- Email: [email protected]
Simultaneous separation of acidic and basic proteins using gemini pyrrolidinium surfactants and hexafluoroisopropanol as dynamic coating additives in capillary electrophoresis
Accepted Manuscript Title: Simultaneous separation of acidic and basic proteins using gemini pyrrolidinium surfactants and hexafluoroisopropanol as dynamic coating additives in capillary electrophoresis Author: Yu Tian Yunfang Li Jie Mei Bo Cai Jinfeng Dong Zhiguo Shi Yuxiu Xiao PII: DOI: Reference:
S0021-9673(15)01169-3 http://dx.doi.org/doi:10.1016/j.chroma.2015.08.020 CHROMA 356771
To appear in:
Journal of Chromatography A
Received date: Revised date: Accepted date:
19-3-2015 8-8-2015 10-8-2015
Please cite this article as: Y. Tian, Y. Li, J. Mei, B. Cai, J. Dong, Z. Shi, Y. Xiao, Simultaneous separation of acidic and basic proteins using gemini pyrrolidinium surfactants and hexafluoroisopropanol as dynamic coating additives in capillary electrophoresis, Journal of Chromatography A (2015), http://dx.doi.org/10.1016/j.chroma.2015.08.020 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.
Simultaneous separation of acidic and basic proteins using gemini
2
pyrrolidinium surfactants and hexafluoroisopropanol as dynamic
3
coating additives in capillary electrophoresis
ip t
1
4
Yu Tian a, Yunfang Li a, Jie Mei a, Bo Cai b, Jinfeng Dong b, Zhiguo Shi b, Yuxiu Xiao a, *
cr
5
us
6 7
a
8
Education), and School of Pharmaceutical Sciences, Wuhan University, Wuhan 430071,
9
China
10
b
11
China
an
Key Laboratory of Combinatorial Biosynthesis and Drug Discovery (Ministry of
d
12
M
College of Chemistry and Molecular Sciences, Wuhan University, Wuhan 430072,
*
14
Tel.: +86 027 68759892; Fax: +86 027 68759850.
16 17 18
Corresponding author at: 185 Donghu Road, Wuchang District, Wuhan 430071, China.
Ac ce p
15
te
13
E-mail address: [email protected] (Y. Xiao).
ABSTRACT
The separation of acidic and basic proteins using CE has been limited in part due to the
19
adsorption of proteins onto the capillary wall. In this work, the efficient control of EOF
20
and the simultaneous separation of acidic and basic proteins are achieved by use of
21
C18-4-C18PB as a dynamic coating additive, which is a representative surfactant for
22
1,1′-(butane-1,s-alkyl)bis(1-alkylpyrrolidinium) bromide (Cn-4-CnPB, n = 10, 12, 14, 16
1
Page 1 of 29
and 18). C18-4-C18PB exhibits a powerful capability in the reversal of EOF, and a low
24
concentration even less than 0.001 mM is sufficient to reverse EOF at the tested pH
25
values (3.0 ~ 9.0). Baseline separation of eight proteins with sharp peaks and high
26
efficiencies (54000–297000 plates/m) is obtained with 30 mM NaH2PO4 buffer (pH 5.0)
27
containing 4 mM C18-4-C18PB. At the same buffer condition, the Cn-4-CnPB with shorter
28
alkyl chain (n=10, 12, 14, 16) can not achieve the same effective protein separation as
29
C18-4-C18PB. However, the combined use of small amounts (less than 0.5% v/v) of
30
hexafluoroisopropanol (HFIP) and Cn-4-CnPB (n=10, 12, 14, 16) as additives can
31
completely separate all eight proteins with high efficiencies of 81000–318000 plates/m.
32
The RSDs of migration time are less than 0.80% and 5.84% for run-to-run and
33
day-to-day assays (n=5), respectively, and the protein recoveries are larger than 90.15%.
34
To the best of our knowledge, this is the first report on the simultaneous separation of
35
acidic and basic proteins using Cn-4-CnPB surfactants or Cn-4-CnPB surfactants
36
combined with HFIP as dynamic coating additives.
38 39 40 41
cr
us
an
M
d
te
Ac ce p
37
ip t
23
Keywords: Capillary electrophoresis; Dynamic coating; Pyrrolidinium gemini surfactant; Hexafluoroisopropanol; Protein separation
1. Introduction
42 43
Protein separation is critical to molecular biology, chemistry, and many other fields.
44
Mass spectrometry, high performance liquid chromatography, SDS polyacrylamide gel
2
Page 2 of 29
electrophoresis, and capillary electrophoresis (CE) have all been described as protein
46
separation tools, but CE is particularly promising due to its low sample volume
47
requirement and high separation efficiency. However, the use of CE for protein
48
separation is limited due to protein adsorption onto the fused silica capillary wall.
49
Adsorption will contaminate the capillary surface and cause a nonuniform ξ-potential
50
across the capillary [1], generating unstable and unpredictable electroosmotic flow
51
(EOF), which results in band broadening or peak asymmetry. As a consequence, protein
52
adsorption gives rise to poor repeatability of migration time, loss of separation efficiency,
53
and low protein recovery [1-4].
an
us
cr
ip t
45
Using a dynamic coating with cationic surfactant additives is one approach to
55
suppress the protein adsorption and to control the EOF. This approach also enables
56
efficient protein separation [5, 6]. It is low cost, easy to implement, and effective at
57
masking the negatively-charged silanol groups on the inner surface of the tube. The use
58
of single / double chain cationic surfactants such as cetyltrimethylammonium bromide
60 61 62
d
te
Ac ce p
59
M
54
(CTAB) / didodecyldimethylammonium bromide (DDAB) has been studied in depth, and the mechanisms by which these surfactants regulate the EOF have been discussed in detail [5-7]. Obviously, both the headgroups and the hydrophobic tails play an important role in the coating formation and the EOF control. On the one hand, the
63
positively-charged
headgroups
provide
electrostatic
attraction
with
the
64
negatively-charged capillary wall, being the driving force for the formation of surfactant
65
coating. Unfortunately, the electrostatic repulsion between the headgroups will bring
66
about a heterogenous coating surface. On the other hand, the number and length of the
3
Page 3 of 29
hydrophobic tails influence the structure of surfactant aggregates on the charged silica
68
surface, the stability of surfactant coating, and the capability of surfactants to control the
69
EOF. For instance, the bilayer coating formed by double-chained DDAB is more
70
homogeneous and with larger surface coverage, and thus more stable than the micellar
71
coating formed by conventional single-chained CTAB; furthermore, DDAB is more
72
effective in the EOF reversal [5]. Based on these findings, cationic gemini surfactants, a
73
new type of cationic surfactants, are expected to provide more effective dynamic
74
coatings for the better EOF control and protein separation compared to single-chain
75
surfactants, in that they have two hydrophobic chains and two positively-charged
76
headgroups covalently attached by an alkyl spacer, which are favorable for the formation
77
of homogeneous bilayer coating.
M
an
us
cr
ip t
67
Over the past decades, cationic gemini surfactants have gained widespread
79
recognition because of their superior properties to corresponding monomer surfactants,
80
such as better water solubility, lower critical micelle concentration (CMC) and higher
82 83 84
te
Ac ce p
81
d
78
surface activity [8-9]. However, they have rarely been utilized as dynamic coating additives for the control of EOF and the suppression of protein adsorption. To our knowledge, only Liu et al [10] have reported a series of relevant researches so far. In 2007, they first reported that a conventional (m-s-m) cationic gemini surfactant
85
[C12H25(CH3)2N(CH2)2N(CH3)2C12H25]2Br (12-2-12) could be used as a dynamic
86
coating additive in CE for the efficient regulation of EOF and suppression of basic
87
protein adsorption. Their results proved that the gemini surfactant is superior to
88
traditional single-chain surfactants in the EOF reversal, and three basic proteins were
4
Page 4 of 29
rapidly separated with satisfactory selectivity, efficiency, repeatability and recovery. In
90
2009, they further employed long-chain gemini surfactants 18-s-18 (s = 3, 4, 5, 6, 8, 10,
91
and 12) as buffer additives to simultaneously separate acidic and basic proteins, and the
92
adsorption of eight proteins was effectively suppressed by 18-5-18 [11]. This group also
93
described the use of m-s-m cationic gemini surfactants as semipermanent coatings for
94
the CE separation of proteins [12, 13]. Recently, our co-workers Cai et al. synthesized
95
and evaluated a new series of cationic gemini surfactants with pyrrolidinium headgroups,
96
1,1′-(butane-1,s-alkyl)bis(1-alkylpyrrolidinium) bromide (Cn-4-CnPB, n = 10, 12, 14, 16
97
and 18), as shown in Fig.1 [14]. It was claimed that Cn-4-CnPB surfactants containing
98
pyrrolidinium groups have lower CMC and superior surface activity compared to
99
traditional gemini surfactants. Besides, Cn-4-CnPB surfactants are similar to certain type
100
ionic liquids in structure [15], and show a great potential of being environmentally
101
friendly solvents. Considering their excellent features, we hope to employ Cn-4-CnPB as
102
dynamic coating additives in CE for EOF control and acid/basic protein separation.
104 105 106
cr
us
an
M
d
te
Ac ce p
103
ip t
89
Alternatively, organic solvents (e.g. methanol, ethanol and acetonitrile) were also
introduced as additives in CE for suppressing protein adsorption [16]. First, the organic solvents could modify the electrostatic interaction and the hydrophobic interaction between proteins and capillary wall by changing the apparent pH and the polarity of
107
buffer solution, respectively. Second, the organic solvents could modify the spatial
108
structure of proteins by influencing the hydration shell of proteins. These modification
109
effects led to the decrease of protein adsorption. Organic solvents are also highly
110
desirable in surfactants involved CE to regulate the separation selectivity [17-19]. Lucy
5
Page 5 of 29
et al reported that high content of methanol and acetonitrile up to 60% v/v could
112
significantly improve the separation selectivity of the dynamic [17] and semipermanent
113
[18] coatings of CTAB, DDAB and other longer double-chain cationic surfactants, but
114
exert little influence on the stability of the coatings. Liu et al [19] demonstrated that
115
methanol or acetonitrile up to 40% v/v could efficiently enhance the resolution and the
116
efficiency of 18-s-18 cationic gemini surfactant semipermanent coatings for separating
117
inorganic anions, without the loss of coating stability.
us
cr
ip t
111
Hexafluoroisopropanol (HFIP) is a perfluorinated alcohol with the unique properties
119
of high hydrophobicity, high ionization and strong hydrogen bond donor resulting from
120
ￚCF3 and ￚOH groups [20, 21]. Our group has currently found that HFIP can more
121
effectively prevent the adsorption of both rigid and soft proteins in CE compared with
122
acetonitrile (the work will be published later). Additionally, it was reported that a small
123
amount (less than 5% v/v) of HFIP addition could significantly improve the separation
124
selectivity of tetradecyltrimethylammonium bromide (TTAB) cationic surfactant micelle
126 127 128 129
M
d
te
Ac ce p
125
an
118
/ CTAB-SDS catanionic vesicle electrokinetic chromatography (EKC) [22, 23]. Given these findings and the unique properties of HFIP, the combined use of HFIP and Cn-4-CnPB surfactants as dynamic coating additives in CE is expected to further prevent the protein adsorption and enhance the separation selectivity for acidic and basic proteins.
130
Herein, in this work, we presented a novel study of Cn-4-CnPB surfactants or
131
Cn-4-CnPB surfactants in combination with HFIP as dynamic coating additives to
132
prevent the protein adsorption and simultaneously separate acidic and basic proteins.
6
Page 6 of 29
Taking C18-4-C18PB, which has the longest alkyl chain and the lowest CMC, as a
134
representative, we first investigated the effects of C18-4-C18PB concentration, buffer pH
135
and buffer concentration on the EOF. Then, the separation of acidic and basic proteins
136
was in detail studied using Cn-4-CnPB surfactants or Cn-4-CnPB surfactants combined
137
with HFIP as dynamic coating additives, respectively. The related mechanisms, by
138
which Cn-4-CnPB surfactants control the EOF and HFIP improves the effect of
139
Cn-4-CnPB coating on the adsorption and separation of proteins, were discussed as well.
us
cr
ip t
133
141
an
140
2. Experimental
d
144
2.1. Chemicals
te
143
M
142
The Cn-4-CnPB (n = 10, 12, 14, 16, and 18) used in this work were synthesized and
146
purified by our co-workers Cai et al. according to published methods [14]. Their
147 148 149 150
Ac ce p
145
chemical structures are shown in Fig. 1. Pepsin (1:3000, porcine), trypsin inhibitor (reagent grade, soybean), cytochrome c (≥95%，bovine heart), ribonuclease A (≥50 units/mg, bovine pancreas), HFIP, and trifluoroethanol (TFE) were purchased from Aladdin Reagent Corporation (Shanghai, China). Hemoglobin (≥99.5%, bovine
151
erythrocytes), α-chymotrypsinogen A (≥99.5%, bovine pancreas), lysozyme (≥99.5%,
152
chicken egg white), and albumin (≥99.5%, bovine serum) were purchased from
153
Shanghai Kayon Biological Technology Co., Ltd. (Shanghai, China). Sodium phosphate
154
monobasic anhydrous (NaH2PO4) and N,N-dimethylformamide (DMF) were purchased
7
Page 7 of 29
from Sinopharm Chemical Regent Co., Ltd. (Shanghai, China). Isopropanol (IPA) was
156
purchased from Kermel Chemical Regent Co., Ltd. (Tianjin, China). Deionized water
157
used was from a Milli-Q water purification system (Millipore, Molsheim, France). The
158
protein samples were freshly prepared in 30 mM NaH2PO4 aqueous solution (pH 7.0)
159
and refrigerated at 4 ± 1°C. Running buffer solutions were freshly prepared by
160
dissolving a certain amount of Cn-4-CnPB in 10-50 mM NaH2PO4 aqueous solutions (pH
161
3.0 ￚ 9.0) with or without the addition of organic modifier (HFIP, TFE or IPA).
