Accepted Manuscript Title: Separation of therapeutic peptides with cyclofructan and glycopeptide based columns in hydrophilic interaction liquid chromatography Author: Yang Shu John C. Lang Zachary S. Breitbach Haixiao Qiu Jonathan Smuts Mayumi Kiyono-Shimobe Mari Yasuda Daniel W. Armstrong PII: DOI: Reference:

S0021-9673(15)00243-5 http://dx.doi.org/doi:10.1016/j.chroma.2015.02.018 CHROMA 356276

To appear in:

Journal of Chromatography A

Received date: Revised date: Accepted date:

31-7-2014 5-2-2015 7-2-2015

Please cite this article as: Y. Shu, J.C. Lang, Z.S. Breitbach, H. Qiu, J. Smuts, M. Kiyono-Shimobe, M. Yasuda, D.W. Armstrong, Separation of therapeutic peptides with cyclofructan and glycopeptide based columns in hydrophilic interaction liquid chromatography, Journal of Chromatography A (2015), http://dx.doi.org/10.1016/j.chroma.2015.02.018 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

4 5

Yang Shua,b, John C. Langa,c, Zachary S. Breitbacha, Haixiao Qiua, Jonathan Smutsa,c,

6

Mayumi Kiyono-Shimobed, Mari Yasudad, Daniel W. Armstrong a,c∗

cr

2

ip t

3

Separation of therapeutic peptides with cyclofructan and glycopeptide based columns in hydrophilic interaction liquid chromatography

1

a Department of Chemistry and Biochemistry, The University of Texas at Arlington, Arlington, TX 76019, USA b College of Life and Health Sciences, Northeastern University, Shenyang 110189, China c AZYP LLC, Arlington, TX 76019, USA d Mitsubishi Chemical Corporation, 1-1-1, Marunouchi, Chiyoda-ku, Tokyo, 100-8251 Japan * Corresponding author. Tel.: 817-272-0632; fax:817-272-0619 E-mail address: [email protected] (D. W. Armstrong).

16

ABSTRACT

17

Three cyclofructan-based, two glycopeptide-based, and one zwitterionic column used in

18

the HILIC mode were assessed within a graphical framework based on different

19

functional characteristics contributing to selectivity. The characteristics of these six

20

HILIC columns are put in the perspective of 33 columns evaluated previously. The

21

isopropyl carbamate modified cyclofructan 6 (CF6) stationary phase, Larihc P, showed

22

reduced component contributions for hydrophilicity and hydrogen bonding relative to the

23

native cyclofructan 6 column (Frulic N).

24

exchange attributed primarily to deprotonation of residual unsubstituted silica with the

25

greater exchange ascribed to the reduced loading of CF6 observed for Larihc P. The

26

cyclofructan 6 column with a polymeric styrene divinylbenzene support (MCI GEL™

27

CRS100) showed distinct selectivities consistent with its decreased cation exchange

28

attributable to its nonionic core. The Chirobiotic T, Chirobiotic V, and ZI-DPPS columns

29

displayed hydrophilicity and ion exchange selectivities similar to other zwitterionic

30

stationary phases. All of the more hydrophilic columns showed excellent separation for

31

the four classes of therapeutic peptides investigated: microbial secondary metabolites

32

used as immune suppressants, synthetic gonadotropin hormones, synthetic cyclic

33

disulfide-linked hormone-regulating hormones, and non-ribosomally derived polycyclic

Ac ce p

te

d

M

an

us

7 8 9 10 11 12 13 14 15

Both Frulic N and Larihc P exhibited cation

1

Page 1 of 34

antibiotics. Resolution provided by these columns and ZIC-HILIC are compared for each

35

class of peptide. Frulic N is primarily suitable for use in the HILIC mode whereas

36

Chirobiotic T, because of its increased efficiency and selectivity, can be useful in both

37

HILIC and reverse phase modes. In some Chirobiotic T applications, addition of low

38

levels of a strong additive (trifluoroacetic acid, formic acid, etc.) to the mobile phase can

39

be beneficial.

40

ionic interaction between analyte and residual charge on the stationary phase improved

41

resolution and selectivity.

ip t

34

cr

In these peptide analyses, a relative weakening of the often-dominant

us

42

Keywords: HILIC, selectivity, cyclofructan stationary phase, glycopeptide stationary

44

phase, peptide drugs

an

43

45 46

1. Introduction

M

47

Hydrophilic interaction liquid chromatography (HILIC), considered a variant of the

49

normal phase modes of liquid chromatography[1-3], is characterized by the combination

50

of a stationary phase with predominantly polar functionalities and a mobile phase that is a

51

mixture of a predominantly aprotic polar organic solvent and water, often less than 30 %.

52

The preferred polar organic solvent is generally non-hydrogen bonding, most frequently

53

acetonitrile. HILIC has been shown to be useful for analyses of an exceedingly wide

54

range of polar analytes, including nucleic acids, nucleotides, and nucleosides; amino

55

acids, peptides, lipo- and glyco-peptides, and proteins of both ribosomal and

56

non-ribosomal origin; sugars, polysaccharides, and carbohydrates; and simple and

57

complex metabolites[1,4-13].

58

HILIC separations are influenced by electrostatic interactions, hydrogen-bonding,

59

dipole-dipole interaction, molecular shape selectivity reflecting steric interactions, and by

60

hydrophobicity accompanied by the exclusion of water.

Ac ce p

te

d

48

The versatility of HILIC is often exceptional[14-17].

61

Several chemometric approaches have been explored to assess analyte interactions

62

and evaluate columns [18-21]. Irgum and coworkers used principal component analysis

63

to identify and segregate patterns of retention and selectivity of more than 20 analytes on

64

more than 20 diverse commercial HILIC columns [22].

One result of the analysis was 2

Page 2 of 34

the rank ordering of the selected probes; the other was the ordering of the columns.

66

two most diverse principal components captured 70-80% of the variation in response.

67

Four groups were segregated, distinguished by the dominant functionality: positively

68

charged and negatively charged, polar nonionic, and zwitterionic functionalities, the latter

69

noted for their extensive solvation [12,22-24]. Positioning alternate HILIC columns

70

within such a framework provides means of assessing dominant interactions with

71

contrasting utility.

cr

72

The

ip t

65

Lucy and coworkers simplified and streamlined the analysis [24,25].

These

investigators selected only three retention factor ratios, two representing different

74

nonionic

75

zwitterionic-sensitive ratio was not included[12,23,26].

76

the performance of the 21 columns investigated by Irgum as well as 8 additional HILIC

77

columns and 4 reverse-phase columns and located them within three grids of retention

78

factors ratios.

79

attributes. The parameters and symbols used in the paper are summarized in Table 1.

of

interaction

and

one

measuring

cation

exchange.

A

These investigators evaluated

an

types

us

73

M

In the study reported here, we utilized that approach to establish baseline

Cyclofructans (CF’s) have a chiral 18-crown-6 structure and consist of β-(2-1) linked

81

D-fructofuranose units. Three columns with cyclofructan based stationary phases have

82

been prepared in our laboratory and reported in the literature.

83

functionalities are summarized in Fig. S1 and include native cyclofructan 6 (CF6)

84

(Frulic™ N) [27-29], an isopropyl carbamate modified cyclofructan 6 stationary phase

85

(Larihc™ CF6-P) found to be excellent for chiral [30] as well as HILIC separations [31],

86

and a 4-chloromethyl-styrene-divinylbenzene resin based CF6 column.[32][33]

Their different

Ac ce p

te

d

80

87

Two macrocyclic glycopeptide based stationary phases, Chirobiotic® T (Teicoplanin)

88

and Chirobiotic® V (Vancomycin), were introduced in the 1990’s [34-37] They consist of

89

aglycon portions of fused macrocyclic rings with a characteristic “basket” shape, yet

90

contrasting stereogenic centers, sugar moieties, and hydrophobicity that contribute to

91

distinctions in selectivity[38-41].

92

The sixth stationary phase assessed is zwitterionic, a phase with 3-P,

93

P-diphenylphosphonium-propylsulfonate covalently bound to silica gel (ZI-DPPS) [42].

94

It possesses a negatively charged sulfonate group, a positively charged quaternary

95

phosphonium group, and a diphenyl component imparting aromaticfunctionality. It 3

Page 3 of 34

96

showed improved retention, high peak efficiency and excellent peak symmetry in

97

separation of -blockers, nucleosides, and water-soluble vitamins [42,43].

98

Our study consists of two parts.

First, small, probe molecule column

characterization is used to assess the predominant modes of separation. This screening is

100

based on selectivity, the ’ s (cf. Table 1), which is just the normal  that is not inverted

101

to be greater than 1, something unnecessary when the ranking is based on a log scale.

