Talanta 137 (2015) 197–203
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
A tunable isoelectric focusing via moving reaction boundary for two-dimensional gel electrophoresis and proteomics Chen-Gang Guo a, Zhi Shang a, Jian Yan a, Si Li a, Guo-Qing Li a, Rong-Zhong Liu b, Ying Qing c, Liu-Yin Fan a, Hua Xiao a,n, Cheng-Xi Cao a,nn a Laboratory of Bioseparation and Analytical Biochemistry, State Key Laboratory of Microbial Metabolism, School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai 200240, China b State Key Laboratory of Oncogenes and Related Genes, Shanghai Cancer Institute, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai 200240, China c Bio-X Institutes, Key Laboratory for the Genetics of Developmental and Neuropsychiatric Disorders, Ministry of Education, Shanghai Jiao Tong University, Shanghai 200240, China
art ic l e i nf o
a b s t r a c t
Article history: Received 21 October 2014 Received in revised form 22 January 2015 Accepted 25 January 2015 Available online 3 February 2015
Routine native immobilized pH gradient isoelectric focusing (IPG-IEF) and two-dimensional gel electrophoresis (2DE) are still suffering from unfortunate reproducibility, poor resolution (caused by protein precipitation) and instability in characterization of intact protein isoforms and posttranslational modifications. Based on the concept of moving reaction boundary (MRB), we firstly proposed a tunable non-IPG-IEF system to address these issues. By choosing proper pairs of catholyte and anolyte, we could achieve desired cathodic and anodic migrating pH gradients in non-IPG-IEF system, effectively eliminating protein precipitation and uncertainty of quantitation existing in routine IEF and 2DE, and enhancing the resolution and sensitivity of IEF. Then, an adjustable 2DE system was developed by combining non-IPG-IEF with polyacrylamide gel electrophoresis (PAGE). The improved 2DE was evaluated by testing model proteins and colon cancer cell lysates. The experiments revealed that (i) a tunable pH gradient could be designed via MRB; (ii) up to 1.65 fold improvement of resolution was achieved via non-IPG-IEF; (iii) the sensitivity of developed techniques was increased up to 2.7 folds; and (iv) up to about 16.4% more protein spots could be observed via the adjustable 2DE as compared with routine one. The developed techniques might contribute to complex proteome research, especially for screening of biological marker and analysis of extreme acidic/alkaline proteins. & 2015 Elsevier B.V. All rights reserved.
Keywords: Isoelectric focusing Moving reaction boundary Proteomics Tunable pH gradient Two-dimensional gel electrophoresis
1. Introduction A series of robust techniques have been developed for proteomics [1,2], functional genomics [3,4] and systemic biology [5]. For example, two-dimensional gel electrophoresis (2DE) [6,7] and nano-liquid chromatography (Nano-LC) [8] are developed for separation of complex protein samples and tryptic peptides, respectively. Powerful nano-LC–MS/MS is proposed for high throughput and sensitive identification of tryptic peptides [1,8]. Meanwhile, many soft ionization techniques (e.g., matrix-assisted laser desorption/ionization [9,10] and electrospray ionization (ESI) [11,12]) are advanced for structural characterization of intact proteins. Generally, there are two classic proteomic strategies. One is the top-down strategy, and the other is the bottom-up
n
Corresponding author. Corresponding author. Tel./fax: þ86 21 3420 5820. E-mail addresses: [email protected] (H. Xiao), [email protected] (C.-X. Cao).
nn
http://dx.doi.org/10.1016/j.talanta.2015.01.038 0039-9140/& 2015 Elsevier B.V. All rights reserved.
approach. Interestingly, 2DE can be the most powerful separation tool for both approaches. In the bottom-up strategy [1,5], protein spots in 2DE are digested into peptides, and then LC–MS/MS is used for the detection of tryptic peptides for protein identification. The bottom-up strategy has many merits, including good sensitivity, high throughput and great convenience as well as automation [12,13]. However, this strategy has its own weaknesses. First, many important data of protein structure are not available at the end because only a sequence of tryptic peptide was detected. Second, different protein isoforms with the same peptide sequences may not be distinguishable because of low sequence coverage of protein. Third, data on post-translational modifications (PTMs) of protein are usually lost. Finally, LC–MS/MS based peptide-centric analyses may complicate protein quantification due to the coexistence of protein isoforms. These defects make structural analyses of intact protein very challenging, especially for the PTMs of proteins [13]. To overcome these defects, different top-down proteomic approaches have been developed [13]. In 2004, Takáts et al.
198
C.-G. Guo et al. / Talanta 137 (2015) 197–203
proposed a novel MS sampling under ambient conditions with desorption electrospray ionization [14]. Qiao et al. [15,16] developed an ambient ionization strategy and ESI for in situ ionization of proteins or peptides in the PAGE and coupled IPG-IEF to ESI-MS directly. Thus, protein samples could be firstly separated by IPGIEF, then in situ detected by rapidly scanning gel strip via ESI-MS. This novel top-down strategy can retain the structure information of intact proteins. Furthermore, this approach avoids gel staining and spot extraction, greatly improved the throughput and reduced labor intensity, running time, contaminant and sample loss [15,16]. The IPG-based IEF has been invented as an excellent electrophoretic technique [17,19] to address the cathodic or anodic drifting of pH gradient in Svensson's IEF [20,21], and used in the routine 2DE method for 30 years. However, the IPG-based IEF and 2DE techniques have been facing with the following challenges. First, many denature reagents and solubilizers (urea, thiourea, CHAPS, dithiothreitol, sugar and sorbitol [18,22,23]) were used for improving the solubility of protein in the routine IPG-IEF, making studies of PTMs and protein isoforms very complex and inconvenient. However, a native IPG-IEF without the use of denature reagents and solubilizers resulted in serious random protein precipitate [24] (Section 3.4), greatly affecting resulting sensitivity, resolution, quantitation and reproducibility and making comparative proteomics (e.g., cancer serum vs. normal control serum) dubious. Second, tunable local-zoomed pH gradient in IEF and 2DE was required for high resolution separation of complex protein isoforms and PTMs. However, once the pH gradient of IPG-IEF was immobilized, it could not be adjusted for high resolution separation of many protein isoforms and PTMs [7,25]. Third, sensitivities of routine IPG-IEF and 2DE are still poor [12,13], making identification of low abundance proteins impossible. All of these issues obviously weakened the power of 2DE on characterization of protein isoforms and PTMs [15,16]. To figure out these issues mentioned above, we have developed the concept of moving reaction boundary (MRB) [26,27] from the prototype ideas of moving reactive front [28,29] and stationary neutralization boundary [30]. The concept of MRB was used to improve the sensitivity of capillary electrophoresis (CE) [31,32], the design of supermolecular boundary electrophoresis [33,34] and the advancement of total protein content titration [35,36]. Furthermore, we developed the MRB-based dynamic theory of IEF [24,26,27,37,38], which quantitatively explained the mechanism of quasi-stable pH gradient in IEF and gave the quantitative illumination on various pH gradient instabilities [39–41] and Hjertén's mobilization [42–44]. However, to the best of our knowledge, a
tunable local-zoomed pH gradient in IEF has never been designed based on the concept of MRB for 2DE. Therefore, the main purposes herein are to design a tunable non-IPG-IEF via the concept of MRB; to evaluate resolution and sensitivity of the developed system by comparing it with the routine IPG-IEF; to create an novel non-IPG-IEF based adjustable 2DE; and to validate the advantage of tunable nonIPG-IEF used as the first dimension of 2DE by using model proteins and lysates of colon cancer samples.
2. Experimental designs and methods The details about materials, instruments and analytical software used herein, as well as procedures of staining, sample preparation and LC–MS/MS, are given in the Supporting Information. 2.1. Tunable pH gradient of non-IPG-IEF Scheme 1 shows the complete design of tunable pH gradient in non-IPG IEF via R value (viz., judgment expressions of MRB) and its combination with 2DE. In our previous works [26,27], we defined the expression for comparing fluxes of protons in an acid and hydroxyl in a base as follows: R¼
mþ cþ κ 1 ðused for R 40Þ m c κ þ
ð1aÞ
m c κ þ ðused for R o0Þ mþ cþ κ
ð1bÞ
R ¼ 1
where, R is the symbol of judgment expression, κ is the conductivity (S/m); m is the ionic mobility (m2 s 1 V 1); c is the equivalent concentration (equiv./l), the subscripts, “ þ” and “ ”, imply the hydrogen and hydroxyl ions, respectively. Herein, Eqs. (1a) and (1b) are used for R4 0 and R o0, respectively. As shown in Scheme 1, the significances of R value are: (i) we can design a cathodic migrating pH gradient of IEF by setting R4 0, leading to obvious enhancement of separation of acidic proteins; (ii) conversely, we may create an anodic migrating pH gradient by using Ro0, resulting in clear improvement of separation of alkaline proteins; and (iii) we can form a non-migrating pH gradient by choosing R ffi 0, increasing separation of neutral proteins. Thus, one can use tunable pH gradient of IEF to 2DE for high resolution separation of acidic, and/or neutral and/or alkaline proteins (Scheme 1). Interestingly, the non-IPG-IEF also has the merit of
Scheme 1. Diagrams of tunable non-IPG-IEF and adjustable 2DE designed via R value of MRB. The red arrows indicate the migration direction of pH gradient.
