Factors Causing Compositional Changes in Soy ... - ACS Publications

Subscriber access provided by Colorado State University | Libraries

Review

Factors Causing Compositional Changes in Soy Protein Hydrolysates and the Effects on Cell Culture Functionality Abhishek J. Gupta, Harry Gruppen, Dominick Maes, Jan-Willem Boots, and Peter A. Wierenga J. Agric. Food Chem., Just Accepted Manuscript • Publication Date (Web): 12 Oct 2013 Downloaded from http://pubs.acs.org on October 13, 2013

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 51

Journal of Agricultural and Food Chemistry

1

Factors Causing Compositional Changes in Soy Protein Hydrolysates and the

2

Effects on Cell Culture Functionality

3

Abhishek J. Gupta

4

Wierenga 1*

1, 2

, Harry Gruppen 1, Dominick Maes 2, Jan-Willem Boots 2, Peter A.

5

6

1

7

The Netherlands

8

2

Wageningen University, Laboratory of Food Chemistry, P.O. Box 17, 6700 AA Wageningen,

FrieslandCampina Domo, P.O. Box 1551, 3800 BN Amersfoort, The Netherlands

9 10

11

12

13

14

* Corresponding author:

15

Dr. ir. Peter A. Wierenga

16

Wageningen University, Laboratory of Food Chemistry, P.O. Box 17, 6700 AA Wageningen,

17

The Netherlands

18

Email: [email protected]; Tel: +31 317 483 786

19 1

ACS Paragon Plus Environment


Page 2 of 51

20

Abstract

21

Soy protein hydrolysates significantly enhance cell growth and recombinant protein production in

22

cell cultures. The extent of this enhancement in cell growth and IgG production is known to vary

23

from batch-to-batch. This can be due to differences in the abundance of different classes of

24

compounds (e.g. peptide content), the quality of these compounds (e.g. glycated peptides), or the

25

presence of specific compounds (e.g. furosine). These quantitative and qualitative differences

26

between batches of hydrolysates result from variation in the seed composition and seed/ meal

27

processing. Although a considerable amount of literature is available which describes these

28

factors, this knowledge has not been combined in an overview yet. The aim of this review is to

29

identify the most dominant factors that affect hydrolysate composition and functionality. While

30

there is a limited influence of variation in the seed composition, the overview shows that the

31

qualitative changes in hydrolysate composition result in the formation of minor compounds (e.g.

32

Maillard reaction products). In pure systems, these compounds have a profound effect on the cell

33

culture functionality. This suggests that presence of these compounds in soy protein hydrolysates

34

may affect hydrolysate functionality as well. This influence on the functionality can be of direct

35

or indirect nature. For instance, some minor compounds (e.g. maillard reaction products) are

36

cytotoxic; while other compounds (e.g. phytates) suppress protein hydrolysis during hydrolysate

37

production resulting in altered peptide composition, and thus, affect the functionality.

38

Keywords

39

Meal Processing. Soybeans.

Enzymatic hydrolysis. Compositional changes. Maillard reaction. Racemization.

2


Page 3 of 51


40

Introduction

41

The performance of a mammalian cell culture system is primarily determined by the product

42

yield (production of recombinant proteins, such as immunoglobulins (IgG) and interferon γ) and

43

product quality (glycosylation of recombinant proteins).1 In many cell culture studies, viable cell

44

growth is used as a second parameter to describe the cell culture performance.1, 2 This is because

45

the recombinant protein production can be affected by cell growth. In some studies, recombinant

46

protein production has been reported to increase with an increase in cell growth,3 while in other

47

studies, recombinant protein production increased following a suppression in the cell growth.4

48

To improve the performance of mammalian cell culture processes, fetal calf serum has been used

49

as a supplement for several decades. This supplement provides growth factors, proteins, lipids,

50

attachment factors, minerals, hormones, and several trace elements that are important for

51

promoting cell growth and enhancing recombinant protein production.1,

52

fetal calf serum has become restricted due to the risk of transmissible diseases like bovine

53

spongiform encephalopathy. As a result, substantial research has been performed to identify