102

Secondly, we apply the recommendations from the probe analysis to columns and mobile

103

phases for use in separation and analysis of therapeutic peptides.

104

therapeutic peptides were investigated, Cyclosporin A and C, microbial secondary

105

metabolites used as immune suppressants; Buserelin, Leuprorelin, Goserelin, and

106

Gonadorelin, synthetic gonadotropin hormones, Oxytocin, Octreotide, and Desmopressin,

107

synthetic cyclic disulfide-linked hormone-regulating hormones; and Daptomycin,

108

Teicoplanin, and Vancomycin, non-ribosomally derived polycyclic antibiotics utilized for

109

penicillin- and other antibiotic-resistant infections. The structures of the five probe

110

molecules and the different peptides from four peptide classes with quite different

111

applications (cf. Section S4) are illustrated in Fig. 1.

112

(D) and isoelectric points of the peptides are provided in Table S1.

cr

ip t

99

116 117 118

us

an

M

d

2. Experimental

Ac ce p

115

Estimated distribution coefficients

te

113 114

Four classes of

2.1. Reagents

Goserelin acetate was purchased from USBiological (Salem, MA). The cyclosporins,

119

the remaining gonadotropin salts, the cyclic hormones, the macrocyclic antibiotics and

120

cytosine, uracil, benzyltrimethylammonium (BTMA), adenosine, adenine, ammonium

121

acetate, acetic acid, trifluoroacetic acid (TFA), triethylamine (TEA), and formic acid were

122

purchased from Sigma-Aldrich (St Louis, MO).

123

as nonstoichiometric acetate salts. Acetonitrile of HPLC grade was obtained from EMD

124

Millipore and water was purified using a Milli-Q Water Purification System (Millipore,

125

Billerica, MA).

All of the gonadotropins were obtained

126 4

Page 4 of 34

127

2.2. HPLC methods

128

All experiments were conducted on Agilent HPLC series 1200 systems (Agilent

130

Technologies, Palo Alto, CA) equipped with a quaternary pump capable of virtually

131

pulsatile free and high-pressure applications, an autosampler, and a multiwavelength

132

UV-Vis detector.

133

version Rev. B.01.03 was used in a Microsoft Windows XP environment. The injection

134

volume was 5μL and analytes were separated under isocratic conditions at 1 mL/min.

135

Separations were carried out at room temperature. The dead time was determined from an

136

abrupt change in baseline associated with unretained solvent.

ip t

129

us

cr

For data acquisition and analysis, Agilent’s Chemstation Software

The three CF6 based columns (Frulic N, Larihc P, MCI GEL™ CRS100) were

138

obtained from AZYP (Arlington, TX, USA). The zwitterionic column (ZI-DPPS) was

139

prepared according to published methods [42]. Two glycopeptide-based columns

140

(Chirobiotic T, Chirobiotic V) were obtained from Advanced Separation Technologies

141

(ASTEC, Whippany, NJ, USA). The ZIC-HILIC was purchased from Merck (Darmstadt,

142

Germany). The dimensions of all columns were 250 mm×4.6 mm i.d. Silica particle size

143

is 5 μm and resins are porous spherical particles with 10-μm diameter. The characteristics

144

of the columns reported here are listed in Table 2.

te

d

M

an

137

146 147 148

Ac ce p

145

2.3. Sample Preparation

All of the peptides investigated were received as powders and were dissolved in an

149

aqueous polar solvent mixture compatible with the mobile phases used for the

150

chromatographic separations.

151

mixtures were dissolved in an 80% / 20% blend of acetonitrile and aqueous ammonium

152

acetate buffer in which the pH of the buffer was adjusted to approximately 4.1.

153

concentrations are nominally 0.5-1 mg/mL, made up to minimum volumes, and freshly

154

prepared for each column evaluation.

155

analysis of pure standards.

156

205 nm for the cyclosporin group, 280 nm for both hormone groups, and 254 nm for the

157

antibiotic group.

Unless indicated otherwise, the peptides and peptide Analyte

Analyte peak positions were identified from

Optimal wavelengths were selected by the class of peptide:

5

Page 5 of 34

158 159

3. Results and discussion

160 161

3.1. Graphical representation of HILIC stationary phases

ip t

162

Lucy et al. investigated the interactions involved in HILIC separations. They

164

classified 33 commercial columns with three 2D plots of selectivity of analyte behavior

165

for five probe molecules.

166

changes in the retention factor ratio '(cytosine,uracil); (2) cation exchange from changes

167

in '(BTMA,cytosine); and (3)“H-bonding”, from changes in '(adenosine,adenine) [24].

168

Column performance was established by the coordinates on 2-dimensional grids of these

169

three ratios. The characteristics and functionalities of 33 previously tested stationary

170

phases reported are listed in Table S2. In the present work, elution behavior of six

171

additional columns, Frulic N (34), Larihc P (35), MCI GEL™ CRS100 (36), Chirobiotic

172

T (37), Chirobiotic V (38), and ZI-DPPS (39) was determined for the five probe

173

molecules following the same approach. A ZIC-HILIC column (40), was used to validate

174

the measurements between the laboratories. The results for the six newly characterized

175

columns are positioned in the earlier grids and an additional one described below.

te

d

M

an

us

The three selectivities were (1) “Hydrophilicity”, assessed as

A conventional assessment of relative hydrophilicity, as measured by the preference

Ac ce p

176

cr

163

177

of a molecule to be dissolved in water, a tendency measured for all forms of the molecule,

178

is its distribution coefficient, D, the octanol:water partition coefficient.

179

quantification of the composite contributions to aqueous solubility.

180

directional effect on the three selectivities, ’ s, which quantify different components of

181

hydrophilic interactions in the aqueous boundary layer of the HILIC columns[10], can be

182

inferred from the differences in the log D’s, provided in Table S5.

183

selectivity ratios explore three mechanisms of hydrophilicity.

184

ranked as ion solvation from ’(BTMA,cyt), hydrogen bonding from ’(ado,adi), and

185

primary vs. secondary nitrogen solvation (N-hydrophilicity) from ’(cyt,ura).

186

five probes explore the nature of the water-enriched environment generated by the

187

distinct surface functionalities and solvent conditions of the mobile phase.

188

Experimental studies of water enrichment near practical HILIC surfaces[12] and

It provides The relative

These three

By magnitude they can be

These

6

Page 6 of 34

molecular dynamics studies of solvent density profiles in the vicinity of model

190

surfaces[10] support the influence of hydrophilic surfaces in maintaining a region

191

significantly distinct from the bulk mobile phase for distance reaching 2 nm and beyond

192

under typical HILIC conditions. The molecular dynamics results suggest ordered and

193

bound layers in close proximity to the surface, where there is a significant loss of the

194

disordering expected in the bulk.

195

region at femtosecond time scales can be gleaned from theory and experiments on bulk

196

water, which are indicative of jump reorientations influenced by incommensurate

197

geometries[44], and these are consistent with effects observed in other systems exhibiting

198

interfaces with water[45,46].

199

functionalization, and mobile phase all influence this aqueous interfacial layer, and

200

interactions with the analyte.

201

molecules.

ip t

189

us

cr

A picture of the types of dynamics occurring in this

In HILIC analyses, the effects of substrate,

an

The cumulative effect is first assessed using pairs of probe

203

M

202

3.1.1 N-Hydrophilicity vs. ion exchange characteristics of HILIC phases

d

204

The non-specific, mixed-mode hydrophilicity of the stationary phase can influence

206

the thickness of the water layer into which the partitioning of analyte can take place and

207

this can be influenced by the level of ligand loading [22,47]. The relative retention of

208

cytosine/uracil represents the nitrogen-dependent increment of hydrophilicity. The higher

209

the relative retention, the more hydrophilic the stationary phase. Ion exchange is also a

210

significant contributor to retention of ionizable solutes. The relative retention of BTMA,

211

a quaternary amine, represents the cation contribution to hydrophilicity. [24] .

Ac ce p

212

te

205

The N-hydrophilicity and ion exchange behavior of Frulic N, Larihc P, MCI GEL™

213

CRS100, Chirobiotic T, Chirobiotic V, and ZI-DPPS are provided on a selectivity grid,

214

’ s, in Fig. 2.

215

3 and identified the caption of Fig. 2.

The symbols and numbers common to Figs. 2-5 are designated in Table

216

From its position on the X-axis, Frulic N shows the third highest N hydrophilicity

217

among the 39 HILIC stationary phases, and this is attributed to the abundance CF6

218

hydroxyl groups in CF’s.