C.-G. Guo et al. / Talanta 137 (2015) 197–203
199
Table 1 Comparisons between the pH gradient design via R values and the experiments in Fig. 1. Comboa
Anolyte
Catholyte
R
Design
Result in Fig. 1
1 2 3 4 5
100 mM Gly 300 mM HCl 800 mM H3PO4 800 mM H3PO4 800 mM H3PO4
100 mM Arg 600 mM (NH2CH2)2 600 mM (NH2CH2)2 200 mM NaOH 100 mM KOH
0.25 0.13 0.01 0.14 0.23
Strong anodic migrating Moderate anodic migrating Stable Weak cathodic migrating Strong cathodic migrating
Strong anodic migrating Moderate anodic migrating Stable Weak cathodic migrating Strong cathodic migrating
a
The numbers herein correspond with A, B, C, D, E in Fig. 1.
high sensitivity, almost the same reproducibility and no protein precipitate comparing with IPG-IEF. In Table 1, we chose five electrolyte pairs to form cathodic, anodic and neutral migrating pH gradients. The R values are also given in Table 1. Prior to non-IPG-IEF run, paper pads were individually immersed into the acid and alkali given in Table 1. After an overnight immersion, the immersed pads bridged the non-IPG strips and the electrodes (Fig. S1). After the rehydration (see below), the strips were quickly placed into focusing tray slot, and the anodic and cathodic pads (Table 1) were respectively set at the anode and cathode (Fig. S1). Then the strips were covered by 1 mL mineral oil to prevent water evaporation. Non-IPG-IEF was run as follows: 250 V for 30 min, followed by a voltage gradient from 250 V to 2000 V (total up to 3000 VHr). Finally, constant 1000 V was applied until IEF end (a total of 8 h). The cooling temperature was set as 17 1C. 2.2. Run of IPG-IEF Unlike non-IPG-IEF (Fig. S1B), no pad was used for IPG-IEF run (Fig. S1A). IPG-IEF was generally performed in accordance with the previous procedure [25]. Briefly, rehydrated IPG strips (see gel strip rehydration) were put into focusing tray slot (Fig. S1A). IPGIEF was run as follows: 250 V for 30 min, 1000 V for 1 h, followed by gradual voltage gradient from 1000 V to 5000 V (total up to 4000 VHr). Finally, constant 1000 V was applied until IEF end (about 4 h). The cooling temperature was set as 17 1C.
Fig. 1. pH gradient migrating experiment of non-IPG-IEF via electrolyte pairs of (A) 100 mM Gly and 100 mM Arg; (B) 300 mM HCl and 600 mM (NH2CH2)2; (C) 800 mM H3PO4 and 600 mM (NH2CH2)2; (D) 800 mM H3PO4 and 200 mM NaOH; and (E) 800 mM H3PO4 and 100 mM KOH. (Phyco: phycocinin, Mb: equine myoglobin, Hb: hemoglobin, Cyt C: cytochrome C.) The conditions were given in Section 2.
SDS) was performed in line with the manual of Bio-Rad ProteomeWorksTM System.
3. Results and discussion 3.1. Tunable local-zoomed pH gradients of non-IPG-IEF
2.3. Gel strip rehydration Rehydration was carried out in a rehydration tray in line with the manufacturer's instruction. For IPG strips, a total volume of 150 μL rehydration solution was added for each IPG strip, which was composed of 30 μL glycerol (20%, v/v), 1.9 μL 40% carrier ampholyte (0.5%, w/v), 2 μL IEF protein sample, and 116 μL Milli-Q grade water. For non-IPG strips, the rehydration solution was consisted of 30 μL glycerol (20%, v/v), 7.5 μL 40% carrier ampholyte (2%, w/v), 2 μL IEF standard protein sample, and 121.6 μL Milli-Q grade water. Finally, the strips were placed on rehydration solution and covered by mineral oil overnight. It ought to be emphasized that the denature reagents and solubilizers (urea, thiourea, CHAPS, dithiothreitol, sugar and sorbitol [18,22,23]) were not used in both the IPG IEF and the non-IPG IEF for the comparison of protein precipitate under the nature conditions. 2.4. SDS‐PAGE SDS-PAGE was performed as previously described with minor modification [7]. Briefly, the strips were equilibrated twice (15 min each time) after IEF runs. The first equilibration solution contained 6 M urea, 75 mM Tris–HCl (pH 8.8), 20% (v/v) glycerol, 2% SDS, 0.002% (w/v) bromophenol blue and 2% dithiothreitol (DTT). The second solution was comprised of 6 M urea, 75 mM Tris–HCl (pH 8.8), 20% (v/v) glycerol, 2% SDS, 0.002% (w/v) bromophenol blue and 2.5% iodoacetamide. Then SDS-PAGE (12% T, 2.6% C and 2%
Table 1 shows that the R values are 0.25, 0.13, 0.01, 0.14 and 0.23 for the electrolyte pairs of Nos. 1, 2, 3, 4 and 5, respectively, and reveals the designs on migrating pH gradients of non-IPG-IEF. Thus, by using the pairs of Nos. 1 and 2, we could design strong and moderate anodic migrating pH gradients, respectively. On the contrary, we might respectively obtain weak and strong anodic migrating pH gradients via the pairs of Nos. 4 and 5. Moreover, by using the pairs of No. 3, we could get a stable pH gradient in nonIPG IEF. The five designs could be experimentally achieved, as demonstrated in Fig. 1. The red dotted line represented the location of the blue phycocyanin band, and the green and black dotted lines indicated the locations of brown Hb/Mb band and red Cyt C band, respectively. It was obvious that the blue, brown and pink standard proteins migrated toward the cathode gradually as R value increased. The qualitative experiments were completely consistent with our designs of migrating pH gradients in Table 1. The moving distances (from the anode) of these chromoproteins in Fig. 1 are given in Table S1. It could be seen that the migration distance of each chromoprotein increased with the increasing of R values of the electrolyte pairs, no matter it was blue or brown proteins, quantitatively indicating that the pH gradient plasticity was intimately correlated to the R value. Meanwhile, although pI (9.6) of Cyt C was close to the cathode end, it could also be observed that the red Cyt C migrated outside the cathode just as shown in Panels D and E of Fig. 1, implying the
200
C.-G. Guo et al. / Talanta 137 (2015) 197–203
Fig. 2. Comparative experiments between IPG-IEF (A, left) and non-IPG-IEF (A, right). (B) Gray levels of model proteins in Panel A. Upper: IPG-IEF, bottom: non-IPG-IEF. (C) Zoomed image of protein peak 1 and 2 in Panel B. (D) Zoomed image of protein peak 4, 5 and 6 in Panel B. (F) Quantitation analysis of the seven separated proteins. The conditions were given in the experimental section. The numbers representing the peak of responding proteins. Phoco: phcocyanin, 1, 2 represent two subunits; Mb1: equine myoglobin minor, Mb2: equine myoglobin; Hb: hemoglobin; Cyt C: cytochrome C.
same results as the above. Fig. S2 reveals the experimental reproducibility of non-IPG-IEF. It could be observed that the three repeated trials were very consistent. The RSD values of the distance between the three proteins of phycocyanin, Mb, Cyt C and the anode were 2.1%, 2.1%, 0.6% for non-IPG-IEF and 3.5%, 3.4% and 1.2% for IPG-IEF, respectively (Table S2). All of the experiments in Figs. 1 and 2 and Tables 1, S1 and S2 along with Fig. S2
quantitatively demonstrated that tunable pH gradient of non-IPG IEF via R values has been achieved. 3.2. Resolution comparison of IPG-IEF and non-IPG-IEF As illuminated in Scheme 1, we could acquire desired pH gradient by adjusting R value to improve the resolution of target
C.-G. Guo et al. / Talanta 137 (2015) 197–203
proteins in non-IPG-IEF. To validate our theory, comparison experiment between IPG-IEF and non-IPG-IEF was performed (Fig. 2). Stable pH gradient created via the electrolyte pairs (800 mM H3PO4 and 600 mM (NH2CH2)2, R¼0.01) was chosen to compare IPG-IEF with non-IPG-IEF. It could be seen in Fig. 2A that the four distinct bands of equine Mb minor, equine Mb, human Hb A and human Hb C could be observed in the center region of nonIPG-IEF. However, for IPG-IEF, only three bands were resolved in the same region. Furthermore, Cyt C was resolved more clearly in cathodic extreme of non-IPG-IEF than IPG-IEF. To further investigate the resolution of these protein bands, we attained gray level vs. distance from the anode curve graph of each protein bands via Meta Morph 7.0 software and Origin 8.0 (Fig. 2B). As shown in Fig. 2B, we could detect up to 9 protein bands in non-IPG-IEF, while only 6 bands were observed in IPG-IEF, which qualitatively implied that higher resolution separation of model proteins had been achieved by non-IPG-IEF. The separation of two phycocyanin subunits (peaks 1 and 2) as well as two proteins (Mb and HbA, corresponding to Peaks 5 and 6, respectively) was compared between IPG-IEF and non-IPG-IEF (Panels C and D). In Panel C, the resolution of Peaks 1 and 2 was 0.86 (IPG-IEF) and 0.88 (non-IPG-IEF), implying slight increase of non-IPG-IEF separation. In Panel D, the resolution of Peaks 5 and 6 was 0.26 (IPG-IEF) and 0.43 (non-IPG-IEF), indicating that the resolution of non-IPG-IEF increased by 65% over IPG-IEF. Thus, these data quantitatively manifested that the separation of neutral proteins was obviously increased via the pH gradient of non-IPGIEF designed by R ffi 0. The 2DE comparative experiments between IPG and non-IPG in Figs. S2 and S3 further verified the above conclusion. The resolution enhancement of non-IPG-IEF may be due to the following two reasons. The first reason is that the broadening pH gradient in non-IPG-IEF results in resolution increase of IEF, especially for target proteins. The second was the case that the better solubility of proteins improve the resolution of non-IPG-IEF system. We can observe that obvious protein precipitates occur in IPG-IEF rather than non-IPG-IEF (marked by the black arrow in Figs. 2A and S2), which reduced the resolution of IPG-IEF, as discussed in Section 3.4. 3.3. Sensitivity comparison between IPG-IEF and non-IPG-IEF Fig. 2E reveals the sensitivity comparison between IPG-IEF and non-IPG-IEF. It can be observed that the OD value of each protein in non-IPG-IEF is higher than that of IPG-IEF, indicating the higher sensitivity of non-IPG-IEF. Through comparing the total OD value of seven proteins in non-IPG-IEF with that in IPG-IEF, we found that the total quantitation of the proteins in non-IPG-IEF was 2.7 folds as high as that of IPG-IEF. It was worth mentioning that some proteins could only be detected in non-IPG-IEF (e.g. Hb C, Cyt C). These quantitative results demonstrated that the sensitivity of non-IPG-IEF was clearly superior to that of IPG-IEF. The comparative experiments of 2DE have further verified the same conclusions (see Section 3.5, Figs. 3 and 4). The sensitive increase in non-IPG IEF is caused by the following three reasons. First, the high resolution separation of proteins can further result in good sensitivity. It is observed in Panels A–D in Fig. 2 that these protein bands or peaks in non-IPG-IEF are obviously narrower than those of IPG-IEF. This indicated that the same loading amount of protein sample might lead to higher protein peak (viz., better sensitivity) in non-IPG-IEF. Second, protein sample had better solubility in non-IPG-IEF. It is seen in Figs. 2A and S2A that random protein precipitates occur in IPG-IEF (marked black arrow), rather than non-IPG-IEF (Section 3.4). This implied that better protein solubility (viz., higher actual sample load) could be achieved in non-IPG-IEF. Hence, better sensitivity
201
Fig. 3. Separation of model proteins (pI 4.45–9.6) by (A) IPG-based and (B) nonIPG-based 2DE with staining of Coomassie brilliant blue. Number of 1–16 representing selected protein spots. The conditions were given in Section 2 and the Online Supporting Information.