54

alternatives for fetal calf serum. Currently, two alternatives have been described in literature: 1)

55

to use chemically defined medium,6-9 or 2) Supplement the basal chemically defined medium

56

with plant protein hydrolysates.10-13 The latter approach has become a common practice in the

57

biopharma industry. In table 1, an overview of the influence of soy protein hydrolysate

58

supplementation to chemically defined medium on cell growth and recombinant protein

59

production is provided. In this overview, the cell growth and recombinant protein production for

60

chemically defined medium is set to 100%. The cell growth and IgG production in hydrolysate-

61

supplemented chemically defined medium ranged from 90-178% and 95-300%, respectively as

62

compared to that in the chemically defined medium (100%) (Table 1).

5

However, the use of

3



Page 4 of 51

63

Commercially, plant protein hydrolysates from several sources such as soy, cotton, rice, wheat

64

gluten, pea, and rapeseed proteins are available for cell culture applications.14-16 The cell growth

65

of the Chinese hamster ovary (CHO, CHO DG44/ dhfr-/-) cells cultivated in HyQ® CDM4-CHO

66

medium supplemented with 16 hydrolysates obtained from different sources varied from 106% to

67

144% relative to that in the chemically defined medium (100%).2 Although, the enhancement in

68

cell growth in culture supplemented with soy protein hydrolysate was the highest, the reason why

69

soy protein hydrolysates performed better than other hydrolysates were not explained.

70

Several cell culture studies have been performed, in which various cell lines (e.g. CHO 320, ME

71

750, and WuT3),11,

72

IMDM:Ham’s F12:NCTC, and 2:1:1 (v/v) of DMEM:F12:RPMI)7, 9, 10 supplemented with soy

73

protein hydrolysates10,

74

comparison of results obtained in these studies is limited, because these chemically defined

75

media and hydrolysates have different chemical compositions and different cell lines have

76

diverse nutritional requirements.1 This comparison is further complicated by the fact that

77

although, the chemically defined media to which hydrolysates are supplemented are called

78

‘chemically defined’, an exact definition, or composition of such a medium is not reported in the

79

scientific literature. This lack of comparability limits the understanding of the role of soy protein

80

hydrolysates in enhancing cell culture functionality. Moreover, this results in wide variability in

81

the functionality of soy protein hydrolysates in cell culture (Table 1). For several studies, while

82

the cell growth was low in hydrolysate supplemented cultures as compared to the chemically

83

defined media, the recombinant protein production was always higher in the former than the

84

latter. The beneficial effects of soy protein hydrolysate are attributed to its complex composition,

85

wherein it contains a large variety of different classes of compounds (e.g. peptides and

12, 17, 18

chemically defined media (e.g. RPMI 1640, 5:5:1 (v/v) of

12, 13, 17

produced using different processes have been tested. The

4


Page 5 of 51


86

carbohydrates). Typically, soy protein hydrolysates (60% peptides/ amino acids and 20%

87

carbohydrates)17 are supplemented to chemically defined medium at 0.1-1.0% (w/v). A typical

88

chemically defined medium like Iscove’s modified Dulbecco’s medium has an amino acid and

89

glucose content of 1.2 g/ L and 4.5 g/ L, respectively.19 The hydrolysate supplementation to

90

chemically defined medium significantly increases the protein content (50-500%) and

91

carbohydrate content (4-44%) of the supplemented medium. As a result, these compounds

92

considerably enhance the cell culture functionality of chemically defined medium. Firstly, these

93

compounds provide carbon and nitrogen to cells resulting in enhanced cell growth and/ or

94

recombinant protein production;16, 20 Secondly, there might be certain specific (key) compounds

95

in hydrolysates that specifically enhance cell growth, recombinant protein production, or both.13,

96

17

97

annotated peptide) were identified as key compounds in soy protein hydrolysates that correlated

98

positively with CHO (CRL-11397) cell growth.13 Lactate, ferulate, syringic acid, galactarate,

99

adenine, and X-198 (an un-annotated peptide) were shown to be the key compounds that

In previous work, phenyllactate, lactate, trigonelline, chiro-inositol, and X-190 (an un-