219

evaluated columns and third only to the Acclaim HILIC-10 amongst all of the forty

It is second only to Chirobiotic V amongst these newly

7

Page 7 of 34

220

columns assessed.

Larihc P exhibits less N-hydrophilicity compared with Frulic N,

221

consistent with some hydroxyls functionalized with isopropyl carbamate and in

222

agreement with characterizations from the Walters test [31].

223

hydrophobic groups into the ligand structure is anticipated to increase organic solvent

224

(e.g. acetonitrile), and diminish water, adsorption[25].

225

average hydrophilicity compared with Chirobiotic V due to its appended hydrophobic

226

side chains covalently bonded to the teicoplanin core, cf. Fig 1E. The location of the

227

ZI-DPPS column on the map is close to the cluster of zwitterionic stationary phases. The

228

hydrophilicity of the MCI GEL™ CRS100 was slightly less than amine or triazole phases

229

even though the CF6 loading of the resin support reaches 0.16 mmol/g [32].

230

often used as a marker of the dead volume in HILIC, was retained on the MCI GEL™

231

CRS100 for several minutes beyond the dead volume, suggesting the influence of

232

exposed aromatic rings of unfunctionalized stationary phase.

Introduction of

Toluene,

an

us

cr

ip t

Chirobiotic T exhibits less

The selectivity of benzyltrimethylammonium chloride, '(BTMA,cytosine), is

234

plotted on the Y-axis and extends from strong cation exchange phases at the top to

235

electrostatic repulsion hydrophilic interaction chromatography, ERLIC, dominated cation

236

repulsion [48] at the bottom. Since the pH of the eluent was maintained at 6.8, the

237

underivatized silica phases would be deprotonated, contributing a diffuse dynamic cation

238

double layer available for cation exchange with the Frulic N, Larihc P, Chirobiotic T,

239

Chirobiotic V, ZI-DPPS columns, all synthesized with porous silica supports. The

240

functional group loading levels of the Frulic N and Larihc P columns are 0.72, and 0.6

241

μmol/m2, respectively, indicative of significant surface coverage. Cation exchange, in

242

relation to the silica gel group and indicated in Fig. 2, is decreased. The

243

'(BTMA,cytosine) is systematically lower for these columns than those with

244

underivatized silica. The HILIC results for

245

with styrene divinylbenzene resin based substrates may be more appropriate for RPLC.

Ac ce p

te

d

M

233

MCI GEL™ CRS100 indicate that columns

246

From the distribution of locations of the columns on the ’(cyt,ura) coordinate in Fig.

247

2, the Chirobiotic V (2) would appear to provide nearly as high selectivity based on

248

N-hydrophilicity as the Acclaim HILIC-10 (column #27 in Table 3), with the highest

249

value of the thirty nine columns evaluated. The column is also nearly comparable to the

250

two adjacent silica columns, LiChrospher SI (60 Ǻ) (#22) and Cogent Silica-C (#23), and 8

Page 8 of 34

251

the polymer-coated silica column, PolySULFOETHYL A (#13). As indicated above and

252

based on ’(cyt,ura) , the Frulic N column (FN) can be expected to provide nearly as

253

high selectivity as the column Chirobiotic V (V).

254

(cf., Section 3.1.4, Section 4, and Table S3) characterizing cumulative retentivity also

255

indicate similarity with, and distinctions from, the silica gel columns.

256

3.1.2 Ion exchange vs. participation in hydrogen bonding of HILIC phases

cr

257

ip t

The derived parameters |kt| and |kni|,

258

In Fig. 3, ’(BTMA,cyt) is plotted on the Y-axis and ’(ado,adi), on the X-axis. The

260

columns can be ranked from those with the greatest hydrogen bonding contribution to

261

those with the least: polymer-coated/polymer substrate (11-13)

an

us

259

262

zwitterionic (1-6)

diol / polyol (9-10)

untreated silica (14-23)

263

reverse phase columns (30-33), where the numbers in parenthesis are the column

264

designations given in the leftmost column of Table 3. The Larihc P column and

265

Chirobiotic columns displayed comparable hydrogen bonding interaction to the cluster of

266

bare silica phases. Frulic N, the native CF6 based stationary phase, possessed somewhat

267

more hydrogen bonding, attributed to its abundance of hydroxyl groups.

te

d

M

amide (7-8)

amine / triazole (24-28)

Based on the solution characteristics of the probes (Table S5, last column), hydrogen

269

bonding might be anticipated to dominate the N-hydrophilicity, and thereby relative

270

column performance. However, from a comparison of the ranges in selectivity in Figs.

271

2 and 3, the effect of N-hydrophilicity is more important, varying over a larger range for

272

these four probe analytes.

273

directionally dependent hydrogen bonding may be less able to take advantage of all of the

274

hydrogen bonding sites on the probe molecules.

275 276

Ac ce p

268

This observation suggests that in the boundary layer the

3.1.3 HILIC-Phase Selectivity/Retentivity Chart

277 278

Fig. 4 represents another selectivity plot for HILIC phases. The hydrophilic

279

retentivity of the HILIC columns is measured by the retention factors of cytosine. The

280

plot of log k(cytosine), rather than the retention ratio of cytosine/uracil, a relative value,

281

reflects the effect of factors such as the pore size and surface area of the phase on the 9

Page 9 of 34

observed retention [24]. The Frulic N column with 150 mm length showed the same

283

HILIC-phase retentivity as Frulic N column with 250 mm length. Similar coordinate

284

positions were observed for ZIC-HILIC columns packed with particles having a size of

285

either 3.5 μm (col #2) or 5 μm (col #1). These results suggest that neither column length

286

nor particle size influence the fundamental HILIC separation notably. High retentivity

287

and selectivity were observed for the Frulic N column, which did have a high loading

288

ratio of CF6, and showed excellent hydrophilicity and modestly weaker ion exchange

289

interaction. The Larihc P column, which had a lower loading ratio of CF6, displayed

290

somewhat reduced hydrophilicity but stronger ion exchange interaction than the Frulic N

291

column. The Chirobiotic T, Chirobiotic V, ZI-DPPS columns showed hydrophilicity and

292

ion exchange characteristics similar to other zwitterionic stationary phases (cols#1-#6)..

293 294

3.1.4 Interaction balance and magnitude

M

295

an

us

cr

ip t

282

The balance of the different interactions also can be assessed using the metrics |kt|

297

and |kni| (defined in Table 1), vector-like sums of relatively independent contributions.

298

The ranking by |kni| and k(BTMA) is provided on the grid in Fig. 5 and numerically for

299

the columns with highest values of the metrics in Table S3. An example of the utility of

300

Fig. 5 is the consistent result observed for the analysis of the gonadotropin releasing

301

hormone agonists using the LiChrospher® SI60-5 column (22) The retention for

302

gonadorelin (Fig. S8), is about twice that for the ZI-DPPS, approximately the ratio of

303

|kni|’s.

304

column, with the ratio of k(BTMA) to |kni| approximately 5, the more closely related

305

agonists, buserelin and goserelin, are not resolved.

306

criticality of concerted and non-dominant interactions in achieving selectivity for

307

multifunctional analytes, notably peptides and proteins.

Ac ce p

te

d

296

However, because of the dominance of cation exchange with the silica gel The observation exemplifies the

308 309

3.1.5 Partitioning mechanism in HILIC mode

310 311

Chromatograms illustrating separation and resolution of a mixture of the five probe

312

analytes on all seven tested columns are provided in Fig. S2. In this series of analyses, 10

Page 10 of 34

retention time of adenosine is shorter than that of adenine for nearly all the tested

314

columns (except ZIC-HILIC and Frulic N), again emphasizing the reduced influence of

315

hydrophilic relative to lipophilic components of retention. The higher partition coefficient

316

of adenine contributed to the longer retention in spite of reduced contribution of

317

hydrogen bonding.

ip t

313

318

3.2 Separation of peptides and protein drugs

cr

319 320

The performance of six HILIC columns – Frulic N, Larihc P, Chirobiotic T,

322

Chirobiotic V, ZI-DPPS, and ZIC-HILIC – has been evaluated for the analysis of four

323

distinct classes of peptides, each containing several structurally and often functionally

324

related compounds. The interdependence of chemical structure, column design, and

325

mobile-phase composition were explored from the framework established by the probe

326

responses on these columns, as described in Section 3.1.

327

the peptide analytes was expected to break out additional distinctions in separation

328

characteristics, applicable for new therapeutics.

