index (CI) of leaves extracted at different pH and temperature.Temperature pH CI Yield w / w RT 4.6 0.16 6.30%8.5 0.34 9.0% 55 °C 4.6 0.25 7.70%8.5 0.73 9.6%

[0266] Phenolics in Quillaja leaves extracted at 5 0 °C and pH 2.7, 4.6 and 8.5.

[0267] To determine the phenolic composition of Quillaja saponaria leaves, extracts were prepared from both fresh and dried material and analyzed by UPLC / MS. Leaves of Q. saponaria (~2 years old) were cultivated at UC Berkeley’s Oxford Tract greenhouses and harvested between December 2023 and January 2024. Dry leaves were air-dried for 10 days at 18-20 °C, milled to 2 mm, and extracted (5 g per 50 mL distilled water) for 2 h at 60 °C. Fresh leaves were ground in liquid nitrogen and extracted similarly (5 g per 25 mL distilled water). Extracts were filtered (0.22 pm) and analyzed in triplicate. UPLC / MS analysis was performed using an Agilent 1260 Infinity II system equipped with a Poroshell 120 EC-C18 column and a mobile phase of 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B) under a gradient elution. Detection was carried out at 210, 280, and 310 nm, with negative ionization mode (100-1200 m / z). All solvents were of analytical or HPLC grade (Fisher Scientific).Reference standards (>95% purity) were obtained from Sigma Aldrich, Acros Organics, MP Biomedicals, TCI America, and PhytoLab GmbH. Quantification of phenolic compounds was based on calibration curves of authentic standards, with derivatives expressed relative to their corresponding parent compounds.

[0268] Table 5 shows the phenolics detected via UPLC / MS for leaves extracted at 50 °C and pH 2.7, 4.6 and 8.5. As summarized in Table 5, the concentration of all analyzed phenolic 30 B 25 -074-2 WOcompounds was substantially higher in green leaves (GL) compared to photobleached leaves (PBL). Phenolic acids decreased with increasing pH, whereas flavonols and phenylpropanoid sucrose esters remained relatively constant. Oleuropein content increased under basic conditions, while at pH 8, the combined effects of temperature and oxygen promoted greater degradation of rutin than quercetin. Importantly, rutin was absent in PBL extracts. The proportion of piscidic acid in the total phenolic pool ranged from 67 % to 86% w / w, averaging about 80% w / w.

[0269] The concentration of phenolic acids appears lower at pH 8.5 than at acidic pH (4.6). However, this is likely not because their solubility is lower. In fact, phenolic acids are more soluble at alkaline pH due to deprotonation of their acidic groups, which results in a charged and more hydrophilic form. The observed lower concentration at pH 8.5 is probably due to alkaline degradation or irreversible complex formation with other molecules (e.g., proteins or other phenolics) that occurs during the extraction process, making them undetectable by the UPLC / MS method (Friedman & Jurgens, 2000; Schäfer et al., 2021). On the other hand, the concentration of flavonols and phenylpropanoid esters remain relatively constant, while the concentration of oleuropein increases. Importantly, the ratio of phenolics to saponins (see bottom line in Table 5), decreases at basic pH. However, this must be balanced against the darker color of the extracts and the complexity of the process due to concomitant extraction of proteins.

[0270] Piscidic acid represents 75-85% w / w of the total phenolics. This agrees with previous reports that indicate that piscidic acid represents 75-87% w / w of all phenolics in products derived from quillaja bark and biomass (excluding leaves) (Maier, Conrad, Carle, et al., 2015). Also, the extracts contain flavonols, specifically rutin and quercetin derivatives, which exhibit a yellow coloration. Rutin degrades fast with basic pH, Fe2+, Cu2+and temperature (Makris & Rossiter, 2000), leading to oxidation and browning. It is postulated that metal ions promote flavonol oxidation through reactive oxygen species formation, whereas browning is due to oxidation and metal-polyphenol interactions. This suggests that using EDTA during the extraction should be advantageous. A third group of compounds are phenylpropanoid sucrose esters, which are also present in bark and commercial extracts (Maier, Conrad, Steingass, et al., 2015). Phenylpropanoid sucrose esters are naturally occurring compounds isolated from various plants and are structurally characterized by a sucrose core connected to one or more Ph-CH=CH-CO- moieties through an ester linkage (Panda et al., 2011). Interestingly, oleuropein is present in notable concentrations. This compound has not been previously identified in quillaja extracts derived from bark or quillaja biomass, likely due to its prevalence in high concentrations solely within leaves, as reported for olive leaves. Importantly, olive leaves 31 B 25 -074-2 WOsubjected to air drying have been shown to oxidize due to PPO activity, generating a brown coloration (De Leonardis et al., 2015).

[0271] Table 5: Phenolic composition of quillaja leaves extracted at 50°C in water at different pH.Compound identity Concentration mg / ml Photo-bleached leaves mg / mlpH 2.7 pH 4.6 pH 8.5 pH 2.7 pH 4.6 pH 8.5 Phenolic acidsShikimic acid 0.000 0.000 0.000 0.000 0.000 0.000 Quinic acid 0.000 0.000 0.000 0.832 1.388 0.101 Piscidic acid* 6.24 6.51 4.40 3.41 3.64 3.26 Coutaric acid 0.009 0.009 0.010 0.073 0.076 0.076 Caffeic acid 1,1 -dimethylallyl ester 0.027 0.027 0.029 0.027 0.028 0.030 p-coumaric acid derivate 0.013 0.012 0.015 0.013 0.012 0.015 l-O-4-(Coumaroyl)-glucose 0.073 0.089 0.061 0.031 0.039 0.043 Sub-total phenolic acids 6.8 7.1 4.3 4.8 4.2 3.4 Flavonols:quercetin / rutin &derivativesQuercetin 3-O-pentosylrutinoside 0.101 0.100 0.103 0.104 0.084 0.114 Quercetin pentoside 0.104 0.098 0.086 0.099 0.084 0.117 Quercetin-pentoside-hexoside 0.110 0.108 0.096 0.089 0.080 0.101 Rhamnosylhexosyl methyl quercetin 0.103 0.090 0.086 0.085 0.066 0.080 Dihexosylquercetin 0.092 0.080 0.074 0.068 0.057 0.070 Rutin 0.152 0.124 0.216 0.102 0.084 0.117 Sub-total flavonols 0.357 0.420 0.392 0.161 0.085 0.117 Phenylpropanoid Sucrose EstersQuillajaside A 0.203 0.196 0.221 0.095 0.069 0.106 Quillajaside A isomer 0.086 0.100 0.092 0.049 0.056 0.074 Quillajaside A isomer 0.008 0.015 0.000 0.012 0.049 0.000 Sub-total Phenylpropanoid Sucrose 0.34 0.311 0.319 0.172 0.174 0.180 EstersOleuropein 0.73 0.77 0.91 0.130 0.126 0.378 Total phenolics mg / ml 8.18 8.57 5.93 5.23 4.56 4.10 Saponin g / 1 7.7 7.7 7.7 8.0 7.7 15.6 Ratio phenolics / saponins 1.06 0.93 0.33 0.68 0.53 0.26

[0272] Overall decolorization with activated charcoal.32 B 25 -074-2 WO

[0273] An initial set of trials explored conventional decolorization methods commonly applied to botanical extracts, specifically the use of activated charcoal. Previous studies have reported that activated charcoal effectively removes chlorophylls and their degradation products, as well as phenolic compounds (Tzima et al., 2020). In this approach, the crude extract was passed through cellulose filters containing activated charcoal. Experimentally, a sample of 500 ml of extract obtained at pH 8.0 and 55 °C for 2h was pumped through a filter unit (10T Series Lab filter, Ertel Alsop), containing cellulose filter impregnated with activated charcoal (MC 35-C, Ertel Alsop, 10 cm diameter) at a flow rate of 10 ml / min with a peristaltic pump. 50 ml fractions were collected and analyzed for saponin by RP-HPLC. Table 6 shows both the saponin and chlorophyll reduction when the extract obtained at 50 °C and pH 8.5 for 2 h was passed through a cellulose filter impregnated with activated charcoal. This treatment produced markedly lighter-colored extracts and as shown in Figure 17, a substantial adsorption of phenolics and their derivatives was observed, as evidenced by the near-complete disappearance of the corresponding initial hydrophilic peaks in the RP-HPLC chromatograms. This behavior is consistent with the well-documented high affinity of activated charcoal for phenolic acids (Dąbrowski et al., 2005). Moreover, the characteristic green-brown hue, attributed to chlorophylls and their degradation products, also decreased, though the reduction estimated as the area of the absorbance at 650-700 nm in the TECAN spectrophotometer, only indicated a reduction of 30% and 28.6% at pH 3.5 and 6.5 respectively. The activated charcoal-treated extracts were subsequently subjected to ultrafiltration (UF), yielding final products with an acceptable dark- yellow coloration.

[0274] However, as summarized in Table 6, the saponin concentration in the activated charcoal treated extracts decreased markedly following charcoal treatment, with overall losses of 25.8 and 33.2% (w / w) relative to the initial saponin content, at pH 3.5 and 6.5, respectively. These results are consistent with previous reports describing the strong adsorption of triterpenoid saponins onto activated charcoal (Mironenko et al., 2024; Mironenko & Selemenev, 2018). Therefore, despite its effectiveness in color reduction, activated charcoal was deemed unsuitable for decolorizing quillaja extracts due to the excessive loss of saponins.

