Use of MALDI-MS Combined with Differential ... - ACS Publications

Subscriber access provided by UNIV OF WESTERN ONTARIO

Article

The use of MALDI-MS combined with differential hydrogen-deuterium exchange for semi-automated protein global conformational screening Gregory F. Pirrone, Heather Wang, Nicole Canfield, Alexander S. Chin, Timothy A Rhodes, and Alexey A. Makarov Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b01590 • Publication Date (Web): 20 Jul 2017 Downloaded from http://pubs.acs.org on July 23, 2017

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.

Analytical 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 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

The use of MALDI-MS combined with differential hydrogen-deuterium exchange for semi-automated protein global conformational screening

1 2 3 4 5

Gregory F. Pirrone 1*, Heather Wang 1, Nicole Canfield 2, Alexander S. Chin 2,

6

Timothy A. Rhodes 2, Alexey A. Makarov 1*

7

Merck & Co., Inc., MRL, 1 Process Research & Development / 2 Pharmaceutical sciences,

8

126 E. Lincoln Ave., Rahway, NJ 07065 USA

9

*

Corresponding authors

10

Tel: (+1) 732-594-7735

11

E-mail: [email protected], [email protected]

12

Merck & Co., Inc.

13

126 East Lincoln Ave.

14

Rahway, NJ 07065, USA

15 16

KEYWORDS: MALDI, HDX, global protein conformational screening, differential

17

hydrogen-deuterium exchange

18

ACS Paragon Plus Environment

1


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

19

Page 2 of 25

Abstract

20

Matrix-assisted laser desorption/ionization (MALDI) coupled with a time-of-

21

flight (TOF) mass-spectrometry (MS) detector is acknowledged to be very useful for

22

analysis of biological molecules. At the same time, hydrogen-deuterium exchange (HDX)

23

is a well-known technique for studying protein higher-order structure. However, coupling

24

MALDI with HDX has been challenging due to undesired back-exchange reactions

25

during analysis. In this report, we survey an approach that utilizes MALDI coupled with

26

an automated sample preparation to compare global conformational changes of proteins

27

under different solution conditions using differential HDX. A non-aqueous matrix was

28

proposed for MALDI sample preparation to minimize undesirable back-exchange. An

29

automated experimental setup based on the use of a liquid-handling robot and automated

30

data acquisition allowed for tracking protein conformational changes as a difference in

31

the number of protons exchanged to deuterons at specified solution conditions.

32

Experimental time points to study the deuteration-labeling kinetics were obtained in a

33

fully automated manner. The use of a non-aqueous matrix solution allowed experimental

34

error to be minimized to within 1% RSD. We applied this newly developed MALDI-

35

HDX workflow to study the effect of several common excipients on insulin folding

36

stability. The observed results were corroborated by literature data and were obtained in a

37

high-throughput and automated manner. The proposed MALDI-HDX approach can also

38

be applied in a high-throughput manner for batch-to-batch higher order structure

39

comparison, as well as for the optimization of protein chemical modification reactions.

40


2

Page 3 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

41


Introduction

42

While small molecule drugs still occupy a significant portion of the

43

pharmaceutical pipeline, the industry’s focus has expanded to peptide and protein-based

44

therapeutics 1. Despite the advantages in specificity and selectivity that peptide and

45

protein-based therapeutics afford, they also introduce a variety of new analytical

46

challenges

47

oxidation and aggregation are a few examples of potential dangers that may arise during

48

the manufacturing, formulation, storage and shipping processes of peptide and protein

49

therapeutics. These modifications can produce immunogenic byproducts, and the need to

50

monitor and characterize the higher order structure of large molecule therapeutics

51

becomes paramount. X-ray crystallography, Circular Dichroism (CD), Nuclear Magnetic

52

Resonance (NMR) and Cryo-EM play a significant role in structural biology and the

53

characterization of peptides and proteins. However, not all proteins are amenable to

54

these techniques due to solubility or size limitations. For example, Circular Dichroism

55

(CD), one of the most common spectroscopic techniques used to study protein higher-

56

order structure in a solution, requires the protein sample to be purified and free from

57

interfering proteins and impurities or optically active buffers (which may contribute to

58

the relevant spectrum )

59

are often used to study purified protein samples in solution. Thus, a methodology that can

60

be applied for studying protein higher-order structure in a mixed sample would be useful.

61

Hydrogen deuterium exchange (HDX) is a well-known technique for studying

62

protein higher-order structure 7-10. The exchange of labile protons to deuterons in amides,

63

alcohols, carboxylic acids, or amines is observable by MS and NMR 11. Since there is an

64

exchangeable NH proton virtually in every peptide bond of a protein (except proline),

65

HDX is a very useful technique for understanding protein higher-order structure and

66

folding. Hydrophobic interior regions of folded proteins have significant hydrogen

67

bonding, which restricts hydrogen/deuteron exchange in those regions, making it possible

68

to evaluate their conformational structure

69

interrogate the conformation and dynamics of proteins that are difficult to study by

70

conventional biophysical methods 14. As such, HDX-MS has seen robust adoption in the

2-4

. As the size of the drugs increases, so does the complexity. Deamidation,

5, 6

. Additionally, NMR and mass-spectrometry (MS) techniques

12, 13

. HDX-MS has become a useful tool to


3


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 25

71

pharmaceutical and biopharmaceutical industries 15, 16. Continuous HDX and pulse HDX

72

are the two main HDX workflows 12, 13 . Both of these HDX approaches suffer from H/D-

73

back-exchange during chromatographic separation and are not efficient in the case of

74

labile protons located on side chains

75

have been demonstrated may not suffer from H/D back-exchange: one uses the high

76

pressure of an LC system in combination with differential deuterium exchange as a direct

77

probe of global protein conformational change by pressure

78

automated HDX approach based on size-exclusion chromatography 18.