100

correlated positively with IgG production.13,

21, 22

101

concentration of the key compounds was linked to the batch-to-batch variability in the cell

102

culture performance of soy protein hydrolysates. A comparison of supplementation of thirty

103

batches of soy protein hydrolysate to chemically defined media to culture CHO (CRL-11397)

104

cells showed that the cell growth and IgG production varied from 148-438% and 117-283%

105

relative to the chemically defined medium (100%), respectively.13 For 29 batches of another soy

106

protein hydrolysate, the recombinant protein production ranged from 52-164% of the average

107

recombinant protein production of all hydrolysates.23

In these studies, the variation in the

5



Page 6 of 51

108

The extent of variability in the cell culture performance between different batches of a

109

hydrolysate also depends on the cell lines used and the test concentration of hydrolysates. The

110

IgG production in CHO cells ranged from 103-138% and that in avian cells ranged from 93-

111

115% as measured for 16 batches of a soy protein hydrolysate.24 When the same set of

112

hydrolysates were supplemented at 0.5% (w/v) to CHO cells, the IgG production ranged from

113

103-138%, while it was 83-138% when hydrolysates were supplemented at 1.5% (w/v)

114

concentration, respectively.24

115

In addition to cell line and hydrolysate concentration, batch-to-batch variation in the hydrolysate

116

functionality is influenced by variability in the gross composition and qualitative changes that

117

occur in the chemical composition of hydrolysates. The gross composition includes the total

118

content and composition of each class of compound present in the hydrolysate, for instance, total

119

peptide content and content of individual peptides. The qualitative changes refer to the chemical

120

modification reactions (e.g. cross-linking, and glycation of proteins) that occur in hydrolysates.

121

These reactions result in the formation of minor compounds, such as furosine, lysinoalanine

122

(LAL), peptide-phytate complexes, and peptide-polyphenol complexes. Conventionally, these

123

compounds are not included in the gross compositional analysis of hydrolysates. Therefore,

124

qualitative changes could result in large differences in the hydrolysate functionality, while the

125

gross composition of different batches of hydrolysate stays identical. Both gross and qualitative

126

compositional changes occur in soy protein hydrolysates due to variations in processing and raw

127

material (seed and/ or meal) composition. While there is a lot of literature available on

128

compositional (gross and qualitative) changes in soybean and its derived products, this

129

knowledge, in relation to cell culture functionality, has not been combined in an overview yet.

6


Page 7 of 51


130

In this review, we first discuss the effects of hydrolysate composition- both gross and qualitative

131

on cell growth and recombinant protein production in cell culture applications. Subsequently, it is

132

discussed how variations in the hydrolysate composition may be induced by variations in the raw

133

material and processing.

134

Chemical composition of soy protein hydrolysates

135

The major compounds present in soy protein hydrolysates are peptides, carbohydrates, and

136

minerals. The typical gross composition of a commercial soy protein hydrolysate produced from

137

defatted soybean meal is 60 ± 5% (w/w) peptides/ amino acids, 20 ± 5% (w/w) carbohydrates and

138

10% (w/w) minerals†.17 While the total carbohydrate and mineral content of soy protein

139

hydrolysate is known, the monosaccharide and mineral composition is not yet reported. For 30

140

batches of another commercial soy protein hydrolysate ‘Proyield Soy SE50MAF-UF’ produced

141

from defatted soybean meal, the protein content varied in a narrow range of 56-58% (w/w).13

142

This suggests that the batch-to-batch variability in the gross composition of a specific hydrolysate

143

produced by a particular manufacturer is relatively small (< 5%).