[0275] Table 6: Saponin and chlorophyll reduction in extracts obtained at pH 8.5, 55°C, 2 h treated with cellulose filters impregnated with activated charcoal (MC 35-C, Ertel-Alsop).Saponin g / 1 Chlorophyll area 650-700 nmpH 3.5 6.5 3.5 6.5Initial 7.3 7.3 3.7 4.3Fl 1.1 3 2.2 4.133 B 25 -074-2 WOF2 3.8 4.5 2.6 3F3 4.8 5 2.6 2.8F4 5.3 6 2.6 2.6F5 5.3 5.9 2.4 3F6 5.6 5.8 2.8 2.7F7 5.8 6 3.1 3.3F8 5.7 6 2.9 3F9 5.7 6 2.5 3.2F10 5.7 6 2.2 3Total recovered 4.9 5.4 2.6 3.1% reduction 33.2% 25.8% 30.0% 28.6%

[0276] Reduction of chlorophy 1 with pH shift and bentonites

[0277] A set of experiments was performed to reduce the chlorophyll content via a pH shift from basic to acid pH, followed by contact with calcium and sodium bentonites, to remove the water-soluble protein chlorophyll complexes (WSPCs). In extractions performed at pH 8.5, 55 °C and 2 h, shifting the pH post-extraction from 8.5 to 3.5 reduced the CI from 0.7 to 0.55. However, this reduction was not sufficient to prevent darkening of the extracts during concentration with UF. Therefore, to remove the remaining chlorophyll the extract was contacted with calcium and sodium bentonites, as they could adsorb the WSPCs. A first set of experiments considered passing the extract as was passed through cellulose filters containing calcium bentonite T5 (Ertel-Alsop, USA, in a similar equipment and methods as those described in Example C. In these experiments, chlorophyll was estimated as the area at 650-700 nm measured by the TECAN spectrophotometer, and saponin concentration with RP-HPLC.

[0278] Table 7 shows that saponin losses with calcium bentonite T5 were negligible, while the chlorophyll area was reduced by about 18%, corresponding to a CI reduction of 20%. However, subsequent diafiltration and concentration with UF still resulted in dark final concentrates.

[0279] Table 7: Saponin and chlorophyll reduction in extracts obtained at pH 8.5, 55°C, 2 h treated with cellulose filters impregnated with calcium bentonite (T5, Ertel-Alsop).Calcium bentonite T5Saponin g / 1 Chlorophyll area 650-700 nm pH 3.5 6.5 3.5 6.5Initial 7.4 7.9 3.9 4.2Fl 7.4 7.6 3.2 3F2 7.4 7.4 3.1 2.834 B 25 -074-2 WOF3 7.4 7.5 3.2 2.6F4 7.3 7.8 2.8 3F5 7.4 7.8 3.3 2.6F6 7 7.7 3.3 3.1F7 7.4 7.9 3.3 3.2F8 7.3 7.9 3.3 3F9 7.3 7.7 3.1 3.1F10 7.5 7.9 3.4 3.17.34 7.72 3.2 3.0% reduction 0.8% 2.3% 17.9% 29.8%

[0280] A second set of experiments was performed using sodium bentonite at different concentrations and measuring the CI after 3 hours of contact at RT. As detailed in Table 8, the CI was reduced from 0.7, to 0.6, 0.4 and 0.19, with 1, 3 and 5 g / 1 of sodium bentonite, respectively. Importantly, there were no saponin losses. Subsequent diafiltration and concentration of the extracts darkened moderately, and when spray dried yielded a light-colored white powder. This was corroborated with a series of experiments, where consistently the final extracts did not darken significantly, provided the CI was < 0.20.

[0281] Table 8: Saponin and chlorophyll reduction in extracts obtained at pH 8.5, 55°C, 2 h treated sodium bentonite added at RT and agitation for 3 hChlorophyll index (CI) Saponin yield, % w / w dry leaves Initial sample 0.7 10.1%Sodium bentonite, 1 g / 1 0.6 10.0%Sodium bentonite, 3 g / 1 0.4 10.1%Sodium bentonite, 5 g / 1 0.19 10.0%

[0282] Removal of phenolics with ultrafiltration.

[0283] To further improve the purity and color of the extracts, additional experiments aimed at minimizing the concentration of phenolic compounds responsible for undesirable browning. Prior studies with quillaja saponin micelles suggest that beyond covalently bound phenolic compounds, the interactions are likely driven by hydrogen bonding between the sugar groups on saponin molecules and phenolic glycosides (Tippel et al., 2017). Moreover, some hydrophobic phenolic derivatives can be incorporated within the hydrophobic core of the micelles.

[0284] To disrupt these interactions and release the phenolic compounds from the micelles, two strategies were tested. The first set of experiments considered adding Ca2+ions to disrupt the associations between the phenolics and the external layers of the micelles, allowing its removal during diafiltration. To test this, extracts treated with sodium bentonite were supplemented with 35 B 25 -074-2 WOCaCl₂, to attain 700-800 mg Ca2+ / l in solution at a pH of about 5. Table 9 shows that after diafiltration and concentration, most of the phenolics were effectively removed, confirming that Ca2+disrupts the association between phenolics and the saponin micelles. For example, piscidic acid, which is the major phenolic compound in quillaja extracts, was reduced by approximately 97% w / w. Also, Figure 17 shows that under these conditions, the initial signal in the RP-HPLC, corresponding to hydrophilic phenolics, is reduced significantly.

[0285] A plausible explanation is that adding calcium at pH 5 disrupts the interaction between saponin micelles and entrapped phenolic compounds, facilitating effective diafiltration. While the saponins form large micelles retained by the ultrafiltration (UF) membrane, smaller phenolic impurities are often encapsulated within these micelles due to hydrophobic and electrostatic forces. The addition of divalent calcium ions (Ca2+) at pH 5 introduces a strong competitive binding force that displaces the phenolics from the micellar structure. This disruption causes the phenolic compounds to be released as smaller molecules into the solution. During the diafiltration process, the large saponin micelles remain trapped by the UF membrane, but the now-freed phenolic compounds and other impurities can efficiently pass through with the permeate, allowing for selective purification of the quillaja saponins.

[0286] A second strategy explored to disrupt the association between the phenolics and the saponin micelles was to perform the diafiltration at pH 7.5-8.0. This is a different mechanism than adding Ca2+, which relies on competitive binding. At pH 7.5-8, the electrostatic repulsion between the highly deprotonated, negatively charged phenolic molecules and the similarly negatively charged saponin micelles increases significantly, effectively pushing the phenolics out of their association with the micelles. Since the large saponin micelles are still retained by the ultrafiltration membrane, the now-separated, smaller phenolic molecules can pass through with the permeate, resulting in a cleaner product. Figure 18 shows that this strategy was equally effective, probably because at this pH both the phenolic acids and the micelles have a negative charge that reduces strong associations.

[0287] Additional experiments examined if the micelles were incorporating hydrophobic phenolic compounds within the micelle core. Figure 19 presents dynamic light scattering analysis (Zetasizer) of the final product in an aqueous buffered solution. The average Z-size of 6.4 nm aligns with previous reports on the hydrodynamic radius of quillaja saponin micelles, typically between 3-6 nm (Mitra & Dungan, 1997). This consistency suggests that the micelles formed do not contain large, solubilized molecules within their hydrophobic core, which are usually present in leaf extracts.36 B 25 -074-2 WO

[0288] Table 9: Reduction of phenolics due to addition of 700-800 mg Ca2+ / l during diafiltration and concentration with 20 kDalton UF membranes. Extract obtained at pH 8.5, 55 °C, 2 h.Sample - mg / mlCompound IdentityExtract Post UF with Ca2+Shikimic acid 0.10 0.10 Quinic acid 0.84 0.19 Piscidic acid 3.15 0.09 Quercetin pentoside 0.40 0.35 Coutaric acid 0.19 0.00 1-O-(4-Coumaroyl)-glucose 0.15 0.00 Quercetin-pentoside-hexoside 0.00 0.00Dihexosylquercetin 0.37 0.35 Caffeic acid 1,1-dimethylallyl ester 0.00 0.00Rutin 0.39 0.00 Quillajaside B 0.03 0.00 Quillajaside A 0.42 0.10 Quillajaside B isomer 0.00 0.00 Quillajaside A isomer 0.17 0.03 Rhamnosylhexosyl methyl quercetin 0.00 0.00Quillajaside B isomer 0.00 0.00 Quillajaside A isomer 0.01 0.00 Quercetin 3-O-pentosylrutinoside 0.00 0.00Oleuropein 0.2 0.10 Total phenolics 6.42 1.31 Total Saponins 7.74 18.33 Ratio phenolics / saponins 0.83 0.07

[0289] Batch extraction at pH 8-8.5, 50-55°C. Production of purified products. Validation studies.

[0290] Based on the positive results obtained in the previous studies, a series of validation experiments were conducted to assess the reproducibility of extractions performed at 50-55 °C and pH 8.0-8.5, followed by chlorophyll reduction with sodium bentonite and diafiltration under either calcium-assisted or pH 8.0 conditions, and final spray drying. The experiments were carried out using dry leaves from two-year-old Quillaja plants cultivated both in a greenhouse 37 B 25 -074-2 WO(Oxford Tract, Rausser College of Natural Resources, UC Berkeley) and in open-field conditions (Ojai, California).

[0291] Table 10 summarizes the yields and CI values from a representative set of experiments (3 repetitions) performed under standardized extraction conditions (pH 8.0-8.5, 50-55 °C, 2 h), followed by bentonite clarification (6 h with agitation), diatomaceous earth (DE) filtration (20 g / 1 Celite, Ertel, nanofiltration at 70 psig for partial concentration (50% volume reduction), diafiltration with 20 kDalton UF membranes (four volumes of water per volume of extract), final concentration with UF, ion removal, pasteurization at 70 °C for 45 minutes, and spray drying.

[0292] Extensive testing established that the chlorophyll index (CI) must be reduced to 0.20-0.25 prior to ultrafiltration (UF) to obtain light-colored spray-dried powders. As shown in Table 10, field-grown plants consistently produced extracts with higher initial CI values than greenhouse-grown plants, requiring proportionally greater bentonite dosages to achieve CI < 0.20-0.25. A consistent trend was also observed wherein extracts following UF exhibited CI values of approximately 0.6-0.75, yet the resulting spray-dried powders were uniformly light-colored (white to pale yellow) and contained >85% w / w saponins in the final product,

[0293] Table 10: Extractions performed at pH 8.0-8.5, 50-55 °C, 2 h (3 repetitions per plant origin)Plant CI post Bentonite CI post CI post UF Final product % w / w saponin extraction g / 1 bentonite color in final product Greenhous 0.45 2-3 0.2 0.6 Cream 86.5% eField 0.6 3-5 0.25 0.75 Cream 86 %

[0294] Batch extraction with buffers and prolonged extraction times

[0295] Despite the positive results obtained with extractions performed at 50-55 °C and pH 8.0-8.5, followed by chlorophyll reduction with sodium bentonite and diafiltration under either calcium-assisted or pH 8.0 conditions, further experiments were performed to improve the scalability of the process. To this end, extraction conditions were investigated to reduce or eliminate the use of chlorophyll reducing agents like bentonite, as this introduces additional process steps and waste streams. As it is described in the following examples, the use of RT, phosphate buffers and prolonged extraction times and use of partially photobleached leaves gave the intended results.