162
Phosphoric acid aqueous solution (1 M) and sodium hydroxide aqueous solution (1 M)
163
were used for pH adjustment (Mettler Toledo pH meter, Model S20 Seven Easy,
164
Shanghai, China). All solutions were filtered through 0.45 µm Nylon membrane filters
165
(Xinya Purifying Equipments, Shanghai, China) prior to use.
M
an
us
cr
ip t
155
169 170 171 172
te
168
2.2. Apparatus
Ac ce p
167
d
166
A Beckman Coulter P/ACE MDQ instrument (Fullerton, CA, USA) equipped with a
photodiode array (PDA) detector was used for capillary electrophoresis. Samples were detected at 214 nm. Electrophoretic data were processed using 32 Karat® Software Version 8.0 (Beckman Coulter, Fullerton, CA, USA), and the data acquisition
173
rate
was
4
Hz.
Experiments
used
bare
fused-silica
capillaries
(Yongnian
174
Photoconductive Fiber Factory, Hebei, China) with an inner diameter of 50 µm and an
175
outer diameter of 365 µm. Their length was 30 cm (20 cm to the detector) in the EOF
176
control studies and 50 cm (40 cm to the detector) for protein separations. Capillaries
8
Page 8 of 29
were maintained at 25 °C. Fresh capillaries were used for different running buffer
178
solutions to avoid hysteresis effects. Capillaries were preconditioned by the following
179
procedures: 1 M HCl for 30 min, water for 20 min, 1 M NaOH for 30 min, water for 20
180
min, and running buffer for 10 min. Between runs, the capillary was rinsed for 5 min
181
with 1 M HCl, 3 min with water, 5 min with 1 M NaOH, 3 min with water, and 5 min
182
with running buffer. Each rinse step was carried out under a high pressure of 20 psi.
us
cr
ip t
177
183
2.3. EOF measurements
an
184 185
The EOF was measured by injecting 0.01% (v/v) DMF aqueous solution for 3 s at
187
0.5 psi with a constant voltage of ±10 kV relative to the direction of the EOF. The
188
magnitude of EOF (µeof) was calculated using the following standard equation [24].
189
µ eof =
191 192 193 194
d
te
Lt Ld t mV
Ac ce p
190
M
186
where Lt and Ld are the total length (30 cm) and the injector-to-detector length (20 cm) of capillary, tm is the migration time for DMF, and V is the applied voltage.
2.4. Protein separations
195
The 0.25 mg/mL protein solutions were injected into the capillary under a pressure
196
of 0.5 psi for 3 s and were separated using a constant applied voltage of -20 kV. The
197
separation efficiency (N) was based on the width at half-height of each peak and was
198
calculated with 32 Karat® Software Version 8.0 (Beckman Coulter, Fullerton, CA, USA). 9
Page 9 of 29
The run-to-run repeatability of migration time was determined for five successive
200
injections. The day-to-day repeatability of migration time was measured over five
201
consecutive days.
ip t
199
The percent recovery used to assess protein adsorption onto the capillary wall was
203
determined as described by Towns and Regnier [25]. Briefly, protein separations were
204
first performed on a 50 cm (40 cm to detector) capillary. The capillary was then reduced
205
to 30 cm (20 cm to detector) and the separations were repeated with the same applied
206
voltage, but with injection and rinse times reduced 0.6-fold. DMF was the internal
207
standard to correct for injection volume variation. We calculated protein recoveries from
208
the ratios of the peak areas obtained in the 50 cm capillary versus those in the 30 cm
209
capillary.
212 213 214 215 216
us
an
M
d te
211
3. Results and discussion
Ac ce p
210
cr
202
3.1. EOF characterization of C18-4-C18PB coating
Among Cn-4-CnPB surfactants studied in this work, C18-4-C18PB with the longest
alkyl chain has the lowest CMC and provides the strongest hydrophobic interaction
217
between their molecules, which are favorable to the formation of homogeneous bilayer
218
coating and the control of EOF. Therefore, we selected C18-4-C18PB as a representative
219
surfactant for Cn-4-CnPB to prepare dynamic coating and then investigated the EOF
220
magnitude—a known indicator of coating stability.
10
Page 10 of 29
It is well known that three factors affecting the EOF are surfactant concentration,
222
buffer pH, and ionic strength. The effects of gemini surfactant concentration and buffer
223
pH on the EOF control were initially investigated. The results are shown in Fig. 2, where
224
a reversed EOF is designated as a positive while a normal EOF as a negative. In the
225
range of the tested pH values (3.0 ~ 9.0), the magnitude of EOF increases with
226
C18-4-C18PB concentration until the reversed EOF tends to reach a plateau over 0.01
227
mM, indicating the formation of surfactants-saturated bilayer coating [7, 10]. However,
228
the minimum C18-4-C18PB concentration required for the EOF reversal is distinctly
229
dependent on the pH. At pH 3.0, only a small amount of silanol groups on the capillary
230
inner surface are ionized (the pKa of silanol is 5.3 [7]), thus a very low concentration of
231
C18-4-C18PB (less than 0.0001 mM) is sufficient to shield all Si–O– groups and
232
thoroughly suppress the normal EOF by coating with the monomeric surfactant [7]. With
233
the concentration of C18-4-C18PB further increasing, the hydrophobic attractive force
234
between the long alkyl chains increases and more C18-4-C18PB molecules adsorb onto
236 237 238
cr
us
an
M
d
te
Ac ce p
235
ip t
221
the monomeric surfactant layer until reaching saturation; herein, the reversed EOF is obtained and then tends to be constant. As the pH increases to 5.0, the relatively higher concentration of C18-4-C18PB (0.001 mM) is required to reverse the EOF because the number of ionized silanols on the capillary wall is increased. Despite this, the plateau
239
magnitude of reversed EOF at pH 5.0 is very close to that at pH 3.0, suggesting that the
240
net positive charge density on the silica wall is almost the same at the two different pH
241
conditions. Possible reason is that although the surfactant molecules adsorbing onto the
242
silica wall to form the monolayer at pH 5.0 are more than those at pH 3.0, the surfactant
11
Page 11 of 29
molecules adsorbing onto the monolayer to form the bilayer are nearly unchanged due to
244
the steric hindrance and electrostatic repulsion between the pyrrolidinium headgroups of
245
C18-4-C18PB [7].
ip t
243
At higher pH of 7.0 and 9.0, the silanol groups on the capillary wall are completely
247
ionized, providing high surface negative charge density. Therefore, the stronger
248
electrostatic attraction between C18-4-C18PB molecules and negative silanol groups takes
249
place and results in more surfactant molecules binding to the capillary wall. As a result,
250
the capillary surface is rapidly saturated, and the EOF is reversed at a lower C18-4-C18PB
251
concentration (less than 0.001 mM) at pH 7.0 and 9.0, as seen in Fig. 2. In addition, at
252
pH 7.0 and 9.0, the magnitude of the reversed EOF at the plateau is smaller than that at
253
pH 3.0 and 5.0. The reason is that the more silanols are ionized at pH 7.0 and 9.0 with
254
no concomitant increase in C18-4-C18PB binding due to capillary surface saturation,
255
leading to the decrease of the net surface positive charges. Similar phenomenon and
256
explanation were described in previous reports using CTAB and a traditional gemini
258 259 260
us
an
M
d
te
Ac ce p
257
cr
246
surfactant (12-2-12) as dynamic coating additives in CE, respectively [10, 26]. The influence of the ionic strength of buffer on the EOF behavior was then
investigated in this paper. In the study of the traditional gemini surfactant (12-2-12) dynamic coating, Tris-HCl, phosphate, and acetate buffers were compared to determine
261
the effect of ionic strength on EOF, and the results showed that the changes in the
262
concentration of phosphate buffer had the greatest effect on the EOF [10]. Based on this,
263
we selected the NaH2PO4 solutions with 0.1 mM C18-4-C18PB and the given pH as CE
264
buffer to investigate the effect of ionic strength on the EOF of C18-4-C18PB coating. The
12
Page 12 of 29
findings are shown in Fig. 3. The EOF decreases evidently as the buffer concentration
266
increases from 10 mM to 20 mM, but it displays no significant change at higher ionic
267
strengths (20-50 mM). On the one hand, as the ionic strength increases, the electrostatic
268
repulsion between the gemini headgroups decreases and the adsorption of gemini
269
surfactants onto the capillary wall increases, thus the reversed EOF should increase.
270
Nevertheless, Fig. 2 shows that at 0.1 mM C18-4-C18PB, the capillary surface is saturated
271
with the C18-4-C18PB bilayer at any pH value. Hence, the increase of the EOF resulting
272
from the increase in ionic strength can be overlooked. On the other hand, as the ionic
273
strength increases, the more anions of phosphate buffer adsorb onto the gemini bilayer,
274
leading to the shrinkage of the gemini bilayer and thus the decrease of the reversed EOF
275
[7, 10]. Moreover, the relatively high stability of the reversed EOF at the range of 20-50
276
mM phosphate buffer concentrations indicates that the shrinking effect of phosphate
277
buffer on the C18-4-C18PB bilayer is limited. That is, the C18-4-C18PB dynamic coating is
278
very stable.
280 281 282
cr
us
an
M
d
te
Ac ce p
279
ip t
265
3.2. C18-4-C18PB coating for protein separation
The representative pyrrolidinium gemini surfactant C18-4-C18PB was employed as a
283
novel dynamic coating additive to separate proteins, and compared with other
284
Cn-4-CnPB (n=10,12,14,16) in suppressing protein adsorption and improving protein
285
separation. To evaluate the performance of the surfactant coatings for protein separation,
286
a mixture of acidic proteins (pepsin, trypsin inhibitor, bovine serum albumin, and
13
Page 13 of 29
hemoglobin) and basic proteins (cytochrome c, ribonuclease A, α-chymotrypsinogen A,
288
and lysozyme) were tested in a single run; the separation selectivity and efficiency, the
289
repeatability of migration time, and the protein recovery were investigated as well.
ip t
287
The effect of C18-4-C18PB concentration (2-5 mM) on protein separation was first
291
examined in 30 mM phosphate buffer (pH 5.0). As shown in Fig. 4A, even at 2 mM
292
C18-4-C18PB, the adsorption of all eight proteins on the capillary inner surface cannot be
293
prevented efficiently and the baseline separation of ɑ-chymotrypsinogen A and
294
ribonuclease A also cannot be achieved, although a very low concentration of
295
C18-4-C18PB can completely suppress the normal EOF (0.001 mM at pH 5.0) and
296
saturate the capillary surface (0.01 mM). Clearly, at 4 mM, all proteins were baseline
297
separated with the best efficiencies ranging from 54000 to 297000 plates/m (Fig. S1A
298
and Table S1). The acidic protein pepsin has a lower efficiency than the other proteins
299
because of its lower pI value (3.2). Pepsin is negatively charged at all studied pH values
300
(5.0-9.0), and therefore can be adsorbed onto the cationic C18-4-C18PB bilayer coating
302 303 304
us
an
M
d
te
Ac ce p
301
cr
290
by electrostatic attraction, leading to a low efficiency. As the C18-4-C18PB concentration further increases to 5 mM, the peaks for cytochrome c and hemoglobin show severe peak tailing and broadening, and the separation efficiencies also decrease for most proteins (Fig. S1A). Possible explanation is that excess C18-4-C18PB may disrupt the
305
spatial structure of proteins [27], and the changes in protein structure may cause stronger
306
protein adsorption [28]. In addition, as seen in Fig. 4A, the analysis time decreases from
307
7.5 minutes to 5.5 minutes as the C18-4-C18PB concentration increasing from 2 mM to 5
308
mM. For example, the electrophoretic mobilities of lysozyme (basic protein, the first
14
Page 14 of 29
eluted) and pepsin (acid protein, the last eluted) increase from 5.26 ± 0.33×10-4 and 2.44
310
± 0.06×10-4 cm2/v.s to 7.22 ± 0.18×10-4 and 3.36 ± 0.11×10-4 cm2/v.s (n=5), respectively.
311
This may be resulted from the slight increase of EOF and the weakened retention of
312
proteins on the capillary wall. The former can be observed clearly in Fig. 2, while the
313
latter is owing to the fact that the proteins are with more positive charges due to their
314
binding with more surfactant molecules in buffer, and thus exhibit stronger electrostatic
315
repulsion for the positive bilayer coating on the capillary wall. The same phenomenon of
316
analysis time reducing with the increase of surfactant concentration was observed in the
317
CE separation of proteins using dynamic coatings of single-chain surfactant CTAB [6]
318
and traditional gemini surfactant 12-2-12 [10].