329

3.2.1 Separation of cyclosporin group

te

330 331

Cyclosporins, Fig. 1B, are a class of cyclic undecapeptides utilized for their

Ac ce p

332

The greater heterogeneity of

d

M

an

us

321

333

anti-immune behavior.

Seven of the amino acids of cyclosporin are N-methylated and

334

the four remaining protonated nitrogen atoms form intramolecular hydrogen bonds with

335

carbonyl groups, contributing to rigidity and reduced water solubility (Table S1). There

336

are only two compounds in this class, differing by only one atom.

337

the relative independence of the distribution coefficient on pH, the similarity in spatial

338

geometry of these compounds might be expected to pose an analytical challenge. Yet, the

339

chromatograms in Fig. 6 illustrate good separation of cyclosporin A from cyclosporin C

340

on the six tested columns, demonstrating the significant consequence of even small

341

changes in interaction. The zwitterionic stationary phases (Chirobiotic T, Chirobiotic V,

342

and ZI-DPPS) showed retention patterns similar to the cyclofructan based stationary

343

phases (Frulic N, Larihc P), though the latter provided the best resolution. The longer

In combination with

11

Page 11 of 34

344

retention observed for Larihc P indicates that just modest hydrophilicity can be important,

345

even when there is reduced preference for the boundary layer [49].

346 347

3.2.2 Separation of gonadotropin-releasing hormones (GnRH’s)

ip t

348

The second group of peptides consisted of human gonadotropin releasing hormone

350

(GnRH, also called gonadorelin) and three linear nominally nonapeptide analogues (Fig.

351

1C). For the four forms investigated, structural differences are limited to substitution of

352

the sixth amino acid residue that is a serine in gonadorelin, and differences in the

353

C-terminal group. The HILIC separations, achieved with all six columns, and reverse

354

phase separations, achieved with only the Chirobiotic, are provided in Fig. 7.. The order

355

of elution was essentially inverted upon changing modes.

an

us

cr

349

. Buserelin and leuprorelin eluted before goserelin, and gonadorelin on all columns in

357

the HILIC mode (Fig. 7 A, B, C, E, G, and H). Correlated with structural differences

358

between buserelin and leuprorelin (ethyl amide for both) and for goserelin and

359

gonadorelin (carboxamide for both) was the stronger retention for the latter. The slight

360

difference in chemical structure between buserelin and leuprorelin is in the 6th residue

361

(Ser(tBu) for buserelin, (d-Leu) for leuprorelin), making this pair the more difficult to

362

resolve. Although the analysis time was slightly long for the ZI-DPPS column, this

363

column is the only one that completely separated buserelin from leuprorelin. Note the

364

seemingly small difference in characteristics of ZI-DPPS and ZIC-HILIC were sufficient

365

to allow only ZI-DPPS to distinguish this most similar pair, and is presumed attributable

366

to its aromaticity. [42].

367

comparable, baseline, separation of these peptides, and in fact inverted the order of

368

elution of buserelin and leuprorelin, signifying the role of interactions other than

369

N-hydrophilicity. It may be useful to diminish interaction with the residual silanol groups

370

on the silica surface by supplementing the mobile phase with TFA or formic acid, and

371

adjusting the pH [39].

Ac ce p

te

d

M

356

Noteworthy, is that neither glycopeptide-based column achieved

372 373

3.2.3 Separation of low molecular weight cyclic hormones (LmwCH’s)

374

12

Page 12 of 34

375

The third group of structurally related peptides (LmwCH’s) to be investigated is a

376

selection of three hormones of cyclic structure, analogues of somatostatin (Fig. 1D), all

377

with a characteristic disulfide bridge. Separation and resolution of the LmwCH’s were

378

optimized and reported in Fig. 8.

379

elution sequence.

380

anticipated greater charge-charge interaction with residual silanols (Table S1) despite its

381

somewhat lesser hydrophilicity[50], which is expected to be shielded by the CF6’s of

382

Frulic N.

For these analytes, all but Frulic N had a consistent

cr

ip t

Octreotide eluted last for these five columns consistent with its

Retention characteristics of the LmwCH’s on the Chirobiotic columns mirrored those

384

observed for the GnRH’s; namely, retention was enhanced at about neutral or somewhat

385

lower pH (here 4.1). Retention times were relatively long. While addition of stronger

386

additives or acids (0.1%TFA/0.1% TEA or 0.1% formic acid) decreased run times

387

dramatically, oxytocin and octreotide then coeluted.

388

Frulic N and ZI-DPPS were rated as first and second, respectively (Section 3.4, Table S4).

an

us

383

M

389

3.2.4 Separation within the cyclic antibiotic peptide group (CAP)

d

390

Based on resolution and run times,

391

Unlike the previous three groups, each of which had similar structural and functional

393

properties, this fourth group consists of three distinct peptides, one pair glycopeptides

394

(vancomycin and teicoplanin) and one pair lipopeptides (teicoplanin and daptomycin)

395

(Fig. 1E). The alkyl chains are of different length (C10 for daptomycin, distribution for

396

teicoplanin) [51]. Vancomycin and teicoplanin are atropisomeric due to their phenolic

397

rotamers. All three are notably different in aqueous solubility and isoelectric points

398

(Table S1). Separations of these CAP’s under optimized mobile phase compositions are

399

provided in Fig. 9. The order of elution is preserved, independent of column or mobile

400

phase. For the most water-soluble, daptomycin, the negative charge at the buffer

401

concentration drives the more rapid elution, whereas for the least water-soluble and

402

lipophilic, teicoplanin, its weak partitioning and its slightly negative charge led to the

403

next most rapid elution.

404

providing some electrostatic attraction for the stationary phase and its absence of a

405

hydrophobic tail.

Ac ce p

te

392

Vancomycin’s slower elution relates to both its positive charge

13

Page 13 of 34

The best separation of the three glycopeptides was achieved using the native

407

cyclofructan-based column, Frulic N, which through its increased interactions with the

408

polar functionalities of the analytes, notably the glycones of the teicoplanin, slowed

409

progression along the column adequately to distinguish the different alkyl chainlengths.

410

Interestingly, retention on the Larihc P column decreased indicating the primary

411

association with these macrocyclic glycopeptides was through their polar functionalities

412

not hydrophobic. Neither the Larihc P nor the ZIC-HILIC, for which the vancomycin

413

retention was extended, provided as adequate a separation of the teicoplanins,

414

distinguishable by hydrophobe chain length and structure [51].

415

efficiency was observed for the Chirobiotic T analysis of teicoplanin, presumably related

416

to the nearly identical functionalities of peptides and stationary phase. It should be noted

417

that the poorer efficiency, and broad peaks, associated with the isocratic separations can

418

be improved significantly with gradient elution.

cr

ip t

406

us

an

M

419 420

3.3 Effect of the mobile phase on separation

d

421 422

A notable increase in

The potential of mobile phase solution conditions to influence retention, efficiency, resolution, and selectivity is well known.

424

glycopeptide [52-55] and aromatic stationary phases[56,57] can be selective for closely

425

related analytes in both HILIC and reversed phase modes. Mobile phase composition

426

and pH are critical to optimizing their performance, and these features are described in

427

the Section 3 of the Supplementary Data.

429 430 431

Previous studies have shown that

Ac ce p

428

te

423

3.4 Column-dependent peptide resolution Evaluation of column resolution from the chromatographic responses illustrated in

432

Figs. 6-9 provides a simple quantitative comparison of these five columns for the

433

four-peptide classes investigated.

434

given class are resolved, the resolution of adjacent peaks has been tabulated (Table S4).

435

The limiting resolution is defined as the smallest value of the resolution for any pair of

436

peptides in a given column’s chromatogram. Conventionally, this will be for the pair of

For the columns in which all of the peptides in a

14

Page 14 of 34

437

analytes whose elution times are nearest one another, and this limiting pair can be

438

different for different columns, since even the order of elution can be column dependent.

439

The most effective column was identified as the one whose limiting resolution is largest,

440

the one identified having the largest limiting resolution (caption for Table S4). In the simplest case, the results from the two cyclosporins indicate the Larihc P

442

column has the highest resolution, with the ZI-DPPS providing the next highest value.

443

For the GnRH’s resolution of the four peptides was achieved by only three of the

444

columns. Buserelin-leuprorelin was the limiting pair of peptides, with the ZI-DPPS and

445

Chirobiotic T providing highest and next highest resolution.

446

three LmwCH’s, and for most of them the limiting pair was octreotide-desmopressin,

447

with the best performing column being Frulic N and second, ZI-DPPS for the shorter run

448

times. If run times were not controlling, then the Chirobiotic T would be selected for its

449

high resolution of both neighboring pairs. For the CAP’s, all peaks were resolved by all

450

columns, but the limiting resolution from ZI-DPPS and Frulic N were superior.