79

12, 13

. Nevertheless, some HDX approaches that

17

, another is a semi-

Matrix-assisted laser desorption/ionization coupled with a time-of-flight MS 19, 20

80

detector was demonstrated to be very useful for analysis of biological molecules

81

There are many reports describing attempts to combine MALDI and HDX to study

82

proteins higher-order structures

83

suffer to an even greater extent from undesired back-exchange reactions during analysis

84

than do HDX methods based on electrospray ionization MS (ESI-MS). For example, in

85

methods based on ESI-HDX, back-exchange values may range from 4 to 30% of the

86

incorporated deuterons; by comparison MALDI-HDX-based techniques may lead to

87

back- exchange values of up to 50%.22, 24-26 Moreover, the percentage of unwanted back-

88

exchange should be estimated based on the total number of back-bone amide hydrogens

89

in an intact peptide or protein; however, sometimes the N-terminal amides are not taken

90

into consideration since back-exchange is accelerated at the N-terminal amino-group,

91

which makes the calculation of unwanted back-exchange significantly under-estimated.

92

The unwanted back-exchange impacts data reproducibility and obstructs structural and

93

kinetic interpretation of the MALDI-HDX experiment. Moreover, artificial deuterium in-

94

exchange may occur when quenching the HDX reaction. Although approaches to

95

minimize the effects of undesirable back-exchanges during sample preparation for

96

MALDI-HDX have been investigated

97

use in combination with HDX),

98

methodology are needed.

21-23

.

. However, the coupling of MALDI to HDX may

27

22, 28-30

(as well as optimization of MALDI matrices for

further studies to develop a robust MALDI-HDX

99

In this study, we report an approach that utilizes MALDI coupled with automated

100

sample preparation to compare global conformational changes of proteins under different

101

solution conditions using differential HDX. A non-aqueous matrix was evaluated to be


4

Page 5 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


102

used for MALDI sample preparation to minimize undesirable D/H back-exchange. The

103

proposed MALDI-HDX approach can be applied in a high-throughput manner for protein

104

global higher order structure comparison during pharmaceutical excipient compatibility

105

studies, batch-to-batch comparison, as well as for the optimization of chemical

106

modifications of proteins.

107


5


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

108 109

Page 6 of 25

Experimental Section Reagents and chemicals

110

Ammonium formate (LC-MS grade), sinapic acid (98%), 3-nitrobenzyl alcohol

111

(98%), deuterium oxide (99.9%), glycerol (>99%), Cavitron (2-Hydroxypropyl)-β-

112

cyclodextrin), ubiquitin (from bovine erythrocytes), insulin (bovine from pancreas &

113

human recombinant, expressed in yeast) and bradykinin acetate were purchased from

114

Sigma-Aldrich Inc. (St. Louis, MO). Ultrapure water was obtained from a Milli-Q

115

Gradient A10 from Millipore (Bedford, MA, USA). Acetonitrile (MeCN), ammonium

116

hydroxide, Tween 80 (polysorbate 80) and trifluoroacetic acid (TFA) HPLC grade were

117

obtained from Fisher Scientific (Fair Lawn, NJ, USA).

118

Experimental conditions

119

Sample solutions of the proteins used for experiments were prepared at ~500 µM

120

in 50 mM ammonium formate buffer at pH 5.5. An Accumet AR-50 pH meter with a

121

standard pH electrode from Fisher Scientific (Fair Lawn, New Jersey) was used for all

122

pH measurements and adjustments. All liquid and protein samples were mixed and

123

dispensed with a Tecan 200 Evo liquid handling robot manufactured by Tecan Group

124

Ltd. (Mannedorf, Switzerland) and was equipped with a 96 tip head. All samples were

125

spotted onto a 384-well MTP Anchorchip target plate from Bruker Corporation (Billerica,

126

Massachusetts). All samples were analyzed using Bruker AutoFlex MALDI-TOF

127

(Billerica, Massachusetts). All spectra were acquired in positive ion mode by summing

128

2500 laser shots. Spectra were analyzed with flexAnalysis 3.0 software using the peak

129

detection algorithm centroid.

130

Solvent accessibility was estimated for the proteins used in this study with the

131

computational chemistry software Discovery Studio v.3.5.0 (Accelrys Software Inc.) with

132

input pdb (Protein Data Bank) files 2MJB and 4I5Z, for ubiquitin, and insulin,

133

respectively (Table 1). The solvent-accessible part of the protein sequence was defined

134

as more than 25% of Solvent Accessible Surface (SAS) modeled by the double cubic

135

lattice method (DCLM)

136

10% of SAS. The number of labile protons was calculated for the solvent-inaccessible

31

. Solvent inaccessibility for a protein was defined as less than


6

Page 7 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


137

part of the protein sequence (as determined by the software), as well as for the whole

138

protein sequence, based on their respective pdb files (Table 1).