M
an
us
cr
ip t
309
To sum up, C18-4-C18PB of 4 mM is regarded as the optimal for the separation of
320
tested eight proteins in 30 mM phosphate buffer (pH 5.0). Liu et al. [11] reported the
321
separation of eight acidic and basic proteins using series of traditional gemini surfactants
322
(18-s-18, 4 ≤ s ≤8) as dynamic coatings, and found that 0.1 mM 18-5-18 in 10 mM
324 325 326
te
Ac ce p
323
d
319
phosphate buffer (pH 3) were the optimized conditions while 0.1 mM 18-4-18 in the same buffer showed a slightly poorer separation. For comparison, we also performed the separation of eight proteins tested in this work (peaks 1-4, 6 and 7 are also the five proteins involved in ref. 11) using 10 mM phosphate buffer (pH 3) with 0.1 mM
327
C18-4-C18PB as running buffer. As seen in Fig. S2, all eight proteins can be detected, and
328
the five proteins have the efficiencies of 15 000-185 000 plates/m, which are lower than
329
those obtained using 18-5-18 with the exception of bovine serum albumin (169 000
330
plates/m for C18-4-C18PB versus 7 900 plates/m for 18-5-18). It is a big regret that Liu et
15
Page 15 of 29
al. [11] did not report the efficiencies of proteins for 18-4-18. Given the poorer ability of
332
18-4-18 than 18-5-18 in protein separation, we believe that C18-4-C18PB should have a
333
comparable capability with 18-4-18 in preventing protein adsorption.
ip t
331
We then investigated the effect of buffer concentration (10-50 mM) on the protein
335
separation in phosphate buffer (pH 5.0) with 4 mM C18-4-C18PB. As shown in Fig. 4B,
336
in 10 mM phosphate buffer, there are no peaks for cytochrome c and pepsin, and the
337
other proteins are detected and well separated but with peak broadening. In the range of
338
20-50 mM, all eight proteins can be detected. From 20 mM to 30 mM, the separation
339
efficiencies of all proteins are improved (e.g., N increases from 105000 to 224000 for
340
ɑ-chymotrypsinogen A, Fig. S1B), suggesting that the increase of ionic strength can
341
suppress protein adsorption. This is perhaps related to the reduction of the ionization
342
degree of proteins and Na+ / H2PO4- competitions with the positively / negatively
343
charged proteins. However, when the buffer concentration exceeds 30 mM, the
344
efficiencies begin to decrease because of increased Joule heat. Thus, 30 mM phosphate
346 347 348
us
an
M
d
te
Ac ce p
345
cr
334
buffer provides optimal separation for the studied proteins. Next, we studied the effect of buffer pH on the protein separation in 30 mM
phosphate buffer with 4 mM C18-4-C18PB. Because the isoelectric point (pI) of the acidic and basic proteins tested is from 3.2 to 11.0, the buffer pH was set in the range of
349
5.0 to 7.0 to investigate the separation ability of C18-4-C18PB dynamic coating for the
350
proteins with different charge (positively-charged, neutral and negatively-charged). As
351
shown in Fig. 4C, all proteins were baseline separated at pH 5.0 and 5.5. Above pH 5.5,
352
the basic proteins (peaks 1-4) were well separated, while the acidic proteins (peaks 5-8)
16
Page 16 of 29
were not resolved. As the acidic proteins has more negative charge at the higher pH
354
values, their difference in charge density declines, leading to unresolved peaks. It is
355
worthy to note that ribonuclease A (pI=9.6, peak 3) migrates slower than cytochrome c
356
(pI=10.2, peak 4) at pH 6.0-7.0, being contrary to the migration sequence of these two
357
proteins at pH 5.0 and 5.5. Possible explanation is as follows. The positive charge
358
density of these two proteins, which is very close to each other, decreases at higher pH
359
values. This can strengthen the interactions between these two proteins and positive
360
surfactant C18-4-C18PB, leading to the variation of the spatial structure of these two
361
proteins and thus their reversed migration. Moreover, separation efficiency is more
362
efficient at pH 5.0 (54000-297000 plates/m) than at pH 5.5 (51000-250000 plates/m)
363
(Fig. S1C).
M
an
us
cr
ip t
353
Because the hydrophobic alkyl chain length influences the micellization capability
365
and the surface activity of Cn-4-CnPB [14], it may affect the feature of Cn-4-CnPB
366
dynamic coatings, and further affect the protein separation. Thus, we would like to
368 369 370
te
Ac ce p
367
d
364
compare the performance of Cn-4-CnPB (n=10,12,14,16,18) for the separation of eight proteins. We first investigated the effects of the concentrations of five gemini surfactants on the protein separation, respectively, and found that the concentrations of Cn-4-CnPB (n=10, 12, 14, 16, 18) needed to obtain the optimal separation for all eight proteins are 8
371
mM, 6 mM, 5 mM, 5 mM, and 4 mM, respectively, as shown in Fig. 4A and Fig. S3.
372
This indicates that the Cn-4-CnPB surfactants with longer alkyl chain have a stronger
373
capability to prevent the protein adsorption. For clear comparison, Fig. 4D further
374
describes the protein separation behaviors of five surfactant dynamic coatings at the
17
Page 17 of 29
same buffer condition (30 mM phosphate buffer at pH 5.0) and 4 mM surfactant
376
concentration, the minimum concentration required to get the optimal separation of eight
377
proteins. As seen, both C10-4-C10PB and C12-4-C12PB coating can not efficiently inhibit
378
the adsorption of all eight proteins because only three basic proteins (lysozyme,
379
ɑ-chymotrypsinogen A, ribonuclease A) and two acidic protein (bovine serum albumin
380
and hemoglobin) can be detected. Besides, ɑ-chymotrypsinogen A and ribonuclease A
381
can not baseline separated, meanwhile hemoglobin and bovine serum albumin show
382
serious peak-tailing on the C10-4-C10PB coating. It is likely that the dynamic coatings
383
formed by both short-chain surfactants at 4 mM are loose, heterogeneous, and
384
surfactant-unsaturated due to their relatively high CMC (5.91 mM for C10-4-C10PB and
385
1.04 mM for C12-4-C12PB), which results from relatively weak hydrophobic interactions
386
between short alkyl chains. That is to say, the part of the capillary surface being exposed
387
to the phosphate buffer at pH 5.0 may be negatively-charged (uncoated) or neutral
388
(monolayer coating). As a result, most proteins adsorb onto the capillary wall by
390 391 392
cr
us
an
M
d
te
Ac ce p
389
ip t
375
electrostatic and hydrophobic interactions, as exemplified by the peak missing of lysozyme or hemoglobin, cytochrome c, trypsin inhibitor, and pepsin. In contrast, for C14-4-C14PB (CMC = 0.142 mM), C16-4-C16PB (CMC = 0.0294
mM), and C18-4-C18PB (CMC = 0.00195 mM) coatings, all eight proteins not only can
393
be detected, but also display acceptable peak shape and separation selectivity (Fig. 4D).
394
We also investigated the repeatability, efficiency, and recovery of protein analysis using
395
the coatings of C14-4-C14PB and C18-4-C18PB (C16-4-C16PB was not included since two
396
proteins were not baseline separated). The parameters were determined in 30 mM
18
Page 18 of 29
phosphate buffer (pH 5.0) with 4 mM C14-4-C14PB or C18-4-C18PB, and the results are
398
listed Table S1. The repeatability is good (RSDs of migration time < 0.79% and < 2.90%
399
for run-to-run and day-to-day assays (n = 5), respectively). The sample proteins show
400
high efficiencies (131000–297000 plates/m, with the exception of pepsin) and high
401
recoveries (90.03 ± 2.38% – 104.11 ± 5.03%). All these findings indicate that the
402
adsorption of eight proteins can be suppressed with high efficiency by longer-chain
403
Cn-4-CnPB (n=14, 16, 18) coatings. This is due to the fact that for these longer-chain
404
surfactants, it is easy for them to form flat-like homogeneous bilayer coatings with
405
saturated surface coverage and high stability.
an
us
cr
ip t
397
407
M
406
3.3. HFIP-modified Cn-4-CnPB coatings for protein separation
te
d
408
Aliphatic organic solvents were widely used to improve the separation selectivity of
410
surfactants-involved CE [17-19]. Recently, it was also reported that a small amount (<
411 412 413 414
Ac ce p
409
5% v/v) of HFIP can efficiently improve the separation selectivity of surfactantsinvolved micelle and vesicle EKC [22, 23]. Especially, we have found that HFIP itself can effectively suppress the adsorption of proteins on capillary wall (the work will be published later). Herein, we studied the combined use of HFIP and Cn-4-CnPB as
415
additives for protein separation to evaluate the modification effect of HFIP on the
416
Cn-4-CnPB coatings.
417
At first, we investigated the effects of different amounts of HFIP addition (0%ￚ1.5%
418
v/v) on the protein separation in 30 mM phosphate buffer (pH 5.0) with 4 mM
19
Page 19 of 29
Cn-4-CnPB. As shown in Fig. 5A and B, for C10-4-C10PB and C12-4-C12PB dynamic
420
coatings, upon the addition of 0.25% and 0.1% (v/v) HFIP, respectively, all eight
421
proteins can be detected, indicating that the protein adsorption is apparently inhibited.
422
However, most proteins show small peak height, significant peak tailing, and rather poor
423
efficiency (Fig. S4A and B), suggesting that these amounts of HFIP are insufficient to
424
effectively prevent protein adsorption. Also, ɑ-chymotrypsinogen A and ribonuclease A
425
can not baseline separated. At 0.5% (v/v) HFIP, all proteins can be resolved with large
426
peak height and high efficiencies of 115000–318000 plate/m by both C10-4-C10PB and
427
C12-4-C12PB coating (Table S2 and S3). As the HFIP amount increases to or over 1.0%
428
(v/v), however, the apparent adsorption and the poor resolution appear again for some
429
proteins. Fig. 5C and D depict the combined use of HFIP and long-chained Cn-4-CnPB
430
(n=14,16) as dynamic coating additives for the protein separation. As seen, baseline
431
separation of ɑ-chymotrypsinogen A and ribonuclease A is readily obtained using
432
0.025% (v/v) HFIP modified Cn-4-CnPB (n=14,16) coating. The best protein separation
434 435 436
cr
us
an
M
d
te
Ac ce p
433
ip t
419
is achieved using 0.1% (v/v) HFIP in combination with C14-4-C14PB and C16-4-C16PB, respectively. Most proteins also show higher efficiencies and recoveries with 0.1% (v/v) HFIP-modified C14-4-C14PB and C16-4-C16PB than with unmodified ones (Fig. S4C and D, Table S1 and Table S4). In the case of C18-4-C18PB, the addition of HFIP larger than
437
0.025% v/v makes running buffer too viscous to perform CE analysis. Even so, the
438
0.025% v/v HFIP-modified C18-4-C18PB coating still gives better resolution for
439
ɑ-chymotrypsinogen A and ribonuclease A, higher efficiency for cytochrome c and
440
pepsin, as well as larger or equivalent recovery for all eight proteins compared to the
20
Page 20 of 29
441
unmodified coating (Fig. S5, Table S1 and Table S6). The effect of the combined use of HFIP and Cn-4-CnPB on the adsorption and
443
separation of proteins can be explained as follows. First, the –O-H group of HFIP shows
444
obvious ionization tendency due to the electron-withdrawing effect of fluorine atom.
445
Thus, HFIP can interact with the positive headgroups of Cn-4-CnPB through electrostatic
446
attraction, screening the electrostatic repulsion between surfactant molecules and
447
increasing the hydrophobic interactions between the alkyl chains of surfactant molecules,
448
which lead to the formation of more compact and more homogeneous surfactant-
449
saturated bilayer coating. Second, because of the hydrophobicity of the –CF3 groups,
450
HFIP can interact with the bilayer coating by hydrophobic interactions, competing
451
against the neutral proteins at pH 5.0. Third, HFIP also can interact with the positive
452
bilayer coating through electrostatic interaction, competing with the negative protein
453
(pepsin) at pH 5.0. All these aspects contribute to the fact that the Cn-4-CnPB in
454
combination with the proper amount of HFIP can more efficiently suppress the
456 457 458
cr
us
an
M
d
te
Ac ce p
455
ip t
442
adsorption and enhance the selectivity for not only positive proteins but neutral and negative proteins as well. However, an excessive amount of HFIP is not benefit for the protein separation because too many positive headgroups of gemini surfactants will be neutralized, which results in the part of silica wall being uncoated with surfactant
459
monolayer. It is very interesting that ɑ-chymotrypsinogen A (peak 2) is eluted after
460
ribonuclease A (peak 3) in Fig. 5B (0.25% v/v HFIP) and Fig. 5C (0.025% and 0.1% v/v
461
HFIP), but peak 2 is eluted before peak 3 in other electrograms. Both
462
a-chymotrypsinogen A and ribonuclease A are globular proteins, and they have similar
21
Page 21 of 29
pI values (8.7 and 9.6) and secondary structure [29]. So their migration times are usually
464
close to each other (Fig. 4 and Fig. 5), which provides a huge possibility for the reversal
465
of their elution order. Meanwhile, it was demonstrated that HFIP could easily bind to
466
proteins through hydrogen bond and hydrophobic interactions, inducing the alterations
467
of protein structure [30]. Thus, upon the addition of different amount of HFIP, the
468
proteins may be in different conformation transition states, possibly contributing to
469
different elution order. From all above, the unique features of HFIP, high ionizing power,
470
high hydrophobicity and strong H-bond donor ability, stemming from the groups of
471
–CF3 and –OH, determine its remarkable effects on the adsorption and separation
472
selectivity of proteins in CE using Cn-4-CnPB surfactants as dynamic coating additives.