was obtained more willingly in non-IPG-IEF than IPG-IEF. Third, protein precipitate after electric focusing reduced the actual protein content in its relevant band or spot in IPG-IEF, as will be revealed below. Consequently, the precipitate further decreased the sensitivity of IPG-IEF. 3.4. Precipitate comparison between IPG-IEF and non-IPG-IEF The comparative experiments in Figs. 2A and S2 evidently demonstrated that there were many random protein precipitates in IPG-IEF, but none in non-IPG-IEF. The phenomenon was similar to the previous observation [24]. The random precipitate might cause serious issues simultaneously. First, the precipitate greatly reduced the reproducibility of IPG-IEF and its 2DE (Fig. S2A). The results in Fig. S3 and Table S3 quantitatively displayed the poor reproducibility of IPG-IEF in contrast to non-IPG-IEF. Second, the precipitates obviously reduced the sensitivity of IPG-IEF and its 2DE due to the actual quantity of focused protein decreased at its pI location. Third, the precipitate, especially the wide flaky precipitate in IPG-IEF gel, would have extremely reduced the resolution of IEF and 2DE (see Figs. 2 and S2A) if the strip was further used for 2DE run (see Section 3.5). Fourth, the precipitate, especially the one out of the strip gel, could greatly affect quantification of proteins in IEF, 2DE and DIGE analysis, making comparative proteomics extremely difficult. The relevant work will be conducted in our further study. From the discussion above, it could be evidently seen that the elimination of protein precipitation in non-IPG-IEF could simultaneously promote the repeatability, resolution and sensitivity of proteins in IEF and 2DE (Figs. 2, S2A and S3). And the higher resolution of proteins could lead to better sensitivity as has been described in Section 3.3. In addition, the non-IPG-IEF benefits the studies of protein isoforms and MPTs. In a routine IPG-IEF, numerous denature solubilizers (e.g. urea, thiourea, CHAPS and DTT) must be used to promote solubility of proteins by reducing protein precipitation
202
C.-G. Guo et al. / Talanta 137 (2015) 197–203
ampholytes used in IEF might have a similar effect of low concentration salt. In IPG-IEF, 0.4% free carrier ampholytes was added during the rehydration. Furthermore, many ampholytes were immobilized within IPG gel strip, reducing hydrophilicity of gel matrix. Consequently, these factors led to the low solubility and precipitate of proteins in IPG-IEF. On the contrary, fairly high concentration of carrier ampholytes (2.0%) was used in non-IPGIEF. Thus, protein precipitate was rarely observed in non-IPG-IEF (Figs. 1, 2 and S2). 3.5. IPG-IEF and non-IPG-IEF based 2DE
Fig. 4. Analyses of proteins from colon cancer cell lines HCT116 by (A)– (C) corresponding Gaussie image of Fig. S4A–C, respectively. Conditions in (A): 800 mM H3PO4 and 200 mM NaOH used as anolyte and catholyte, respectively; Conditions in (C): 300 mM HCl and 600 mM (NH2CH2)2 used as anolyte and catholyte, respectively. The other conditions were given in detail in Section 2 and the Online Supporting Information. The yellow “þ ” indicates protein spots detected by PD Quest software. Red arrow refers to the migrating direction. The letters of ‘a–d’ are corresponding to these selected protein spots for comparison.
[18,22,23]. But these solubilizers can destroy quaternary structure of protein, making the studies of isoforms and MPTs difficult [15,16]. Conversely, the non-IPG-IEF is native milieu, which did not destroy the quaternary structure of protein, suitable for separation of protein isoforms and MPTs. Finally, the non-IPG-IEF was simple and low cost due to no denature reagents were used (e. g. urea, thiourea, CHAPS and DTT) for increasing the solubility of proteins [17]. The detail mechanism of protein precipitates in IPG-IEF (Figs. 2 and S2) is still unknown. It may be caused by the following reasons. It is well known that high concentration of salt (e.g., higher than about 1.5 M) can reduce protein solubility and lead to protein precipitate [43]. However, low concentration range salt (e.g., less than about 1.5 M) can increase protein solubility [45]. In other words, under low concentration of salt, protein solubility will increase with the increase of salt concentration. Free carrier
3.5.1. Separation of model protein subunits We compared the 2DE images of IPG system with non-IPG through analyzing standard proteins. In Fig. 3, the upper pattern and bottom pattern was obtained via IPG-IEF and non-IPG-IEF, respectively. The labeled protein spots in dotted boxes in green, red, yellow and black were selected to assess the resolution and sensitivity of 2DE. We could found that all model proteins could be resolved by IPG- and non-IPG-based 2DE. However, due to the superior pH gradient, the non-IPG-based 2DE could resolve more protein subunits than the IPG-based 2DE, as shown in the red, yellow and black dotted box of Fig. 3B. It could be seen from red dotted box that the non-IPG-based 2DE could resolve eight subunits of hemoglobin A (spots 1, 2, 3, 4) and C (spots 5, 6, 7, 8), illustrating higher separation power of protein isoforms. The protein spots in non-IPG-IEF were successfully identified by NanoLC/Q-TOFMS (Table S4). On the contrary, the same protein spots surrounded by a red ring in the IPG-based 2DE gathered together and could not be discriminated. Meanwhile, identical results could be observed in yellow ellipse ring, demonstrating that protein spots could be more efficiently resolved in the non-IPG-based 2DE (see Nos. 9, 10 and 11). These protein spots were identified by NanoLC/Q-TOF-MS as well (Table S4). It could be seen that three subunits of lectin protein were recognized in the non-IPG-based 2DE, while they were very blurry in the IPG-based 2DE. Similar results could be seen in protein spots 14 and 15. The deep capability of our approach to resolve different protein subunits demonstrated the separation power of non-IPG-based 2DE. Additionally, low abundance protein spots 12, 13, 16 were selected to study the sensitivities of IPG-based and non-IPGbased 2DE systems (Table S5). The gray intensity of spots 12, 13, 16 increased by 18.5%, 29.5%, 66.3% in non-IPG-based 2DE compared to that of IPG-based 2DE, respectively. p Value of these three protein spots were 0.022, 0.003, 0.0005 (n ¼3). Each of the p values was less than 0.05, which implied significant difference of intensity between IPG-based and non-IPG-based 2DE systems. 3.5.2. Adjustable 2DE via tunable non-IPG-IEF Herein, 800 mM H3PO4 and 200 mM NaOH were used as anolyte and catholyte, respectively, representing a cathode migrating protein spots in 2DE (Fig. 4A). Meanwhile, 300 mM HCl and 600 mM (NH2CH2)2 were chosen to perform anode migrating protein spots (Fig. 4C). A routine IPG-IEF-based 2DE was carried out as a control (Fig. 4B). Fig. 4 shows the Gaussian 2DE graphs using colon cancer cell lines HCT116 as samples. In order to investigate the resolution of protein spots in three 2DE systems, we compared the difference of relative position of four typical protein spots (marked as red letter a, b, c, d) through calculating the horizontal distances between them. It could be seen that horizontal distances between protein spot a and b (1.24 cm) in Fig. 4A were expanded significantly when comparing with that (0.85 cm) in Fig. 4B. However, the distance between spot c and d (1.54 cm) in Fig. 4A was narrower than that (2.37 cm) in Fig. 4B.