[0296] In these examples, extractions were performed in a 20-1 stainless steel tank (Keg-Outlet, USA), equipped with a 100 pm sieve and a discharge valve at the bottom, a variable speed agitator and an electric mantle to maintain the temperature. The pH was maintained using an automatic pH controller (ECD, Anaheim, California) and 0.5 M NaOH. The water to dry leaf38 B 25 -074-2 WOmass ratio was 15:1. The leaves were grounded to 1 mm with a cutting milling equipment (Retsch, model SM400XL).

[0297] Following extraction, the bottom valve was connected to a peristaltic pump (Watson Marlow model 630UN / RE, Massachusetts, USA), that pumped the extract through a 75 pm inline filter, and then to a cellulose filter unit (BuonVino, SuperJet Filter, USA) equipped with 10 pm filters. Ultrafiltration to remove non-saponin compounds via concentration and diafiltration was performed in a system equipped with spiral 20 kDalton polyethersulfone -based ultrafiltration membranes (Synder, Vacaville, California, USA. The system was operated at 30 psig. Diafiltration was performed with deionized water, maintaining the level of the UF retentate reservoir constant and pumping water at the same rate of filtrate removal. Following ultrafiltration, the extracts were deionized in a water softener system (Pentair Pentek, USA). The extracts were pasteurized at 70 °C for 45 min and then spray dried in a bench-scale spray drier at 170 °C (Buchi, Mini Spray Dryer S-300). At different points during the process the °Brix and the A670were measured to determine the chlorophyll index. Samples were also used to determine saponin concentration using RP-HPLC or UPLC / MS when saponin identification was required or when analyzing phenolics.

[0298] Phosphate buffers of 100 mM, 75 mM, and 150 mM were prepared by mixing appropriate proportions of monobasic (NaH₂PO₄) and dibasic (Na₂HPO₄) sodium phosphate salts to achieve the desired molarity and pH (typically 6.8-7.4). Each buffer was prepared in deionized water, adjusted to the target pH using small volumes of 1 M NaOH or HCl, and brought to final volume before filtration.

[0299] Plants used in the next 5 examples were obtained from 2-year-old plants growing in a greenhouse (Oxford Tract, Rausser College of Natural Resources, UC Berkeley).

[0300] Effect of pH and Temperature and prolonged extraction time on Saponin Yield and Color Development

[0301] The optimization of extraction conditions aimed at minimizing chlorophyll solubilization and preventing phenolic degradation led to an unexpected finding. As shown in Table 11, increasing the extraction pH from its equilibrium value of 4.6 to 7.0 (adjusted with 0.5 M NaOH) resulted in a significant increase in saponin yield, from 6.3% to 7.8% w / w after 12 hours. Notably, the color index (CI) at pH 7 remained low at 0.22.

[0302] Table 12 further compares these results with extractions performed at pH 8.5 under different temperatures and extraction times. At pH 4.6 and room temperature (RT) for 12 hours, the saponin yield was 6.3% w / w with a CI of 0.25. When the pH was increased to 8.5 and the extraction carried out at RT for 2 hours, the saponin yield rose to 7.7% w / w, with only a modest increase in CI to 0.30. This indicates that pH elevation alone enhances saponin solubilization 39 B 25 -074-2 WOwithout substantial color development. However, when the same alkaline condition (pH 8.5) was combined with elevated temperature (55 °C) for 2 hours, the saponin yield further increased to 9.4% w / w, but the extract darkened significantly (CI = 0.55). This -83% increase in color index, despite shorter extraction time, confirms that higher temperature promotes unwanted pigment formation and degradation reactions.

[0303] These observations suggest that prolonged extraction time under mild conditions (e.g., pH 7, RT) can compensate for the reduced kinetics of saponin solubilization typically achieved under harsher conditions (high pH or elevated temperature), while effectively preventing the color formation associated with pigment degradation.

[0304] Table 11: Comparison of saponin yields (% w / w based on dry leaves) at RT and pH 4.6, 7 and 8.5 (12 h) vs 55 °C, pH 8.5, 2 h.Solvent pH Temperature Saponin yield CI Time, h Water 4.6 RT 6.30% 0.25 127 RT 7.8% 0.22 128.5 RT 7.70% 0.3 255 °C 9.4% 0.55 2

[0305] Table 12: Kinetic of saponin yields (% w / w based on dry leaves) and CI at pH 4.6 and pH 7pH 4.6- equilibrium pH pH 7 with 0.5 NaOHTime Saponin yield CI Saponin yield CI1 3.1% 0.1 3.7% 02 3.4% 0.14 4.7% 0.044 4.1% 0.16 5.7% 0.068 4.8% 0.22 7.1% 0.1612 6.3% 0.25 8% 0.2222 6.8% 0.28 8.7% 0.23

[0306] Room-Temperature Phosphate Buffer Extraction versus High-Temperature, Basic pH Extraction

[0307] Based on the positive outcomes from Example H, extractions were next performed using a 100 mM phosphate buffer to maintain the pH within 7.0-7.2. This approach minimized the need for direct addition of 0.5 M NaOH, which can generate localized regions of very high pH near the addition point despite continuous agitation.

[0308] As shown in Table 13, extraction with 100 mM phosphate buffer at pH 7.2, room temperature, and 12-hour duration yielded 8.2% w / w saponins with a color index (CI) of 0.20 — an improvement over extractions where pH was controlled manually with NaOH. This yield was 40 B 25 -074-2 WOonly about 15% lower than that obtained under the conventional high-temperature process (9.4% at 55 °C, pH 8.5, 2 h) but produced extracts with markedly superior color quality (CI 0.20 vs 0.55), representing a 64% reduction in color intensity. For commercial implementation, this trade-off is economically favorable, given the elimination of bentonite clarification, lower waste disposal costs, and the premium value of lighter-colored, refined extracts.

[0309] Table 13: Comparison of extraction at pH 8.5, 55°C, 2 h vs 100 mM phosphate buffer, pH 7.2, 12 h.Extraction conditions Saponin CIyield w / w%Water, pH 8.5, 55°C, 2 h 9.4% 0.55100 mM phosphate buffer, pH 7.2, 12 h 8.2% 0.2

[0310] Phosphate Buffer Concentration Optimization

[0311] Systematic optimization of phosphate buffer concentration at pH 7.0 and room temperature (12-hour extraction) revealed an optimal ionic strength range for maximizing saponin yield while minimizing color development. As shown in Table 14, extraction with 75 mM phosphate produced a saponin yield of 7.6% w / w and a color index (CI) of 0.21. Increasing the buffer concentration to 100 mM improved the yield to 7.8% w / w while maintaining a light color (CI 0.1 ). However, further increasing the phosphate concentration to 150 mM reduced the yield to 7.2% and slightly increased color intensity (CI 0.21), suggesting that excessive ionic strength favors the co-extraction of colored compounds. Therefore, the 100 mM phosphate buffer represents the optimal condition, achieving the highest saponin yield with minimal color formation. This trend is visually evident in Figure 20, where extracts obtained with 75 and 150 mM buffers appear darker and more reddish, while the 100 mM extract exhibits a light-yellow hue.

[0312] Table 14: Optimization of phosphate buffer concentration. Extractions at RT, 12 h pH 7Solvent Saponin yield CI75 mM phosphate 7.6% 0.21100 mM phosphate 7.80% 0.19150 mM phosphate 7.2% 0.21

[0313] Saponin extraction kinetics with 100 mM phosphate buffer, RT.

[0314] Table 15 and Figure 21 present the kinetics of saponin extraction (expressed as % w / w based on dry leaf weight) along with the corresponding chlorophyll index (CI) for extractions conducted using 100 mM phosphate buffer at pH 7.2 (average of three experiments). The results indicate that the optimal extraction time is approximately 12 hours, at which point about 94%41 B 25 -074-2 WOw / w of the total extractable saponins are recovered relative to a 17-hour extraction. Extending the extraction beyond this point is not recommended, as it may promote microbial growth without yielding significant additional saponin recovery. The CI values remain essentially constant over the extraction period, confirming limited solubilization of chlorophyll and other colored compounds.

[0315] Table 15: Kinetics of saponin extraction (% w / w yield based on dry leaves) and chlorophyll index (CI) at pH 7.2 adjusted with 100 mM phosphate bufferHours % saponin CI2 6.0% 0.104.5 7.8% 0.108 8.2% 0.112 9.0% 0.1017 9.6% 0.11

[0316] Phosphate versus Bicarbonate Buffer Systems

[0317] The critical importance of buffer ion chemistry was demonstrated through direct comparison of phosphate and bicarbonate buffers at identical pH 7.2 and room temperature with 12-hour extraction. As shown in Table 16, with 100 mM phosphate buffer, saponin yield was 8.0% (CI 0.20). Using a 100 mM bicarbonate buffer under identical conditions resulted in significantly lower saponin yield of 7.0% (CI 0.20). Despite achieving identical color indices, the phosphate buffer produced 14% higher saponin yield. This difference is attributable to the polyvalent nature of phosphate (HPO-at pH 7), which enables more effective selective precipitation of protein-phenolic and protein-chlorophyll complexes compared to the monovalent bicarbonate (HCCh-), while simultaneously maintaining superior ionic conditions for saponin solubilization.

[0318] Moreover, Table 16 shows extractions at pH 7.2, room temperature, and 12-hour extraction time with 25 mM Na2HPO4-2H2O supplemented with 90 mM NaCl, a cost-effective alternative to standard 100 mM phosphate. This formulation achieved saponin yield of 8.10% (CI 0.20), essentially identical to the standard 100 mM phosphate buffer (8.0%, CI 0.20), while maintaining equivalent ionic strength (12-13 mS / cm). This demonstrates that the critical parameter is not phosphate concentration per se, but rather the total ionic strength combined with the presence of polyvalent phosphate ions, enabling industrial operation with substantially reduced reagent costs without sacrificing extraction performance.