451

the Frulic N column was capable of providing better resolution of the different

452

teicoplanin alkane analogs than even the Chirobiotic T column.

M

an

us

More columns resolved the

Notably,

d

4. Conclusions

te

453 454

cr

ip t

441

Several probe or screening characteristics of HILIC columns were demonstrated to

455

assist in selection of columns for specific applications.

457

retention factors, their ratios (’ s), and average interaction strengths (|kt| and |kni|).

458

These were complemented by design parameters associated with the columns (Tables 2 &

459

S2).

460

largest limiting resolution (Table S4) for the four different classes of peptides amongst

461

the seven columns evaluated, six newly characterized HILIC and the ZIC-HILIC column.

462

The ZI-DPPS column provided the best resolution for two classes of peptides, the

463

gonadotropin releasing hormones and the cyclic antibiotic peptides.

464

provided the best resolution for the cyclosporins and the Frulic N provided the best

465

resolution for the cyclic hormone group. The MCI GEL™ CRS100 appeared to have

466

more potential for RP than for HILIC separations.

467

necessary in order to achieve high surface coverage of resin-based columns.

Ac ce p

456

They include probe-based

These characteristics contributed to the analysis of the columns providing the

The Larihc P

Special techniques appear to be For 15

Page 15 of 34

468

example, recently Svec and coworkers have shown that gold nanoparticles attached to a

469

resin can serve as effective anchors for imparting functionalities useful in HILIC

470

separations [58] For these four classes of larger more complex analytes the diversity of interactions

472

were shown to distinguish column performance from that of the probe behavior.

473

Notably, from the probe analysis Chirobiotic V and Frulic N would appear to contrast

474

most greatly with interaction highly favoring the Chirobiotic V column; yet for two of the

475

classes of peptides, Frulic N provided resolution superior to Chirobiotic V.

476

class the Chirobiotics provided improved resolution at the price of increased retention.

477

Chirobiotic T showed higher selectivity for the pH sensitive peptides, the GnRH’s, than

478

Chirobiotic V.

479

Chirobiotic T appears to function well in both HILIC and reverse phase modes. Improved

480

selectivity was observed for the Chirobiotic columns when the charged surface of the

481

silica support was masked by a small counterion.

482

even the definitions of the components of “hydrophilicity”, could be sharpened based on

483

correlated thermodynamic and spectroscopic characterizations.

te

485

Ac ce p

486

488

In the future, such distinctions, and

d

M

an

Frulic N is primarily suitable for HILIC mode separations whereas

484

487

In a third

us

cr

ip t

471

Acknowledgments

489

The authors gratefully acknowledge Professor Irgum and Professor Lucy and their

490

coworkers for sharing their original data used in this work. Mitsubishi Chemical

491

Corporation is gratefully acknowledged for providing partial funding of this research.

492

DWA thanks the Robert A. Welch Foundation (Y-0026) for its partial support. Further, the

493

authors appreciate partial financial support from the China Scholarship Council, the

494

Natural Science Foundation of China (No. 21105008) and the Chinese Fundamental

495

Research Funds for the Central Universities (N110805001). The content is solely the

496

responsibility of the authors and does not necessarily represent the official views of these

497

funding agencies.

16

Page 16 of 34

498 499 500

References

501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547

[1]

[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]

ip t

cr

us

an

[6]

M

[5]

d

[4]

te

[3]

Ac ce p

[2]

A.J. Alpert, Hydrophilic-interaction chromatography for the separation of peptides, nucleic acids and other polar compounds, Journal of Chromatography A 499 (1990) 177-196. J. Pesek, M.T. Matyska, A comparison of two separation modes: HILIC and aqueous normal phase chromatography, (2007). A. Berthod, S.S. Chang, J.P. Kullman, D.W. Armstrong, Practice and mechanism of HPLC oligosaccharide separation with a cyclodextrin bonded phase, Talanta 47 (1998) 1001-1012. D.W. Armstrong, H.L. Jin, Evaluation of the liquid chromatographic separation of monosaccharides, disaccharides, trisaccharides, tetrasaccharides, deoxysaccharides and sugar alcohols with stable cyclodextrin bonded phase columns, Journal of chromatography 462 (1989) 219-232. A.J. Alpert, M. Shukla, A.K. Shukla, L.R. Zieske, S.W. Yuen, M.A.J. Ferguson, A. Mehlert, M. Pauly, R. Orlando, Hydrophilic-interaction chromatography of complex carbohydrates, Journal of Chromatography A 676 (1994) 191-202. W. Naidong, Bioanalytical liquid chromatography tandem mass spectrometry methods on underivatized silica columns with aqueous/organic mobile phases, Journal of Chromatography B 796 (2003) 209-224. P. Hemström, K. Irgum, Hydrophilic interaction chromatography, Journal of Separation Science 29 (2006) 1784-1821. C. Wang, C. Jiang, D.W. Armstrong, Considerations on HILIC and polar organic solvent-based separations: use of cyclodextrin and macrocyclic glycopetide stationary phases, Journal of Separation Science 31 (2008) 1980-1990. T. Ikegami, K. Tomomatsu, H. Takubo, K. Horie, N. Tanaka, Separation efficiencies in hydrophilic interaction chromatography, Journal of Chromatography A 1184 (2008) 474-503. S.M. Melnikov, A. Höltzel, A. Seidel-Morgenstern, U. Tallarek, Adsorption of Water–Acetonitrile Mixtures to Model Silica Surfaces, The Journal of Physical Chemistry C 117 (2013) 6620-6631. S.M. Melnikov, A. Höltzel, A. Seidel-Morgenstern, U. Tallarek, Composition, Structure, and Mobility of Water− Acetonitrile Mixtures in a Silica Nanopore Studied by Molecular Dynamics Simulations, Analytical chemistry 83 (2011) 2569-2575. N.P. Dinh, T. Jonsson, K. Irgum, Water uptake on polar stationary phases under conditions for hydrophilic interaction chromatography and its relation to solute retention, Journal of Chromatography A 1320 (2013) 33-47. G. Greco, T. Letzel, Main Interactions and Influences of the Chromatographic Parameters in HILIC Separations, J Chromatogr Sci 51 (2013) 684-693. P.J. Boersema, S. Mohammed, A.J.R. Heck, Hydrophilic interaction liquid chromatography (HILIC) in proteomics, Anal Bioanal Chem 391 (2008) 151-159. A.H. Honore, M. Thorsen, T. Skov, Liquid chromatography-mass spectrometry for metabolic footprinting of co-cultures of lactic and propionic acid bacteria, Anal Bioanal Chem 405 (2013) 8151-8170. C.T. Mant, Z.Q. Jiang, B.E. Boyes, R.S. Hodges, An improved approach to hydrophilic interaction chromatography of peptides: Salt gradients in the presence of high isocratic acetonitrile concentrations, Journal of Chromatography A 1277 (2013) 15-25. G. Kahsay, H. Song, A. Van Schepdael, D. Cabooter, E. Adams, Hydrophilic interaction chromatography (HILIC) in the analysis of antibiotics, Journal of Pharmaceutical and Biomedical Analysis 87 (2014) 142-154. L.R. Snyder, H. Poppe, Mechanism of solute retention in liquid-solid chromatography and the role of the mobile phase in affecting separation - competition versus sorption, Journal of Chromatography 184 (1980) 363-413. M.H. Abraham, J.C. McGowan, The use of characteristic volumes to measure cavity terms in 17

Page 17 of 34

[27] [28] [29]

[30] [31] [32] [33] [34] [35] [36] [37] [38] [39]

ip t

cr

[26]

us

[25]

an

[23] [24]