C.-G. Guo et al. / Talanta 137 (2015) 197–203
These results indicated that a superior separation power existed in the acidic region of the cathode migrating system. Similarly, horizontal distance between spot c and d (2.90 cm) in Fig. 4C was widened comparing to that (2.37 cm) in Fig. 4B. Especially, the distance between spots b and d had a distinct broaden in Fig. 4C (4.54 cm) than in Fig. 4B (3.51 cm), revealing a superior separation power existed in the alkaline region of the anode migrating system. The results mentioned above were completely consistent with our design of IEF and its 2DE in Scheme 1. This is to say that the tunable 2DE could enhance the resolution of proteins in 2D by broadening the desired pH gradient region (e.g., cathodic region or anodic one). Subsequently, we investigated the sensitivity of IPG- and nonIPG-IEF based 2DE system by counting the proteins detected by PD Quest software in Gaussian 2DE graphs. 287 and 280 proteins were detected in Fig. 4A and C, respectively, while only 262 proteins were detected in Fig. 4B. The number of detected protein spots increased by 9.5% and 6.9%, respectively, indicating that nonIPG-based 2DE was more sensitive than IPG one. This echoed the result of IEF in Section 3.3. The relevant mechanism has been discussed in the above section. Therefore, we concluded that enhanced solubility of low abundance proteins led to increase of sensitivity or number of detected protein spots in 2DE.
4. Conclusion In this study, a novel non-IPG-IEF was designed through adjusting the R value of MRB. In this novel system, by choosing a proper pair of catholyte and anolyte, we could adjust the cathodic or anodic migrating pH gradient for non-IPG-IEF. The experiments revealed that the migrating pH gradient in Fig. 1 was completely consistent with R value based design (Table 1). Meanwhile, the sensitivity in non-IPG-IEF was also enhanced (increased by up to 270%) due to elimination of protein precipitation. Subsequently, non-IPG-based 2DE was developed and applied for separation of model proteins. For the proteome analysis of colon cancer cells, tunable non-IPG-IEF based 2DE resolved up to about 16.4% (combination of Figs. 4A and C) more protein spots than IPGbased 2DE. Thus, tunable non-IPG-IEF demonstrated superior resolution, repeatability and sensitivity than routine IPG-IEF approach, and benefited the separation of protein isoforms and PTMs. In addition, the non-IPG IEF was simple and low cost. Although there are so many advantages for non-IPG-IEF system, some drawbacks also drew our attention. For instance, we could not quantitatively know the accurate displacement from the anode corresponding to each protein band when using the electrolyte pairs with specific R value. In addition, some issues still exist when interfacing the electrode paper pad with the gel strip during the experiment. This problem will greatly affect the homogeneity of individual protein band. We plan to address these issues in our future work.
Acknowledgments This work was supported by grants from the National Natural Science Foundation of China (Nos. 21035004, 21275099, 21305087, and 21475086), the National Key Development of Scientific Instruments (No. 2011YQ030139) and the Key Scientific Project of Shanghai Jiao Tong University (Nos. YG2010ZD209 and YG2014QN21). H.X. is supported by the Recruitment Program of Global Youth Experts of
203
China and National High-tech R&D Program of China (863 Program, No. 2014AA020545).
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.talanta.2015.01.038. References [1] M. Mann, R.C. Hendrickson, A. Pandey, Annu. Rev. Biochem. 70 (2001) 437. [2] Y. Ge, B.G. Lawhorn, M. ElNaggar, E. Strauss, J.H. Park, T.P. Begley, F.W. McLafferty, J. Am. Chem. Soc. 124 (2002) 672. [3] K.L. Williams, D.F. Hochstrasser, Proteome Research: New Frontiers in Functional Genomics, Springer, Verlag, Berlin Heidelberg, New York (1997) 10. [4] L. Zamdborg, R.D. LeDuc, K.J. Glowacz, Y.B. Kim, V. Viswanathan, I.T. Spaulding, B.P. Early, E.J. Bluhm, S. Babai, N.L. Kelleher, Nucleic Acids Res. 35 (2007) W701. [5] R. Aebersold, M. Mann, Nature 422 (2003) 198. [6] P.H. O'Farrell, J. Biol. Chem. 250 (1975) 4007. [7] T. Rabilloud, Proteome Research: Two-Dimensional Gel Electrophoresis and Identification Methods, Springer, Verlag, Berlin Heidelberg, New York, 2000. [8] Y. Shen, N. Tolic, C. Masselon, L. Pasa-Tolic, D.N. Camp, K.K. Hixson, R. Zhao, G.A. Anderson, R.D. Smith, Anal. Chem. 76 (2004) 144. [9] M. Karas, F. Hillenkamp, Anal. Chem. 60 (1988) 2299. [10] K. Tanaka, H. Waki, Y. Ido, S. Akita, Y. Yoshida, T. Yoshida, T. Matsuo, Rapid Commun. Mass Spectrom. (1988) 151–153. [11] T.R. Covey, A.P. Bruins, J.D. Heninon, Org. Mass Spectrom. 23 (1988) 178. [12] J.B. Fenn, M. Mann, C.K. Meng, S.F. Wong, C.M. Whitehouse, Science 246 (1989) 64. [13] S. Wu, N.M. Lourette, N. Tolic, R. Zhao, E.W. Robinson, A.V. Tolmachev, R.D. Smith, L. Pasa-Tolic, J. Proteome Res. 8 (2009) 1347. [14] Z. Takats, J.M. Wiseman, B. Gologan, R.G. Cooks, Science 306 (2004) 471. [15] L. Qiao, R. Sartor, N. Gasilova, Y. Lu, E. Tobolkina, B. Liu, H.H. Girault, Anal. Chem. 84 (2012) 7422. [16] L. Qiao, E. Tobolkina, B. Liu, H.H. Girault, Anal. Chem. 85 (2013) 4745. [17] B. Bjellqvist, K. Ek, P.G. Righetti, E. Gianazza, A. Görg, R. Westermeier, W. Postel, J. Biochem. Biophys. Methods 6 (1982) 317. [18] P.G. Righetti, Immobilized pH Gradient: Theory and Methodology, Elsevier, Amsterdam, New York, Oxford (1990) 8–13. [19] P.G. Righetti, Isoelectric Focusing: Theory, Methodology and Applications, first edn., Elsevier Biomedical Press, Amsterdam, New York, Oxford, 1983. [20] H. Svensson, Acta Chem. Scand. 15 (1961) 325. [21] H. Svensson, Acta Chem. Scand. 16 (1962) 456. [22] P.G. Righetti, E. Fasoli, S.C. Righetti, in: Jan Christer Janson (Ed.), Protein Purification: Principles, High Resolution Methods, and Applications, John Wiley & Sons, Inc., 2011, pp. 379–409. [23] P.G. Righetti, A. Bossi, Anal. Biochem. 247 (1997) 1. [24] C.G. Guo, S. Li, H.Y. Wang, D. Zhang, G.Q. Li, J. Zhang, L.Y. Fan, C.X. Cao, Talanta 111 (2013) 20. [25] A. Görg, C. Obermaier, G. Boguth, W. Weiss, Electrophoresis 20 (1999) 712–717. [26] C.X. Cao, J. Chromatogr. A 183 (1998) 173–177. [27] C.X. Cao, L.Y. Fan, W. Zhang, Analyst 133 (2008) 1139. [28] J. Deman, W. Rigole, J. Phys. Chem. 74 (1970) 1122. [29] J. Deman, Anal. Chem. 42 (1970) 321. [30] J. Pospichal, M. Deml, P. Bocek, J. Chromatogr. A 638 (1993) 179. [31] C.X. Cao, Y.Z. He, M. Li, Y.T. Qian, M.F. Gao, L.H. Ge, S.L. Zhou, L. Yang, Q.S. Qu, Anal. Chem. 74 (2002) 4167. [32] C.X. Cao, W. Zhang, W.H. Qin, S. Li, W. Zhu, W. Liu, Anal. Chem. 77 (2005) 955. [33] L. Fan, W. Yan, C. Cao, W. Zhang, Q. Chen, Anal. Chim. Acta 650 (2009) 111. [34] J. Dong, S. Li, H. Wang, Q. Meng, L. Fan, H. Xie, C. Cao, W. Zhang, Anal. Chem. 85 (2013) 5884. [35] H.Y. Wang, C.Y. Guo, C.G. Guo, L.Y. Fan, L. Zhang, C.X. Cao, Anal. Chim. Acta 774 (2013) 92. [36] H. Wang, Y. Shi, J. Yan, J. Dong, S. Li, H. Xiao, H. Xie, L.Y. Fan, C.X. Cao, Anal. Chem. 86 (2014) 2888–2894. [37] H. Liang, Y. Chen, L.J. Tian, L. Zhang, Electrophoresis 30 (2009) 3134. [38] H. Liang, L.F. Ou Yang, Q. Liu, L. Zhang, L.J. Tian, Y. Chen, J. Sep. Sci. 34 (2011) 1212–1219. [39] A. Chrambach, P. Doerr, G.R. Finlayson, L.E. Miles, R. Sherins, D. Rodbard, Ann. N. Y. Acad. Sci. 209 (1973) 44. [40] H. Rilbe, in: B.J. Radola, D. Graeslin (Eds.), Proceedings of the International Symposium, Walter de Gruyter & Co., Hamburg, Germany, 1977, pp. 35–50. [41] I.K.O.K., A. Murel, J. Chromatogr. A 174 (1979) 1. [42] S. Hjertén, M. Zhu, J. Chromatogr. A 346 (1985) 265. [43] S. Hjertén, J.L. Liao, K.Q. Yao, J. Chromatogr. 387 (1987) 127. [44] F. Kilar, S. Hjertén, Electrophoresis 10 (1989) 23. [45] A.A. Green, J. Biol. Chem. 95 (1932) 47–66.