139

Results and Discussion

140

Utilizing

robotics

for

automated

sample

preparation

and

141

experimentation

142

One of the major advantages of Matrix Assisted Laser Desorption Ionization

143

Mass Spectrometry (MALDI-MS) is that technique can be easily adopted for high-

144

throughput analysis (HTA). MALDI instrumentation provides the ability to use 96, 384

145

or 1536 well-plates for HTA. However, one of the major challenges of using MALDI in

146

high-throughput analysis is to generate and apply samples to the well-plate in a quick,

147

robust and reproducible manner. To address this challenge, we developed an automation

148

protocol using a Tecan Evo 200 liquid handling robot to generate and plate samples for

149

MALDI analysis in our study. An overview of the experimental workflow for this study

150

is shown in Figure 1. Protein samples of interest were equilibrated in a buffered stock

151

solution at target conditions (i.e. near neutral pH and 25⁰C). Each individual sample was

152

transferred to a single well in a 384 well plate. To initiate the continuous labeling

153

reaction, 2 µl of each sample was diluted into 20 µl of a deuterated labeling buffer by the

154

Tecan liquid handling robot. This 1:10 dilution favors the forward exchange process and

155

is routinely used in conventional HDX-MS protocols 14 . The samples were continuously

156

labeled for predetermined time periods. In this proof-of-concept study, 1, 10 and 60

157

minutes labeling time periods were used. To terminate the labeling reaction, each sample

158

was mixed 1:1 with a saturated matrix solution (5 mg/mL of sinapic acid in acetonitrile

159

with 0.1% 3-nitrobenzyl alcohol). 1 µl of each mixed sample was immediately spotted

160

and dried on a 384-well MALDI plate prior to MS analysis.

161

All sample handling steps were executed by the Tecan liquid handling robot. For

162

this proof-of-concept workflow, the Tecan liquid handling robot was equipped with a 96

163

tip head. Given the dimensions of the 384-well MALDI plate, this allows for a maximum

164

of 96 individual labeling experiments with one undeuterated control and three time points

165

(1, 10 and 60 minutes). Each labeling experiment was run simultaneously and less than


7


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 25

166

50 µl was required per sample. Note: this workflow can be seamlessly customized to

167

adopt higher throughput experimentation by changing to a 384 Tecan tip-head and to a

168

1536 well MALDI plate target.

169

Assessing back-exchange

170

Assessing deuterium back-exchange is critical for any HDX-MS method. If the

171

total loss of deuterium label is significant, the interpretation of HDX experimental results

172

becomes very difficult. Furthermore, unwanted back-exchange significantly reduces

173

replicate-to-replicate reproducibility since labeling may not be uniform. There are two

174

major sources of back-exchange for a conventional HDX-MS experiment: 1) incomplete

175

quenching of the intrinsic D/H exchange and 2) the chromatographic separation; both

176

steps involve introducing the newly deuterated peptides to an aqueous environment

177

which starts exchange of deuterons back to protons

178

these losses by performing the quenching and chromatography at pH 2.5 and 0⁰ C, where

179

the intrinsic H/D exchange (forward and backward) is at a minimum rate

180

unwanted back-exchange values for a conventional HDX-MS experiment are

181

approximately 20-30 percent 14.

14

. Several groups have mitigated 14

. Typical

182

Completely removing the chromatography step does not remedy the back-

183

exchange issues for HDX-MS. Other groups performing MALDI-HDX experiments have

184

reported significant back exchange in their data as well (up to 50%)

185

reports were based on the use of completely or partially aqueous quench solutions (i.e.

186

0.1% TFA aqueous) when applying the MALDI matrix. As noted, our MALDI-HDX

187

approach utilizes an acetonitrile and sinapic acid matrix as quench solution to minimize

188

the intrinsic H/D exchange. Our goal was to drastically reduce or eliminate back-

189

exchange by using a non-aqueous environment as the quench solution. The addition of

190

0.1% 3-nitrobenzyl alcohol helped to ionize the analyte protein in the absence of water

191

and results in higher signal intensity.

30

. However, these

192

Previously, the use of small test peptides without conformational restrictions

193

(such as bradykinin) to determine the extent of back exchange in HDX-MS experiments

194

was reported

195

between the back-exchange level and the length of the peptides, and it was concluded that

32

. It was also previously reported that there is no observable correlation


8

Page 9 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


196

intrinsic exchange rates are not related to the amino acid composition of the peptides

197

using MALDI-HDX experiment 33. We used bradykinin as a control peptide to assess the

198

amount of back-exchange in our MALDI-HDX workflow. A maximally labeled control

199

sample was generated by diluting ~1 mg of bradykinin with 1 mL of pure D2O, heating to

200

40 ⁰C and shaking overnight. A corresponding unlabeled control was generated in a

201

similar fashion, except H2O was the primary diluent. Both samples were analyzed using

202

the automated MALDI-HDX workflow described above and the resulting mass spectra

203

are shown in Figure 2. The observed centroid mass for the unlabeled bradykinin control

204

was 1062 Da (Figure 2, top panel) with a theoretical target [M+H] mass of 1062 Da.