[0319] Table 16: Phosphate buffer vs bicarbonate buffer & low-cost phosphate / NaCl buffer Solvent pH Temperature Saponin yield CI Time, h42 B 25 -074-2 WO100 mM phosphate 7.2 RT 8% 0.2 12 Phosphate / NaCl 7.2 RT 8.10% 0.2 12 Bicarbonate 7.2 RT 7.00% 0.2 12

[0320] Intermediate Temperature Extraction Protocol with Phosphate / NaCl Buffer

[0321] An enhanced extraction protocol incorporating a brief elevated-temperature step followed by an extended room-temperature extraction was developed to further improve saponin yield while limiting color formation. In this method, extraction begins at 50 °C for 10-30 minutes and is then cooled to room temperature to continue for 6-12 hours. The initial elevatedtemperature period promotes rapid saponin solubilization and partial disruption of leaf cellular structures, capturing the kinetic advantages of higher temperature. The subsequent roomtemperature extraction allows continued diffusion of saponins without prolonged exposure to conditions that favor degradation or color development.

[0322] Experiments performed with two-year-old field-grown plants from Ojai, California — known to produce extracts with higher CT values — showed that this approach increased saponin yield but also raised the CI from 0.34 to 0.64 (Table 17 and Figure 22). Nonetheless, the modified procedure accelerated extraction kinetics, enabling shorter operational cycles.

[0323] Importantly, field-grown quillaja leaves exposed to direct sunlight for 3-10 days postharvest undergo partial photobleaching, substantially reducing their initial chlorophyll content. As shown in Table 18, when the intermediate-temperature extraction protocol (50 °C initial stage followed by room-temperature continuation) was applied to such photobleached leaves, the resulting extracts exhibited very low chlorophyll levels (CI = 0.1), eliminating the need for bentonite clarification. This integration of a simple agricultural pre-treatment with the optimized extraction process removes the requirement for chemical clarification and associated waste streams while maintaining saponin yields comparable to conventional high-temperature methods.

[0324] Table 17: Comparison of saponin yields (% w / w based on dry leaves) extracted with phosphate / NaCl buffer at Room temperature (RT) vs initial 30 minutes at 50 °C, followed by extraction at RT.RT, pH 7.2 50 °C, pH 7.2, 30 min; RTTime % saponin CI % saponin CI1.5 4.4% 0.12 5.7% 0.193 5.0% 0.17 6.6% 0.376 6.2% 0.26 7.4% 0.449 7.3% 0.33 7.8% 0.5043 B 25 -074-2 WO12 7.3% 0.34 8.0% 0.64

[0325] Table 18: Integration of Modified Extraction Procedure (50 °C initial stage followed by room-tcmpcraturc continuation) with leaves exposed to sunlight for 5-10 days post pruning.pH Temperature Saponin yield % w / w CI Time 7.2 RT 6.80% 0.05 12 7.2 30' 50 °C, RT 8.0% 0.11 12

[0326] Pi ot plant batch extraction experiments. Comparison of extractions at pH 8.5, 2 h vs pH 7 with buffer, RT and 12 h

[0327] To validate bench-scale findings at manufacturing scale, a comprehensive pilot-plant study compared two extraction processes using a 100-liter extraction unit equipped with a 100 pm sieve, bottom discharge valve, and mechanical agitation. Deionized water was sparged with nitrogen to reduce dissolved oxygen and supplemented with 1 g / L ascorbic acid and 0.1 g / L EDTA as antioxidants. Water was pre-heated to approximately 60°C before leaf addition at a 12:1 water-to-dry-leaf ratio by weight. Leaves were obtained from 2-year old shrubs grown in Ojai, California. Temperature was maintained via external electric mantle heating, and pH was controlled through automated addition of 0.5 M NaOH. The results are shown in Table 19.

[0328] Process 1 (High-Temperature / Bentonite): As shown in Table 19, extraction at pH 8-8.5 and 55°C for 2.5 hours yielded 11.0% w / w saponin with CI 0.87. Extract was continuously removed at 2 L / min via peristaltic pump to a refrigerated tank where pH was adjusted to 3.8 using 100 g / L citric acid. Following cooling below 25°C, bentonite (2.5 g / L) was added and stirred for 6 hours, then settled at 8°C for 12 hours. Supernatant was combined with 20 g / L diatomaceous earth (CELITE) and filtered through a pilot-scale frame filter with turbidity monitoring. The filtered extract was concentrated via nanofiltration at 70 psig and below 15°C (50% volume reduction), then purified using pilot-scale ultrafiltration with 20 kDalton membranes. During ultrafiltration, pH was elevated to 5, and 5 g / L calcium was added to enhance phenolic removal. Extract underwent diafiltration with 3.5 volumes of deionized water, polishing through 0.45 pm cellulose filters, and deionization. The extract was heat sterilized in a water bath and spray dried.

[0329] Process 2 (Room- Temperature Buffer): As shown in Table 19, extraction at pH 7 and room temperature for 12 hours yielded 9.40% w / w saponin with CI 0.12. Extract was transferred directly to the refrigerated tank without pH adjustment or bentonite addition, proceeding directly to nanofiltration concentration and identical ultrafiltration purification, polishing, sterilization, and spray drying.

[0330] Pilot-plant results demonstrated that Process 1 yields higher saponin extraction (11.0% w / w versus 9.4% w / w) but requires substantial bentonite treatment to reduce color from CI 0.8744 B 25 -074-2 WOto 0.2 post-pH adjustment. Process 2 achieved significantly lower initial color (CI 0.12), eliminating bentonite addition and diatomaceous earth filtration. Critically, color development patterns diverged between the two concentration methods: nanofiltration did not increase color intensity in either process (Process 1: CI 0.2— >0.23; Process 2: CI 0.11 — >0.12), as nanofiltration excludes phenolic molecules while concentrating other extract components. The retained phenolics functioned as natural antioxidants, minimizing oxidative color development. In contrast, ultrafiltration significantly increased color in both processes (Process 1: CI 0.23^0.70; Process 2: CI 0.12 — >0.65) as ultrafiltration removed protective phenolics while retaining oxidation-prone molecules.

[0331] Process 1 experienced 30.9% w / w overall saponin losses post-extraction, compared to 22.6% w / w for Process 2, representing a 26% reduction in losses for the room-temperature approach. Both processes achieved comparable final saponin purity in spray-dried powder (85.5% and 86.5%, respectively), substantially exceeding typical commercial extracts (60-65% w / w) and approaching vaccine-grade specifications (>90% w / w). Process 2 yielded a white-to-light cream spray-dried powder, while Process 1 produced light-yellow powder.

[0332] Operationally, Process 2 required approximately 4x longer extraction time (9 hours versus 2.5 hours), necessitating expanded extraction tank capacity. However, Process l’s bentonite treatment introduced 6 hours of stirring plus 12 hours of settling (18 hours total postextraction contact time), offsetting the apparent extraction advantage. Process 2 eliminated bentonite procurement, diatomaceous earth filtration supplies, waste disposal costs, and extended tank cooling requirements, while Process 1 's higher extraction yield partially offset these cost advantages. All pilot-plant experiments were performed in triplicate with consistent results.

[0333] Table 19: Comparison of high pH-T process, short time vs process at RT, with phosphate buffer- long extraction time, pilot-plant scale.pH 8, 55 °C, 2.5 h pH 7, RT, 12 h Process step Saponin yield CI Saponin yield CI Extraction 11.0% 0.87 9.40% 0.12 pH to 3.5 0.76 0.1 Bentonite, 3.5 g / 1 0.2 - Post bentonite removal 10.2% - Post Filtration, 10 pm 9.8% 0.19 9.1% 0.11 Post nanofiltration 9.7% 0.23 8.70% 0.12 Post UF 8.0% 0.7 7.7% 0.6545 B 25 -074-2 WOPost pasteurization, 75 °C, 1 h 8.0% 0.75 8% 0.71 Post spray- drying 7.6% 7.3%Overall losses post extraction 30.9% 22.6%Color spray dried powder Light yellow White light cream Saponin content final powder 85.5% 86.5%

[0334] Water-based Batch Extraction, Leading to Semi-refined Product

[0335] 10 kg of ground (1- 3 mm) dried green leaves obtained from 2-year old plants grown in the Oxford Tract, Rausser College of Natural Resources, UC Berkeley were treated with 100 L of water containing ascorbic acid and EDTA for 1.5 hours at pH 8.5 and 50 °C with agitation. The leaves were separated using an internal sieve, and the extract recovered. 80 L of fresh water was added to wash the occluded solution. A total of 150 L of extract and wash were recovered.

[0336] The pH of the combined extracts was adjusted to 3.5 and cooled to 20-40 °C. Sodium bentonite (3 g / L) was added and agitated for 3 hours. The solution was filtered through a plateframe filter with diatomaceous earth. The extract underwent nanofiltration for concentration and passed through a 0.45 pm filter. The final extract was spray-dried, with a final yield based on dry leaves of 9% w / w. Overall saponin losses of the saponins recovered during extraction and washing were 15% w / w. The final product was amber with a saponin content > 90% w / w determined by RP-HPLC. The final product was dark amber with a saponin content > 30% w / w determined by RP-HPLC.

[0337] Water-based Batch Extraction Leading to Refined Product

[0338] 10 kg of ground (1- 5 mm) dried green leaves obtained from 2-year old plants grown in the Oxford Tract, Rausser College of Natural Resources, UC Berkeley were treated with 100 L of water containing ascorbic acid and EDTA for 1.5 hours at pH 8.5 and 50 °C with agitation. The leaves were separated using an internal sieve, and the extract recovered. 80 L of fresh water was added to wash the occluded solution. A total of 150 L of extract and wash were recovered.

[0339] The pH of the combined extracts was adjusted to 3.5 and cooled to 20-40 °C. Sodium bentonite (3 g / L) was added and agitated for 3 hours. The solution was filtered through a plateframe filter with diatomaceous earth. The extract underwent nanofiltration for concentration, followed by diafiltration with ultrafiltration membranes (10-100 kDalton) in the presence of divalent ions. The final extract was spray-dried, with a final yield based on dry leaves was 7% w / w. Overall saponin losses of the saponins recovered during extraction and washing were 20% w / w. The final product was amber in color, with a saponin content > 90% w / w determined by RP-HPLC.