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
A tunable isoelectric focusing via moving reaction boundary for two-dimensional gel electrophoresis and proteomics Chen-Gang Guo a, Zhi Shang a, Jian Yan a, Si Li a, Guo-Qing Li a, Rong-Zhong Liu b, Ying Qing c, Liu-Yin Fan a, Hua Xiao a,n, Cheng-Xi Cao a,nn a Laboratory of Bioseparation and Analytical Biochemistry, State Key Laboratory of Microbial Metabolism, School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai 200240, China b State Key Laboratory of Oncogenes and Related Genes, Shanghai Cancer Institute, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai 200240, China c Bio-X Institutes, Key Laboratory for the Genetics of Developmental and Neuropsychiatric Disorders, Ministry of Education, Shanghai Jiao Tong University, Shanghai 200240, China
art ic l e i nf o
a b s t r a c t
Article history: Received 21 October 2014 Received in revised form 22 January 2015 Accepted 25 January 2015 Available online 3 February 2015
Routine native immobilized pH gradient isoelectric focusing (IPG-IEF) and two-dimensional gel electrophoresis (2DE) are still suffering from unfortunate reproducibility, poor resolution (caused by protein precipitation) and instability in characterization of intact protein isoforms and posttranslational modifications. Based on the concept of moving reaction boundary (MRB), we firstly proposed a tunable non-IPG-IEF system to address these issues. By choosing proper pairs of catholyte and anolyte, we could achieve desired cathodic and anodic migrating pH gradients in non-IPG-IEF system, effectively eliminating protein precipitation and uncertainty of quantitation existing in routine IEF and 2DE, and enhancing the resolution and sensitivity of IEF. Then, an adjustable 2DE system was developed by combining non-IPG-IEF with polyacrylamide gel electrophoresis (PAGE). The improved 2DE was evaluated by testing model proteins and colon cancer cell lysates. The experiments revealed that (i) a tunable pH gradient could be designed via MRB; (ii) up to 1.65 fold improvement of resolution was achieved via non-IPG-IEF; (iii) the sensitivity of developed techniques was increased up to 2.7 folds; and (iv) up to about 16.4% more protein spots could be observed via the adjustable 2DE as compared with routine one. The developed techniques might contribute to complex proteome research, especially for screening of biological marker and analysis of extreme acidic/alkaline proteins. & 2015 Elsevier B.V. All rights reserved.
Keywords: Isoelectric focusing Moving reaction boundary Proteomics Tunable pH gradient Two-dimensional gel electrophoresis
1. Introduction A series of robust techniques have been developed for proteomics [1,2], functional genomics [3,4] and systemic biology [5]. For example, two-dimensional gel electrophoresis (2DE) [6,7] and nano-liquid chromatography (Nano-LC) [8] are developed for separation of complex protein samples and tryptic peptides, respectively. Powerful nano-LC–MS/MS is proposed for high throughput and sensitive identification of tryptic peptides [1,8]. Meanwhile, many soft ionization techniques (e.g., matrix-assisted laser desorption/ionization [9,10] and electrospray ionization (ESI) [11,12]) are advanced for structural characterization of intact proteins. Generally, there are two classic proteomic strategies. One is the top-down strategy, and the other is the bottom-up
n
Corresponding author. Corresponding author. Tel./fax: þ86 21 3420 5820. E-mail addresses: [email protected] (H. Xiao), [email protected] (C.-X. Cao).
nn
http://dx.doi.org/10.1016/j.talanta.2015.01.038 0039-9140/& 2015 Elsevier B.V. All rights reserved.
approach. Interestingly, 2DE can be the most powerful separation tool for both approaches. In the bottom-up strategy [1,5], protein spots in 2DE are digested into peptides, and then LC–MS/MS is used for the detection of tryptic peptides for protein identification. The bottom-up strategy has many merits, including good sensitivity, high throughput and great convenience as well as automation [12,13]. However, this strategy has its own weaknesses. First, many important data of protein structure are not available at the end because only a sequence of tryptic peptide was detected. Second, different protein isoforms with the same peptide sequences may not be distinguishable because of low sequence coverage of protein. Third, data on post-translational modifications (PTMs) of protein are usually lost. Finally, LC–MS/MS based peptide-centric analyses may complicate protein quantification due to the coexistence of protein isoforms. These defects make structural analyses of intact protein very challenging, especially for the PTMs of proteins [13]. To overcome these defects, different top-down proteomic approaches have been developed [13]. In 2004, Takáts et al.
198
C.-G. Guo et al. / Talanta 137 (2015) 197–203
proposed a novel MS sampling under ambient conditions with desorption electrospray ionization [14]. Qiao et al. [15,16] developed an ambient ionization strategy and ESI for in situ ionization of proteins or peptides in the PAGE and coupled IPG-IEF to ESI-MS directly. Thus, protein samples could be firstly separated by IPGIEF, then in situ detected by rapidly scanning gel strip via ESI-MS. This novel top-down strategy can retain the structure information of intact proteins. Furthermore, this approach avoids gel staining and spot extraction, greatly improved the throughput and reduced labor intensity, running time, contaminant and sample loss [15,16]. The IPG-based IEF has been invented as an excellent electrophoretic technique [17,19] to address the cathodic or anodic drifting of pH gradient in Svensson's IEF [20,21], and used in the routine 2DE method for 30 years. However, the IPG-based IEF and 2DE techniques have been facing with the following challenges. First, many denature reagents and solubilizers (urea, thiourea, CHAPS, dithiothreitol, sugar and sorbitol [18,22,23]) were used for improving the solubility of protein in the routine IPG-IEF, making studies of PTMs and protein isoforms very complex and inconvenient. However, a native IPG-IEF without the use of denature reagents and solubilizers resulted in serious random protein precipitate [24] (Section 3.4), greatly affecting resulting sensitivity, resolution, quantitation and reproducibility and making comparative proteomics (e.g., cancer serum vs. normal control serum) dubious. Second, tunable local-zoomed pH gradient in IEF and 2DE was required for high resolution separation of complex protein isoforms and PTMs. However, once the pH gradient of IPG-IEF was immobilized, it could not be adjusted for high resolution separation of many protein isoforms and PTMs [7,25]. Third, sensitivities of routine IPG-IEF and 2DE are still poor [12,13], making identification of low abundance proteins impossible. All of these issues obviously weakened the power of 2DE on characterization of protein isoforms and PTMs [15,16]. To figure out these issues mentioned above, we have developed the concept of moving reaction boundary (MRB) [26,27] from the prototype ideas of moving reactive front [28,29] and stationary neutralization boundary [30]. The concept of MRB was used to improve the sensitivity of capillary electrophoresis (CE) [31,32], the design of supermolecular boundary electrophoresis [33,34] and the advancement of total protein content titration [35,36]. Furthermore, we developed the MRB-based dynamic theory of IEF [24,26,27,37,38], which quantitatively explained the mechanism of quasi-stable pH gradient in IEF and gave the quantitative illumination on various pH gradient instabilities [39–41] and Hjertén's mobilization [42–44]. However, to the best of our knowledge, a
tunable local-zoomed pH gradient in IEF has never been designed based on the concept of MRB for 2DE. Therefore, the main purposes herein are to design a tunable non-IPG-IEF via the concept of MRB; to evaluate resolution and sensitivity of the developed system by comparing it with the routine IPG-IEF; to create an novel non-IPG-IEF based adjustable 2DE; and to validate the advantage of tunable nonIPG-IEF used as the first dimension of 2DE by using model proteins and lysates of colon cancer samples.
2. Experimental designs and methods The details about materials, instruments and analytical software used herein, as well as procedures of staining, sample preparation and LC–MS/MS, are given in the Supporting Information. 2.1. Tunable pH gradient of non-IPG-IEF Scheme 1 shows the complete design of tunable pH gradient in non-IPG IEF via R value (viz., judgment expressions of MRB) and its combination with 2DE. In our previous works [26,27], we defined the expression for comparing fluxes of protons in an acid and hydroxyl in a base as follows: R¼
mþ cþ κ 1 ðused for R 40Þ m c κ þ
ð1aÞ
m c κ þ ðused for R o0Þ mþ cþ κ
ð1bÞ
R ¼ 1
where, R is the symbol of judgment expression, κ is the conductivity (S/m); m is the ionic mobility (m2 s 1 V 1); c is the equivalent concentration (equiv./l), the subscripts, “ þ” and “ ”, imply the hydrogen and hydroxyl ions, respectively. Herein, Eqs. (1a) and (1b) are used for R4 0 and R o0, respectively. As shown in Scheme 1, the significances of R value are: (i) we can design a cathodic migrating pH gradient of IEF by setting R4 0, leading to obvious enhancement of separation of acidic proteins; (ii) conversely, we may create an anodic migrating pH gradient by using Ro0, resulting in clear improvement of separation of alkaline proteins; and (iii) we can form a non-migrating pH gradient by choosing R ffi 0, increasing separation of neutral proteins. Thus, one can use tunable pH gradient of IEF to 2DE for high resolution separation of acidic, and/or neutral and/or alkaline proteins (Scheme 1). Interestingly, the non-IPG-IEF also has the merit of
Scheme 1. Diagrams of tunable non-IPG-IEF and adjustable 2DE designed via R value of MRB. The red arrows indicate the migration direction of pH gradient.