205

The observed centroid mass for the maximally labeled bradykinin sample was 1078 Da

206

(Figure 2, bottom panel). Since bradykinin has 17 labile proton sites

207

measurement suggests that the majority of these sites are occupied with deuterium

208

throughout the liquid handling, mixing, plating and drying steps that are required in this

209

MALDI-HDX workflow. (Note that using aprotic quenching solvent may allow taking

210

into consideration back-bone as well as side-chain labile hydrogen positions). We

211

attribute this low back-exchange to the use of a non-aqueous matrix solution that acts as a

212

quenching step for this process. The ∆HDX number did not change for the labeled

213

bradykinin sample over 20 hours on the MALDI plate (Figure SI-2). Unwanted back- or

214

forward-exchange was estimated in a similar manner using a partially labeled (after 10

215

minutes) insulin sample (Figure SI-3). The ∆HDX number did not change for the

216

partially labeled insulin sample over 25 hours on the MALDI plate (plate was kept in the

217

instrument under vacuum conditions). To assess unwanted forward-exchange, unlabeled

218

control samples of bradykinin, insulin and ubiquitin were used in the automated MALDI-

219

HDX workflow; there were no detectable unwanted forward-exchange based on the

220

studied workflow (Figure SI-4).

221

34

, this

Monitoring of induced conformational changes in proteins

222

One of the principal objectives for developing this MALDI-HDX workflow was

223

to be able to compare global conformational differences between peptides and proteins in

224

a high throughput manner. Pharmaceutical excipients and formulations can influence the

225

dynamics and conformation of peptide and protein based drugs


35

. The ability to rapidly

9


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 25

226

investigate a wide range of variables and identify any induced structural changes is

227

highly desirable. To assess this workflow’s ability to observe conformational changes,

228

insulin and ubiquitin were investigated. These proteins have been modeled and studied

229

previously by our group and others

230

organic solvent content of the equilibration and labeling buffers were modified to induce

231

conformational changes prior to analysis. Note the design of HDX experiment requires

232

consideration for HDX intrinsic-rate as a function of solution parameters such as pH,

233

temperature, and deuterium isotope concentration. It was previously reported that the

234

slowest rate of H/D exchange occurs at pH 2.3-2.5

235

exchange rate constant for alanine tripeptide provided by Bai et al.36, the estimated time

236

to have 90% H/D exchanged at pH 2.0 and 20°C is about 123 seconds, and at pH 5.5 is

237

about 18 seconds. To estimate the effect of acetonitrile concentration in deuterium oxide

238

buffer solution on labeling in the studied workflow, we used 30:70 D2O (at pH 2): MeCN

239

(as worst case scenario) and bradykinin as control peptide. It was demonstrated the

240

peptide incorporated significant amounts of deuterium within 1 minute and it was

241

maximally labeled within 60 minutes (Figure SI-5).

18

and serve as model compounds. The pH and

36-38

. For example, using the H/D

242

Bovine insulin was labeled for 1, 10 and 60 minutes using the protocol described

243

above, and the average deuterium uptake (following continuous labeling in pH 5.5 buffer

244

modified with varying percentages of acetonitrile (0, 10, 30 and 70%)) was measured and

245

compared (Figure 3, examples of MS spectra Figure SI-1). All measurements were

246

recorded and averaged from four independent labeling replicates

247

for each data point was consistently less than 1%, indicating a high degree of robustness

248

for this workflow (Table SI-1). There were distinct differences in deuterium

249

incorporation as a function of organic solvent in the labeling buffers. In the absence of

250

acetonitrile, the average observed mass for insulin increased by 4.9, 5.9 and 17.1 Da for

251

the 1, 10 and 60 minute time points, respectively (Figure 3, Table SI-1), indicating

252

initial protection in insulin is due to a folded solution conformation. The rate of

253

deuteration is accelerated as the percentage of organic solvent is increased in the labeling

254

buffer. With 10% acetonitrile added to the labeling buffer, the average observed masses

255

in insulin increased by 7.0 and 8.5 Da at the 1 and 10 minute time points (Figure 3,

256

Table SI-1). Interestingly, insulin labeled with a 70% acetonitrile buffer results in a 15.0


14

. The %RSD values

10

Page 11 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


257

Da increase within 1 minute, a nearly two-fold increase in deuteration (Figure 3, Table

258

SI-1). Taken together, these data suggest that the structure of insulin is beginning to

259

unfold in the presence of 10% acetonitrile in the labeling buffer. This is indicated by the

260

increased deuteration at the two early time points relative to the 0% acetonitrile sets.

261

Unfolding of insulin results in an increase of solvent accessibility of labile protons of the

262

protein and allows for more rapid labeling with deuterium. This global deprotection is

263

exacerbated with higher organic content, as indicated by the increased deuteration

264

observed at the early time points in the 30%, 50% and 70% acetonitrile labeling

265

experiments (Figure 3, Table SI-1). These data are consistent with previously reported

266

findings on insulin

267

detecting global conformational changes in proteins in a robust and high throughput

268

manner.