M

[22]

d

[21]

reversed phase liquid-chromatography, Chromatographia 23 (1987) 243-246. F. Gritti, G. Guiochon, Mass transfer mechanism in hydrophilic interaction chromatography, Journal of Chromatography A 1302 (2013) 55-64. D.V. McCalley, Study of the selectivity, retention mechanisms and performance of alternative silica-based stationary phases for separation of ionised solutes in hydrophilic interaction chromatography, Journal of Chromatography A 1217 (2010) 3408-3417. N.P. Dinh, T. Jonsson, K. Irgum, Probing the interaction mode in hydrophilic interaction chromatography, Journal of Chromatography A 1218 (2011) 5880-5891. R.G. Laughlin, The aqueous phase behavior of surfactants, Academic Press London, 1994. M.E.A. Ibrahim, Y. Liu, C.A. Lucy, A simple graphical representation of selectivity in hydrophilic interaction liquid chromatography, Journal of Chromatography A 1260 (2012) 126-131. S. Noga, S. Bocian, B. Buszewski, Hydrophilic interaction liquid chromatography columns classification by effect of solvation and chemometric methods, Journal of Chromatography A 1278 (2013) 89-97. O. Azzaroni, A.A. Brown, W.T. Huck, UCST Wetting Transitions of Polyzwitterionic Brushes Driven by Self‐Association, Angewandte Chemie 118 (2006) 1802-1806. H. Qiu, L. Loukotková, P. Sun, E. Tesařová, Z. Bosáková, D.W. Armstrong, Cyclofructan 6 based stationary phases for hydrophilic interaction liquid chromatography, Journal of Chromatography A 1218 (2011) 270-279. R.M. Woods, D.C. Patel, Y. Lim, Z.S. Breitbach, H. Gao, C. Keene, G. Li, L. Kürti, D.W. Armstrong, Enantiomeric separation of biaryl atropisomers using cyclofructan based chiral stationary phases, Journal of Chromatography A (2014). N.L. Padivitage, E. Dodbiba, Z.S. Breitbach, D.W. Armstrong, Enantiomeric separations of illicit drugs and controlled substances using cyclofructan‐based (LARIHC) and cyclobond I 2000 RSP HPLC chiral stationary phases, Drug testing and analysis (2013). P. Sun, D.W. Armstrong, Effective enantiomeric separations of racemic primary amines by the isopropyl carbamate-cyclofructan6 chiral stationary phase, Journal of Chromatography A 1217 (2010) 4904-4918. P. Kozlík, V. Šímová, K. Kalíková, Z. Bosáková, D.W. Armstrong, E. Tesařová, Effect of silica gel modification with cyclofructans on properties of hydrophilic interaction liquid chromatography stationary phases, Journal of Chromatography A 1257 (2012) 58-65. H. Qiu, M. Kiyono-Shimobe, D.W. Armstrong, Native/derivatized cyclofructan 6 bound to resins via “click” chemistry as stationary phases for achiral/chiral separations, Journal of Liquid Chromatography & Related Technologies 37 (2014) 2302-2326. H. Qiu, D.W. Armstrong, M. Kiyono-Shimobe, Chromatographic separation material, 2014, USPTO, D.W. Armstrong, Y.B. Tang, S.S. Chen, Y.W. Zhou, C. Bagwill, J.R. Chen, Macrocyclic Antibiotics as a New Class of Chiral Selectors for Liquid-Chromatography, Anal Chem 66 (1994) 1473-1484. A. Berthod, Y.B. Liu, C. Bagwill, D.W. Armstrong, Facile liquid chromatographic enantioresolution of native amino acids and peptides using a teicoplanin chiral stationary phase, Journal of Chromatography A 731 (1996) 123-137. K.H. Ekborg-Ott, Y. Liu, D.W. Armstrong, Highly enantioselective HPLC separations using the covalently bonded macrocyclic antibiotic, ristocetin A, chiral stationary phase, Chirality 10 (1998) 434-483. M.P. Gasper, A. Berthod, U.B. Nair, D.W. Armstrong, Comparison and Modeling Study of Vancomycin, Ristocetin A, and Teicoplanin for CE Enantioseparations, Analytical Chemistry 68 (1996) 2501-2514. A. Berthod, T.L. Xiao, Y. Liu, R.D. McCulla, W.S. Jenks, D.W. Armstrong, Separation of chiral sulfoxides by liquid chromatography using macrocyclic glycopeptide chiral stationary phases (vol 955, pg 53, 2002), Journal of Chromatography A 1047 (2004) 163-163. B. Zhang, R. Soukup, D.W. Armstrong, Selective separations of peptides with sequence deletions, single amino acid polymorphisms, and/or epimeric centers using macrocyclic glycopeptide liquid chromatography stationary phases, Journal of Chromatography A 1053 (2004) 89-99.

te

[20]

Ac ce p

548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599

18

Page 18 of 34

[44] [45] [46] [47]

[48]

[50] [51] [52] [53] [54] [55]

[56] [57] [58]

ip t

te

d

[49]

cr

[43]

us

[42]

an

[41]

R.J. Soukup-Hein, J. Schneiderheinze, P. Mehelic, D.W. Armstrong, LC and LC-MS separation of peptides on macrocyclic gly copeptide stationary phases: Diastereomeric series and large peptides, Chromatographia 66 (2007) 461-468. L. Sipos, I. Ilisz, Z. Pataj, Z. Szakonyi, F. Fulop, D.W. Armstrong, A. Peter, High-performance liquid chromatographic enantioseparation of monoterpene-based 2-amino carboxylic acids on macrocyclic glycopeptide-based phases, Journal of Chromatography A 1217 (2010) 6956-6963. H.X. Qiu, E. Wanigasekara, Y. Zhang, T. Tran, D.W. Armstrong, Development and evaluation of new zwitterionic Hydrophilic interaction liquid chromatography stationary phases based on 3-P,P-diphenylphosphonium-propylsulfonate, Journal of Chromatography A 1218 (2011) 8075-8082. H.X. Qiu, D.W. Armstrong, A. Berthod, Thermodynamic studies of a zwitterionic stationary phase in hydrophilic interaction liquid chromatography, Journal of Chromatography A 1272 (2013) 81-89. D. Laage, G. Stirnemann, F. Sterpone, J.T. Hynes, Water jump reorientation: from theoretical prediction to experimental observation, Accounts of chemical research 45 (2011) 53-62. J.C. Lang, R.D. Morgan, Nonionic surfactant mixtures. I. Phase equilibria in C10E4-H2O and closed-loop coexistence, The Journal of Chemical Physics 73 (1980) 5849-5861. C. Tanford, The Hydrophobic Effect: Formation of Micelles and Biological Membranes 2d Ed, J. Wiley., 1980. W. Bicker, J. Wu, H. Yeman, K. Albert, W. Lindner, Retention and selectivity effects caused by bonding of a polar urea-type ligand to silica: A study on mixed-mode retention mechanisms and the pivotal role of solute–silanol interactions in the hydrophilic interaction chromatography elution mode, Journal of Chromatography A 1218 (2011) 882-895. A.J. Alpert, Electrostatic Repulsion Hydrophilic Interaction Chromatography for Isocratic Separation of Charged Solutes and Selective Isolation of Phosphopeptides, Analytical Chemistry 80 (2008) 62-76. N. El Tayar, A.E. Mark, P. Vallat, R.M. Brunne, B. Testa, W.F. van Gunsteren, Solvent-dependent conformation and hydrogen-bonding capacity of cyclosporin A: evidence from partition coefficients and molecular dynamics simulations, Journal of Medicinal Chemistry 36 (1993) 3757-3764. P. Buchwald, N. Bodor, Octanol–water partition of nonzwitterionic peptides: Predictive power of a molecular size‐based model, Proteins: Structure, Function, and Bioinformatics 30 (1998) 86-99. A. Bernareggi, A. Borghi, M. Borgonovi, L. Cavenaghi, P. Ferrari, K. Vékey, M. Zanol, L. Zerilli, Teicoplanin metabolism in humans, Antimicrobial agents and chemotherapy 36 (1992) 1744-1749. P. Hägglund, J. Bunkenborg, F. Elortza, O.N. Jensen, P. Roepstorff, A new strategy for identification of N-glycosylated proteins and unambiguous assignment of their glycosylation sites using HILIC enrichment and partial deglycosylation, Journal of proteome research 3 (2004) 556-566. P. Jandera, Stationary and mobile phases in hydrophilic interaction chromatography: a review, Analytica chimica acta 692 (2011) 1-25. A. Madhavi, G. Reddy, M. Suryanarayana, A. Naidu, Chiral separation of (r, r)-tadalafil and its enantiomer in bulk drug samples and pharmaceutical dosage forms by chiral RP-LC, Chromatographia 67 (2008) 633-638. M.M. Warnke, Z.S. Breitbach, E. Dodbiba, J.A. Crank, T. Payagala, P. Sharma, E. Wanigasekara, X. Zhang, D.W. Armstrong, Positive mode electrospray ionization mass spectrometry of bisphosphonates using dicationic and tricationic ion-pairing agents, Analytica Chimica Acta 633 (2009) 232-237. M.F. Wahab, M.E. Ibrahim, C.A. Lucy, Carboxylate Modified Porous Graphitic Carbon: A New Class of Hydrophilic Interaction Liquid Chromatography Phases, Analytical chemistry 85 (2013) 5684-5691. L. Pereira, HILIC-MS Sensitivity without Silica, LC GC North America 29 (2011) 262-269. Y.Q. Lv, Z.X. Lin, F. Svec, Hypercrosslinked Large Surface Area Porous Polymer Monoliths for Hydrophilic Interaction Liquid Chromatography of Small Molecules Featuring Zwitterionic

M

[40]

Ac ce p

600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652

19

Page 19 of 34

653 654 655 656 657 658 659

Figure Captions Fig. 1. Structures of the probes and peptides. (A) Cytosine, uracil, benzyltrimethyl

660

ammonium chloride (BTMA), adenosine, and adenine Probes. (B) Cyclosporins A and

661

C.