473
This is further confirmed by the comparison studies of HFIP with other organic
474
modifiers (TFE and IPA), as shown in Fig. S5 and S6.
477 478 479 480
cr
us
an
M
d
te
476
In summary, 0.5% v/v HFIP modifier for Cn-4-CnPB (n=10, 12), 0.1% v/v HFIP modifier for Cn-4-CnPB (n=14, 16), and 0.025% v/v HFIP modifier for Cn-4-CnPB (n=18)
Ac ce p
475
ip t
463
produce the best effectiveness in inhibiting the protein adsorption and improving the protein separation, respectively. Table S2-Table S6 list the repeatability, efficiency and recovery of the protein analysis, derived using the optimal amount of HFIP and 4 mM Cn-4-CnPB (n=10-18) as additives in 30 mM phosphate buffer (pH 5.0). As seen, the
481
RSDs of migration time are less than 0.80% and 5.84% for run-to-run and day-to-day
482
assays (n = 5), respectively, the separation efficiencies higher than 133 000 plates/m
483
with the exception of pepsin (>81 000), and the recoveries are larger than 90.15%,
484
displaying satisfactory analytical results. Considering the fact that the large amounts of
22
Page 22 of 29
485
aliphatic organic solvents (generally larger than 10% v/v, even up to 60% v/v) are
486
required
487
surfactants-based CE for analytes [17-19], we have to say it is a miracle that such a
488
small amount of HFIP takes so great positive influence on the capability of Cn-4-CnPB
489
dynamic coatings in suppressing protein adsorption and improving protein separation.
improve
the
separation
selectivity
and
of
cationic
cr us
490 491
efficiency
ip t
to
4. Conclusions
an
492
In this study, we developed a novel CE method to effectively control EOF,
494
efficiently suppress protein adsorption, and simultaneously separate acidic and basic
495
proteins using the cationic gemini pyrrolidinium surfactants Cn-4-CnPB and HFIP as
496
buffer additives. The hydrophobic chain length of the Cn-4-CnPB can distinctly affect the
497
separation selectivity and efficiency. Very small amounts of HFIP addition can promote
498
the formation of more homogeneous and more compact bilayer coating, and further
500 501 502 503
d
te
Ac ce p
499
M
493
improve the separation ability of Cn-4-CnPB for proteins. We strongly believe that this simple but effective method is promising in the CE analysis of proteins and peptides. A series of related researches are still going on in our laboratory, such as the effects of the spacer group length of Cn-4-CnPB surfactants on the protein separation, Cn-4-CnPB semi-permanent coatings and their applications in CE separation.
504 505
Acknowledgements
506
23
Page 23 of 29
This work was supported by the National Nature Science Foundation of China
508
(Grant No. 81373045 and 21177099) and the Fundamental Research Funds for the
509
Central Universities (Grant No. 2014306020204).
ip t
507
510
References
cr
511 512
[1] M. Gllges, M.H. Kleemlss, G. Schomburg, Capillary zone electrophoresis
514
separations of basic and acidic proteins using poly(vinyl alcohol) coatings in fused
515
silica capillaries, Anal. Chem. 66 (1994) 2038-2046.
an
us
513
[2] N. González, C. Elvira, J.S. Román, A. Cifuentes, New physically adsorbed
517
polymer coating for reproducible separations of basic and acidic proteins by
518
capillary electrophoresis, J. Chromatogr. A 1012 (2003) 95-101.
M
516
[3] T.-F. Jiang, Y.-L. Gu, B. Liang, J.-B. Li, Y.-P. Shi, Q.-Y. Ou, Dynamically coating
520
the capillary with 1-alkyl-3-methylimidazolium-based ionic liquids for separation
521
of basic proteins by capillary electrophoresis, Anal. Chim. Acta 479 (2003)
523 524 525 526
te
Ac ce p
522
d
519
249-254.
[4] J. Li, H.-F. Han, Q. Wang, X. Liu, S.-X. Jiang, Polymeric ionic liquid as a dynamic coating additive for separation of basic proteins by capillary electrophoresis, Anal. Chim. Acta 674 (2010) 243-248.
[5] J.E. Melanson, N.E. Baryla, C.A. Lucy, Dynamic capillary coatings for
527
electroosmotic flow control in capillary electrophoresis, Trends Anal. Chem. 20
528
(2001) 365-374.
529
[6] Q. Liu, Y.-M. Yang, Y. Huang, C.-F. Pan, Z. Nie, S.-Z. Yao, Separation of acidic and
530
basic proteins by CE with CTAB additive and its applications in peptide and protein
531
profiling, Electrophoresis 30 (2009) 2151-2158. 24
Page 24 of 29
532
[7] C.A. Lucy, R.S. Underhill, Characterization of the cationic surfactant induced
533
reversal of electroosmotic flow in capillary electrophoresis, Anal. Chem. 68 (1996)
534
300-305.
537 538
ip t
536
[8] F.M. Menger, C.A. Littau, Gemini surfactants: a new class of self-assembling molecules, J. Am. Chem. Soc. 115 (1993) 10083–10090.
[9] F.M. Menger, J.S. Keiper, Gemini surfactants, Angew. Chem. Int. Ed. 39 (2000)
cr
535
1906-1920.
[10] Q. Liu, Y.-Q. Li, F. Tang, L. Ding, S.-Z. Yao, Cationic gemini surfactant as dynamic
540
coating in CE for the control of EOF and wall adsorption, Electrophoresis 28 (2007)
541
2275-2282.
an
us
539
[11] Q. Liu, Y.-Q. Li, Y.-M. Yang, S.-Z. Yao, Separation of acidic and basic proteins by
543
capillary electrophoresis using gemini surfactants and gemini-capped nanoparticles
544
as buffer additives, Sci China Ser B-Chem. 52 (2009) 1666-1676.
d
[12] Q. Liu, J.-B. Yuan, Y.-Q. Li, S.-Z. Yao, Long-chained gemini surfactants for
te
545
M
542
semipermanent
wall
coatings
547
Electrophoresis 29 (2008) 871-879.
in
capillary
electrophoresis
of
proteins,
552
Ac ce p
546
553
[15] S. Tang, S.-J. Liu, Y. Guo, X. Liu, S.-X. Jiang, Recent advances of ionic liquids and
548
[13] Q. Liu, Y.-M. Yang, S.-Z. Yao, Enhanced stability of surfactant-based
549
semipermanent wall coatings in capillary electrophoresis using oppositely charged
550 551
surfactant, J. Chromatogr. A 1187 (2008) 260-266.
[14] B. Cai, X.-F. Li, Y. Yang, J.-F. Dong, Surface properties of gemini surfactants with pyrrolidinium head groups, J. Colloid Interface Sci. 370 (2012) 111-116.
554
polymeric
ionic
liquids
in
capillary
electrophoresis
555
electrochromatography, J. Chromatogr. A 1357 (2014) 147-157.
and
capillary
25
Page 25 of 29
556
[16] A. Staub, S. Comte, S. Rudaz, J.-L. Veuthey, J. Schappler, Use of organic solvent to
557
prevent protein adsorption in CE-MS experiments, Electrophoresis 31 (2010)
558
3326-3333. [17] A.G. Diress, C.A. Lucy, Electroosmotic flow reversal for the determination of
560
inorganic anions by capillary electrophoresis with methanol-water buffers, J.
561
Chromatogr. A 1027 (2004) 185-191.
563
cr
[18] A.G. Diress, M.M. Yassine, C.A. Lucy, Semipermanent capillary coatings in mixed organic-water solvents for CE, Electrophoresis 28 (2007) 1189-1196.
us
562
ip t
559
[19] Q. Liu, Y.-Q. Li, L.-H. Yao, S.-Z. Yao, Use of gemini surfactants as semipermanent
565
capillary coatings in aqueous-organic solvents for capillary electrophoretic
566
separation of inorganic anions, J. Sep. Sci. 32 (2009) 4148-4154.
M
an
564
[20] J.-F. Berrien, M. Ourévitch, G. Morgant, N.E. Ghermani, B. Crousse, D.
568
Bonnet-Delpon, A crystalline H-bond cluster of hexafluoroisopropanol (HFIP) and
569
piperidine structure determination by X ray diffraction, J. Fluorine Chem. 128
570
(2007) 839-843.
te
d
567
[21] B. Czarnik-Matusewicz, S. Pilorz, L.-P. Zhang, Y.-Q. Wu, Structure of
572
hexafluoroisopropanol–water mixture studied by FTIR-ATR spectra and selected
576
Ac ce p
571
577
[23] Y. Tian, Y.-F. Li, J. Mei, B. Deng, Y.-X. Xiao, Hexafluoroisopropanol-modified
573
chemometric methods, J. Mol. Struct. 883 – 884 (2008) 195-202.
574
[22] C.-X. Fu, M.G. Khaledi, Selectivity patterns in micellar electrokinetic
575
chromatography characterization of fluorinated and aliphatic alcohol modifiers by micellar selectivity triangle, J. Chromatogr. A 1216 (2009) 1901-1907.
578
cetyltrimethylammoniumbromide/sodium
dodecyl
sulfate
vesicles
as
a
579
pseudostationary phasein electrokinetic chromatography, J. Chromatogr. A 1404
580
(2015) 131-140.
26
Page 26 of 29
581
[24] K.K.-C. Yeung, C.A. Lucy, Suppression of electroosmotic flow and prevention of
582
wall adsorption in capillary zone electrophoresis using zwitterionic surfactants,
583
Anal. Chem. 69 (1997) 3435-3441.
585
[25] J.K. Towns, F.E. Regnier, Capillary electrophoretic separations of proteins using nonionic surfactant coatings, Anal. Chem. 63 (1991) 1126-1132.
ip t
584
[26] G.M. Janini, K.C. Chan, J.A. Barnes, G.M. Muschik, H.J. Issaq, Separation of
587
pyridinecarboxylic acid isomers and related compounds by capillary zone
588
electrophoresis effect of cetyltrimethylammonium bromide on electroosmotic flow
589
and resolution, J. Chromatogr. A 653 (1993) 321-327.
an
us
cr
586
[27] R.R. Ogorzalekloo, N. Dales, P.C. Andrews, Surfactant effects on protein structure
591
examined by electrospray ionization mass spectrometry, Protein Sci. 3 (1994)
592
1975-1983.
M
590
[28] C.A. Lucy, A.M. MacDonald, M.D. Gulcev, Review Non-covalent capillary
594
coatings for protein separations in capillary electrophoresis, J. Chromatogr. A 1184
595
(2008) 81-105.
te
d
593
[29] J.L. Koenig, B.G. Frushour, Raman scattering of chymotrypsinogen A, ribonuclease,
597
and ovalbumin in the aqueous solution and solid state, Biopolymers 11 (1972)
598 599 600
Ac ce p
596
2505-2520.
[30] D.S. Dwyer, R.J. Bradley, Chemical properties of alcohols and their protein binding sites, Cell. Mol. Life Sci. 57 (2000) 265-275.
601
27
Page 27 of 29
Figure Captions
602
Fig. 1. Chemical structures of cationic gemini surfactants with pyrrolidinium head
603
groups (Cn-4-CnPB, n = 10, 12, 14, 16, and 18).
604
Fig. 2. The relationship between EOF and the concentration of C18-4-C18PB at various
605
buffer pH values. Running buffer: 0–1 mM C18-4-C18PB in 10 mM NaH2PO4.
606
Fig. 3. Effect of the concentration of phosphate buffer on EOF. Running buffer: 0.1 mM
607
C18-4-C18PB in10-50 mM NaH2PO4 at pH 5.0, 7.0 and 9.0.
608
Fig. 4. Effects of C18-4-C18PB concentration (A), buffer concentration (B), buffer pH (C),
609
and the hydrophobic chain length of Cn-4-CnPB (n = 10, 12, 14, 16, or 18) (D) on the
610
separation of proteins. Peak designation: 1. lysozyme, 2. α-chymotrypsinogen A, 3.
611
ribonuclease A, 4. cytochrome c, 5. hemoglobin, 6. bovine serum albumin, 7. trypsin
612
inhibitor, and 8. pepsin. Running buffer: 2–5 mM C18-4-C18PB in 30 mM NaH2PO4 (pH
613
5) (A), 4 mM C18-4-C18PB in 10-50 mM NaH2PO4 (pH 5) (B), 4 mM C18-4-C18PB in 30
614
mM NaH2PO4 (pH 5-7) (C), and 4 mM Cn-4-CnPB (n = 10, 12, 14, 16, or 18) in 30 mM
616 617 618
cr
us
an
M
d
te
Ac ce p
615
ip t
601
NaH2PO4 (pH 5) (D).
Fig. 5. Effect of HFIP content (v/v) on protein separation in 30 mM NaH2PO4 (pH 5.0) with 4 mM C10-4-C10PB (A), C12-4-C12PB (B), C14-4-C14PB (C), or C16-4-C16PB (D). Peak designations are the same as in Fig. 4.
619 620
Highlights for JCA-15-479 revision
621 622 623
Gemini surfactants (Cn-4-nPB) were used as novel dynamic coating additives in CE.