18

and validate that the MALDI-HDX workflow is capable of

269

Ubiquitin was interrogated in a similar fashion (0, 10, 30, 50 and 70%

270

acetonitrile) and at two different pHs (2.5 and 5.5). The resulting deuterium incorporation

271

into ubiquitin under these conditions was measured and compared (Figure 4, examples

272

of MS spectra Figure SI-1). As before, four independent labeling replicates were

273

measured for each condition

274

measurement, indicating a highly robust workflow. In the absence of acetonitrile, the

275

average observed mass of ubiquitin increased by 29.5, 37.2 and 47.8 Da for the 1, 10 and

276

60 minutes time points at pH 2.5, respectively (Figure 4 A, Table SI-2). Ubiquitin

277

labeled in buffer containing 10% acetonitrile has significantly slower deuterium uptake

278

with an average observed mass increase of 21.9 at the 1 minute time point, 7.6 Da less

279

than the 0% acetonitrile labeling state. Further increasing the acetonitrile content reduces

280

the labeling at the early time points for 30, 50 and 70% acetonitrile (Figure 4 A, Table

281

SI-2). Despite the initial reduction in deuteration at the early time points, ubiquitin is

282

labeled to a similar extent by 60 minutes regardless of acetonitrile content. These data

283

suggest that acetonitrile compacts the molecule resulting in initial protection, as reported

284

by the 1 and 10 minute time points. However, the ubiquitin remains dynamic and is able

285

to incorporate similar amounts of deuterium irrespective of the percentage of acetonitrile

286

in the labeling buffer. This trend was also observed at pH 5.5. Ubiquitin labeled with no

287

acetonitrile in the labeling buffer showed an average deuterium incorporation of 12.8 and

14

and the %RSD values were under 1% for each


11


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 25

288

13.8 Da for the 1 and 10 minutes time points respectively. This contrasts with the average

289

deuterium incorporation for 70% acetonitrile, with increases of 5.8 and 10.9 Da for the

290

same time points (Figure 4 B, Table SI-3). The facilitation of compact folding by

291

acetonitrile was corroborated by literature data

292

be used for protein denaturation and for disrupting hydrogen bonding.

293

internal hydrogen bonding present in the solvent accessible areas of the ubiquitin,

294

trifluoroacetic acid (TFA) was used as chaotropic additive. The addition of TFA into

295

labeling buffer helped to increase labeling of ubiquitin after 10 minutes (at room

296

temperature) up to 56% (Table SI-4) of total solvent accessible labile protons (Table 1).

297

Using chaotropic additive in the labeling solution could be helpful for experimental

298

measurement of total number of labile protons in the structure. Overall, this MALDI-

299

HDX workflow can capture solvent-induced structural transitions within different

300

proteins and at different pHs.

39, 40

. Chaotropic additives are known to 41

To disrupt

301

After validating that our system could be used to monitor solvent-induced

302

conformational changes, we shifted our focus to possible conformational changes

303

induced by the presence of pharmaceutical excipients. It is common for therapeutics to be

304

combined with additives or excipients to increase stability or bioavailability of active

305

pharmaceutical ingredients

306

structure have been studied prior 44. We sought to apply our MALDI-HDX workflow to

307

rapidly screen for effects from excipients on higher-order structure in human insulin.

308

Insulin (10 mg/mL) was mixed with 1:1 stock solutions of glycerol (10% v/v after

309

mixing), Tween 80 (20 mg/mL after mixing), Cavitron (200 mg/mL after mixing) and

310

water (control) and mixed-end-over end at 25 ⁰C for 24 hours (Note: all samples were

311

prepared in 50 mM ammonium formate buffer at pH 5.5). The following day, these

312

mixtures were then labeled and analyzed by our MALDI-HDX workflow. As described

313

previously, we assayed multiple labeling buffer conditions. 0% acetonitrile was used to

314

interrogate any intrinsic effects that the excipients may induce on the higher order

315

structure of insulin, while 70% acetonitrile was used to assess how these excipients may

316

stabilize insulin in denaturing conditions. Association with various excipients resulted in

317

increased deuterium uptake in some cases and decreased deuterium uptake in others.

318

Insulin is largely destabilized by Tween 80 at these concentrations, showing increases of

42, 43

and the effects of excipients on protein higher order


12

Page 13 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


319

18.0 and 22.3 Da for the 1 and 10 minute time points respectively (Figure 5 A),

320

significantly higher relative to the control sample. Labeling in 70% acetonitrile did not

321

induce any further destabilization as insulin was labeled to a similar extent during the

322

time course (Figure 5 B). Glycerol had the opposite effect at these concentrations,

323

largely stabilizing insulin. In the presence of glycerol, insulin showed modest increases of

324

1.4, 2.7 and 3.7 Da for the 1, 10 and 60 minute time points (Figure 5 A), suggesting

325

significant protection as well as reduced dynamics, relative to the control. These trends

326

are maintained even in the presence of 70% acetonitrile, showing increases of 2.7, 10.5

327

and 13.0 Da throughout the time course, significantly lower than the control sample.

328

Cavitron had more subtle effects on insulin structure, showing slightly reduced deuterium

329

incorporation at 10 minutes compared to the control (7.6 vs 12.2 Da), but similar uptake

330

for the 1 and 60 minute time points. This suggests that insulin is modestly protected in

331

the presence of Cavitron at these concentrations, but remains nearly as dynamic in

332

solution as the control sample (Figure 5 A). The stabilization effects of Cavitron are

333

more apparent when labeling with 70% acetonitrile showing reduced deuterium uptake at

334

all-time points relative to the control sample (Figure 5 B).