C.-G. Guo et al. / Talanta 137 (2015) 197–203
199
Table 1 Comparisons between the pH gradient design via R values and the experiments in Fig. 1. Comboa
Anolyte
Catholyte
R
Design
Result in Fig. 1
1 2 3 4 5
100 mM Gly 300 mM HCl 800 mM H3PO4 800 mM H3PO4 800 mM H3PO4
100 mM Arg 600 mM (NH2CH2)2 600 mM (NH2CH2)2 200 mM NaOH 100 mM KOH
0.25 0.13 0.01 0.14 0.23
Strong anodic migrating Moderate anodic migrating Stable Weak cathodic migrating Strong cathodic migrating
Strong anodic migrating Moderate anodic migrating Stable Weak cathodic migrating Strong cathodic migrating
a
The numbers herein correspond with A, B, C, D, E in Fig. 1.
high sensitivity, almost the same reproducibility and no protein precipitate comparing with IPG-IEF. In Table 1, we chose five electrolyte pairs to form cathodic, anodic and neutral migrating pH gradients. The R values are also given in Table 1. Prior to non-IPG-IEF run, paper pads were individually immersed into the acid and alkali given in Table 1. After an overnight immersion, the immersed pads bridged the non-IPG strips and the electrodes (Fig. S1). After the rehydration (see below), the strips were quickly placed into focusing tray slot, and the anodic and cathodic pads (Table 1) were respectively set at the anode and cathode (Fig. S1). Then the strips were covered by 1 mL mineral oil to prevent water evaporation. Non-IPG-IEF was run as follows: 250 V for 30 min, followed by a voltage gradient from 250 V to 2000 V (total up to 3000 VHr). Finally, constant 1000 V was applied until IEF end (a total of 8 h). The cooling temperature was set as 17 1C. 2.2. Run of IPG-IEF Unlike non-IPG-IEF (Fig. S1B), no pad was used for IPG-IEF run (Fig. S1A). IPG-IEF was generally performed in accordance with the previous procedure [25]. Briefly, rehydrated IPG strips (see gel strip rehydration) were put into focusing tray slot (Fig. S1A). IPGIEF was run as follows: 250 V for 30 min, 1000 V for 1 h, followed by gradual voltage gradient from 1000 V to 5000 V (total up to 4000 VHr). Finally, constant 1000 V was applied until IEF end (about 4 h). The cooling temperature was set as 17 1C.
Fig. 1. pH gradient migrating experiment of non-IPG-IEF via electrolyte pairs of (A) 100 mM Gly and 100 mM Arg; (B) 300 mM HCl and 600 mM (NH2CH2)2; (C) 800 mM H3PO4 and 600 mM (NH2CH2)2; (D) 800 mM H3PO4 and 200 mM NaOH; and (E) 800 mM H3PO4 and 100 mM KOH. (Phyco: phycocinin, Mb: equine myoglobin, Hb: hemoglobin, Cyt C: cytochrome C.) The conditions were given in Section 2.
SDS) was performed in line with the manual of Bio-Rad ProteomeWorksTM System.
3. Results and discussion 3.1. Tunable local-zoomed pH gradients of non-IPG-IEF
2.3. Gel strip rehydration Rehydration was carried out in a rehydration tray in line with the manufacturer's instruction. For IPG strips, a total volume of 150 μL rehydration solution was added for each IPG strip, which was composed of 30 μL glycerol (20%, v/v), 1.9 μL 40% carrier ampholyte (0.5%, w/v), 2 μL IEF protein sample, and 116 μL Milli-Q grade water. For non-IPG strips, the rehydration solution was consisted of 30 μL glycerol (20%, v/v), 7.5 μL 40% carrier ampholyte (2%, w/v), 2 μL IEF standard protein sample, and 121.6 μL Milli-Q grade water. Finally, the strips were placed on rehydration solution and covered by mineral oil overnight. It ought to be emphasized that the denature reagents and solubilizers (urea, thiourea, CHAPS, dithiothreitol, sugar and sorbitol [18,22,23]) were not used in both the IPG IEF and the non-IPG IEF for the comparison of protein precipitate under the nature conditions. 2.4. SDS‐PAGE SDS-PAGE was performed as previously described with minor modification [7]. Briefly, the strips were equilibrated twice (15 min each time) after IEF runs. The first equilibration solution contained 6 M urea, 75 mM Tris–HCl (pH 8.8), 20% (v/v) glycerol, 2% SDS, 0.002% (w/v) bromophenol blue and 2% dithiothreitol (DTT). The second solution was comprised of 6 M urea, 75 mM Tris–HCl (pH 8.8), 20% (v/v) glycerol, 2% SDS, 0.002% (w/v) bromophenol blue and 2.5% iodoacetamide. Then SDS-PAGE (12% T, 2.6% C and 2%
Table 1 shows that the R values are 0.25, 0.13, 0.01, 0.14 and 0.23 for the electrolyte pairs of Nos. 1, 2, 3, 4 and 5, respectively, and reveals the designs on migrating pH gradients of non-IPG-IEF. Thus, by using the pairs of Nos. 1 and 2, we could design strong and moderate anodic migrating pH gradients, respectively. On the contrary, we might respectively obtain weak and strong anodic migrating pH gradients via the pairs of Nos. 4 and 5. Moreover, by using the pairs of No. 3, we could get a stable pH gradient in nonIPG IEF. The five designs could be experimentally achieved, as demonstrated in Fig. 1. The red dotted line represented the location of the blue phycocyanin band, and the green and black dotted lines indicated the locations of brown Hb/Mb band and red Cyt C band, respectively. It was obvious that the blue, brown and pink standard proteins migrated toward the cathode gradually as R value increased. The qualitative experiments were completely consistent with our designs of migrating pH gradients in Table 1. The moving distances (from the anode) of these chromoproteins in Fig. 1 are given in Table S1. It could be seen that the migration distance of each chromoprotein increased with the increasing of R values of the electrolyte pairs, no matter it was blue or brown proteins, quantitatively indicating that the pH gradient plasticity was intimately correlated to the R value. Meanwhile, although pI (9.6) of Cyt C was close to the cathode end, it could also be observed that the red Cyt C migrated outside the cathode just as shown in Panels D and E of Fig. 1, implying the
200
C.-G. Guo et al. / Talanta 137 (2015) 197–203
Fig. 2. Comparative experiments between IPG-IEF (A, left) and non-IPG-IEF (A, right). (B) Gray levels of model proteins in Panel A. Upper: IPG-IEF, bottom: non-IPG-IEF. (C) Zoomed image of protein peak 1 and 2 in Panel B. (D) Zoomed image of protein peak 4, 5 and 6 in Panel B. (F) Quantitation analysis of the seven separated proteins. The conditions were given in the experimental section. The numbers representing the peak of responding proteins. Phoco: phcocyanin, 1, 2 represent two subunits; Mb1: equine myoglobin minor, Mb2: equine myoglobin; Hb: hemoglobin; Cyt C: cytochrome C.
same results as the above. Fig. S2 reveals the experimental reproducibility of non-IPG-IEF. It could be observed that the three repeated trials were very consistent. The RSD values of the distance between the three proteins of phycocyanin, Mb, Cyt C and the anode were 2.1%, 2.1%, 0.6% for non-IPG-IEF and 3.5%, 3.4% and 1.2% for IPG-IEF, respectively (Table S2). All of the experiments in Figs. 1 and 2 and Tables 1, S1 and S2 along with Fig. S2
quantitatively demonstrated that tunable pH gradient of non-IPG IEF via R values has been achieved. 3.2. Resolution comparison of IPG-IEF and non-IPG-IEF As illuminated in Scheme 1, we could acquire desired pH gradient by adjusting R value to improve the resolution of target
C.-G. Guo et al. / Talanta 137 (2015) 197–203
proteins in non-IPG-IEF. To validate our theory, comparison experiment between IPG-IEF and non-IPG-IEF was performed (Fig. 2). Stable pH gradient created via the electrolyte pairs (800 mM H3PO4 and 600 mM (NH2CH2)2, R¼0.01) was chosen to compare IPG-IEF with non-IPG-IEF. It could be seen in Fig. 2A that the four distinct bands of equine Mb minor, equine Mb, human Hb A and human Hb C could be observed in the center region of nonIPG-IEF. However, for IPG-IEF, only three bands were resolved in the same region. Furthermore, Cyt C was resolved more clearly in cathodic extreme of non-IPG-IEF than IPG-IEF. To further investigate the resolution of these protein bands, we attained gray level vs. distance from the anode curve graph of each protein bands via Meta Morph 7.0 software and Origin 8.0 (Fig. 2B). As shown in Fig. 2B, we could detect up to 9 protein bands in non-IPG-IEF, while only 6 bands were observed in IPG-IEF, which qualitatively implied that higher resolution separation of model proteins had been achieved by non-IPG-IEF. The separation of two phycocyanin subunits (peaks 1 and 2) as well as two proteins (Mb and HbA, corresponding to Peaks 5 and 6, respectively) was compared between IPG-IEF and non-IPG-IEF (Panels C and D). In Panel C, the resolution of Peaks 1 and 2 was 0.86 (IPG-IEF) and 0.88 (non-IPG-IEF), implying slight increase of non-IPG-IEF separation. In Panel D, the resolution of Peaks 5 and 6 was 0.26 (IPG-IEF) and 0.43 (non-IPG-IEF), indicating that the resolution of non-IPG-IEF increased by 65% over IPG-IEF. Thus, these data quantitatively manifested that the separation of neutral proteins was obviously increased via the pH gradient of non-IPGIEF designed by R ffi 0. The 2DE comparative experiments between IPG and non-IPG in Figs. S2 and S3 further verified the above conclusion. The resolution enhancement of non-IPG-IEF may be due to the following two reasons. The first reason is that the broadening pH gradient in non-IPG-IEF results in resolution increase of IEF, especially for target proteins. The second was the case that the better solubility of proteins improve the resolution of non-IPG-IEF system. We can observe that obvious protein precipitates occur in IPG-IEF rather than non-IPG-IEF (marked by the black arrow in Figs. 2A and S2), which reduced the resolution of IPG-IEF, as discussed in Section 3.4. 3.3. Sensitivity comparison between IPG-IEF and non-IPG-IEF Fig. 2E reveals the sensitivity comparison between IPG-IEF and non-IPG-IEF. It can be observed that the OD value of each protein in non-IPG-IEF is higher than that of IPG-IEF, indicating the higher sensitivity of non-IPG-IEF. Through comparing the total OD value of seven proteins in non-IPG-IEF with that in IPG-IEF, we found that the total quantitation of the proteins in non-IPG-IEF was 2.7 folds as high as that of IPG-IEF. It was worth mentioning that some proteins could only be detected in non-IPG-IEF (e.g. Hb C, Cyt C). These quantitative results demonstrated that the sensitivity of non-IPG-IEF was clearly superior to that of IPG-IEF. The comparative experiments of 2DE have further verified the same conclusions (see Section 3.5, Figs. 3 and 4). The sensitive increase in non-IPG IEF is caused by the following three reasons. First, the high resolution separation of proteins can further result in good sensitivity. It is observed in Panels A–D in Fig. 2 that these protein bands or peaks in non-IPG-IEF are obviously narrower than those of IPG-IEF. This indicated that the same loading amount of protein sample might lead to higher protein peak (viz., better sensitivity) in non-IPG-IEF. Second, protein sample had better solubility in non-IPG-IEF. It is seen in Figs. 2A and S2A that random protein precipitates occur in IPG-IEF (marked black arrow), rather than non-IPG-IEF (Section 3.4). This implied that better protein solubility (viz., higher actual sample load) could be achieved in non-IPG-IEF. Hence, better sensitivity
201
Fig. 3. Separation of model proteins (pI 4.45–9.6) by (A) IPG-based and (B) nonIPG-based 2DE with staining of Coomassie brilliant blue. Number of 1–16 representing selected protein spots. The conditions were given in Section 2 and the Online Supporting Information.
