Preparation of mixed-mode stationary phase for separation of peptides and proteins in high performance liquid chromatography

Porous silica particles were prepared by sol–gel method with some modification to get wide-pore particles. These particles were derivatized with N-phenylmaleimide-methylvinylisocyanate (PMI) and styrene by reversible addition fragmentation chain transfer (RAFT) polymerization to prepare N-phenylmaleimide embedded polystyrene (PMP) stationary phases. Narrow bore stainless steel column (100 × 1.8 mm i.d) was packed by slurry packing method. The chromatographic performance of PMP column was evaluated for the separation of synthetic peptides mixture composed of five peptides (Gly-Tyr, Gly-Leu-Tyr, Gly-Gly-Tyr-Arg, Tyr-Ile-Gly-Ser-Arg, Leucine enkephalin) and tryptic digest of human serum albumin (HAS) respectively. Number of theoretical plates as high as 280,000 plates/m were obtained for peptides mixture at optimum elution condition. Separation performance of the developed column was compared with commercial Ascentis Express RP-Amide column and it was observed that separation performance of PMP column was better than commercial column in terms of separation efficiency and resolution.

In recent years the biopharmaceutical industry has become an expanding global market with a substantially increased market share. The analysis of peptides and proteins is highly desired with the explosive growth of biopharmaceutical industry [1][2][3] . Several impurities are produced during peptide synthesis in addition to the target peptides, therefore chromatographic purification is required to get peptides of the required purity. Analysis and characterization of proteins in bodily fluids, tissues and cells is a highly challenging task owing to the presence of large number of potentially detectable species in a single sample. Even though mass spectrometry is an effective tool for peptides and proteins sequencing but if such samples are injected at once to mass spectrometer, the separation will be not satisfactory. This issue can be mitigated by the implementation of liquid chromatography (LC) separation prior to MS analysis, which will reduce the number of analytes entering to the mass spectrometer Synthesis of polar-embedded RP stationary phase. A stationary phase containing both polar groups and polystyrene chain was prepared by the surface modification of silica particles with a pre-synthesized ligand phenylmaleimide-methylvinylisocyanate (PCMP) followed by radial polymerization with styrene. The preparation procedure is described below.
Functionalization of silica particles with PMCP. Dried silica particles (2 g) were dispersed in dry toluene (100 mL), stirred and sonicated for 10 min in 500 mL round bottom flask. PMCP (10 mg) was dissolved in toluene and added drop-wise into the reaction flask through a dropping funnel. The mixture was refluxed at 100 °C for 8 h, filtered and washed with acetone and dried at 60 °C for 3 h. Then, PMCP bonded silica particles (100 g) was dissolved in toluene (200 ml), then 4-hydroxy-TEMPO (2 mL) was added into it in the presence of 100 µL dibutyl-tin-dilaurate as a catalyst. The mixture was stirred at 50 °C for 8 h, filtered and dried at 50 °C for 3 h. www.nature.com/scientificreports/ Radical Polymerization of Styrene. Styrene (1 mL), benzoyl peroxide BPO (0.5 mL) and TEMPO-PMCP attached silica particles (1.5 g) were dispersed in toluene and purged with nitrogen. The polymerization of styrene was carried out at 100 °C for 12 h. The resultant product was washed with methanol and dried at 60 °C overnight. The overall reaction scheme is given in the Fig. 1.
Characterization of the stationary phase. The samples were degassed at 393 K for 1 h to obtain a residual pressure of less than 10 −3 Torr. The amount of N 2 adsorbed at a relative pressure of P/P0 = 0.99 was used to determine the total pore volume. The morphology of bare and ligand bonded silica particles was examined  www.nature.com/scientificreports/ with scanning electron microscope (Hitachi High Technologies, Tokyo, Japan). The dried samples (bare silica and ligand bonded silica particles) were put on aluminum stubs using adhesive carbon tape. A Q150T sputter coater was used to coat gold over the sample and 5 nm Au layer was deposited over the sample. This enhances the efficiency of the process using low voltages and giving a fine-grain, cool sputtering. A Thermo Electron (Waltham, MA, USA) Flash EA1112 elemental analyzer was used to carry out elemental analysis. A Malvern (Worcestershire, UK) Mastersizer 2000 particle size analyzer was used to get particle size distribution. Bare silica particles and ligand bonded silica particles (5 mg each) were dispersed in 5 mL isopropyl alcohol, sonicate for 10 min, vortexes for 5 min, and subjected to the optical bench of Mastersizer. Thermo gravimetric analysis were carried out over a temperature range of 30 to 800 °C at 5 °C per minute.
Column packing. Glass-lined stainless-steel narrow-bore column of dimensions (100 × 1.8 mm i.d.) was packed using slurry packing method, the same procedure was applied as used in Ref. 31 . A stainless steel column (glass lined, 100 × 1.8 mm i.d.) with an outlet union containing 1 µm screen frit was connected to the slurry packer (Alltech Deerfield, IL, USA). The slurry of stationary phase was prepared by suspending the 150 mg stationary phase in 1.2 mL methanol, and fed into the reservoir column. Methanol was used as slurry solvent as well as pushing solvent. The columns were packed by applying a pressure sequence of 100 MP for 10 min, 80 MP for 15 min, and 60 MP for 30 min. Mechanical vibration was applied with two GC column vibrators (Alltech, Deerfield, IL, USA) during packing to ensure the homogenous packing of column. Slurry packer was closed and the pressure was released slowly to prevent any damage inside the column. Column was disconnected from the slurry packing set-up and another union was connected to the inlet and connected to LC system to check its performance. Columns evaluation. PMP column was evaluated for the separation of peptides mixture and tryptic digest of HSA respectively. PMP column was checked for the separation of peptides mixture and tryptic digest of HSA and the results were compared with Ascentis Express RP-Amide column. Numbers of theoretical plates were calculated by the following equation:

HPLC analysis.
where w 1/2 is the bandwidth at half height and tr is the retention time.

Results and discussion
Morphology of the stationary phase. The SEM images of bare silica particles and ligand bonded silica particles are shown in Fig. 2. SEM images of bare silica particles (A,B) show that these particles are spherical in shape as compared to our previous study where the particles were elongated or having an irregular symmetry. The surface of ligand bonded silica particles (C, D) are smoother than bare silica particles which may be owing to the coating of polystyrene chain over the surface of silica particles.
Particle size distribution. The particle size distribution of bare silica particles and ligand bonded silica particles is shown in Fig. 3(A). The volume-based particle size distribution curves show that the size of silica particles increases after chemical modification (Fig. 3A). The particles size distribution data of silica particles of the current study and previous study are compared in Table 1(A). The volume-based particle size d(0.5) of PMP is 3.36 μm while d(0.5) value of our previous study was 3.05 μm (polystyrene bound silica particles) 34 . The particle size distribution of this batch is narrow as compared to our previous study 34 due to change in the ratio of PEG, urea, TMOS and acetic acid in reaction mixture. The particles size of PMP phase is little bigger than polystyrene bound silica particles phase of our previous study. It means that the surface functionalization of silica particles with styrene only have deposited a polystyrene layer (0.97 µm) over the silica surface while in PMP phase the layer thickness is 1.38 µm.
Pore size distribution of bare and ligand bound silica particles. The pore size, pore volume and surface area of silica particles of the current this study are given in Table 1(B). The PSD plots of bare silica particles and ligand bonded silica particles are shown in Fig. 3(B). The results are comparable to our previous study 34 . www.nature.com/scientificreports/ The pore size of bare and ligand bound silica particles is 310 Å and 241 Å respectively, which indicated that after chemical modification the pore size decreased by 69 Å as shown in Table 1(B) and the curve's shift can be seen in Fig. 3(B). Similarly, the pore volume of silica particles decreased from 0.67 to 0.58 cm 3 /g after chemical modification. The specific surface area of silica particles of current study is 116 m 2 /g which is comparable to that of our previous study (124 m 2 /g). The surface area (m 2 /g) of silica particles also decreased from 116 m 2 /g to 105 m 2 /g after chemical modification as shown in Table 1(B).
Characterization of PMP stationary phase. The elemental analysis results of the stationary phase is shown in Table 2. The carbon load of current stationary phase is 6.35%, which is less than those of our previous studies, (polystyrene bonded silica particles, which were 7.93% 35 and 10.21 %) 42 . The carbon load of the current stationary phase is low because some polar ligands such as phenylmaleimide-methylvinylisocyanate (PCMP) and 4-hydroxy-TEMPO were used in addition to styrene in the preparation of current SP. The nitrogen weight %  www.nature.com/scientificreports/ of current stationary phase is 2.21% while that of previous studies were 0.17 35 and 0.85% 42 respectively. It means that the wt % of nitrogen is high in the current stationary phase owing to phenylmaleimide. Similarly the carbon loads of product (4) and (5) are 2.7% and 2.9% respectively, while the carbon load of final product (6) is 6.35% as shown in Table 2. Thermo gravimetric analysis (TGA) was carried out for PMP stationary phase to check the weight loss, and the TGA curve is shown in Fig. 4. The TGA curve shows the weight loss of 8.6% which is in good agreement with the carbon load (6.35%) since the ligand contains not only C but also N, O, and H. Phenylmaleimide-methylvinylisocyanate ligand was chosen for surface modification of silica particles because it have a polar phenylmaleimide group and vinyl isocyanate group. Vinyl isocyanate group can be further reacted with styrene via living radical polymerization. The second reason was to insert a group which have moderate interaction with analyte and do not have strong electrostatic interactions between analytes and stationary phase since the phenylmaleimide moiety have no virtual charge at normal pH. The polarity of the stationary phase  Table 2. Elemental analysis of polystyrene bound silica particles. a Corresponds to the product (4) in Fig. 1 (reaction scheme). b Corresponds to the product (5) in Fig. 1 (reaction scheme). c Corresponds to the product (6) in Fig. 1 Table 3 and the chromatograms are shown in Fig. 5A. Fast analysis was carried out at high flow rate (700 μL/min) on PMP column and five peptides were eluted within one minute with quite good N value 13,500 ± 330 per column (100 × 1.8 mm i.d.) corresponding to 135,000 plates/m (Fig. 5B). Three columns of the same dimensions (100 × 1.8 mm i.d. were packed from three different batches of PMP stationary phases to check the reproducibility. The same test mixture was separated on each column using optimum elution conditions and the number of theoretical plates N and retention times of analytes were recorded for each column. The reproducibility data of PMP columns are shown in Table 4. There is a good correlation in reproducibility of PMP columns with very low %RSD values as shown in Table 3.  