28
Page 28 of 29
HFIP can improve the effect of Cn-4-nPB on the adsorption and separation of proteins.
Neutral and charged proteins were well separated with 115000–318000 plates/m.
ip t
624 625 626 627 628 629 630 631
Ac ce p
te
d
M
an
us
cr
632
29
Page 29 of 29
S0021-9673(15)01169-3 http://dx.doi.org/doi:10.1016/j.chroma.2015.08.020 CHROMA 356771
To appear in:
Journal of Chromatography A
Received date: Revised date: Accepted date:
19-3-2015 8-8-2015 10-8-2015
Please cite this article as: Y. Tian, Y. Li, J. Mei, B. Cai, J. Dong, Z. Shi, Y. Xiao, Simultaneous separation of acidic and basic proteins using gemini pyrrolidinium surfactants and hexafluoroisopropanol as dynamic coating additives in capillary electrophoresis, Journal of Chromatography A (2015), http://dx.doi.org/10.1016/j.chroma.2015.08.020 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.
Simultaneous separation of acidic and basic proteins using gemini
2
pyrrolidinium surfactants and hexafluoroisopropanol as dynamic
3
coating additives in capillary electrophoresis
ip t
1
4
Yu Tian a, Yunfang Li a, Jie Mei a, Bo Cai b, Jinfeng Dong b, Zhiguo Shi b, Yuxiu Xiao a, *
cr
5
us
6 7
a
8
Education), and School of Pharmaceutical Sciences, Wuhan University, Wuhan 430071,
9
China
10
b
11
China
an
Key Laboratory of Combinatorial Biosynthesis and Drug Discovery (Ministry of
d
12
M
College of Chemistry and Molecular Sciences, Wuhan University, Wuhan 430072,
*
14
Tel.: +86 027 68759892; Fax: +86 027 68759850.
16 17 18
Corresponding author at: 185 Donghu Road, Wuchang District, Wuhan 430071, China.
Ac ce p
15
te
13
E-mail address: [email protected] (Y. Xiao).
ABSTRACT
The separation of acidic and basic proteins using CE has been limited in part due to the
19
adsorption of proteins onto the capillary wall. In this work, the efficient control of EOF
20
and the simultaneous separation of acidic and basic proteins are achieved by use of
21
C18-4-C18PB as a dynamic coating additive, which is a representative surfactant for
22
1,1′-(butane-1,s-alkyl)bis(1-alkylpyrrolidinium) bromide (Cn-4-CnPB, n = 10, 12, 14, 16
1
Page 1 of 29
and 18). C18-4-C18PB exhibits a powerful capability in the reversal of EOF, and a low
24
concentration even less than 0.001 mM is sufficient to reverse EOF at the tested pH
25
values (3.0 ~ 9.0). Baseline separation of eight proteins with sharp peaks and high
26
efficiencies (54000–297000 plates/m) is obtained with 30 mM NaH2PO4 buffer (pH 5.0)
27
containing 4 mM C18-4-C18PB. At the same buffer condition, the Cn-4-CnPB with shorter
28
alkyl chain (n=10, 12, 14, 16) can not achieve the same effective protein separation as
29
C18-4-C18PB. However, the combined use of small amounts (less than 0.5% v/v) of
30
hexafluoroisopropanol (HFIP) and Cn-4-CnPB (n=10, 12, 14, 16) as additives can
31
completely separate all eight proteins with high efficiencies of 81000–318000 plates/m.
32
The RSDs of migration time are less than 0.80% and 5.84% for run-to-run and
33
day-to-day assays (n=5), respectively, and the protein recoveries are larger than 90.15%.
34
To the best of our knowledge, this is the first report on the simultaneous separation of
35
acidic and basic proteins using Cn-4-CnPB surfactants or Cn-4-CnPB surfactants
36
combined with HFIP as dynamic coating additives.
38 39 40 41
cr
us
an
M
d
te
Ac ce p
37
ip t
23
Keywords: Capillary electrophoresis; Dynamic coating; Pyrrolidinium gemini surfactant; Hexafluoroisopropanol; Protein separation
1. Introduction
42 43
Protein separation is critical to molecular biology, chemistry, and many other fields.
44
Mass spectrometry, high performance liquid chromatography, SDS polyacrylamide gel
2
Page 2 of 29
electrophoresis, and capillary electrophoresis (CE) have all been described as protein
46
separation tools, but CE is particularly promising due to its low sample volume
47
requirement and high separation efficiency. However, the use of CE for protein
48
separation is limited due to protein adsorption onto the fused silica capillary wall.
49
Adsorption will contaminate the capillary surface and cause a nonuniform ξ-potential
50
across the capillary [1], generating unstable and unpredictable electroosmotic flow
51
(EOF), which results in band broadening or peak asymmetry. As a consequence, protein
52
adsorption gives rise to poor repeatability of migration time, loss of separation efficiency,
53
and low protein recovery [1-4].
an
us
cr
ip t
45
Using a dynamic coating with cationic surfactant additives is one approach to
55
suppress the protein adsorption and to control the EOF. This approach also enables
56
efficient protein separation [5, 6]. It is low cost, easy to implement, and effective at
57
masking the negatively-charged silanol groups on the inner surface of the tube. The use
58
of single / double chain cationic surfactants such as cetyltrimethylammonium bromide
60 61 62
d
te
Ac ce p
59
M
54
(CTAB) / didodecyldimethylammonium bromide (DDAB) has been studied in depth, and the mechanisms by which these surfactants regulate the EOF have been discussed in detail [5-7]. Obviously, both the headgroups and the hydrophobic tails play an important role in the coating formation and the EOF control. On the one hand, the
63
positively-charged
headgroups
provide
electrostatic
attraction
with
the
64
negatively-charged capillary wall, being the driving force for the formation of surfactant
65
coating. Unfortunately, the electrostatic repulsion between the headgroups will bring
66
about a heterogenous coating surface. On the other hand, the number and length of the
3
Page 3 of 29
hydrophobic tails influence the structure of surfactant aggregates on the charged silica
68
surface, the stability of surfactant coating, and the capability of surfactants to control the
69
EOF. For instance, the bilayer coating formed by double-chained DDAB is more
70
homogeneous and with larger surface coverage, and thus more stable than the micellar
71
coating formed by conventional single-chained CTAB; furthermore, DDAB is more
72
effective in the EOF reversal [5]. Based on these findings, cationic gemini surfactants, a
73
new type of cationic surfactants, are expected to provide more effective dynamic
74
coatings for the better EOF control and protein separation compared to single-chain
75
surfactants, in that they have two hydrophobic chains and two positively-charged
76
headgroups covalently attached by an alkyl spacer, which are favorable for the formation
77
of homogeneous bilayer coating.
M
an
us
cr
ip t
67
Over the past decades, cationic gemini surfactants have gained widespread
79
recognition because of their superior properties to corresponding monomer surfactants,
80
such as better water solubility, lower critical micelle concentration (CMC) and higher
82 83 84
te
Ac ce p
81
d
78
surface activity [8-9]. However, they have rarely been utilized as dynamic coating additives for the control of EOF and the suppression of protein adsorption. To our knowledge, only Liu et al [10] have reported a series of relevant researches so far. In 2007, they first reported that a conventional (m-s-m) cationic gemini surfactant
85
[C12H25(CH3)2N(CH2)2N(CH3)2C12H25]2Br (12-2-12) could be used as a dynamic
86
coating additive in CE for the efficient regulation of EOF and suppression of basic
87
protein adsorption. Their results proved that the gemini surfactant is superior to
88
traditional single-chain surfactants in the EOF reversal, and three basic proteins were
4
Page 4 of 29
rapidly separated with satisfactory selectivity, efficiency, repeatability and recovery. In
90
2009, they further employed long-chain gemini surfactants 18-s-18 (s = 3, 4, 5, 6, 8, 10,
91
and 12) as buffer additives to simultaneously separate acidic and basic proteins, and the
92
adsorption of eight proteins was effectively suppressed by 18-5-18 [11]. This group also
93
described the use of m-s-m cationic gemini surfactants as semipermanent coatings for
94
the CE separation of proteins [12, 13]. Recently, our co-workers Cai et al. synthesized
95
and evaluated a new series of cationic gemini surfactants with pyrrolidinium headgroups,
96
1,1′-(butane-1,s-alkyl)bis(1-alkylpyrrolidinium) bromide (Cn-4-CnPB, n = 10, 12, 14, 16
97
and 18), as shown in Fig.1 [14]. It was claimed that Cn-4-CnPB surfactants containing
98
pyrrolidinium groups have lower CMC and superior surface activity compared to
99
traditional gemini surfactants. Besides, Cn-4-CnPB surfactants are similar to certain type
100
ionic liquids in structure [15], and show a great potential of being environmentally
101
friendly solvents. Considering their excellent features, we hope to employ Cn-4-CnPB as
102
dynamic coating additives in CE for EOF control and acid/basic protein separation.
104 105 106
cr
us
an
M
d
te
Ac ce p
103
ip t
89
Alternatively, organic solvents (e.g. methanol, ethanol and acetonitrile) were also
introduced as additives in CE for suppressing protein adsorption [16]. First, the organic solvents could modify the electrostatic interaction and the hydrophobic interaction between proteins and capillary wall by changing the apparent pH and the polarity of
107
buffer solution, respectively. Second, the organic solvents could modify the spatial
108
structure of proteins by influencing the hydration shell of proteins. These modification
109
effects led to the decrease of protein adsorption. Organic solvents are also highly
110
desirable in surfactants involved CE to regulate the separation selectivity [17-19]. Lucy
5
Page 5 of 29
et al reported that high content of methanol and acetonitrile up to 60% v/v could
112
significantly improve the separation selectivity of the dynamic [17] and semipermanent
113
[18] coatings of CTAB, DDAB and other longer double-chain cationic surfactants, but
114
exert little influence on the stability of the coatings. Liu et al [19] demonstrated that
115
methanol or acetonitrile up to 40% v/v could efficiently enhance the resolution and the
116
efficiency of 18-s-18 cationic gemini surfactant semipermanent coatings for separating
117
inorganic anions, without the loss of coating stability.
us
cr
ip t
111
Hexafluoroisopropanol (HFIP) is a perfluorinated alcohol with the unique properties
119
of high hydrophobicity, high ionization and strong hydrogen bond donor resulting from
120
ￚCF3 and ￚOH groups [20, 21]. Our group has currently found that HFIP can more
121
effectively prevent the adsorption of both rigid and soft proteins in CE compared with
122
acetonitrile (the work will be published later). Additionally, it was reported that a small
123
amount (less than 5% v/v) of HFIP addition could significantly improve the separation
124
selectivity of tetradecyltrimethylammonium bromide (TTAB) cationic surfactant micelle
126 127 128 129
M
d
te
Ac ce p
125
an
118
/ CTAB-SDS catanionic vesicle electrokinetic chromatography (EKC) [22, 23]. Given these findings and the unique properties of HFIP, the combined use of HFIP and Cn-4-CnPB surfactants as dynamic coating additives in CE is expected to further prevent the protein adsorption and enhance the separation selectivity for acidic and basic proteins.
130
Herein, in this work, we presented a novel study of Cn-4-CnPB surfactants or
131
Cn-4-CnPB surfactants in combination with HFIP as dynamic coating additives to
132
prevent the protein adsorption and simultaneously separate acidic and basic proteins.
6
Page 6 of 29
Taking C18-4-C18PB, which has the longest alkyl chain and the lowest CMC, as a
134
representative, we first investigated the effects of C18-4-C18PB concentration, buffer pH
135
and buffer concentration on the EOF. Then, the separation of acidic and basic proteins
136
was in detail studied using Cn-4-CnPB surfactants or Cn-4-CnPB surfactants combined
137
with HFIP as dynamic coating additives, respectively. The related mechanisms, by
138
which Cn-4-CnPB surfactants control the EOF and HFIP improves the effect of
139
Cn-4-CnPB coating on the adsorption and separation of proteins, were discussed as well.
us
cr
ip t
133
141
an
140
2. Experimental
d
144
2.1. Chemicals
te
143
M
142
The Cn-4-CnPB (n = 10, 12, 14, 16, and 18) used in this work were synthesized and
146
purified by our co-workers Cai et al. according to published methods [14]. Their
147 148 149 150
Ac ce p
145
chemical structures are shown in Fig. 1. Pepsin (1:3000, porcine), trypsin inhibitor (reagent grade, soybean), cytochrome c (≥95%，bovine heart), ribonuclease A (≥50 units/mg, bovine pancreas), HFIP, and trifluoroethanol (TFE) were purchased from Aladdin Reagent Corporation (Shanghai, China). Hemoglobin (≥99.5%, bovine
151
erythrocytes), α-chymotrypsinogen A (≥99.5%, bovine pancreas), lysozyme (≥99.5%,
152
chicken egg white), and albumin (≥99.5%, bovine serum) were purchased from
153
Shanghai Kayon Biological Technology Co., Ltd. (Shanghai, China). Sodium phosphate
154
monobasic anhydrous (NaH2PO4) and N,N-dimethylformamide (DMF) were purchased
7
Page 7 of 29
from Sinopharm Chemical Regent Co., Ltd. (Shanghai, China). Isopropanol (IPA) was
156
purchased from Kermel Chemical Regent Co., Ltd. (Tianjin, China). Deionized water
157
used was from a Milli-Q water purification system (Millipore, Molsheim, France). The
158
protein samples were freshly prepared in 30 mM NaH2PO4 aqueous solution (pH 7.0)
159
and refrigerated at 4 ± 1°C. Running buffer solutions were freshly prepared by
160
dissolving a certain amount of Cn-4-CnPB in 10-50 mM NaH2PO4 aqueous solutions (pH
161
3.0 ￚ 9.0) with or without the addition of organic modifier (HFIP, TFE or IPA).