was obtained more willingly in non-IPG-IEF than IPG-IEF. Third, protein precipitate after electric focusing reduced the actual protein content in its relevant band or spot in IPG-IEF, as will be revealed below. Consequently, the precipitate further decreased the sensitivity of IPG-IEF. 3.4. Precipitate comparison between IPG-IEF and non-IPG-IEF The comparative experiments in Figs. 2A and S2 evidently demonstrated that there were many random protein precipitates in IPG-IEF, but none in non-IPG-IEF. The phenomenon was similar to the previous observation [24]. The random precipitate might cause serious issues simultaneously. First, the precipitate greatly reduced the reproducibility of IPG-IEF and its 2DE (Fig. S2A). The results in Fig. S3 and Table S3 quantitatively displayed the poor reproducibility of IPG-IEF in contrast to non-IPG-IEF. Second, the precipitates obviously reduced the sensitivity of IPG-IEF and its 2DE due to the actual quantity of focused protein decreased at its pI location. Third, the precipitate, especially the wide flaky precipitate in IPG-IEF gel, would have extremely reduced the resolution of IEF and 2DE (see Figs. 2 and S2A) if the strip was further used for 2DE run (see Section 3.5). Fourth, the precipitate, especially the one out of the strip gel, could greatly affect quantification of proteins in IEF, 2DE and DIGE analysis, making comparative proteomics extremely difficult. The relevant work will be conducted in our further study. From the discussion above, it could be evidently seen that the elimination of protein precipitation in non-IPG-IEF could simultaneously promote the repeatability, resolution and sensitivity of proteins in IEF and 2DE (Figs. 2, S2A and S3). And the higher resolution of proteins could lead to better sensitivity as has been described in Section 3.3. In addition, the non-IPG-IEF benefits the studies of protein isoforms and MPTs. In a routine IPG-IEF, numerous denature solubilizers (e.g. urea, thiourea, CHAPS and DTT) must be used to promote solubility of proteins by reducing protein precipitation
202
C.-G. Guo et al. / Talanta 137 (2015) 197–203
ampholytes used in IEF might have a similar effect of low concentration salt. In IPG-IEF, 0.4% free carrier ampholytes was added during the rehydration. Furthermore, many ampholytes were immobilized within IPG gel strip, reducing hydrophilicity of gel matrix. Consequently, these factors led to the low solubility and precipitate of proteins in IPG-IEF. On the contrary, fairly high concentration of carrier ampholytes (2.0%) was used in non-IPGIEF. Thus, protein precipitate was rarely observed in non-IPG-IEF (Figs. 1, 2 and S2). 3.5. IPG-IEF and non-IPG-IEF based 2DE
Fig. 4. Analyses of proteins from colon cancer cell lines HCT116 by (A)– (C) corresponding Gaussie image of Fig. S4A–C, respectively. Conditions in (A): 800 mM H3PO4 and 200 mM NaOH used as anolyte and catholyte, respectively; Conditions in (C): 300 mM HCl and 600 mM (NH2CH2)2 used as anolyte and catholyte, respectively. The other conditions were given in detail in Section 2 and the Online Supporting Information. The yellow “þ ” indicates protein spots detected by PD Quest software. Red arrow refers to the migrating direction. The letters of ‘a–d’ are corresponding to these selected protein spots for comparison.
[18,22,23]. But these solubilizers can destroy quaternary structure of protein, making the studies of isoforms and MPTs difficult [15,16]. Conversely, the non-IPG-IEF is native milieu, which did not destroy the quaternary structure of protein, suitable for separation of protein isoforms and MPTs. Finally, the non-IPG-IEF was simple and low cost due to no denature reagents were used (e. g. urea, thiourea, CHAPS and DTT) for increasing the solubility of proteins [17]. The detail mechanism of protein precipitates in IPG-IEF (Figs. 2 and S2) is still unknown. It may be caused by the following reasons. It is well known that high concentration of salt (e.g., higher than about 1.5 M) can reduce protein solubility and lead to protein precipitate [43]. However, low concentration range salt (e.g., less than about 1.5 M) can increase protein solubility [45]. In other words, under low concentration of salt, protein solubility will increase with the increase of salt concentration. Free carrier
3.5.1. Separation of model protein subunits We compared the 2DE images of IPG system with non-IPG through analyzing standard proteins. In Fig. 3, the upper pattern and bottom pattern was obtained via IPG-IEF and non-IPG-IEF, respectively. The labeled protein spots in dotted boxes in green, red, yellow and black were selected to assess the resolution and sensitivity of 2DE. We could found that all model proteins could be resolved by IPG- and non-IPG-based 2DE. However, due to the superior pH gradient, the non-IPG-based 2DE could resolve more protein subunits than the IPG-based 2DE, as shown in the red, yellow and black dotted box of Fig. 3B. It could be seen from red dotted box that the non-IPG-based 2DE could resolve eight subunits of hemoglobin A (spots 1, 2, 3, 4) and C (spots 5, 6, 7, 8), illustrating higher separation power of protein isoforms. The protein spots in non-IPG-IEF were successfully identified by NanoLC/Q-TOFMS (Table S4). On the contrary, the same protein spots surrounded by a red ring in the IPG-based 2DE gathered together and could not be discriminated. Meanwhile, identical results could be observed in yellow ellipse ring, demonstrating that protein spots could be more efficiently resolved in the non-IPG-based 2DE (see Nos. 9, 10 and 11). These protein spots were identified by NanoLC/Q-TOF-MS as well (Table S4). It could be seen that three subunits of lectin protein were recognized in the non-IPG-based 2DE, while they were very blurry in the IPG-based 2DE. Similar results could be seen in protein spots 14 and 15. The deep capability of our approach to resolve different protein subunits demonstrated the separation power of non-IPG-based 2DE. Additionally, low abundance protein spots 12, 13, 16 were selected to study the sensitivities of IPG-based and non-IPGbased 2DE systems (Table S5). The gray intensity of spots 12, 13, 16 increased by 18.5%, 29.5%, 66.3% in non-IPG-based 2DE compared to that of IPG-based 2DE, respectively. p Value of these three protein spots were 0.022, 0.003, 0.0005 (n ¼3). Each of the p values was less than 0.05, which implied significant difference of intensity between IPG-based and non-IPG-based 2DE systems. 3.5.2. Adjustable 2DE via tunable non-IPG-IEF Herein, 800 mM H3PO4 and 200 mM NaOH were used as anolyte and catholyte, respectively, representing a cathode migrating protein spots in 2DE (Fig. 4A). Meanwhile, 300 mM HCl and 600 mM (NH2CH2)2 were chosen to perform anode migrating protein spots (Fig. 4C). A routine IPG-IEF-based 2DE was carried out as a control (Fig. 4B). Fig. 4 shows the Gaussian 2DE graphs using colon cancer cell lines HCT116 as samples. In order to investigate the resolution of protein spots in three 2DE systems, we compared the difference of relative position of four typical protein spots (marked as red letter a, b, c, d) through calculating the horizontal distances between them. It could be seen that horizontal distances between protein spot a and b (1.24 cm) in Fig. 4A were expanded significantly when comparing with that (0.85 cm) in Fig. 4B. However, the distance between spot c and d (1.54 cm) in Fig. 4A was narrower than that (2.37 cm) in Fig. 4B.