[0340] Continuous Process, Leading to Semi-refined Product46 B 25 -074-2 WO

[0341] Twenty kilograms of freshly harvested leaves, including stems, were processed using a commercial garden chipper shredder. The leaves were obtained from 2-year old plants grown in the Oxford Tract, Rausser College of Natural Resources, UC Berkeley. The shredded material was then mixed with 60 liters of water and agitated for 2 hours at room temperature. Following this, the mixture was passed through a dewatering screw press (Vincent Corp).

[0342] The recovered extract exhibited abundant foam and a dark green color. To reduce its chlorophyll content, the extract was allowed to settle overnight and subsequently treated with 2.5 g / L of sodium bentonite for 2 hours under agitation. Diatomaceous earth was then added, and the extract was filtered using a small plate-and- frame filter. The final extract had a clear yellow color, with an overall yield of 7% based on dry material.

[0343] This yield was comparable to that of a batch extraction process, where thoroughly milled leaves (processed in a Ninja blender) were extracted at 65°C for 2 hours. Interestingly, the saponin content in the soluble fraction was 45% w / w for the screw -pressed leaves, compared to 32% w / w for the batch-extracted leaves. This suggests that the screw-pressed method results in a lower concentration of non-saponin compounds, an advantage for subsequent refinement processes.

[0344] Bench Scale Two-Stage Process

[0345] The two-stage process consists of two sequential extractions performed on the same batch of Quillaja leaves. Stage 1, conducted at room temperature, primarily extracts saponins and phenolic compounds and yields purified products without the need for chlorophyll removal using bentonite. Stage 2, carried out under basic pH and elevated temperature, further disrupts the leaf microstructure to release additional saponins but simultaneously promotes the coextraction of chlorophyll and proteins, requiring bentonite treatment prior to ultrafiltration (UF) purification. Both stages produce more refined and higher-purity saponin extracts than those obtained through a single-stage batch extraction.

[0346] Stage 1 (Extraction at Room Temperature):

[0347] One kilogram of dried Quillaja leaves harvested from 2-year old shrubs grown in Ojai, California, previously exposed to sunlight for ten days and milled to a 1 mm particle size (RETSCH mill), was loaded into the extraction tank. The leaves were contacted with 10 L of water containing ascorbic acid (1 g / L) and EDTA (0.1 g / L) at room temperature (RT) and pH 4.5, with agitation for 2 hours. The extract was separated through a 150 pm sieve.

[0348] The resulting extract yielded 6.6% w / w saponins (based on dry leaf weight) and exhibited a CI of 0.03, making bentonite treatment unnecessary. The extract was filtered through 10 pm cellulose filters and subsequently purified by ultrafiltration (UF) using 20 kDa47 B 25 -074-2 WOmembranes. Diafiltration was performed with five volumes of water at pH 4.8 in the presence of CaCb (5 g / L).

[0349] Figure 23 shows the RP-HPLC chromatograms of the Stage 1 extract before and after UF. The initial chromatogram indicated a high content of phenolic compounds (early eluting peaks), which were effectively removed during UF in the presence of CaCF. Following diafiltration, the retentate was concentrated to approximately 30 g / L saponins, passed through an ion-exchange column to remove residual calcium, polished using 0.45 pm cellulose wine filters (Buon Vino, USA), and freeze-dried. The final Stage 1 product was a light cream-colored powder with a saponin purity of 91% w / w.

[0350] Stage 2 (Hot Alkaline Extraction):

[0351] The residual leaf cake from Stage 1 was reused in the same extraction tank.Approximately 8 L of hot water (65 °C) was added, and the tank temperature was maintained at 50 °C using a heating jacket. Agitation was applied at 50 rpm, and the pH was adjusted to 8.0-8.5 using 0.5 M NaOH. Extraction proceeded for 1.5 hours, after which the extract was separated through a 150 pm sieve.

[0352] This extract yielded 6 % w / w saponins (based on dry leaf weight) and had a CI of 0.23. To reduce color intensity, bentonite (3 g / L) was added, lowering the CI to 0.03 prior to downstream purification. The same DSP sequence as in Stage 1 was applied — UF with diafiltration (20 kDa membranes), ion-exchange polishing, and final fi ll rat ion (0.45 pm).

[0353] Figure 24 shows the RP-HPLC chromatograms for the Stage 2 extract, indicating reduced carly-cluting peaks compared to Stage 1. This confirmed that most phenolic compounds were extracted during the first stage and further reduced during UF in the presence of CaCL. The concentrated extract was freeze-dried, yielding a white powder with a saponin purity of 92% w / w. Figure 25 shows Stage- 1 extract leading to light-colored freeze-dried products with > 90% w / w saponin purity.

[0354] Pilot-Plant Two-Stage Extraction Process for Quillaja Leaf Saponins

[0355] The two-stage process consists of two sequential extractions performed on the same batch of Quillaja leaves. Stage 1, conducted at room temperature, primarily extracts saponins and phenolic compounds and yields purified products without the need for chlorophyll removal using bentonite. Stage 2, carried out under basic pH and elevated temperature, further disrupts the leaf microstructure to release additional saponins but simultaneously promotes the coextraction of chlorophyll and proteins, requiring bentonite treatment prior to ultrafiltration (UF) purification. Both stages produce more refined and higher-purity saponin extracts than those obtained through a single-stage batch extraction.

[0356] Stage 1 (Percolation at Room Temperature):48 B 25 -074-2 WO

[0357] Extraction is performed in a HO L cold-brew extractor tank equipped with a 150 pm sieve at the bottom. Ten kilograms of leaves, harvested from 2-year old shrubs grown in Ojai, California and previously exposed to sunlight for 10 days and milled to 1 mm (RETSCH mill), are loaded into the tank. Ambient-temperature water is distributed over the leaf bed through a shower head while the percolate is displaced through the sieve using a peristaltic pump (Watson-Marlow) at 1 L / min. The extract is filtered through a 10 pm lenticular filter (Ertel-Alsop M104) and collected in a nanofiltration tank. Percolation continues until approximately 12 bed volumes of water have passed through, yielding about 100 L of extract.

[0358] To prevent oxidation, ascorbic acid (1 g / L) and EDTA (0.1 g / L) are added immediately after collection. Nanofiltration is carried out at 1.6 gal / min and 70 psi for approximately 2 hours, producing about 70 L of permeate and increasing the soluble solids concentration from 30 g / L to 70 g / L. The concentrate is transferred to an ultrafiltration (UF) unit equipped with 20 kDa membranes (7.2 in2total area). Prior to UF, the pH is adjusted to 4.5-5.0 with 0.5 M NaOH, and ascorbic acid, EDTA, and CaCL (5 g / L) are added to minimize oxidation and disrupt phenolicsaponin micelles. Diafiltration is performed with four volumes of water at a flow rate of 15 L / min and 40 psi for about 3 hours, maintaining constant retentate volume. The extract is then concentrated to ~30 g / L saponins, passed through an ion-exchange column to remove calcium, polished using 0.45 pm cellulose wine filters (Buon Vino, USA), and freeze-dried to complete Stage 1.

[0359] Stage 2 (Hot Alkaline Extraction):

[0360] The same extracted leaf cake is reused in the same extractor. Approximately 40-60 L of hot water (65 °C) is added, and the tank temperature is maintained at 45-50 °C using a heating jacket. Agitation is applied at 50 rpm, and the pH is raised to 8.0-8.5 with 0.5 M NaOH.Extraction proceeds for 1.5 hours. The extract (~60 L) is then pumped into a holding tank, acidified to pH 3.8 with HC1, and treated with bentonite (2.5 g / L). The mixture is kept under refrigeration with agitation overnight to promote chlorophyll adsorption. After settling, the supernatant (-75-80% of total volume) is filtered through a 10 pm lenticular filter (Ertel-Alsop Ml 04) and transferred to the UF system.

[0361] The filtrate is adjusted to pH 4.5 and treated again with ascorbic acid, EDTA, and CaCh as in Stage 1. Diafiltration and concentration are performed under the same UF conditions (15 L / min, 40 psi, 2-3 hours) to reach -30 g / L saponins. The concentrated extract is then polished through an ion-exchange column and 0.45 pm cellulose filters (Buon Vino, USA), followed by freeze-drying to obtain the final Stage 2 product.

[0362] The CI obtained from Stage 1 was only 0.01, so no bentonite was added. Figure 26 shows the RP-HPLC chromatograms that Stage 1 extract had a high content of phenolic49 B 25 -074-2 WOcompounds (early part of the RP-HPLC chromatogram), but that these were effectively removed in the UF process using CaC12. The final product was a white powder. The saponin yield of Stage 1 was 4.3% w / w, based on the dry leaves. The contents of saponins fractions QH-A and QH-C were 14.1% w / w and 17.3% w / w (based on total saponins), respectively. The saponin content of the final powder was 91 % w / w.

[0363] As shown in Figure 27, the initial peak is lower for Stage 2 extraction, as most of the phenolics were extracted during Stage 1. The CI obtained from Stage 2 was 0.36, which dropped to 0.12 following treatment with 2.5 g / 1 of bentonite. Treatment with UF with CaC12 effectively removed most of the phenolic compounds. The saponin yield of Stage 1 was 4.5% w / w, based on the dry leaves. The contents of saponins fractions QH-A and QH-C were 16.6% w / w and 18.1% w / w (based on total saponins), respectively. The saponin content of the final powder was 90 % w / w.

[0364] Key properties of leaf extracts: foam and emulsifying

[0365] For food applications, foaming capacity is a key functional property of quillaja extracts. A simple comparative foam test, as described by (San Martin & Briones, 2000), was used to evaluate the samples. Extracts obtained from leaves consistently exhibited a foam index of 170 mL after 30 seconds and 160 mL after 30 minutes, comparable to the performance of commercial partially purified products, such as Andean QDP Ultra Organic (Desert King International, USA), at equivalent saponin concentrations (35 mg / L).