162
Phosphoric acid aqueous solution (1 M) and sodium hydroxide aqueous solution (1 M)
163
were used for pH adjustment (Mettler Toledo pH meter, Model S20 Seven Easy,
164
Shanghai, China). All solutions were filtered through 0.45 µm Nylon membrane filters
165
(Xinya Purifying Equipments, Shanghai, China) prior to use.
M
an
us
cr
ip t
155
169 170 171 172
te
168
2.2. Apparatus
Ac ce p
167
d
166
A Beckman Coulter P/ACE MDQ instrument (Fullerton, CA, USA) equipped with a
photodiode array (PDA) detector was used for capillary electrophoresis. Samples were detected at 214 nm. Electrophoretic data were processed using 32 Karat® Software Version 8.0 (Beckman Coulter, Fullerton, CA, USA), and the data acquisition
173
rate
was
4
Hz.
Experiments
used
bare
fused-silica
capillaries
(Yongnian
174
Photoconductive Fiber Factory, Hebei, China) with an inner diameter of 50 µm and an
175
outer diameter of 365 µm. Their length was 30 cm (20 cm to the detector) in the EOF
176
control studies and 50 cm (40 cm to the detector) for protein separations. Capillaries
8
Page 8 of 29
were maintained at 25 °C. Fresh capillaries were used for different running buffer
178
solutions to avoid hysteresis effects. Capillaries were preconditioned by the following
179
procedures: 1 M HCl for 30 min, water for 20 min, 1 M NaOH for 30 min, water for 20
180
min, and running buffer for 10 min. Between runs, the capillary was rinsed for 5 min
181
with 1 M HCl, 3 min with water, 5 min with 1 M NaOH, 3 min with water, and 5 min
182
with running buffer. Each rinse step was carried out under a high pressure of 20 psi.
us
cr
ip t
177
183
2.3. EOF measurements
an
184 185
The EOF was measured by injecting 0.01% (v/v) DMF aqueous solution for 3 s at
187
0.5 psi with a constant voltage of ±10 kV relative to the direction of the EOF. The
188
magnitude of EOF (µeof) was calculated using the following standard equation [24].
189
µ eof =
191 192 193 194
d
te
Lt Ld t mV
Ac ce p
190
M
186
where Lt and Ld are the total length (30 cm) and the injector-to-detector length (20 cm) of capillary, tm is the migration time for DMF, and V is the applied voltage.
2.4. Protein separations
195
The 0.25 mg/mL protein solutions were injected into the capillary under a pressure
196
of 0.5 psi for 3 s and were separated using a constant applied voltage of -20 kV. The
197
separation efficiency (N) was based on the width at half-height of each peak and was
198
calculated with 32 Karat® Software Version 8.0 (Beckman Coulter, Fullerton, CA, USA). 9
Page 9 of 29
The run-to-run repeatability of migration time was determined for five successive
200
injections. The day-to-day repeatability of migration time was measured over five
201
consecutive days.
ip t
199
The percent recovery used to assess protein adsorption onto the capillary wall was
203
determined as described by Towns and Regnier [25]. Briefly, protein separations were
204
first performed on a 50 cm (40 cm to detector) capillary. The capillary was then reduced
205
to 30 cm (20 cm to detector) and the separations were repeated with the same applied
206
voltage, but with injection and rinse times reduced 0.6-fold. DMF was the internal
207
standard to correct for injection volume variation. We calculated protein recoveries from
208
the ratios of the peak areas obtained in the 50 cm capillary versus those in the 30 cm
209
capillary.
212 213 214 215 216
us
an
M
d te
211
3. Results and discussion
Ac ce p
210
cr
202
3.1. EOF characterization of C18-4-C18PB coating
Among Cn-4-CnPB surfactants studied in this work, C18-4-C18PB with the longest
alkyl chain has the lowest CMC and provides the strongest hydrophobic interaction
217
between their molecules, which are favorable to the formation of homogeneous bilayer
218
coating and the control of EOF. Therefore, we selected C18-4-C18PB as a representative
219
surfactant for Cn-4-CnPB to prepare dynamic coating and then investigated the EOF
220
magnitude—a known indicator of coating stability.
10
Page 10 of 29
It is well known that three factors affecting the EOF are surfactant concentration,
222
buffer pH, and ionic strength. The effects of gemini surfactant concentration and buffer
223
pH on the EOF control were initially investigated. The results are shown in Fig. 2, where
224
a reversed EOF is designated as a positive while a normal EOF as a negative. In the
225
range of the tested pH values (3.0 ~ 9.0), the magnitude of EOF increases with
226
C18-4-C18PB concentration until the reversed EOF tends to reach a plateau over 0.01
227
mM, indicating the formation of surfactants-saturated bilayer coating [7, 10]. However,
228
the minimum C18-4-C18PB concentration required for the EOF reversal is distinctly
229
dependent on the pH. At pH 3.0, only a small amount of silanol groups on the capillary
230
inner surface are ionized (the pKa of silanol is 5.3 [7]), thus a very low concentration of
231
C18-4-C18PB (less than 0.0001 mM) is sufficient to shield all Si–O– groups and
232
thoroughly suppress the normal EOF by coating with the monomeric surfactant [7]. With
233
the concentration of C18-4-C18PB further increasing, the hydrophobic attractive force
234
between the long alkyl chains increases and more C18-4-C18PB molecules adsorb onto
236 237 238
cr
us
an
M
d
te
Ac ce p
235
ip t
221
the monomeric surfactant layer until reaching saturation; herein, the reversed EOF is obtained and then tends to be constant. As the pH increases to 5.0, the relatively higher concentration of C18-4-C18PB (0.001 mM) is required to reverse the EOF because the number of ionized silanols on the capillary wall is increased. Despite this, the plateau
239
magnitude of reversed EOF at pH 5.0 is very close to that at pH 3.0, suggesting that the
240
net positive charge density on the silica wall is almost the same at the two different pH
241
conditions. Possible reason is that although the surfactant molecules adsorbing onto the
242
silica wall to form the monolayer at pH 5.0 are more than those at pH 3.0, the surfactant
11
Page 11 of 29
molecules adsorbing onto the monolayer to form the bilayer are nearly unchanged due to
244
the steric hindrance and electrostatic repulsion between the pyrrolidinium headgroups of
245
C18-4-C18PB [7].
ip t
243
At higher pH of 7.0 and 9.0, the silanol groups on the capillary wall are completely
247
ionized, providing high surface negative charge density. Therefore, the stronger
248
electrostatic attraction between C18-4-C18PB molecules and negative silanol groups takes
249
place and results in more surfactant molecules binding to the capillary wall. As a result,
250
the capillary surface is rapidly saturated, and the EOF is reversed at a lower C18-4-C18PB
251
concentration (less than 0.001 mM) at pH 7.0 and 9.0, as seen in Fig. 2. In addition, at
252
pH 7.0 and 9.0, the magnitude of the reversed EOF at the plateau is smaller than that at
253
pH 3.0 and 5.0. The reason is that the more silanols are ionized at pH 7.0 and 9.0 with
254
no concomitant increase in C18-4-C18PB binding due to capillary surface saturation,
255
leading to the decrease of the net surface positive charges. Similar phenomenon and
256
explanation were described in previous reports using CTAB and a traditional gemini
258 259 260
us
an
M
d
te
Ac ce p
257
cr
246
surfactant (12-2-12) as dynamic coating additives in CE, respectively [10, 26]. The influence of the ionic strength of buffer on the EOF behavior was then
investigated in this paper. In the study of the traditional gemini surfactant (12-2-12) dynamic coating, Tris-HCl, phosphate, and acetate buffers were compared to determine
261
the effect of ionic strength on EOF, and the results showed that the changes in the
262
concentration of phosphate buffer had the greatest effect on the EOF [10]. Based on this,
263
we selected the NaH2PO4 solutions with 0.1 mM C18-4-C18PB and the given pH as CE
264
buffer to investigate the effect of ionic strength on the EOF of C18-4-C18PB coating. The
12
Page 12 of 29
findings are shown in Fig. 3. The EOF decreases evidently as the buffer concentration
266
increases from 10 mM to 20 mM, but it displays no significant change at higher ionic
267
strengths (20-50 mM). On the one hand, as the ionic strength increases, the electrostatic
268
repulsion between the gemini headgroups decreases and the adsorption of gemini
269
surfactants onto the capillary wall increases, thus the reversed EOF should increase.
270
Nevertheless, Fig. 2 shows that at 0.1 mM C18-4-C18PB, the capillary surface is saturated
271
with the C18-4-C18PB bilayer at any pH value. Hence, the increase of the EOF resulting
272
from the increase in ionic strength can be overlooked. On the other hand, as the ionic
273
strength increases, the more anions of phosphate buffer adsorb onto the gemini bilayer,
274
leading to the shrinkage of the gemini bilayer and thus the decrease of the reversed EOF
275
[7, 10]. Moreover, the relatively high stability of the reversed EOF at the range of 20-50
276
mM phosphate buffer concentrations indicates that the shrinking effect of phosphate
277
buffer on the C18-4-C18PB bilayer is limited. That is, the C18-4-C18PB dynamic coating is
278
very stable.
280 281 282
cr
us
an
M
d
te
Ac ce p
279
ip t
265
3.2. C18-4-C18PB coating for protein separation
The representative pyrrolidinium gemini surfactant C18-4-C18PB was employed as a
283
novel dynamic coating additive to separate proteins, and compared with other
284
Cn-4-CnPB (n=10,12,14,16) in suppressing protein adsorption and improving protein
285
separation. To evaluate the performance of the surfactant coatings for protein separation,
286
a mixture of acidic proteins (pepsin, trypsin inhibitor, bovine serum albumin, and
13
Page 13 of 29
hemoglobin) and basic proteins (cytochrome c, ribonuclease A, α-chymotrypsinogen A,
288
and lysozyme) were tested in a single run; the separation selectivity and efficiency, the
289
repeatability of migration time, and the protein recovery were investigated as well.
ip t
287
The effect of C18-4-C18PB concentration (2-5 mM) on protein separation was first
291
examined in 30 mM phosphate buffer (pH 5.0). As shown in Fig. 4A, even at 2 mM
292
C18-4-C18PB, the adsorption of all eight proteins on the capillary inner surface cannot be
293
prevented efficiently and the baseline separation of ɑ-chymotrypsinogen A and
294
ribonuclease A also cannot be achieved, although a very low concentration of
295
C18-4-C18PB can completely suppress the normal EOF (0.001 mM at pH 5.0) and
296
saturate the capillary surface (0.01 mM). Clearly, at 4 mM, all proteins were baseline
297
separated with the best efficiencies ranging from 54000 to 297000 plates/m (Fig. S1A
298
and Table S1). The acidic protein pepsin has a lower efficiency than the other proteins
299
because of its lower pI value (3.2). Pepsin is negatively charged at all studied pH values
300
(5.0-9.0), and therefore can be adsorbed onto the cationic C18-4-C18PB bilayer coating
302 303 304
us
an
M
d
te
Ac ce p
301
cr
290
by electrostatic attraction, leading to a low efficiency. As the C18-4-C18PB concentration further increases to 5 mM, the peaks for cytochrome c and hemoglobin show severe peak tailing and broadening, and the separation efficiencies also decrease for most proteins (Fig. S1A). Possible explanation is that excess C18-4-C18PB may disrupt the
305
spatial structure of proteins [27], and the changes in protein structure may cause stronger
306
protein adsorption [28]. In addition, as seen in Fig. 4A, the analysis time decreases from
307
7.5 minutes to 5.5 minutes as the C18-4-C18PB concentration increasing from 2 mM to 5
308
mM. For example, the electrophoretic mobilities of lysozyme (basic protein, the first
14
Page 14 of 29
eluted) and pepsin (acid protein, the last eluted) increase from 5.26 ± 0.33×10-4 and 2.44
310
± 0.06×10-4 cm2/v.s to 7.22 ± 0.18×10-4 and 3.36 ± 0.11×10-4 cm2/v.s (n=5), respectively.
311
This may be resulted from the slight increase of EOF and the weakened retention of
312
proteins on the capillary wall. The former can be observed clearly in Fig. 2, while the
313
latter is owing to the fact that the proteins are with more positive charges due to their
314
binding with more surfactant molecules in buffer, and thus exhibit stronger electrostatic
315
repulsion for the positive bilayer coating on the capillary wall. The same phenomenon of
316
analysis time reducing with the increase of surfactant concentration was observed in the
317
CE separation of proteins using dynamic coatings of single-chain surfactant CTAB [6]
318
and traditional gemini surfactant 12-2-12 [10].