www.nature.com/scientificreports/ Separation of tryptic digest of Human Serum Albumin (HSA). PMP column (100 × 1.8 mm i.d) was evaluated for the separation of tryptic digest of human serum albumin in high performance liquid chromatography. The chromatogram in Fig. 6 shows that the sample is separated well with quite good resolution. The HSA digest was analyzed using flow rate of 100 µL/min, mobile phase 70/30 acetonitrile/water with 0.1% TFA. The HSA digest   www.nature.com/scientificreports/ has been separated into 17 peaks as shown in the chromatogram (Fig. 6) which corresponds to 17 peptides. The separation efficiency of individual peaks from HSA digest are calculated and the values are given in Table 5.
Permeability of PMP column. Column permeability was calculated using the following equation.
where L is the column length, η is the viscosity of the mobile phase, ΔP is the column back pressure and u is mobile phase linear velocity. The permeability of PMP column was 2.5 × 10 −14 m 2 at 25 μL/min using 60/40 v/v ACN/water. The permeability of the PMP column (100 × 1.8 mm i.d) is similar to that of our previous study Ref. 34 . The permeability of the columns packed with superficially porous particles were noted to be 1.7 × 10 −15 for 1.3 μm particles, 3.1 × 10 −15 for 1.7 μm particles, 5.2 × 10 −15 for 2.6 μm particles and 2.5 × 10 −14 m 2 for 5 μm particles 43 . So, the permeability PMP phase is similar to the permeability of 5 μm core-shell particles.
Porosity of PMP column. Porosity was calculated by the following equation using methanol and chloroform as filling solvents where Wx is the weight of the column filled with chloroform and Wy is the weight of column filled with methanol, ρ is the density of the solvent. Density of methanol (ρ = 0.7866) and that of chloroform (ρ = 1.484). The total porosity of SILICA PARTICLES -C18 column (100 × 1.8 mm i.d) 34 and C18-Urea column 31 of our previous study were found to be 0.63 and 0.55 respectively. It means that the presence of urea ligand has decreased the permeability of stationary phase. On the other hand the total porosity of PMP column (100 × 1.8 mm i.d) was 0.60. The permeability PMP column is less than C18 bonded silica particles packed column because in C18 types stationary phases the C18 ligands attached to silica particles in the form of linear chains while in polystyrene types stationary phases relatively thick layer of polymer form around silica particles. In a typical experiment the column porosity was calculated as, Van Deemter plots. Figure 7A, B shows the van Deemter plot of PMP column (100 × 1.8 mm i.d) and Ascentis Express RP-Amide column (100 × 1.8 mm i.d) respectively using the same elution conditions i.e. 60/40 ACN/H 2 O with 0.1% TFA. Separation of the selected peptides mixture (Gly-Tyr, Gly-Leu-Tyr, Gly-Gly-Tyr-Arg, Tyr-Ile-Gly-Ser-Arg, Leucine Enkephalin) was carried out at various flow rates ranging from 20 µL/min to 800 µL/min on both columns. The minimum HETP values at optimum flow rate (80 µL/min) for PMP column and Ascentis Express RP-Amide column were found to be 2.6 µm and 3.9 µm respectively. The HETP values shows that the separation efficiency of PMP column (100 × 1.8 mm i.d) is much better than commercially available Ascentis Express RP-Amide column (100 × 1.8 mm i.d). The van Deemter plot in Fig. 7(A) shows that the decrease in N value with increase in flow rate is not significantly high as compared to our previous study 34 . Better separation efficiency of the PMP column (100 × 1.8 mm i.d) in comparison to Ascentis Express RP-Amide column is based on the improvement in particles shape, size and sophisticated column packing procedure used in the current work 34 .

Conclusions
A polar embedded polystyrene stationary phases was prepared and evaluated for the separation of synthetic peptides mixture and tryptic digest of human serum albumin (HAS) in high performance liquid chromatography. The chromatographic performance of PMP column for peptides mixture is excellent in terms of separation efficiency and resolution. The enhanced separation performance of PMP column is owing to several reasons such as particle size and pore size of silica particles, controlled synthesis of the stationary phase and sophisticated column packing. In addition to high separation efficiency, column back pressure was low at high flow rates, which is an additional advantage of this stationary phase. The PMP column exhibited good repeatability and could be applied for the analysis of peptides mixture and tryptic digest of various proteins. We intend to apply this column for the separation of natural products, bioactive compounds from medicinal plants and fungi extracts in liquid chromatography. In future the PMP column will also be evaluated for the separation of proteins and monoclonal antibodies.