144

In addition to peptides, carbohydrates, and minerals, several compounds like phytates ( 10 kDa) and less-

212

extensively hydrolyzed (18% peptides < 1 kDa; 26% peptides > 10 kDa) rapeseed protein

213

hydrolysates supplemented to chemically defined media showed cell growth of 133% and 85%,

214

respectively.30 Similar observations were reported when rapeseed protein hydrolysates made with

215

Esperase 7.5L®, Neutrase 0.8L®, and Orientase 90N® were tested in cell culture. Esperase is a

216

serine protease from Bacillus lentus, Neutrase is a neutral metallo-protaese from Bacillus

217

amyloliquefaciens, and Orientase is a serine protease from Bacillus subtilis.30 All of the enzymes

218

had low specificities.30 The activity of the enzymes was standardized using a hemoglobin

219

standard allowing comparison of different enzymes. The hydrolysates had similar molecular size

220

distribution of peptides but the cell growth observed was not similar. For e.g. the hydrolysate 10


Page 11 of 51


221

made with Esperase enhanced the cell growth 1.5 times higher than hydrolysates made with

222

Neutrase.30 This showed that hydrolysis with different enzymes resulted in formation of different

223

peptides which affected the functionality of hydrolysates.

224

(b)

225

carbohydrates differ in the rates at which they are consumed by cells and hence, the maximum

226

achievable cell growth differs.31,

227

concentration were tested in a media formulation used for CHO-TF-70R cells. While the average

228

cell concentration after 140 hours of culturing in media containing glucose or mannose was 1.2 *

229

106 cells/ mL, it was 0.7 * 106 cells/ mL in media with galactose or fructose.31 Furthermore, the

230

consumption rate of glucose or mannose was much faster than galactose or fructose. After 150

231

hours of culturing, 12.5 mM of glucose/ mannose and 2.5mM of galactose/ fructose was

232

consumed by CHO-TF-70R cells.31 In a similar study, Barngrover et al. (1985) studied the

233

influence of fructose and galactose concentration on cell growth of Vero African green monkey

234

cells. After 6 days culture, the cell growth was 27.5, 23.8, 11.3, and 16.3 * 105 cells/ mL in 5, 10,

235

15, and 20 mM of galactose in L-15 (Leibovitz) medium.33 On the other hand, the cell growth

236

was 23.8, 27.5, 23.8, and 16.8 * 105 cells/ mL in 5, 10, 15, and 20 mM of fructose concentration

237

in L-15 (Leibovitz) medium.33 Therefore, a strategic selection of carbohydrates and concentration

238

in cell culture can be used to enhance the recombinant protein production.34 While the maximum

239

cell growth and tissue plasminogen activator production in 20 mM glucose medium was 1.6 * 106

240

cells/ mL and 3.9 µg/ mL, it was 1.9 * 106 cells/ mL and 4.5 µg/ mL in 5 mM glucose and 20 mM

241

galactose medium.34

242

(c)

243

has been investigated in few studies. In HeLa35 and chick embryo cell culture36 racemized D-

244

amino acids did not inhibit or enhance cell growth. Similarly, Naylor et al. (1976) tested

Carbohydrates: Carbohydrates are a source of carbon that supports cell growth. Different

32

Glucose, mannose, galactose, and fructose at 20mM

Racemized amino acids/ peptides: The influence of racemized amino acids on cell growth

11



Page 12 of 51

245

racemized amino acids in several mammalian cell lines (e.g. LM (TK-), A9, HTC+, 6TG-11,

246

CHO, and B16) and reported that racemized amino acids did not affect cell growth.37

247

(d)

248

phytates,40 and MRP41, 42 have been shown to chelate cations. Cations have an important role in

249

mammalian cell culture, for e.g. magnesium controls the activities of many glycolytic enzymes in

250

the Krebs cycle,43 and calcium and zinc are essential for cell growth and differentiation, apoptosis

251

and recombinant protein production.44, 45 For instance, the supplementation of phytic acid (0.27

252

mM to 2.7 mM) resulted in decreased BALB/c mouse 3T3 fibroblast cell growth (93% to 45%)

253

as compared to the non-supplemented control (100%).46

254

(e)

255

amino acids (e.g. lysine), as the MRP cannot be metabolized in cell cultures. Moreover, Maillard

256

reaction also results in formation of new compounds, such as pyrazines and hydroxymethyl

257

furfural, which affect cell cultures. Pyrazine and its derivatives at 1% (w/v) concentration

258

induced genotoxicity in the range of 0.4-29% in CHO cells.47 More than 1% (w/v) pyrazine