M
an
us
cr
ip t
309
To sum up, C18-4-C18PB of 4 mM is regarded as the optimal for the separation of
320
tested eight proteins in 30 mM phosphate buffer (pH 5.0). Liu et al. [11] reported the
321
separation of eight acidic and basic proteins using series of traditional gemini surfactants
322
(18-s-18, 4 ≤ s ≤8) as dynamic coatings, and found that 0.1 mM 18-5-18 in 10 mM
324 325 326
te
Ac ce p
323
d
319
phosphate buffer (pH 3) were the optimized conditions while 0.1 mM 18-4-18 in the same buffer showed a slightly poorer separation. For comparison, we also performed the separation of eight proteins tested in this work (peaks 1-4, 6 and 7 are also the five proteins involved in ref. 11) using 10 mM phosphate buffer (pH 3) with 0.1 mM
327
C18-4-C18PB as running buffer. As seen in Fig. S2, all eight proteins can be detected, and
328
the five proteins have the efficiencies of 15 000-185 000 plates/m, which are lower than
329
those obtained using 18-5-18 with the exception of bovine serum albumin (169 000
330
plates/m for C18-4-C18PB versus 7 900 plates/m for 18-5-18). It is a big regret that Liu et
15
Page 15 of 29
al. [11] did not report the efficiencies of proteins for 18-4-18. Given the poorer ability of
332
18-4-18 than 18-5-18 in protein separation, we believe that C18-4-C18PB should have a
333
comparable capability with 18-4-18 in preventing protein adsorption.
ip t
331
We then investigated the effect of buffer concentration (10-50 mM) on the protein
335
separation in phosphate buffer (pH 5.0) with 4 mM C18-4-C18PB. As shown in Fig. 4B,
336
in 10 mM phosphate buffer, there are no peaks for cytochrome c and pepsin, and the
337
other proteins are detected and well separated but with peak broadening. In the range of
338
20-50 mM, all eight proteins can be detected. From 20 mM to 30 mM, the separation
339
efficiencies of all proteins are improved (e.g., N increases from 105000 to 224000 for
340
ɑ-chymotrypsinogen A, Fig. S1B), suggesting that the increase of ionic strength can
341
suppress protein adsorption. This is perhaps related to the reduction of the ionization
342
degree of proteins and Na+ / H2PO4- competitions with the positively / negatively
343
charged proteins. However, when the buffer concentration exceeds 30 mM, the
344
efficiencies begin to decrease because of increased Joule heat. Thus, 30 mM phosphate
346 347 348
us
an
M
d
te
Ac ce p
345
cr
334
buffer provides optimal separation for the studied proteins. Next, we studied the effect of buffer pH on the protein separation in 30 mM
phosphate buffer with 4 mM C18-4-C18PB. Because the isoelectric point (pI) of the acidic and basic proteins tested is from 3.2 to 11.0, the buffer pH was set in the range of
349
5.0 to 7.0 to investigate the separation ability of C18-4-C18PB dynamic coating for the
350
proteins with different charge (positively-charged, neutral and negatively-charged). As
351
shown in Fig. 4C, all proteins were baseline separated at pH 5.0 and 5.5. Above pH 5.5,
352
the basic proteins (peaks 1-4) were well separated, while the acidic proteins (peaks 5-8)
16
Page 16 of 29
were not resolved. As the acidic proteins has more negative charge at the higher pH
354
values, their difference in charge density declines, leading to unresolved peaks. It is
355
worthy to note that ribonuclease A (pI=9.6, peak 3) migrates slower than cytochrome c
356
(pI=10.2, peak 4) at pH 6.0-7.0, being contrary to the migration sequence of these two
357
proteins at pH 5.0 and 5.5. Possible explanation is as follows. The positive charge
358
density of these two proteins, which is very close to each other, decreases at higher pH
359
values. This can strengthen the interactions between these two proteins and positive
360
surfactant C18-4-C18PB, leading to the variation of the spatial structure of these two
361
proteins and thus their reversed migration. Moreover, separation efficiency is more
362
efficient at pH 5.0 (54000-297000 plates/m) than at pH 5.5 (51000-250000 plates/m)
363
(Fig. S1C).
M
an
us
cr
ip t
353
Because the hydrophobic alkyl chain length influences the micellization capability
365
and the surface activity of Cn-4-CnPB [14], it may affect the feature of Cn-4-CnPB
366
dynamic coatings, and further affect the protein separation. Thus, we would like to
368 369 370
te
Ac ce p
367
d
364
compare the performance of Cn-4-CnPB (n=10,12,14,16,18) for the separation of eight proteins. We first investigated the effects of the concentrations of five gemini surfactants on the protein separation, respectively, and found that the concentrations of Cn-4-CnPB (n=10, 12, 14, 16, 18) needed to obtain the optimal separation for all eight proteins are 8
371
mM, 6 mM, 5 mM, 5 mM, and 4 mM, respectively, as shown in Fig. 4A and Fig. S3.
372
This indicates that the Cn-4-CnPB surfactants with longer alkyl chain have a stronger
373
capability to prevent the protein adsorption. For clear comparison, Fig. 4D further
374
describes the protein separation behaviors of five surfactant dynamic coatings at the
17
Page 17 of 29
same buffer condition (30 mM phosphate buffer at pH 5.0) and 4 mM surfactant
376
concentration, the minimum concentration required to get the optimal separation of eight
377
proteins. As seen, both C10-4-C10PB and C12-4-C12PB coating can not efficiently inhibit
378
the adsorption of all eight proteins because only three basic proteins (lysozyme,
379
ɑ-chymotrypsinogen A, ribonuclease A) and two acidic protein (bovine serum albumin
380
and hemoglobin) can be detected. Besides, ɑ-chymotrypsinogen A and ribonuclease A
381
can not baseline separated, meanwhile hemoglobin and bovine serum albumin show
382
serious peak-tailing on the C10-4-C10PB coating. It is likely that the dynamic coatings
383
formed by both short-chain surfactants at 4 mM are loose, heterogeneous, and
384
surfactant-unsaturated due to their relatively high CMC (5.91 mM for C10-4-C10PB and
385
1.04 mM for C12-4-C12PB), which results from relatively weak hydrophobic interactions
386
between short alkyl chains. That is to say, the part of the capillary surface being exposed
387
to the phosphate buffer at pH 5.0 may be negatively-charged (uncoated) or neutral
388
(monolayer coating). As a result, most proteins adsorb onto the capillary wall by
390 391 392
cr
us
an
M
d
te
Ac ce p
389
ip t
375
electrostatic and hydrophobic interactions, as exemplified by the peak missing of lysozyme or hemoglobin, cytochrome c, trypsin inhibitor, and pepsin. In contrast, for C14-4-C14PB (CMC = 0.142 mM), C16-4-C16PB (CMC = 0.0294
mM), and C18-4-C18PB (CMC = 0.00195 mM) coatings, all eight proteins not only can
393
be detected, but also display acceptable peak shape and separation selectivity (Fig. 4D).
394
We also investigated the repeatability, efficiency, and recovery of protein analysis using
395
the coatings of C14-4-C14PB and C18-4-C18PB (C16-4-C16PB was not included since two
396
proteins were not baseline separated). The parameters were determined in 30 mM
18
Page 18 of 29
phosphate buffer (pH 5.0) with 4 mM C14-4-C14PB or C18-4-C18PB, and the results are
398
listed Table S1. The repeatability is good (RSDs of migration time < 0.79% and < 2.90%
399
for run-to-run and day-to-day assays (n = 5), respectively). The sample proteins show
400
high efficiencies (131000–297000 plates/m, with the exception of pepsin) and high
401
recoveries (90.03 ± 2.38% – 104.11 ± 5.03%). All these findings indicate that the
402
adsorption of eight proteins can be suppressed with high efficiency by longer-chain
403
Cn-4-CnPB (n=14, 16, 18) coatings. This is due to the fact that for these longer-chain
404
surfactants, it is easy for them to form flat-like homogeneous bilayer coatings with
405
saturated surface coverage and high stability.
an
us
cr
ip t
397
407
M
406
3.3. HFIP-modified Cn-4-CnPB coatings for protein separation
te
d
408
Aliphatic organic solvents were widely used to improve the separation selectivity of
410
surfactants-involved CE [17-19]. Recently, it was also reported that a small amount (<
411 412 413 414
Ac ce p
409
5% v/v) of HFIP can efficiently improve the separation selectivity of surfactantsinvolved micelle and vesicle EKC [22, 23]. Especially, we have found that HFIP itself can effectively suppress the adsorption of proteins on capillary wall (the work will be published later). Herein, we studied the combined use of HFIP and Cn-4-CnPB as
415
additives for protein separation to evaluate the modification effect of HFIP on the
416
Cn-4-CnPB coatings.
417
At first, we investigated the effects of different amounts of HFIP addition (0%ￚ1.5%
418
v/v) on the protein separation in 30 mM phosphate buffer (pH 5.0) with 4 mM
19
Page 19 of 29
Cn-4-CnPB. As shown in Fig. 5A and B, for C10-4-C10PB and C12-4-C12PB dynamic
420
coatings, upon the addition of 0.25% and 0.1% (v/v) HFIP, respectively, all eight
421
proteins can be detected, indicating that the protein adsorption is apparently inhibited.
422
However, most proteins show small peak height, significant peak tailing, and rather poor
423
efficiency (Fig. S4A and B), suggesting that these amounts of HFIP are insufficient to
424
effectively prevent protein adsorption. Also, ɑ-chymotrypsinogen A and ribonuclease A
425
can not baseline separated. At 0.5% (v/v) HFIP, all proteins can be resolved with large
426
peak height and high efficiencies of 115000–318000 plate/m by both C10-4-C10PB and
427
C12-4-C12PB coating (Table S2 and S3). As the HFIP amount increases to or over 1.0%
428
(v/v), however, the apparent adsorption and the poor resolution appear again for some
429
proteins. Fig. 5C and D depict the combined use of HFIP and long-chained Cn-4-CnPB
430
(n=14,16) as dynamic coating additives for the protein separation. As seen, baseline
431
separation of ɑ-chymotrypsinogen A and ribonuclease A is readily obtained using
432
0.025% (v/v) HFIP modified Cn-4-CnPB (n=14,16) coating. The best protein separation
434 435 436
cr
us
an
M
d
te
Ac ce p
433
ip t
419
is achieved using 0.1% (v/v) HFIP in combination with C14-4-C14PB and C16-4-C16PB, respectively. Most proteins also show higher efficiencies and recoveries with 0.1% (v/v) HFIP-modified C14-4-C14PB and C16-4-C16PB than with unmodified ones (Fig. S4C and D, Table S1 and Table S4). In the case of C18-4-C18PB, the addition of HFIP larger than
437
0.025% v/v makes running buffer too viscous to perform CE analysis. Even so, the
438
0.025% v/v HFIP-modified C18-4-C18PB coating still gives better resolution for
439
ɑ-chymotrypsinogen A and ribonuclease A, higher efficiency for cytochrome c and
440
pepsin, as well as larger or equivalent recovery for all eight proteins compared to the
20
Page 20 of 29
441
unmodified coating (Fig. S5, Table S1 and Table S6). The effect of the combined use of HFIP and Cn-4-CnPB on the adsorption and
443
separation of proteins can be explained as follows. First, the –O-H group of HFIP shows
444
obvious ionization tendency due to the electron-withdrawing effect of fluorine atom.
445
Thus, HFIP can interact with the positive headgroups of Cn-4-CnPB through electrostatic
446
attraction, screening the electrostatic repulsion between surfactant molecules and
447
increasing the hydrophobic interactions between the alkyl chains of surfactant molecules,
448
which lead to the formation of more compact and more homogeneous surfactant-
449
saturated bilayer coating. Second, because of the hydrophobicity of the –CF3 groups,
450
HFIP can interact with the bilayer coating by hydrophobic interactions, competing
451
against the neutral proteins at pH 5.0. Third, HFIP also can interact with the positive
452
bilayer coating through electrostatic interaction, competing with the negative protein
453
(pepsin) at pH 5.0. All these aspects contribute to the fact that the Cn-4-CnPB in
454
combination with the proper amount of HFIP can more efficiently suppress the
456 457 458
cr
us
an
M
d
te
Ac ce p
455
ip t
442
adsorption and enhance the selectivity for not only positive proteins but neutral and negative proteins as well. However, an excessive amount of HFIP is not benefit for the protein separation because too many positive headgroups of gemini surfactants will be neutralized, which results in the part of silica wall being uncoated with surfactant
459
monolayer. It is very interesting that ɑ-chymotrypsinogen A (peak 2) is eluted after
460
ribonuclease A (peak 3) in Fig. 5B (0.25% v/v HFIP) and Fig. 5C (0.025% and 0.1% v/v
461
HFIP), but peak 2 is eluted before peak 3 in other electrograms. Both
462
a-chymotrypsinogen A and ribonuclease A are globular proteins, and they have similar
21
Page 21 of 29
pI values (8.7 and 9.6) and secondary structure [29]. So their migration times are usually
464
close to each other (Fig. 4 and Fig. 5), which provides a huge possibility for the reversal
465
of their elution order. Meanwhile, it was demonstrated that HFIP could easily bind to
466
proteins through hydrogen bond and hydrophobic interactions, inducing the alterations
467
of protein structure [30]. Thus, upon the addition of different amount of HFIP, the
468
proteins may be in different conformation transition states, possibly contributing to
469
different elution order. From all above, the unique features of HFIP, high ionizing power,
470
high hydrophobicity and strong H-bond donor ability, stemming from the groups of
471
–CF3 and –OH, determine its remarkable effects on the adsorption and separation
472
selectivity of proteins in CE using Cn-4-CnPB surfactants as dynamic coating additives.