335

It should be noted that the concentration of Tween 80 used in this study is well

336

above the critical micelle concentration 45 and it has been reported that above the critical

337

micelle concentration, Tween 80 can displace insulin adsorbed to a hydrophobic surface

338

by releasing protein in large clusters 48

46, 47

. Glycerol has been reported to prevent insulin

49

339

precipitation

340

Cavitron and insulin

341

on insulin physical stability at the studied concentrations are consistent with previous

342

literature findings and further validate our MALDI-HDX workflow for high-throughput

343

screening of proteins and peptides.

344

and increased stability . Similar stability findings have been reported for 50

. Taken together, these observed trends of the excipients’ impact

Conclusions

345

MALDI-TOF-MS coupled with differential hydrogen–deuterium exchange was

346

demonstrated in this report to be a practical approach for comparing global protein

347

conformational changes in solution. An automated experimental setup based on the use of

348

a TECAN liquid handling robot and automated MALDI data acquisition allowed for


13


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 25

349

tracking protein conformational changes following perturbations to a particular

350

acetonitrile concentration in solution in a multivariate experiment. In this setup, the

351

∆HDX trends-over-time showed global protein conformational changes as a difference in

352

the number of protons exchanged to deuterons at specified solution conditions.

353

Experimental time points to study deuteration labeling kinetics were obtained in a fully

354

automated manner. The kinetic points and sample transfer to the MALDI plate was

355

performed by a TECAN liquid handling robot. The non-aqueous matrix solution was

356

designed to be used for MALDI sample preparation to eliminate undesirable D/H back-

357

exchange. The use of a non-aqueous matrix solution minimized experimental error to

358

within 1% RSD. We applied the newly developed MALDI-HDX workflow to study the

359

effect of several common excipients on insulin folding stability. The observed results

360

were corroborated by literature data and were obtained in a high-throughput and fully

361

automated manner. The proposed MALDI-HDX approach can also be applied in a high-

362

throughput manner for batch-to-batch higher order structure comparison, as well as for

363

the optimization of chemical modifications of proteins.

364

Acknowledgements

365

The authors would like to thank Dr. Ian Mangion and Dr. Yun Mao for fruitful

366

discussions and helpful suggestions.

367


14

Page 15 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411


References (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27)

Walsh, G. 2014, 32, 992-1000. Berkowitz, S. A.; Engen, J. R.; Mazzeo, J. R.; Jones, G. B. 2012, 11, 527-540. Sandra, K.; Vandenheede, I.; Sandra, P. 2013, 1335, 81-103. Sathish, J. G.; Sethu, S.; Bielsky, M.-C.; de Haan, L.; French, N. S.; Govindappa, K.; Green, J.; Griffiths, C. E. M.; Holgate, S.; Jones, D. 2013, 12, 306-324. Greenfield, N. J. Analytical biochemistry 1996, 235, 1-10. Provencher, S. W.; Gloeckner, J. Biochemistry 1981, 20, 33-37. Englander, S. W. Journal of the American Society for Mass Spectrometry 2006, 17, 1481-1489. Hentze, N.; Mayer, M. P. Journal of Visualized Experiments 2013, 50839. Leurs, U.; Mistarz, U. H.; Rand, K. D. European Journal of Pharmaceutics and Biopharmaceutics 2015, 93, 95-109. Majumdar, R.; Middaugh, C. R.; Weis, D. D.; Volkin, D. B. Journal of pharmaceutical sciences 2015, 104, 327-345. Englander, S. W.; Mayne, L. Annual review of biophysics and biomolecular structure 1992, 21, 243-265. Engen, J. R. Analytical chemistry 2009, 81, 7870-7875. Konermann, L.; Pan, J.; Liu, Y.-H. Chemical Society Reviews 2011, 40, 12241234. Engen, J. R.; Wales, T. E. 2015, 8, 127-148. Houde, D.; Berkowitz, S. A.; Engen, J. R. 2011, 100, 2071-2086. Campobasso, N.; Huddler, D. 2015, 25, 3771-3776. Makarov, A. A.; Schafer, W. A.; Helmy, R. Analytical chemistry 2015, 87, 23962402. Makarov, A. A.; Helmy, R. Journal of Chromatography A 2016, 1431, 224-230. Karas, M.; Bachmann, D.; Bahr, U. e.; Hillenkamp, F. International journal of mass spectrometry and ion processes 1987, 78, 53-68. Burlingame, A. L.; Boyd, R. K.; Gaskell, S. J. Analytical chemistry 1996, 68, 599652. Mandell, J. G.; Falick, A. M.; Komives, E. A. Proceedings of the National Academy of Sciences 1998, 95, 14705-14710. Mandell, J. G.; Falick, A. M.; Komives, E. A. Analytical Chemistry 1998, 70, 3987-3995. Woofter, R. T.; Maurer, M. C. Archives of biochemistry and biophysics 2011, 512, 87-95. Ghaemmaghami, S.; Fitzgerald, M. C.; Oas, T. G. Proceedings of the National Academy of Sciences 2000, 97, 8296-8301. Hoofnagle, A. N.; Resing, K. A.; Ahn, N. G. Annual review of biophysics and biomolecular structure 2003, 32, 1-25. Pingerelli, P. L.; Ozols, V. V.; Saleem, H.; Anderson, C. R.; Burns, R. S. European Journal of Mass Spectrometry 2009, 15, 739-746. Kipping, M.; Schierhorn, A. Journal of mass spectrometry 2003, 38, 271-276.