259

concentration was toxic and completely inhibited cell growth. Another MRP, hydroxymethyl

260

furfural, at 0.1% (w/v) concentration exhibited similar genotoxic and growth inhibition effects in

261

chick-embryo fibroblasts.48

262

(f)

263

investigated by supplementing growth medium with sunflower meal hydrolysates containing 1%

264

and 6% (w/w) of polyphenols. While the cell growth and recombinant protein production for the

265

former hydrolysate was 9 g biomass/ L and 180 units of streptokinase/ mL-1 hour-1, it was only 2

266

g biomass/ L and 40 units of streptokinase/ mL-1 hour-1 for the latter, respectively.49 This clearly

267

showed that high levels of polyphenol in hydrolysates inhibited cell growth and productivity.

268

Although a large but similar extent of variability is reported for polyphenol content in sunflower

Chelating compounds (LAL, phytates, and MRP): In human and animal studies, LAL,38, 39

Maillard reaction products (MRP): The Maillard reaction results in loss of essential

Peptide-polyphenol complexes: The influence of peptide-polyphenol complexes was

12


Page 13 of 51


269

(1.4-6.1%)49-52 and soybean meal (0.02-9%)53-55, the polyphenol composition between the two

270

meals differ greatly. While the sunflower meal primarily contains chlorogenic acid, caffeic and

271

quinic acids, soybean meals are rich in isoflavones such as daidzein, glycitein, and genistein. In

272

chemically defined cell cultures, influence of soy isoflavones on the cell culture functionality has

273

also been investigated. The supplementation with genistein at > 10µM concentration (2.7 µg/mL)

274

to the cell culture media resulted in drastic reduction of CHO cell viability, and apoptosis and

275

genotoxicity was induced.56 In addition to cytotoxicity, genotoxic effects on cell cultures due to

276

polyphenols are reported. A 40-fold increase in chlorogenic acid concentration from 0.01 mg/ mL

277

to 0.4 mg/ mL resulted in 40-fold increase in genotoxicity in CHO cells.57 These effects increased

278

further in the presence of metal ions like manganese and copper.57

279

In addition to a direct effect, the above-mentioned minor compounds influence the cell culture

280

functionality by affecting hydrolysate composition. These compounds interfere with the

281

production of peptides during enzymatic hydrolysis in hydrolysate production (Figure 1).

282

Production of soy protein hydrolysates

283

Soy protein hydrolysates for cell culture applications are produced by enzymatic hydrolysis of

284

defatted soybean meal, soy protein concentrate, or protein isolate, followed by enzyme

285

inactivation, coarse-filtration, ultra-filtration, and spray drying (Figure 2).58 The industrial

286

enzyme preparations that are used for hydrolysis may contain carbohydrase side activities. Such

287

side activities in Neutrase (0.025% on protein basis) resulted in release of 14% neutral sugars

288

from the water unextractable solids fraction obtained from toasted soybean meal.59 The

289

hydrolysis reaction is stopped by inactivating the enzyme using a heat treatment. After enzyme

290

inactivation, the solution is coarse-filtered to remove large and insoluble impurities, mainly

13



Page 14 of 51

291

polysaccharides, non-hydrolyzed proteins, and aggregated peptides. Coarse filtration is usually

292

performed using filter aids, which are available in wide range of particle sizes (median particle

293

size from 1.5-26 µm).60 The coarse filtration can be substantially hindered by polysaccharides

294

and protein-lysophospholipids complexes.61 After coarse filtration, the hydrolysate solution is

295

ultra-filtered, typically using a 10 kDa membrane, to remove large molecular weight compounds,

296

such

297

lipopolysaccharides, which are a major component of the cell wall membrane of most gram

298

negative bacteria. The presence of endotoxin in hydrolysate adversely affects its functionality. In

299

B-9 cells, presence of 20 ng/ mL endotoxins resulted in 30% reduced production of a

300

recombinant human protein called Mullerian inhibiting substance.62 After ultra-filtration, the

301

hydrolysate is composed of only water-soluble compounds (including the minor compounds

302

described in previous sections). Finally, the hydrolysate solution is concentrated and spray dried.