473
This is further confirmed by the comparison studies of HFIP with other organic
474
modifiers (TFE and IPA), as shown in Fig. S5 and S6.
477 478 479 480
cr
us
an
M
d
te
476
In summary, 0.5% v/v HFIP modifier for Cn-4-CnPB (n=10, 12), 0.1% v/v HFIP modifier for Cn-4-CnPB (n=14, 16), and 0.025% v/v HFIP modifier for Cn-4-CnPB (n=18)
Ac ce p
475
ip t
463
produce the best effectiveness in inhibiting the protein adsorption and improving the protein separation, respectively. Table S2-Table S6 list the repeatability, efficiency and recovery of the protein analysis, derived using the optimal amount of HFIP and 4 mM Cn-4-CnPB (n=10-18) as additives in 30 mM phosphate buffer (pH 5.0). As seen, the
481
RSDs of migration time are less than 0.80% and 5.84% for run-to-run and day-to-day
482
assays (n = 5), respectively, the separation efficiencies higher than 133 000 plates/m
483
with the exception of pepsin (>81 000), and the recoveries are larger than 90.15%,
484
displaying satisfactory analytical results. Considering the fact that the large amounts of
22
Page 22 of 29
485
aliphatic organic solvents (generally larger than 10% v/v, even up to 60% v/v) are
486
required
487
surfactants-based CE for analytes [17-19], we have to say it is a miracle that such a
488
small amount of HFIP takes so great positive influence on the capability of Cn-4-CnPB
489
dynamic coatings in suppressing protein adsorption and improving protein separation.
improve
the
separation
selectivity
and
of
cationic
cr us
490 491
efficiency
ip t
to
4. Conclusions
an
492
In this study, we developed a novel CE method to effectively control EOF,
494
efficiently suppress protein adsorption, and simultaneously separate acidic and basic
495
proteins using the cationic gemini pyrrolidinium surfactants Cn-4-CnPB and HFIP as
496
buffer additives. The hydrophobic chain length of the Cn-4-CnPB can distinctly affect the
497
separation selectivity and efficiency. Very small amounts of HFIP addition can promote
498
the formation of more homogeneous and more compact bilayer coating, and further
500 501 502 503
d
te
Ac ce p
499
M
493
improve the separation ability of Cn-4-CnPB for proteins. We strongly believe that this simple but effective method is promising in the CE analysis of proteins and peptides. A series of related researches are still going on in our laboratory, such as the effects of the spacer group length of Cn-4-CnPB surfactants on the protein separation, Cn-4-CnPB semi-permanent coatings and their applications in CE separation.
504 505
Acknowledgements
506
23
Page 23 of 29
This work was supported by the National Nature Science Foundation of China
508
(Grant No. 81373045 and 21177099) and the Fundamental Research Funds for the
509
Central Universities (Grant No. 2014306020204).
ip t
507
510
References
cr
511 512
[1] M. Gllges, M.H. Kleemlss, G. Schomburg, Capillary zone electrophoresis
514
separations of basic and acidic proteins using poly(vinyl alcohol) coatings in fused
515
silica capillaries, Anal. Chem. 66 (1994) 2038-2046.
an
us
513
[2] N. González, C. Elvira, J.S. Román, A. Cifuentes, New physically adsorbed
517
polymer coating for reproducible separations of basic and acidic proteins by
518
capillary electrophoresis, J. Chromatogr. A 1012 (2003) 95-101.
M
516
[3] T.-F. Jiang, Y.-L. Gu, B. Liang, J.-B. Li, Y.-P. Shi, Q.-Y. Ou, Dynamically coating
520
the capillary with 1-alkyl-3-methylimidazolium-based ionic liquids for separation
521
of basic proteins by capillary electrophoresis, Anal. Chim. Acta 479 (2003)
523 524 525 526
te
Ac ce p
522
d
519
249-254.
[4] J. Li, H.-F. Han, Q. Wang, X. Liu, S.-X. Jiang, Polymeric ionic liquid as a dynamic coating additive for separation of basic proteins by capillary electrophoresis, Anal. Chim. Acta 674 (2010) 243-248.
[5] J.E. Melanson, N.E. Baryla, C.A. Lucy, Dynamic capillary coatings for
527
electroosmotic flow control in capillary electrophoresis, Trends Anal. Chem. 20
528
(2001) 365-374.
529
[6] Q. Liu, Y.-M. Yang, Y. Huang, C.-F. Pan, Z. Nie, S.-Z. Yao, Separation of acidic and
530
basic proteins by CE with CTAB additive and its applications in peptide and protein
531
profiling, Electrophoresis 30 (2009) 2151-2158. 24
Page 24 of 29
532
[7] C.A. Lucy, R.S. Underhill, Characterization of the cationic surfactant induced
533
reversal of electroosmotic flow in capillary electrophoresis, Anal. Chem. 68 (1996)
534
300-305.
537 538
ip t
536
[8] F.M. Menger, C.A. Littau, Gemini surfactants: a new class of self-assembling molecules, J. Am. Chem. Soc. 115 (1993) 10083–10090.
[9] F.M. Menger, J.S. Keiper, Gemini surfactants, Angew. Chem. Int. Ed. 39 (2000)
cr
535
1906-1920.
[10] Q. Liu, Y.-Q. Li, F. Tang, L. Ding, S.-Z. Yao, Cationic gemini surfactant as dynamic
540
coating in CE for the control of EOF and wall adsorption, Electrophoresis 28 (2007)
541
2275-2282.
an
us
539
[11] Q. Liu, Y.-Q. Li, Y.-M. Yang, S.-Z. Yao, Separation of acidic and basic proteins by
543
capillary electrophoresis using gemini surfactants and gemini-capped nanoparticles
544
as buffer additives, Sci China Ser B-Chem. 52 (2009) 1666-1676.
d
[12] Q. Liu, J.-B. Yuan, Y.-Q. Li, S.-Z. Yao, Long-chained gemini surfactants for
te
545
M
542
semipermanent
wall
coatings
547
Electrophoresis 29 (2008) 871-879.
in
capillary
electrophoresis
of
proteins,
552
Ac ce p
546
553
[15] S. Tang, S.-J. Liu, Y. Guo, X. Liu, S.-X. Jiang, Recent advances of ionic liquids and
548
[13] Q. Liu, Y.-M. Yang, S.-Z. Yao, Enhanced stability of surfactant-based
549
semipermanent wall coatings in capillary electrophoresis using oppositely charged
550 551
surfactant, J. Chromatogr. A 1187 (2008) 260-266.
[14] B. Cai, X.-F. Li, Y. Yang, J.-F. Dong, Surface properties of gemini surfactants with pyrrolidinium head groups, J. Colloid Interface Sci. 370 (2012) 111-116.
554
polymeric
ionic
liquids
in
capillary
electrophoresis
555
electrochromatography, J. Chromatogr. A 1357 (2014) 147-157.
and
capillary
25
Page 25 of 29
556
[16] A. Staub, S. Comte, S. Rudaz, J.-L. Veuthey, J. Schappler, Use of organic solvent to
557
prevent protein adsorption in CE-MS experiments, Electrophoresis 31 (2010)
558
3326-3333. [17] A.G. Diress, C.A. Lucy, Electroosmotic flow reversal for the determination of
560
inorganic anions by capillary electrophoresis with methanol-water buffers, J.
561
Chromatogr. A 1027 (2004) 185-191.
563
cr
[18] A.G. Diress, M.M. Yassine, C.A. Lucy, Semipermanent capillary coatings in mixed organic-water solvents for CE, Electrophoresis 28 (2007) 1189-1196.
us
562
ip t
559
[19] Q. Liu, Y.-Q. Li, L.-H. Yao, S.-Z. Yao, Use of gemini surfactants as semipermanent
565
capillary coatings in aqueous-organic solvents for capillary electrophoretic
566
separation of inorganic anions, J. Sep. Sci. 32 (2009) 4148-4154.
M
an
564
[20] J.-F. Berrien, M. Ourévitch, G. Morgant, N.E. Ghermani, B. Crousse, D.
568
Bonnet-Delpon, A crystalline H-bond cluster of hexafluoroisopropanol (HFIP) and
569
piperidine structure determination by X ray diffraction, J. Fluorine Chem. 128
570
(2007) 839-843.
te
d
567
[21] B. Czarnik-Matusewicz, S. Pilorz, L.-P. Zhang, Y.-Q. Wu, Structure of
572
hexafluoroisopropanol–water mixture studied by FTIR-ATR spectra and selected
576
Ac ce p
571
577
[23] Y. Tian, Y.-F. Li, J. Mei, B. Deng, Y.-X. Xiao, Hexafluoroisopropanol-modified
573
chemometric methods, J. Mol. Struct. 883 – 884 (2008) 195-202.
574
[22] C.-X. Fu, M.G. Khaledi, Selectivity patterns in micellar electrokinetic
575
chromatography characterization of fluorinated and aliphatic alcohol modifiers by micellar selectivity triangle, J. Chromatogr. A 1216 (2009) 1901-1907.
578
cetyltrimethylammoniumbromide/sodium
dodecyl
sulfate
vesicles
as
a
579
pseudostationary phasein electrokinetic chromatography, J. Chromatogr. A 1404
580
(2015) 131-140.
26
Page 26 of 29
581
[24] K.K.-C. Yeung, C.A. Lucy, Suppression of electroosmotic flow and prevention of
582
wall adsorption in capillary zone electrophoresis using zwitterionic surfactants,
583
Anal. Chem. 69 (1997) 3435-3441.
585
[25] J.K. Towns, F.E. Regnier, Capillary electrophoretic separations of proteins using nonionic surfactant coatings, Anal. Chem. 63 (1991) 1126-1132.
ip t
584
[26] G.M. Janini, K.C. Chan, J.A. Barnes, G.M. Muschik, H.J. Issaq, Separation of
587
pyridinecarboxylic acid isomers and related compounds by capillary zone
588
electrophoresis effect of cetyltrimethylammonium bromide on electroosmotic flow
589
and resolution, J. Chromatogr. A 653 (1993) 321-327.
an
us
cr
586
[27] R.R. Ogorzalekloo, N. Dales, P.C. Andrews, Surfactant effects on protein structure
591
examined by electrospray ionization mass spectrometry, Protein Sci. 3 (1994)
592
1975-1983.
M
590
[28] C.A. Lucy, A.M. MacDonald, M.D. Gulcev, Review Non-covalent capillary
594
coatings for protein separations in capillary electrophoresis, J. Chromatogr. A 1184
595
(2008) 81-105.
te
d
593
[29] J.L. Koenig, B.G. Frushour, Raman scattering of chymotrypsinogen A, ribonuclease,
597
and ovalbumin in the aqueous solution and solid state, Biopolymers 11 (1972)
598 599 600
Ac ce p
596
2505-2520.
[30] D.S. Dwyer, R.J. Bradley, Chemical properties of alcohols and their protein binding sites, Cell. Mol. Life Sci. 57 (2000) 265-275.
601
27
Page 27 of 29
Figure Captions
602
Fig. 1. Chemical structures of cationic gemini surfactants with pyrrolidinium head
603
groups (Cn-4-CnPB, n = 10, 12, 14, 16, and 18).
604
Fig. 2. The relationship between EOF and the concentration of C18-4-C18PB at various
605
buffer pH values. Running buffer: 0–1 mM C18-4-C18PB in 10 mM NaH2PO4.
606
Fig. 3. Effect of the concentration of phosphate buffer on EOF. Running buffer: 0.1 mM
607
C18-4-C18PB in10-50 mM NaH2PO4 at pH 5.0, 7.0 and 9.0.
608
Fig. 4. Effects of C18-4-C18PB concentration (A), buffer concentration (B), buffer pH (C),
609
and the hydrophobic chain length of Cn-4-CnPB (n = 10, 12, 14, 16, or 18) (D) on the
610
separation of proteins. Peak designation: 1. lysozyme, 2. α-chymotrypsinogen A, 3.
611
ribonuclease A, 4. cytochrome c, 5. hemoglobin, 6. bovine serum albumin, 7. trypsin
612
inhibitor, and 8. pepsin. Running buffer: 2–5 mM C18-4-C18PB in 30 mM NaH2PO4 (pH
613
5) (A), 4 mM C18-4-C18PB in 10-50 mM NaH2PO4 (pH 5) (B), 4 mM C18-4-C18PB in 30
614
mM NaH2PO4 (pH 5-7) (C), and 4 mM Cn-4-CnPB (n = 10, 12, 14, 16, or 18) in 30 mM
616 617 618
cr
us
an
M
d
te
Ac ce p
615
ip t
601
NaH2PO4 (pH 5) (D).
Fig. 5. Effect of HFIP content (v/v) on protein separation in 30 mM NaH2PO4 (pH 5.0) with 4 mM C10-4-C10PB (A), C12-4-C12PB (B), C14-4-C14PB (C), or C16-4-C16PB (D). Peak designations are the same as in Fig. 4.
619 620
Highlights for JCA-15-479 revision
621 622 623
Gemini surfactants (Cn-4-nPB) were used as novel dynamic coating additives in CE.
28
Page 28 of 29
HFIP can improve the effect of Cn-4-nPB on the adsorption and separation of proteins.
Neutral and charged proteins were well separated with 115000–318000 plates/m.
ip t
624 625 626 627 628 629 630 631
Ac ce p
te
d
M
an
us
cr
632
29
Page 29 of 29