662

gonadorelin.

663

Cyclic antibiotic peptides: daptomycin, teicoplanin, and vancomycin.

664

Fig. 2.

665

the six evaluated HILIC phases: Frulic N (circled FN), Larihc P (circled LP),

666

MCI-GELTM CRS100 (circled R), Chirobiotic T (circled T), Chirobiotic V (circled V),

667

and ZI-DPPS (circled Z) are easily located.

668

selectivities ’s defined in Table 1, of the columns evaluated by Lucy and coworkers [18],

669

and by Irgum and coworkers [17] are included for reference.

670

are small filled blue circles(●); the amine or triazole phases, open blue circles (○); amide

671

phases, open squares (□); the zwitterionic phases, crosses (+); the RPLC phases, blue x’s

672

(x); polymer or polymer-coated silica, diamonds (◆); diol or polyol, triangles (▲).The

ip t

Functionalities Attached to Gold Nanoparticles Held in Layered Structure, Anal Chem 84 (2012) 8457-8460.

(C) Gonadotropin releasing hormones: buserelin, leuprorelin, goserelin, and

cr

(D) Cyclic hormones: oxytocin, octreotide, and desmopressin.

(E)

an

us

Plot of ion exchange character, on a log scale, vs. hydrophilicity character for

The uncoated silica phases

Ac ce p

te

d

M

Retention factor ratios, rectified

T

evaluated67340

FN

R

LP

Z

V

674

newly

columns are designated with letters enclosing circles at the

676

designated locations in the chart: Frulic

677

CRS100

678

zwitterionic

679

column

680

cross-referenced in Table 3.

681

mL/min; eluent, 5 mM ammonium acetate buffer, pH 6.8, in 80% ACN; UV detection at

682

254 nm.

683

Fig. 3.

684

’s,

685

Fig. 4.

, Chirobiotic T

column

ZI-DPPS .

N

, Larihc P

, MCI GEL

, Chirobiotic V ,

and

, a a

ZIC-HILIC

The other columns are identified by number, and

The experimental conditions included: flow rate, 1.0

Plot of ion exchange character, on a log scale, vs. H-bond formation selectivities,

for the six evaluated HILIC phases. For key and conditions, see Fig. 2. The HILIC-Phase selectivity chart as a plot of ion exchange character on a log 20

Page 20 of 34

686

scale vs. s log10

of the retention factor of cytosine for the six evaluated HILIC phases.

687

For key and conditions, see Fig. 2.

688

Fig. 5.

689

scale, vs. the sum of the polar contributions to retention, measured as the vector length of

690

the cumulative contributions from the polar components of retention factors, |kni|, defined

691

in Table 1. For key and conditions, see Fig. 2.

692

Fig. 6.

693

the six tested columns. Mobile phase composition: ACN/20 mM ammonium acetate

694

buffer with the buffer ratio as indicated; pH 4.1; flow rate, 1 mL/min; UV detection, 205

695

nm. Analytes: 1, cyclosporine A; 2, cyclosporine C.

696

Fig. 7.

697

the six tested columns. Buffer used in mobile phase: 20 mM ammonium acetate buffer;

698

pH 6.5; flow rate, 1 mL/min; UV detection, 280 nm. Analytes: 1, buserelin; 2, leuprorelin;

699

3, goserelin; 4, gonadorelin.

700

*0.1%TFA and 0.1% TEA were added to the mobile phase.

701

Fig 8.

702

tested columns. Mobile phase composition, ACN/20 mM ammonium acetate buffer; pH

703

4.1; flow rate, 1 mL/min; UV detection, 280 nm. Analytes: 1, oxytocin; 2, octreotide; 3,

704

desmopressin.

705

* The retention time of compound 2 is 67 min.

706

Fig. 9.

707

tested columns.

708

enclosed by the red boxes, is exhibited by the six columns. Buffer used in mobile phase,

709

20 mM ammonium acetate buffer; pH 4.1; flow rate, 1 mL/min; UV detection, 254 nm.

710

Analytes: 1, daptomycin; 2, teicoplanin; 3, vancomycin.

711

*0.1%TFA and 0.1% TEA were added to the mobile phase.

ip t

Plot of cation exchange, measured as the retention factor of BTMA on a log

us

cr

Chromatograms of separation of cyclosporine A (1) and cyclosporine C (2) on

M

an

Chromatograms of separation of gonadotropin-releasing hormone agonists on

Ac ce p

te

d

Chromatograms of separation of oxytocin/octreotide/desmopressin on the six

Chromatograms of separation of the three cyclic antibiotic peptides on the six Different resolution of the lipid functionality of the teicoplanin,

712 713

21

Page 21 of 34

Figure 1

R

2. Uracil

3. Benzyltrimethyl- 4. Adenosine ammonium chloride (BTMA)

H

2. Cyclosporin C

OH

5. Adenine

(B) Cyclosporins

us

cr

(A) Probes

ip t

1. Cytocine

1. Cyclosporin A

an

(C) Linear gonadotropin-releasing hormones

NH O

O HN

S

O H2N

NH O

NH

O

N

NH O

O H3C CH3

O NH2

Ac ce p

1. Oxytocin

1. Daptomycin

NH

O

d

S

H2N

S

O

S

NH

OH

OH

HN

O

O

H2N

H3C

NH2

te

HO

NH

OH

O

CH3 NH

M

H C O3

O

CH3

NH O

NH

O

O

OH

H2N

NH

O O

NH NH2

NH

NH

NH

H2N

S S

NH N

NH NH

NH

O

O O

O O

NH

NH2 NH

O

O

NH2 O

O

2. Octreotide

NH

3. Desmopressin

(D) Cyclic hormones

2. Teicoplanin

3. Vancomycin

(E) Cyclic Antibiotic Peptides

Fig. 1. 1

Page 22 of 34

Figure 2

Cation Exchange vs. N-Hydrophilicity

10.00

18 17

16

7

1.00

12 2 1 4

LP

Z

3

5

13

FN

8

an

9

10

27

11

M

α'(BTMA,cyt)

32

40 6

31

V

15

us

30

19 21 20

29

22

23

cr

T

33

ip t

14

28

Ac ce p

te

d

0.10

24 26 25

R

0.01

0.00

1.00

2.00

3.00

4.00

α'(cyt,ura)

Fig. 2.

1

Page 23 of 34

Figure 3

ip t

Cation Exchange vs. Hydrogen Bonding

10.00

cr

14 18

T

V

15

31

an

32

16 17

LP

Z 30

20 21 29

us

33

22

23

19

M

FN

1

7

Ac ce p

5

13 8

10 27

11

28

te

0.10

12

9

d

α'(BTMA,cyt)

4

1.00

3

40 6

2

24

26

25

R

0.01

0.30

0.60

0.90

1.20

α'(ado,adi)

Fig. 3. 1

Page 24 of 34

Figure 4

ip t

Cation Exchange vs. Cytosine

18 16

17

21

33

32

19

20

31

LP Z

6

2

3

5

1

1.00

V

15

40 12

T

us

29

22

23

an

30

4

M

9

FN

13 8

7

10 27

11

Ac ce p

0.10

te

d

α'(BTMA,cyt)

cr

14

10.00

0.01 -1.100

-0.700

28 26

24 25

R

-0.300

0.100

0.500

Log10 (k(cyt))

Fig. 4. 1

Page 25 of 34

Figure 5

Cation Exchange vs. Mean Polar Effects

ip t

22

23

V 21

T LP

15

1

8

7

an

4

FN

9

1.00

11

d

10

M

k(BTMA)

3

5

2

13

6

Z

40

cr

16

14

us

10.00

24

Ac ce p

te

0.10

28

25

R

0.01

0.0

2.0

4.0

6.0

|kni|

Fig. 5.

1

Page 26 of 34

Figure 6

A

2 FRULIC N

ACN/ buffer=95/5

LARIHC P

ACN/ buffer=95/5

CHIROBIOTIC T

ACN/ buffer=95/5

CHIROBIOTIC V

ACN/ buffer=95/5

ZI-DPPS

ACN/ buffer=95/5

ZIC-HILIC

ACN/ buffer=98/2

1 B

2 1

D 1 E

F

2 1

1

2

3

4 5 6 t (min)

7

8

9

an

2

1

0

us

2

M

Absorbance (mv)

C

cr

1

ip t

2

Ac ce p

te

d

Fig. 6.