303

Conventionally, commercial soy protein hydrolysates are characterized for molecular weight

304

distribution of peptides. Other analyses may include determination of protein, ash, moisture, and

305

amino acids.63

306

Compounds affecting enzymatic hydrolysis of the soybean meal

307

The enzymatic hydrolysis of the soybean meal is affected by the minor compounds following

308

three mechanisms: firstly, due to processing treatments, the substrate protein may get modified

309

(e.g. MRP and cross-linked proteins). Secondly, compounds present in the soybean meal (e.g.

310

protease inhibitors) may inhibit the enzyme used for hydrolysis. Thirdly, some compounds (e.g.

311

polyphenols and phytates) can both modify the substrate and inhibit the enzyme (Figure 1).

312

Since systematic data on soy protein hydrolysis in relation to these mechanisms are not available,

313

references that describe effects on other types of proteins are included (Table 2). In addition, data

as

partially

hydrolyzed

peptides

and

endotoxins.58

Endotoxins

are

complex

14


Page 15 of 51


314

on in vitro protein digestibility is not available and therefore, results from in vivo animal nutrition

315

studies are used to illustrate the effects of these mechanisms on protein digestibility. These

316

effects are then used as an indication to state that similar effects can be expected in in vitro

317

protein hydrolysis during hydrolysate production.

318

Specific substrate modification

319

(a)

320

been shown to occur in amino acids and peptides. In these studies, combinations of extreme

321

experimental conditions (pH 2-12; temperature = 25 °C-130 °C; heating time = 1-24 hours) have

322

been used to induce racemization.64, 65 However, these processing conditions never occur during

323

hydrolysate production. The extent of racemization has not been investigated under

324

representative conditions (neutral pH; temperature = 50 °C-100 °C; heating time = 0.5-1 hour).

325

This is important because, the occurrence of even very small extent of racemization in proteins

326

and/or peptides greatly affects its hydrolysis.66 In vitro hydrolysis of synthetic racemized and

327

non-racemized tripeptides (Ala-L-Glu-Ala, Ala-D-Glu-Ala, Ala-L-Asp-Ala, Ala- D-Asp-Ala,

328

Val-L-Asp-Val, Val-D-Asp-Val, Ala-L-Phe-Leu, Ala-D-Phe-Leu, Ala-Met-Ala.HCl, Ala-D-Met-

329

Ala.HCl, Val-Met-Phe.HCl, and Val-D-Met-Phe.HCl) using intestinal peptidases was

330

investigated. While, tripeptides containing L-amino acid residues were completely hydrolyzed,

331

the tripeptides containing D-amino acid residues were not hydrolyzed at all.67, 68 In another study,

332

peptic hydrolysis of benzyloxycarbonyl-L-Ala-L-Phe-L-Tyr, benzyloxycarbonyl-L-Ala-L-Phe-L-

333

Leu-L-Ala, benzyloxycarbonyl-L-Ala-L-Gly-L-Phe-L-Tyr and their corresponding D-isomers

334

was studied. While the rate of hydrolysis was 22, 95, and 584 nmoles/minute/mg pepsin for L-

335

isomers, it was 1, 7, and 4 nmoles/minute/mg pepsin for D-isomers, respectively.69

Racemization: In several studies, racemization, i.e. conversion of L-form to D-form, has

15



Page 16 of 51

336

(b)

Lysinoalanine (LAL) formation: The processing treatments used during oil extraction

337

from soybeans, production of the meal and hydrolysates can lead to formation of compounds

338

such as LAL, lanthionine, dehydroalanine, and β-aminoalanine. The cross-linking reaction and

339

formation of these compounds is well described in literature.70, 71 Soy proteins (Defatted soy flour

340

precipitated at pH 4.5, washed with water and freeze dried) were heat treated at 100 °C-120 °C,

341

pH 6.5 for 1-3 hours. The LAL content in these heat processed soy proteins was

Factors Causing Compositional Changes in Soy ... - ACS Publications

Recommend Documents