[0366] For applications in foods, beverages, and drug delivery systems, it is essential that the extracts effectively stabilize emulsions — particularly nanocmulsions with droplet sizes in the 20-200 nm range (Jaiswal et al., 2015). Figure 28 presents the dynamic light scattering (DES) analysis of a 10% oil-in-water emulsion stabilized with 0.5% quillaja saponins (equivalent to 50 mg saponin / g oil). The average droplet (Z-average) size of 188 nm is consistent with literature reports for emulsions stabilized by quillaja saponins from biomass (Chen et al., 2017; Yang et al., 2013), despite the use of lower saponin concentrations (0.5% vs. 1.0-1.5% w / w).

[0367] ENVIRONMENTAL IMPACT OF THE INVENTION

[0368] This invention significantly reduces the environmental strain caused by harvesting old, wild-grown trees in Chile, creating a more sustainable supply chain. Unlike traditional methods that require 10-15 years for plantation-grown biomass or over 30 years for wild-grown trees in Chile, this process utilizes plants after just two years of growth, in a renewable way, making the supply of quillaja saponins much more sustainable and more plentiful.

[0369] Additionally, the plant used is drought-resistant and well-suited to water-scarce regions. Its water requirements are approximately 1 acre-foot of water, far less than the 4 acre-feet of water needed for crops like almonds and avocados in California, making it an efficient50 B 25 -074-2 WOalternative for farmers in arid areas. This opens new business opportunities for farmers, while increasing the availability of a highly valued food ingredient.

[0370] The products derived from this process have desirable qualities for food and pharmaceutical applications, including a lighter color, milder taste, and active ingredients with reduced chemical variability, enhancing their appeal across a range of uses.

[0371] In addition, the resulting products have better properties than existing commercial products derived from quillaja biomass, exclusively produced in Chile.

[0372] Table 20 demonstrates the impact of this novel process on land use efficiency. Currently, over 20,000 tons of air-dried quillaja biomass is consumed annually, yielding approximately 1.4% w / w saponin, or around 280 tons of saponins per year. The table compares this demand with the land requirements for traditional quillaja biomass plantations (orchard-style) versus the leaf-based production method described in this invention.

[0373] In traditional biomass plantations, each cultivated tree is expected to yield about 35 kg of dry biomass after 10 years. Meeting annual demand would require harvesting approximately 460 hectares per year. This would mean planting 460 hectares annually over a 10-year cycle, totaling 4,600 hectares to achieve a sustainable biomass supply.

[0374] By contrast, leaf production involves planting shrubs in dense hedges (plants-50-80 cm apart), with around 5,000 trees per hectare. Assuming each shrub produces at least 5 kg of fresh biomass (or 2.5 kg dry leaves), each hectare is expected to yield 12 tons of dry leaves. With a conservative saponin yield of 6.5% w / w, meeting the full annual saponin demand would require less than 400 hectares — 11 times less land than traditional biomass plantations (4,600 hectares).

[0375] Table 20: Land usage of a traditional plantation to produce biomass and bark, vs hedgelike plantation for the exploitation of quillaja leaves. Base case: demand of 280 tons of saponin / yearDemand tons saponin / yearTons dry biomass 20,000% yield saponins / dry biomass 1.40%Tons saponin / year 280Plantations for saponins from biomassTrees / ha 1,250Kg dry biomass / tree, @ 10 y 35Tons biomass / ha, @ 10 y 44Ha needed, @ 10 y 45751 B 25 -074-2 WOTotal ha needed, 10 y rotation 4,571Plantation for saponins from leavesShrubs / ha 4762Kg dry leaves / shrub-year 2.5Tons leaves / ha-year 12% yield saponins / dry leaves 6.5%Tons saponins / hectare 0.77Ha harvested / year 362

[0376] REFERENCES

[0377] Avila, G., Aljaro, M. E., Araya, S., Montenegro, G., & Kummerow, J. (1975). The seasonal cambium activity of Chilean and Californian shrubs. American Journal of Botany, 62(5), 473. https: / / doi.org / 10.2307 / 2441954

[0378] Bijlsma, J., de Bruijn, W. J. C., Velikov, K. P., & Vincken, J.-P. (2022). Unravelling discolouration caused by iron-flavonoid interactions: Complexation, oxidation, and formation of networks. Food Chemistry, 370(131292), 131292.

[0379] Buchner, N., Krumbein, A., Rohn, S., & Kroh, L. W. (2006). Effect of thermal processing on the flavonols rutin and quercetin. Rapid Communications in Mass Spectrometry: RCM, 20(21), 3229-3235. https: / / doi.org / 10.1002 / rcm.2720

[0380] Chen, X.-W., Wang, J.-M., Guo, J., Wan, Z.-L„ Yin, S.-W.. & Yang, X.-Q. (2017). Hierarchical high internal phase emulsions and transparent oleogels stabilized by quillaja saponin-coated nanodroplets for color performance. Food & Function, 8(2), 823-831. https: / / doi.org / 10.1039 / c6fo01752e

[0381] Cleland, J. L., Kensil, C. R., Lim, A., Jacobsen, N. E., Basa, L., Spellman, M., Wheeler, D. A., Wu, J. Y., & Powell, M. F. (1996). Isomerization and formulation stability of the vaccine adjuvant QS-21. Journal of Pharmaceutical Sciences, 85(1), 22-28. https: / / doi.org / 10.1021 / js9503136

[0382] De Leonardis, A., Macciola, V., Cuomo, F., & Lopez, F. (2015). Evidence of oleuropein degradation by olive leaf protein extract. Food Chemistry, 175, 568-574.https: / / doi.org / 10.1016 / j.foodchem.2014.12.016

[0383] Donoso, S., Pena, K., Pacheco, C., Luna, G., & Aguirre, A. (2011). Respuesta fisiologica y de crecimiento en plantas de Quillaja saponaria y Cryptocarya alba sometidas a restriccion hidrica. Bosque (Valdivia), 32(2), 187-195. https: / / doi.org / 10.4067 / s0717-92002011000200009

[0384] Fang, Y„ & Xiao, H. (2021). The transport of triterpenoids. Biotechnology Notes (Amsterdam, Netherlands), 2, 11-17. https: / / doi.org / 10.1016 / j.biotno.2021.03.00152 B 25 -074-2 WO

[0385] Friedman, M., & Jurgens, H. S. (2000). Effect of pH on the stability of plant phenolic compounds. Journal of Agricultural and Food Chemistry, 48(6), 2101-2110. https: / / doi.org / 10.1021 / jf990489j

[0386] Girr, P., & Paulsen, H. (2021). How water-soluble chlorophyll protein extracts chlorophyll from membranes. Biochimica et Biophysica Acta. Biomembranes, 1863(C), 183479. https: / / doi.org / 10.1016 / j.bbamem.2020.183479

[0387] Imuetinyan, H., Agi, A., Gbadamosi, A., & Junin, R. (2022). Extraction, characterization and evaluation of saponin-based natural surfactant for enhanced oil recovery. Arabian Journal of Geosciences. 15(3). https: / / doi.org / 10.1007 / sl2517-021-09275-6

[0388] Indrasti, D., Andarwulan, N., Purnomo, E. H., & Wulandari, N. (2018). Stability of chlorophyll as natural colorant: A review for Suji (Dracaena angustifolia (Medik.) Roxb.) leaves’ case. Current Research in Nutrition and Food Science Journal, 6(3), 609-625. https: / / doi. org / 10.12944 / c mfsj.6.3.04

[0389] Jaiswal, M., Dudhe, R., & Sharma, P. K. (2015). Nanoemulsion: an advanced mode of drug delivery system. 3 Biotech, 5(2), 123-127. https: / / doi.org / 10.1007 / sl3205-014-0214-0

[0390] Kaczorek, E., Smulek, W., Zdarta, A., Sawczuk, A., & Zgola-Grzeskowiak, A. (2016). Influence of saponins on the biodegradation of halogenated phenols. Ecotoxicology and Environmental Safety, 131, 127-134. https: / / doi. Org / 10.1016 / j.ecoenv.2016.05.015

[0391] Kamstrup, S., San Martin, R., Doberti, A., Grande, H., & Dalsgaard, K. (2000).Preparation and characterisation of quillaja saponin with less heterogeneity than Quil-A.Vaccine, 78(21), 2244-2249. https: / / doi.org / 10.1016 / s0264-410x(99)00560-5

[0392] Kensil, C. R., Patel, U., Lennick, M., & Marciani, D. (1991). Separation and characterization of saponins with adjuvant activity from Quillaja saponaria Molina cortex. The Journal of Immunology, 146(2), 431-437. https: / / doi. Org / 10.4049 / jimmunol.146.2.431

[0393] Kite, G. C., Howes, M.-J. R., & Simmonds, M. S. J. (2004). Metabolomic analysis of saponins in crude extracts of Quillaja saponaria by liquid chromatography / mass spectrometry for product authentication. Rapid Communications in Mass Spectrometry': RCM, 78(23), 2859-2870. https: / / doi.org / l 0.1002 / rcm.1698

[0394] Magedans, Y. V., Yendo, A. C., Costa, F. de, Gosmann, G., & Fett-Neto, A. G. (2019). Foamy matters: an update on Quillaja saponins and their use as immunoadjuvants. Future Medicinal Chemistry, 77(12), 1485-1499. https: / / doi.org / 10.4155 / fmc-2018-0438

[0395] Maier, C., Conrad, J., Carle, R., Weiss, J., & Schweiggert, R. M. (2015). Phenolic constituents in commercial aqueous Quillaja (Quillaja saponaria Molina) wood extracts. Journal of Agricultural and Food Chemistry, 63(6), 1756-1762. https: / / doi.org / 10.1021 / jf506277p53 B 25 -074-2 WO

[0396] Maier, C., Conrad, J., Steingass, C. B., Beifuss, U., Carle, R., & Schweiggert, R. M. (2015). Quillajasides A and B: New phenylpropanoid sucrose esters from the inner bark of Quillaja saponaria Molina. Journal of Agricultural and Food Chemistry', 55(40), 8905-8911. https: / / doi.org / 10.1021 / acs.jafc.5b03532

[0397] Makris, D. P., & Rossiter, J. T. (2000). Heat-induced, metal-catalyzed oxidative degradation of quercetin and rutin (Quercetin 3-O-rhamnosylglucoside) in aqueous model systems. Journal of Agricultural and Food Chemistry, 48(9), 3830-3838.https: / / doi. org / 10.1021 / j fOOO 1280

[0398] Method for producing saponins from plants. (2022). US Patent, 11, 954-B956.