C.-G. Guo et al. / Talanta 137 (2015) 197–203
These results indicated that a superior separation power existed in the acidic region of the cathode migrating system. Similarly, horizontal distance between spot c and d (2.90 cm) in Fig. 4C was widened comparing to that (2.37 cm) in Fig. 4B. Especially, the distance between spots b and d had a distinct broaden in Fig. 4C (4.54 cm) than in Fig. 4B (3.51 cm), revealing a superior separation power existed in the alkaline region of the anode migrating system. The results mentioned above were completely consistent with our design of IEF and its 2DE in Scheme 1. This is to say that the tunable 2DE could enhance the resolution of proteins in 2D by broadening the desired pH gradient region (e.g., cathodic region or anodic one). Subsequently, we investigated the sensitivity of IPG- and nonIPG-IEF based 2DE system by counting the proteins detected by PD Quest software in Gaussian 2DE graphs. 287 and 280 proteins were detected in Fig. 4A and C, respectively, while only 262 proteins were detected in Fig. 4B. The number of detected protein spots increased by 9.5% and 6.9%, respectively, indicating that nonIPG-based 2DE was more sensitive than IPG one. This echoed the result of IEF in Section 3.3. The relevant mechanism has been discussed in the above section. Therefore, we concluded that enhanced solubility of low abundance proteins led to increase of sensitivity or number of detected protein spots in 2DE.
4. Conclusion In this study, a novel non-IPG-IEF was designed through adjusting the R value of MRB. In this novel system, by choosing a proper pair of catholyte and anolyte, we could adjust the cathodic or anodic migrating pH gradient for non-IPG-IEF. The experiments revealed that the migrating pH gradient in Fig. 1 was completely consistent with R value based design (Table 1). Meanwhile, the sensitivity in non-IPG-IEF was also enhanced (increased by up to 270%) due to elimination of protein precipitation. Subsequently, non-IPG-based 2DE was developed and applied for separation of model proteins. For the proteome analysis of colon cancer cells, tunable non-IPG-IEF based 2DE resolved up to about 16.4% (combination of Figs. 4A and C) more protein spots than IPGbased 2DE. Thus, tunable non-IPG-IEF demonstrated superior resolution, repeatability and sensitivity than routine IPG-IEF approach, and benefited the separation of protein isoforms and PTMs. In addition, the non-IPG IEF was simple and low cost. Although there are so many advantages for non-IPG-IEF system, some drawbacks also drew our attention. For instance, we could not quantitatively know the accurate displacement from the anode corresponding to each protein band when using the electrolyte pairs with specific R value. In addition, some issues still exist when interfacing the electrode paper pad with the gel strip during the experiment. This problem will greatly affect the homogeneity of individual protein band. We plan to address these issues in our future work.
Acknowledgments This work was supported by grants from the National Natural Science Foundation of China (Nos. 21035004, 21275099, 21305087, and 21475086), the National Key Development of Scientific Instruments (No. 2011YQ030139) and the Key Scientific Project of Shanghai Jiao Tong University (Nos. YG2010ZD209 and YG2014QN21). H.X. is supported by the Recruitment Program of Global Youth Experts of
203
China and National High-tech R&D Program of China (863 Program, No. 2014AA020545).
Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.talanta.2015.01.038. References [1] M. Mann, R.C. Hendrickson, A. Pandey, Annu. Rev. Biochem. 70 (2001) 437. [2] Y. Ge, B.G. Lawhorn, M. ElNaggar, E. Strauss, J.H. Park, T.P. Begley, F.W. McLafferty, J. Am. Chem. Soc. 124 (2002) 672. [3] K.L. Williams, D.F. Hochstrasser, Proteome Research: New Frontiers in Functional Genomics, Springer, Verlag, Berlin Heidelberg, New York (1997) 10. [4] L. Zamdborg, R.D. LeDuc, K.J. Glowacz, Y.B. Kim, V. Viswanathan, I.T. Spaulding, B.P. Early, E.J. Bluhm, S. Babai, N.L. Kelleher, Nucleic Acids Res. 35 (2007) W701. [5] R. Aebersold, M. Mann, Nature 422 (2003) 198. [6] P.H. O'Farrell, J. Biol. Chem. 250 (1975) 4007. [7] T. Rabilloud, Proteome Research: Two-Dimensional Gel Electrophoresis and Identification Methods, Springer, Verlag, Berlin Heidelberg, New York, 2000. [8] Y. Shen, N. Tolic, C. Masselon, L. Pasa-Tolic, D.N. Camp, K.K. Hixson, R. Zhao, G.A. Anderson, R.D. Smith, Anal. Chem. 76 (2004) 144. [9] M. Karas, F. Hillenkamp, Anal. Chem. 60 (1988) 2299. [10] K. Tanaka, H. Waki, Y. Ido, S. Akita, Y. Yoshida, T. Yoshida, T. Matsuo, Rapid Commun. Mass Spectrom. (1988) 151–153. [11] T.R. Covey, A.P. Bruins, J.D. Heninon, Org. Mass Spectrom. 23 (1988) 178. [12] J.B. Fenn, M. Mann, C.K. Meng, S.F. Wong, C.M. Whitehouse, Science 246 (1989) 64. [13] S. Wu, N.M. Lourette, N. Tolic, R. Zhao, E.W. Robinson, A.V. Tolmachev, R.D. Smith, L. Pasa-Tolic, J. Proteome Res. 8 (2009) 1347. [14] Z. Takats, J.M. Wiseman, B. Gologan, R.G. Cooks, Science 306 (2004) 471. [15] L. Qiao, R. Sartor, N. Gasilova, Y. Lu, E. Tobolkina, B. Liu, H.H. Girault, Anal. Chem. 84 (2012) 7422. [16] L. Qiao, E. Tobolkina, B. Liu, H.H. Girault, Anal. Chem. 85 (2013) 4745. [17] B. Bjellqvist, K. Ek, P.G. Righetti, E. Gianazza, A. Görg, R. Westermeier, W. Postel, J. Biochem. Biophys. Methods 6 (1982) 317. [18] P.G. Righetti, Immobilized pH Gradient: Theory and Methodology, Elsevier, Amsterdam, New York, Oxford (1990) 8–13. [19] P.G. Righetti, Isoelectric Focusing: Theory, Methodology and Applications, first edn., Elsevier Biomedical Press, Amsterdam, New York, Oxford, 1983. [20] H. Svensson, Acta Chem. Scand. 15 (1961) 325. [21] H. Svensson, Acta Chem. Scand. 16 (1962) 456. [22] P.G. Righetti, E. Fasoli, S.C. Righetti, in: Jan Christer Janson (Ed.), Protein Purification: Principles, High Resolution Methods, and Applications, John Wiley & Sons, Inc., 2011, pp. 379–409. [23] P.G. Righetti, A. Bossi, Anal. Biochem. 247 (1997) 1. [24] C.G. Guo, S. Li, H.Y. Wang, D. Zhang, G.Q. Li, J. Zhang, L.Y. Fan, C.X. Cao, Talanta 111 (2013) 20. [25] A. Görg, C. Obermaier, G. Boguth, W. Weiss, Electrophoresis 20 (1999) 712–717. [26] C.X. Cao, J. Chromatogr. A 183 (1998) 173–177. [27] C.X. Cao, L.Y. Fan, W. Zhang, Analyst 133 (2008) 1139. [28] J. Deman, W. Rigole, J. Phys. Chem. 74 (1970) 1122. [29] J. Deman, Anal. Chem. 42 (1970) 321. [30] J. Pospichal, M. Deml, P. Bocek, J. Chromatogr. A 638 (1993) 179. [31] C.X. Cao, Y.Z. He, M. Li, Y.T. Qian, M.F. Gao, L.H. Ge, S.L. Zhou, L. Yang, Q.S. Qu, Anal. Chem. 74 (2002) 4167. [32] C.X. Cao, W. Zhang, W.H. Qin, S. Li, W. Zhu, W. Liu, Anal. Chem. 77 (2005) 955. [33] L. Fan, W. Yan, C. Cao, W. Zhang, Q. Chen, Anal. Chim. Acta 650 (2009) 111. [34] J. Dong, S. Li, H. Wang, Q. Meng, L. Fan, H. Xie, C. Cao, W. Zhang, Anal. Chem. 85 (2013) 5884. [35] H.Y. Wang, C.Y. Guo, C.G. Guo, L.Y. Fan, L. Zhang, C.X. Cao, Anal. Chim. Acta 774 (2013) 92. [36] H. Wang, Y. Shi, J. Yan, J. Dong, S. Li, H. Xiao, H. Xie, L.Y. Fan, C.X. Cao, Anal. Chem. 86 (2014) 2888–2894. [37] H. Liang, Y. Chen, L.J. Tian, L. Zhang, Electrophoresis 30 (2009) 3134. [38] H. Liang, L.F. Ou Yang, Q. Liu, L. Zhang, L.J. Tian, Y. Chen, J. Sep. Sci. 34 (2011) 1212–1219. [39] A. Chrambach, P. Doerr, G.R. Finlayson, L.E. Miles, R. Sherins, D. Rodbard, Ann. N. Y. Acad. Sci. 209 (1973) 44. [40] H. Rilbe, in: B.J. Radola, D. Graeslin (Eds.), Proceedings of the International Symposium, Walter de Gruyter & Co., Hamburg, Germany, 1977, pp. 35–50. [41] I.K.O.K., A. Murel, J. Chromatogr. A 174 (1979) 1. [42] S. Hjertén, M. Zhu, J. Chromatogr. A 346 (1985) 265. [43] S. Hjertén, J.L. Liao, K.Q. Yao, J. Chromatogr. 387 (1987) 127. [44] F. Kilar, S. Hjertén, Electrophoresis 10 (1989) 23. [45] A.A. Green, J. Biol. Chem. 95 (1932) 47–66.