1

Page 27 of 34

Figure 7

4

1

3

4

D

CHIROBIOTIC T

ACN/ water*=80/20

CHIROBIOTIC T

ACN/ water*=20/80

CHIROBIOTIC V

ACN/ water*=87/13

4

3+1 2 4

1 3

4

1

G

2 1 2

1+2

3

3 4

5

10

15 t (min)

M

0

ACN/ buffer=70/30

3 2

F

H

LARIHC P

4

cr

2

ACN/ buffer=70/30

us

Absorbance (mv)

C

E

3

1+2

B

FRULIC N

ip t

3

1+2

CHIROBIOTIC V

ACN/ water*=7/93

ZI-DPPS

ACN/ buffer=70/30

ZIC-HILIC

ACN/ buffer=70/30

4

an

A

20

25

Ac ce p

te

d

Fig. 7.

1

Page 28 of 34

Figure 8

1

3

3

ACN/ buffer=70/30

3

E

2

3

CHIROBIOTIC V ACN/ buffer=50/50

2

1

3 2

ZI-DPPS

ZIC-HILIC

1 3 10 t (min)

15

ACN/ buffer=70/30

ACN/ buffer=70/30

20

Fig. 8.

Ac ce p

te

d

5

an

1

CHIROBIOTIC T ACN/ buffer=50/50

us

1

M

Absorbance (mv)

C*

0

LARIHC P

cr

1

F

ACN/ buffer=75/25

2

B

D

FRULIC N

ip t

2

A

1

Page 29 of 34

2

A

3 FRULIC N

B

2

an

1 2 3

M

Absorbance (mv)

C

2

3

1 E

2

ACN/ water*=60/40

CHIROBIOTIC V

ACN/ water*=80/20

ZI-DPPS

ACN/ buffer=60/40

d 2

1

5

10

Ac ce p

0

te

F

CHIROBIOTIC T

3

1

15

ACN/ buffer=70/30

LARIHC P

3

1

ACN/ buffer=70/30

us

1

D

cr

ip t

Figure 9

ZIC-HILIC

20

25 t (min)

30

35

ACN/ buffer=70/30

3 40

45

Fig. 9.

1

Page 30 of 34

Table 1. Definitions and symbols used in the paper.

| kt | | kni |

M an

ed

|k|

length of vector k (or quasi-vector, if components are not quite orthogonal) length of vector k corresponding to all 5 analytes length of vector k corresponding to 4 analytes, all of the polar analytes, but excluding that for cation exchange

ce pt

'(A,B)

Description retention time of dead volume retention time of analyte i retention factor of analyte i retention factor ratio ( ≥1) retention factor ratio of compounds A and B (also selectivity or separation factor) rectified selectivity of compounds A and B (RS)

cr Definition

k = (ti/t0)-1=( ti-t0)/t0 (A,B)=k(A)/k(B) if k(A) > k(B) (A,B)=k(B)/k(A) if k(B) > k(A) '(A,B)= k(A)/k(B) without constraint '(A,B)=(A,B) if k(A) > k(B) '(A,B)=(A,B)-1 if k(B) > k(A)

| kt |=(k(BMTA)2+ k(cyt)2+ k(ura)2+ k(ado)2+ (adi)2)0.5 | kni |=(k(cyt)2+ k(ura)2+ k(ado)2+ (adi)2)0.5

Ac

Symbol t0 ti k(i)  (A,B)

us

Tables.

ip t

Table 1

Note the log ('(A,B)) is symmetric about 1.

Page 31 of 34

cr

Tables.

ip t

Table 2

us

Table 2. The characteristics of the three CF6 based columns (Frulic N, Larihc P, CF6-CMS), two glycopeptide based columns (Chirobiotic T, Chirobiotic V), one zwitterionic column (ZI-DPPS), and Merck’s ZIC-HILIC column used in the current

Column

Support

Functionality

M an

investigations. Particle size (μm)

Pore size (Å)

Surface area (m2/g)

Column Column length diameter (mm) (mm)

Loading ratio (μmol/m2)

5

100

440

250

4.6

0.72

5

100

440

250

4.6

0.6

10

170

360

250

4.6

.44

silica

cyclofructan 6

Larihc P

silica

MCI-GELTM CRS100

CMS*

isopropyl carbamate cyclofructan 6 cyclofructan 6

Chirobiotic T

silica

teicoplanin

5

100

440

250

4.6

-

Chirobiotic V

silica

vancomycin

5

100

440

250

4.6

-

ZI-DPPS

silica

3-P,P-diphenylphosphoniumpropylsulfonate propyl sulphobetaine

5

100

440

250

4.6

1.93

5

200

135

250

4.6

-

ce pt

Ac

ZIC-HILIC

ed

Frulic N

Silica

* CMS: Chloromethyl modified styrene divinyl benzene

Page 32 of 34

Table 3

Tables.

ip t

Table 3. Key to the symbols, and numbering of the liquid chromatography columns in characterized in Figs. 2-5. The last column should assist in location of the column in the characterizations provided in the figures. The number designations provided in column 3 (Column) are those from the earlier papers [20,22]. Column Zwitterionic Phases 1. ZIC-HILIC 100x4.6, 5µm, 200Å 2. ZIC-HILIC 150x4.6, 3.5µm, 200Å 3. ZIC-HILIC 150x4.6, 3.5µm, 100Å 4. ZIC-pHILIC 50x4.6, 5µm 5. Nucleodur 100x4.6, 5µm, 100Å 6. Shiseido 100x4.6, 5µm, 100Å

7 8

2  

Amides 7. Tosoh Amide 80 100x4.6, 5µm 8. Tosoh Amide 80 50x4.6, 3µm

9 10

3  

Diol or Polyol 10. LiChrospher Diol 100x4, 5µm, 100Å 11. Luna 5u HILIC 100x4.6, 5µm (514356-6)

0.89 0.51

11 12 13

4   

Polymer-Coated/ Polymer Substrate 9. PolyHYDROXYETHYL A 100x4.6, 5µm, 100Å 27. Acclaim Trinity P1 12. PolySULFOETHYL A 100x4.6, 5µm, 100Å

0.34 2.00 1.28

5          

Unloaded Silica Phases 13. Chromolith SI 100x4.6, 5µm, 200Å 14. Atlantis HILIC SILICA 100x4.6, 5µm, 200Å 15. Purospher STAR SI 125x4, 5µm, 120Å 22. Atlantis HILIC 23. Onyx silica monolith 24. Zorbax HILIC plus 26. Zorbax RRHD HILIC plus 16. LiChrospher SI 100x4, 5µm, 100Å 17. LiChrospher SI 100x4, 5µm, 60Å 18. Cogent Silica-C 100x4.6, 4µm

11.48 4.41 6.02 6.30 8.00 5.10 5.10 5.05 6.52 6.30

d

M

an

us

α’(BTMA,cytosine)

te

Ac ce p

14 15 16 17 18 19 20 21 22 23

cr

1 2 3 4 5 6

Series / Symbol 1      

Col #

1.55 1.68 1.61 1.14 1.36 2.15 1.09 1.06

Page 33 of 34

7

20. Purospher NH2 100x4, 5µm, 100Å 19. LiChrospher NH2 100x4, 5µm, 100Å 28. Cosmosil HILIC 29. Acclaim HILIC-10 21. Tosoh NH2 50x4.6, 3µm

cr

α’(BTMA,cytosine)

Column

Amine or Triazole

Monolithic

0.01 0.05 0.08 0.40 0.11



25. Silica monolith coated with AS9-SC

4.80

30 31 32 33

8    

Reverse Phase 30. Zorbax Eclipse XDB-C18 31. XBridge C18 32. YMC Pro C18 33. Zorbax SB-aq

1.70 1.70 1.60 4.85

34 35 36 37 38 39 40

New FN LP R Z T V 40

d

M

29

te

Columns Evaluated

Frulic N Larihc P Resin-cyclofructan ZI-DPPS Teicoplanin Vancomycin 1*. ZIC-HILIC 250x4.6,5µm, 200Å a

Ac ce p

a

6     

us

24 25 26 27 28

Series / Symbol

an

Col #

ip t

Table 3. (cont’d)

1.23 3.21 0.04 2.97 5.21 4.30 2.23

A modified version of column #1, 2.5 times as long.

Page 34 of 34