[0399] Mironenko, N. V., & Selemenev, V. F. (2018). Sorption interactions in an surfactant-triterpenoid saponin-activated carbon system. Russian Journal of Physical Chemistry A, 92(11), 2296-2301. https: / / doi.org / 10.1134 / s0036024418110274

[0400] Mironenko, N. V., Selemenev, V. F„ Shkutina, I. V., & Ishchenko, U. S. (2024).Sorption of triterpene glycoside on different brands of active carbon. Russian Journal of Physical Chemistry A, 98(10), 2324-2332. https: / / doi.org / 10.1134 / s0036024424701486

[0401] Mitra, S., & Dungan, S. R. (1997). Micellar properties of Quillaja saponin. 1. Effects of temperature, salt, and pH on solution properties. Journal of Agricultural and Food Chemistry, 45(5), 1587-1595. https: / / doi.org / 10.1021 / jf960349z

[0402] Mokaya, R., Jones, W., Davies, M. E., & Whittle, M. E. (1994). The mechanism of chlorophyll adsorption on acid-activated clays. Journal of Solid State Chemistry, 111(1), 157-163. https: / / doi.org / 10.1006 / jssc.1994.1212

[0403] Nord, L. I., & Kenne, L. (1999). Separation and structural analysis of saponins in a bark extract from Quillaja saponaria Molina. Carbohydrate Research, 320(1-2), 70-81. https: / / doi.org / 10.1016 / s0008-6215(99)00134-2

[0404] Padilla Iglesias, L., Gonzalez Castro, J., Aranguiz Arze-Vargas, V., & Otero Peredo, R. (2021). Production of biomass in ultra-high-density plantations. US11,254,699

[0405] Padilla Iglesias, L„ & Valencia Michaud, A. (2022). Method for obtaining saponins from plants. US 11,510,954

[0406] Panda, P., Appalashetti, M., & Judeh, Z. M. A. (2011). Phenylpropanoid sucrose esters: plant-derived natural products as potential leads for new therapeutics. Current Medicinal Chemistry, 18(21), 3234-3251. https: / / www.ncbi.nlm.nih.gov / pubmed / 21671860

[0407] Reichert, C. L., Salminen, H., & Weiss, J. (2019). Quillaja saponin characteristics and functional properties. Annual Review of Food Science and Technology, 10(1), 43-73. https: / / doi. org / 10.1146 / annurev-food-032818-12201054 B 25 -074-2 WO

[0408] San Martin, R., & Briones, R. (2000). Quality control of commercial quillaja (Quillaja saponaria Molina) extracts by reverse phase HPLC. Journal of the Science of Food and Agriculture, 80(14), 2063-2068. https: / / doi.org / 10.1002 / 1097-0010(200011)80:14<2063::aid-jsfa750>3.3.co;2-u

[0409] Schefer, S., Oest, M., & Rohn, S. (2021). Interactions between phenolic acids, proteins, and carbohydrates-influence on dough and bread properties. Foods (Basel, Switzerland), 10( 11), 2798. https: / / doi.org / 10.3390 / foodsl0112798

[0410] Schlotterbeck, T., Castillo-Ruiz, M., Cañon-Jones, H., & Martin, R. S. (2015). The use of leaves from young trees of Quillaja saponaria (Molina) plantations as a new source of saponins. Economic Botany, 69(3), 262-272. https: / / doi.org / 10.1007 / sl2231-015-9320-0

[0411] Tamayo Tenorio, A., Boom, R. M., & van der Goot, A. J. (2017). Understanding leaf membrane protein extraction to develop a food-grade process. Food Chemistry, 217, 234-243. https: / / doi. org / 10.1016 / j.foodchem.2016.08.093

[0412] Tenorio, A. T., de Jong, E. W. M., Nikiforidis, C. V., Boom, R. M., & van der Goot, A. J. (2017). Interfacial properties and emulsification performance of thylakoid membrane fragments. Soft Matter, 13(3), 608-618. https: / / doi.org / 10.1039 / c6sm02195f

[0413] Tippel, J., Gies, K., Harbaum-Piayda, B., Steffen-Heins, A., & Drusch, S. (2017).Composition of Quillaja saponin extract affects lipid oxidation in oil-in-water emulsions. Food Chemistry, 221, 386-394. https: / / doi. Org / 10.1016 / j.foodchem.2016.10.055

[0414] Tzima, K., Brunton, N. P., & Rai, D. K. (2020). Evaluation of the impact of chlorophyll removal techniques on polyphenols in rosemary and thyme by-products. Journal of Food Biochemistry, 44(3), el3148. https: / / doi.org / 10.1111 / jfbc.13148

[0415] Wallace, F., Bennadji, Z., Ferreira, F„ & Olivaro, C. (2017). Analysis of an immunoadjuvant saponin fraction from Quillaja brasiliensis leaves by electrospray ionization ion trap multiple-stage mass spectrometry. Phytochemistry Letters, 20, 228-233.https: / / doi. org / 10.1016 / j.phytol.2017.04.020

[0416] Yang, Y„ Leser, M. E„ Sher, A. A., & McClements, D. J. (2013). Formation and stability of emulsions using a natural small molecule surfactant: Quillaja saponin (Q-Naturale®). Food Hydrocolloids, 30(2), 589-596. https: / / doi. Org / 10.1016 / j.foodhyd.2012.08.008

[0417] Yasuda, M., & Tabata, M. (2021). Effect of emulsifiers on the discoloration of chlorophyll and their potential for use in green beverages. Journal of Food Science, 86(7), 3033-3045. https: / / doi.org / 10.1111 / 1750-3841.15782

[0418] Zhang, C„ Sanders, J. P. M„ Xiao, T. T„ & Brains, M. E. (2015). How does alkali aid protein extraction in green tea leaf residue: A basis for integrated biorefinery of leaves. PloS One, 10(7). e0133046. https: / / doi.org / 10.1371 / journal.pone.013304655 B 25 -074-2 WO

[0419] ENUMERATED INVENTIONS

[0420] Overarching Inventions

[0421] 1) Innovative extraction and downstream purification processes to obtain saponins from Quillaja saponaria leaves, as well as all the aerial biomass that generates when harvesting leaves, with high content of saponins, suitable for all applications including food emulsifier and foaming agent, feed additive, vaccine adjuvants, and nanoemulsions for drug delivery.

[0422] EXTRACTION INVENTIONS

[0423] 2) Use of water, water-ethanol, or water-methanol as solvents to extract fresh, dried, or dried photobleached milled leaves at a solvent to dry leaves mass ratio of 5-20.

[0424] 3) Preferential use of water as the extraction solvent with additives including ascorbic acid and EDTA to reduce oxidization, as well as magnesium sulfate, to minimize the loss of magnesium ion in chlorophyll, thereby minimizing its degradation to colored compounds such as pheophytin.

[0425] 4) Maximization of saponin yields by extracting at room temperature and pH 6.5 or greater.

[0426] 5) Maximization of saponin yields by extracting at 50-60°C and basic pH (8-8.5) to disrupt the microstructure of the aerial biomass.

[0427] 6) Extraction with aqueous buffer at pH 6.5-8, room temperature, and prolonged extraction times of 0.5-16 hours, rendering a process that generates minimum waste streams.

[0428] 7) Maximization of saponin extraction yields with lower co-extraction of non-saponin compounds by using a phosphate buffer at an optimum concentration of 80-120 mM, pH 6.5-8.

[0429] 8) Maximization of saponin extraction yields and lowering of extraction times when using buffers by initially extracting at 45-55°C for 10-30 minutes, followed by extraction at room temperature for 0.5-12 hours.

[0430] 9) Low-cost phosphate / NaCl buffer formulation, industrially scalable substitute for 80-120 mM phosphate buffer.

[0431] 10) Three scalable extraction and downstream purification processes — batch, continuous, and two-stage — all of which produce refined, semi-refined, and highly refined products, with saponin contents based on extracted solubles of 20-30% w / w, 75-85% w / w, and >90% w / w, respectively.

[0432] 11) Two-stage extraction process to minimize the simultaneous degradation of colored compounds such as phenolics and chlorophyll, wherein Stage 1 comprises percolation or batch extraction at room temperature and acid pH that preferentially solubilizes saponin and phenolics with minimum chlorophyll co-extraction, followed by Stage 2 that comprises hot alkaline extraction of the same leaf batch to disrupt the material microstructure and solubilize saponins 56 B 25 -074-2 WOand water-soluble protein-chlorophyll complexes (WSPCs) with minimal phenolic degradation. The extracts of both stages render light-colored products and can be purified to high saponin levels >90% w / w.

[0433] 12) Fraction enrichment in the extracts for QH-A, the fraction rich in less-toxic saponin QS-7, by extracting at basic pH (8-8.5) and 20-60°C.

[0434] 13) Photobleached- leaf pre -treatment with sunlight exposure (3-10 days) or artificial light to reduce chlorophyll before extraction.

[0435] 14) Use of photobleached leaves with any of the three extraction processes to produce light-colored extracts with minimal chlorophyll content.

[0436] DOWNSTREAM PURIFICATION (DSP) INVENTIONS— Removal of Non-Saponin Compounds (primarily colored compounds: chlorophyll and phenolics and their degradation products)

[0437] 15) Innovative downstream purification processes for all three extraction processes to reduce non-saponin compounds such as chlorophyll, phenolics, sugars and salts that are coextracted with saponins and to produce semi-refined, refined, and highly refined products either as liquid concentrates or spray-dried products.

[0438] 16) Downstream purification process comprising initial water removal via nanofiltration, reverse osmosis, or ultrafiltration, followed by removal of non-saponin compounds through sequential chlorophyll and phenolic reduction, followed by pasteurization of final products via ultra-high short-time (UHST) heat treatment or 0.2 pm filters prior to spray drying or packaging of liquid concentrates.

[0439] 17) Chlorophyll Index (CI) method as a rapid, normalized color metric for process control, with empirical determination that CI must be reduced to <0.25 prior to downstream purification to obtain light-colored products.

[0440] 18) Reduction of chlorophyll content by reducing pH to <4 to promote the precipitation of water-soluble protein-chlorophyll complexes (WSPCs) and thus effectively reducing the chlorophyll content of the extracts.