15


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455

(28) (29) (30) (31) (32) (33) (34) (35) (36) (37)

(38) (39) (40) (41) (42) (43) (44) (45) (46) (47) (48) (49) (50)

Page 16 of 25

Villanueva, J.; Canals, F.; Villegas, V.; Querol, E.; Aviles, F. X. FEBS letters 2000, 472, 27-33. Bache, N.; Rand, K. D.; Roepstorff, P.; Jorgensen, T. J. D. Analytical chemistry 2008, 80, 6431-6435. Lemaire, P.; Debois, D.; Smargiasso, N.; Quinton, L.; Gabelica, V.; De Pauw, E. A. Rapid Communications in Mass Spectrometry 2013, 27, 1837-1846. Eisenhaber, F.; Lijnzaad, P.; Argos, P.; Sander, C.; Scharf, M. Journal of Computational Chemistry 1995, 16, 273-284. Zhang, Z.; Zhang, A.; Xiao, G. Analytical chemistry 2012, 84, 4942-4949. Nazabal, A.; Hornemann, S.; Aguzzi, A.; Zenobi, R. Journal of mass spectrometry 2009, 44, 965-977. Mao, D.; Douglas, D. Journal of the American Society for Mass Spectrometry 2003, 14, 85-94. Zarmpi, P.; Flanagan, T.; Meehan, E.; Mann, J.; Fotaki, N. European Journal of Pharmaceutics and Biopharmaceutics 2017, 111, 1-15. Bai, Y.; Milne, J. S.; Mayne, L.; Englander, S. W. Proteins 1993, 17, 75. Englander, J. J.; Del Mar, C.; Li, W.; Englander, S. W.; Kim, J. S.; Stranz, D. D.; Hamuro, Y.; Woods, V. L. Proceedings of the National Academy of Sciences 2003, 100, 7057-7062. Wang, L.; Lane, L. C.; Smith, D. L. Protein Science 2001, 10, 1234-1243. Figueroa, I. D.; Russell, D. H. Journal of the American Society for Mass Spectrometry 1999, 10, 719-731. Pierson, N. A.; Makarov, A. A.; Strulson, C. A.; Mao, Y.; Mao, B. Journal of Chromatography A 2017, 1496, 51-57. Zhang, Y.; Cremer, P. S. Current opinion in chemical biology 2006, 10, 658-663. Abrantes, C. G.; Duarte, D.; Reis, C. P. Journal of pharmaceutical sciences 2016, 105, 2019-2026. Markl, D.; Zeitler, J. A. Pharmaceutical Research 2017, 1-28. Wei, H.; Ahn, J.; Yu, Y. Q.; Tymiak, A.; Engen, J. R.; Chen, G. Journal of the American Society for Mass Spectrometry 2012, 23, 498-504. Wan, L. S. C.; Lee, P. F. S. Journal of pharmaceutical sciences 1974, 63, 136137. Gunning, P. A.; Mackie, A. R.; Gunning, A. P.; Woodward, N. C.; Wilde, P. J.; Morris, V. J. Biomacromolecules 2004, 5, 984-991. Wu, C.-Y.; Benet, L. Z. Pharmaceutical research 2005, 22, 11-23. Blackshear, P. J.; Rohde, T. D.; Palmer, J. L.; Wigness, B. D.; Rupp, W. M.; Buchwald, H. Diabetes Care 1983, 6, 387-392. Amar-Yuli, I.; Azulay, D.; Mishraki, T.; Aserin, A.; Garti, N. Journal of colloid and interface science 2011, 364, 379-387. Zhang, L.; Zhu, W.; Song, L.; Wang, Y.; Jiang, H.; Xian, S.; Ren, Y. International journal of molecular sciences 2009, 10, 2031-2040.


16

Page 17 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

456


Figure Captions:

457 458

Figure 1. MALDI-HDX Workflow. Continuous deuterium labeling is initiated by

459

diluting an aliquot of protein stock solution into a deuterated labeling buffer. Each sample

460

is quenched by mixing 1:1 with a saturated non-aqeous matrix solution and immediately

461

spotted onto a clean MALDI plate. The sample is rapidly dried under a vacuum prior to

462

MS analysis. All sample handling and mixing is accomplished by an automated TECAN

463

Evo 200 liquid handling robot.

464

Figure 2. Assessing back-exchange. Comparison of undeuterated and maximally

465

deuterated bradykinin controls that have been plated and analyzed using the MALDI-

466

HDX workflow. The centroided masses of each spectra are listed above. The majority of

467

the labile proton sites are still occupied with deuterium in the maximally deuterated

468

control indicating little back exchange from the workflow

469

Figure 3. Effects of acetonitrile on global protein conformation. The average deuterium

470

uptake into insulin following continuous labeling in buffer modified with varying

471

percentages of acetonitrile. Increasing the organic content results in higher deuterium

472

uptake at shorter time points, suggesting global protein unfolding and deprotection. Error

473

bars represents the standard deviation of four individual labeling replicates. The statistics

474

for these data are shown in Table SI-1.

475

Figure 4. Effects of acetonitrile on global protein conformation. The average deuterium

476

uptake into ubiquitin following continuous labeling in buffer modified with varying

477

percentages of acetonitrile. Buffer pHs 2.5 (panel A) and 5.5 (panel B) were investigated.

478

Error bars represent the standard deviation of four individual labeling replicates. The

479

statistics for these data are shown in Table SI-2 and SI-3.

480

Figure 5. Effects of excipients on insulin solution conformation. Insulin was mixed with

481

excipients and labeled in buffers containing 0% (A) and 70% MeCN (B). Tween 80 has a

482

destabilizing effect relative to the control. Glycerol and Cavitron have a stabilizing effect

483

in both native-like and stressed conditions.

484 485


17


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 25

486 487 488

Tables Table 1. Characteristics of the compounds used in this study Experimental total number of Number of labile Calculated solvent total number protons*** accessible of labile labile protons* ** In 0.1%TFA/ protons 1.6 M GndCl at pH 7

Compound name

Molecular weight, Dalton

Number of amino acids

Number of solvent inaccessible labile protons**

Bradykinin

1060

9

0

17

17

17/17

Insulin

5734

51

15

76

91

91/91

Ubiquitin

8563

76

33

121

154

154/145

489 490 491 492 493 494 495 496 497 498 499 500

*

Calculated based on pdb files 2MJB, 4I5Z respectively for ubiquitin, insulin (not bradykinin 34) ** Solvent accessibility was calculated in Discovery Studio v.3.5.0 (Accelrys Software Inc.): solvent inaccessible is defined as less than 10% of solvent accessible surface (SAS), accessible is more than 25% of SAS (also included 10%-25%) based on approach by the DCLM method 31. *** Total number of labile protons were measured based on H/D exchange at 0.1%TFA or 1.6 M GndCl at pH 7 for 3 hours at 45°C, than measured by SEC-HDX at pH 5.4 using deuterium oxide in the mobile phase 18.


18

Page 19 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517


Supporting Information: Table SI-1. Insulin deuterium labeling statistics acquired from four independent labeling replicates; Table SI-2. Ubiquitin deuterium labeling statistics acquired at pH 2.5 from four independent labeling replicates; Table SI-3. Ubiquitin deuterium labeling statistics acquired at pH 5.5 from four independent labeling replicates; Table SI-4. Ubiquitin deuterium labeling after 10 minutes in the presence of trifluoroacetic acid; Figure SI-1. Examples of spectra obtained using MALDI-HDX workflow; Figure SI-2. Assessment of back-exchange of dried sample on MALDI plate; Figure SI-3. Assessment of back- or forward-exchange of dried sample on MALDI plate; Figure SI-4. Assessment of forward-exchange during sample preparation; Figure SI-5. Assessment of deuterium labeling at ambient temperature in 30:70 D2O (pH2): MeCN


19


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

518 519

Page 20 of 25

For Table of Contents Only (Table of Contents/Abstract Graphics)

520


20

Page 21 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1. MALDI-HDX Workflow. Continuous deuterium labeling is initiated by diluting an aliquot of protein stock solution into a deuterated labeling buffer. Each sample is quenched by mixing 1:1 with a saturated non-aqeous matrix solution and immediately spotted onto a clean MALDI plate. The sample is rapidly dried under a vacuum prior to MS analysis. All sample handling and mixing is accomplished by an automated TECAN Evo 200 liquid handling robot. 190x254mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2. Assessing back-exchange. Comparison of undeuterated and maximally deuterated bradykinin controls that have been plated and analyzed using the MALDI-HDX workflow. The centroided masses of each spectra are listed above. The majority of the labile proton sites are still occupied with deuterium in the maximally deuterated control indicating little back exchange from the workflow 190x254mm (300 x 300 DPI)


Page 22 of 25

Page 23 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. Effects of acetonitrile on global protein conformation. The average deuterium uptake into insulin following continuous labeling in buffer modified with varying percentages of acetonitrile. Increasing the organic content results in higher deuterium uptake at shorter time points, suggesting global protein unfolding and deprotection. Error bars represents the standard deviation of four individual labeling replicates. The statistics for these data are shown in Table SI-1. 190x254mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4. Effects of acetonitrile on global protein conformation. The average deuterium uptake into ubiquitin following continuous labeling in buffer modified with varying percentages of acetonitrile. Buffer pHs 2.5 (panel A) and 5.5 (panel B) were investigated. Error bars represent the standard deviation of four individual labeling replicates. The statistics for these data are shown in Table SI-2 and SI-3. 190x254mm (300 x 300 DPI)


Page 24 of 25

Page 25 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. Effects of excipients on insulin solution conformation. Insulin was mixed with excipients and labeled in buffers containing 0% (A) and 70% MeCN (B). Tween 80 has a destabilizing effect relative to the control. Glycerol and Cavitron have a stabilizing effect in both native-like and stressed conditions. 190x254mm (300 x 300 DPI)


Use of MALDI-MS Combined with Differential ... - ACS Publications

Recommend Documents