[0441] 19) Treatment with sodium bentonite (1-5 g / L) under acidic pH (3.5-4.5) to adsorb water-soluble protein-chlorophyll complexes (WSPCs), thus effectively reducing the chlorophyll content of the extracts.

[0442] 20) Reduction or full removal of low molecular weight compounds using diafiltration with hydrophilic ultrafiltration membranes with molecular weight cutoffs within 10-100 kDalton, preferentially 20-30 kDalton.

[0443] 21) Diafiltration with hydrophilic ultrafiltration membranes using cation-assisted disruption of phenolic-saponin micelles with Mg2+, Ca2+, or Zn2+cations, or operating at pH 7.8- 57 B 25 -074-2 WO8.2 to disrupt phenolic-saponin micelle interactions, thereby facilitating the removal of phenolic compounds.

[0444] PRODUCT INVENTIONS

[0445] 22) Final products derived from leaves with higher saponin levels than existing products derived from bark and biomass for non-pharmacological applications (80-85% w / w versus 60-65% w / w saponin content).

[0446] 23) Final products derived from leaves with comparable saponin profile to existing products derived from bark for pharmacological applications, with >90% w / w saponin concentration.

[0447] 24) Use of phenolic compounds to detect patent infringement: Quillaja leaves contain unique phenolics not found in Quillaja bark, including quercetin derivatives and oleuropein, which serve as compositional "markers" unique to leaf-derived extracts.

[0448] ASPECTS

[0449] Innovative processes to produce semi refined, refined and highly refined saponins from quillaja leaves, using quillaja plant material that is primarily leaves, e.g. greater than 80, 90, 95 or 99% w / w leaves.

[0450] Three scalable methods are detailed:!) Batch extraction, 2) Continuous extraction, and 3) Two-stage extraction / purification, with sequential removal of phenolics (stage 1) and chlorophyll (stage 2), to produce light-colored final extracts with minimal saponin losses and minimal color degradation products.

[0451] Significant enhancement of saponin recovery yields operating at basic pH and high temperatures.

[0452] Use of magnesium ions to minimize chlorophyll degradation when extracting at basic pH and high temperature.

[0453] Use of pH shifts, activated charcoal and bentonites to reduce chlorophyll content in the extracts.

[0454] Use of diafiltration with ultrafiltration membranes in the presence of divalent ions or basic pH to disrupt the bonds between saponins and phenolics, effectively removing phenolics via diafiltration with water.

[0455] Use of phenolics to detect patent infringement: quillaja leaves have unique phenolics that are not found in quillaja bark, e.g., quercetin (and derivatives), and oleuropein. These compounds can serve as “markers” unique to leaf derived extracts58 B 25 -074-2 WO

Claims

CLAIMS1. A method for producing saponins, comprisinga) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaja leaves;b) treating the quillaja leaves with an aqueous solution comprising an antioxidant and a chelating agent at a pH between 7.5 and 8.5 and a temperature between 45-65 °C;c) separating the treated quillaja leaves after step b) to obtain an extract;d) contacting the extract from step c) with a nutritionally-inert sorbant, such as calcium bentonite, sodium bentonite or activated charcoal;e) filtering the extract from step d) in the presence of divalent cations or at a pH between 7.5 and 8 to yield a purified saponin composition.

2. The method of claim 1, to produce semi refined, refined and highly refined saponins from leaves.

3. The method of claim 1, comprising a) batch extraction, b) continuous extraction, or c) two-stage extraction / purification, with sequential removal of phenolics (stage 1) and chlorophyll (stage 2), to produce light-colored final extracts with minimal saponin losses and minimal color degradation products.

4. The method of claim 1, providing significant enhancement of saponin recovery yields operating at basic pH and high temperatures.

5. The method of claim 1, comprising use of magnesium ions to minimize chlorophyll degradation when extracting at basic pH and high temperature.

6. The method of claim 1, comprising use of pH shifts, activated charcoal and bentonites to reduce chlorophyll content in the extract.

7. The method of claim 1, comprising use of diafiltration with ultrafiltration membranes in the presence of divalent ions or the use of basic pH to disrupt the bonds between saponins and phenolics, effectively removing phenolics via diafiltration with water.

8. A method of determining the source of a saponin composition, the method comprising detecting phenolic markers in the composition, wherein quillaja leaves have unique phenolic 59 B 25 -074-2 WOcompounds that are not found in quillaja bark, e.g., quercetin (and derivatives), and oleuropein, wherein these compounds serve as markers specific to leaf derived extracts.

9. A method for producing quillaja saponins from leaves, comprising processing steps to produce semi-refined quillaja extracts from leaves, substantially as shown in Fig. 3.

10. A method for producing quillaja saponins from leaves, comprising processing steps to produce refined quillaja extracts from leaves, substantially as shown in Fig. 4.

11. A method for producing quillaja saponins from leaves, comprising an extraction / purification process flowsheet to produce refined quillaja extracts using buffers, substantially as shown in Fig. 5.

12. A method for producing quillaja saponins from leaves, comprising processing steps to produce refined quillaja extracts from leaves using a continuous process, substantially as shown in Fig. 12.

13. A method for producing quillaja saponins from leaves, comprising a two-stage extraction / purification process, substantially as shown in Fig. 14.

14. A method for producing saponins, comprising:a) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaja leaves:b) treating the quillaja leaves with an aqueous solution comprising an antioxidant and a chelating agent at a pH between 7.5 and 8.5 and a temperature between 45-65 °C;c) separating the treated quillaja leaves after step b) to obtain an extract;d) contacting the extract from step c) with a nutritionally-inert sorbant, such as calcium bentonite, sodium bentonite or activated charcoal;e) filtering the extract from step d) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a purified saponin composition.

15. The method of claim 14, wherein step b) is performed for a duration between 1 hour and 3 hours.60 B 25 -074-2 WO16. The method of claim 14 or 15, further comprising:washing the separated leaves from step c) with water to obtain a wash extract;combining the wash extract with the extract in step c); andcontaining the combined extract with the bentonite in step d).

17. A method for producing saponins, comprising:a) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaja leaves;b) treating the quillaja leaves for 1-3 h with an aqueous solution comprising an antioxidant and a chelating agent at a pH between 4 and 5.5 and at room temperature;c) separating the treated quillaja leaves after step b) to obtain a first extract;d) filtering the first extract from step c) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a first purified saponin composition;e) subjecting the separated, treated quillaja leaves from step c) to a hot alkaline extraction at a pH between 7.5 and 8.5 and at a temperature between 45 °C and 65 °C to obtain a second extract;f) contacting the second extract from step e) with a nutritionally-inert sorbant, such as calcium bentonite, sodium bentonite, or activated charcoal;g) filtering the second extract from step f) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a second purified saponin composition.

18. The method of claim 17, wherein step e) is performed for a duration between 0.5 hours and 3 hours.

19. The method of claim 17 or 18, further comprising:combining the first purified saponin composition and the second purified saponin composition.

20. A method for producing saponins, comprising:a) providing a composition that is mostly (preferably at least 80 or 90 or 95 or 99% wt / wt) quillaja leaves;b) treating the quillaja leaves with an aqueous solution comprising an antioxidant and a chelating agent and buffer at a pH between 6.5 and 8 and at a temperature between 15 °C and 30 °C;61 B 25 -074-2 WOc) separating the treated quillaja leaves after step b) to obtain an extract;d) filtering the extract from step c) using an ultrafiltration membrane having a molecular weight cut off between 10 kDa and 100 kDa in the presence of divalent cations or at a pH between 7.5 and 8 to yield a purified saponin composition.

21. The method of claim 20, wherein the buffer comprises phosphate ions.

22. The method of claim 20 or 21, wherein the buffer can be used alone or in combination with sodium chloride.

23. The method of any one of claims 20 to 22, wherein step b) is performed for a duration between 0.5 hours and 16 hours.

24. The method of any one of claims 20 to 23, further comprising:drying the purified saponin composition to obtain saponin-enriched powder.

25. The method of any one of claims 14 to 24, wherein the quillaja leaves are fresh, dried, partially dried, intact, chopped, shredded, ground, milled, or photobleached.

26. The method of any one of claims 14 to 25, wherein the antioxidant is a natural antioxidant.

27. The method of any one of claims 14 to 25, wherein the antioxidant comprises ascorbic acid, cysteine or citric acid, or any combination thereof.

28. The method of any one of claims 14 to 27, wherein the chelating agent comprises ethylenediaminetetraacetic acid (EDTA), citric acid or phosphoric acid, or any combination thereof.

29. The method of any one of claims 14 to 28, wherein the divalent cations are calcium, zinc or magnesium cations.

30. The method of any one of claims 14 to 29, wherein the method is performed as a batch process.62 B 25 -074-2 WO31. The method of any one of claims 14 to 29, wherein the method is performed as a continuous process.

32. The method of any one of claims 14 to 31, wherein the chlorophyll index of each extract obtained in the method is below 0.25.

33. A purified saponin composition obtained according to a method of any one of claims 1 to 32.

34. A purified saponin composition, having a saponin content of at least 80% w / w, and comprising detectable amounts of phenolic markers present in quillaja leaves from which the purified saponin composition is obtained, wherein the total phenolic markers present in the composition is no more than about 5 % w / w.

35. The composition of claim 34, wherein the ratio of phenolic markers to saponins is between 0.01 and 0.15 w / w.

36. A purified saponin composition comprising saponins and phenolic markers, wherein the ratio of phenolic markers to saponins is between 0.01 and 0.15 w / w.

37. The composition of any one of claims 34 to 36, wherein the phenolic markers comprise: quercetin pcntosidc; dihcxosylqucrcctin; or oleuropein, or any combination thereof.

38. The composition of any one of claims 34 to 36, comprising: between 0.5 and 2 % w / w quercetin pentoside; between 0.5 and 2 % w / w dihexosylquercetin; or between 0.1 and 1 % w / w oleuropein, or any combination thereof.

39. The composition of any one of claims 34 to 38, wherein the composition is in powder form.

40. The composition of claim 39, wherein the composition has a Hunter Lab Lightness (L) value of at least about 75.0.63 B 25 -074-2 WO

Images