Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
Contents lists available at ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba
Evaluation of new superficially porous particles with carbon core and nanodiamond–polymer shell for proteins characterization Balázs Bobály a , Davy Guillarme b , Szabolcs Fekete b,∗ a b
Budapest University of Technology and Economics, Department of Inorganic and Analytical Chemistry, Szt. Gellért tér 4, 1111 Budapest, Hungary School of Pharmaceutical Sciences, University of Geneva, University of Lausanne, Bd d’Yvoy 20, 1211 Geneva 4, Switzerland
a r t i c l e
i n f o
Article history: Received 16 September 2014 Received in revised form 18 November 2014 Accepted 20 November 2014 Available online 27 November 2014 Keywords: Superficially porous Peak capacity FLARE® Aeris® Protein separation
a b s t r a c t A new superficially porous material possessing a carbon core and nanodiamond–polymer shell and pore size of 180 A˚ was evaluated for the analysis of large proteins. Because the stationary phase on this new support contains a certain amount of protonated amino groups within the shell structure, the resulting retention mechanism is most probably a mix between reversed phase and anion exchange. However, under the applied conditions (0.1–0.5% TFA in the mobile phase), it seemed that the main retention mechanism for proteins was hydrophobic interaction with the C18 alkylchains on this carbon based material. In this study, we demonstrated that there was no need to increase mobile phase temperature, as the peak capacity was not modified considerably between 30 and 80 ◦ C for model proteins. Thus, the risk of thermal on-column degradation or denaturation of large proteins is not relevant. Another important difference compared to silica-based materials is that this carbon-based column requires larger amount of TFA, comprised between 0.2 and 0.5%. Finally, it is important to mention that selectivity between closely related proteins (oxidized, native and reduced forms of Interferon ␣-2A variants) could be changed mostly through mobile phase temperature. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Thanks to its high resolving power, reversed phase liquid chromatography (RPLC) is an important and promising technique in the separation of biological macromolecules [1–3]. Due to the development of instruments [4,5] and column technology [6–10], the past decade served significant improvements in the separation of proteins under reversed phase conditions. Sub-2 m fully porous particles and core–shell materials possessing large pore size in the range of 170–300 A˚ provide outstanding performance when analyzing proteins larger than 10–15 kDa [9]. This is explained by the excellent kinetic performance produced by small fully porous particles and by the so-called core–shell advantages [11] when using superficially porous materials. Except for kinetic performance improvements, recently packed column developments focused on the introduction of various stationary phase chemistries (e.g. C3 C12, pentafluoro-phenyl, phenyl-hexyl, biphenyl, etc.), assumed to serve different selectivity from the conventional C18 phases.
∗ Corresponding author. Tel.: +41 36 30 395 6657; fax: +41 22 379 6808. E-mail address: [email protected] (S. Fekete). http://dx.doi.org/10.1016/j.jpba.2014.11.039 0731-7085/© 2014 Elsevier B.V. All rights reserved.
Surface modifications have been executed on the traditional silica or hybrid-based materials, which are the workhorses of liquid chromatography [12]. Recently, an alternative, carbon–nanodiamond based C18 superficially porous material has been introduced to the market under the trademark FLARE® from Diamond Analytics. This new column technology could be particularly interesting in the separation of large analytes, since it possesses an average pore size of 180 A˚ and offers the benefits of core–shell particle technology. Natural diamond microparticles have been first reported as chromatographic packing in 1973 [13]. However, the low surface area and inhomogeneous adsorption properties of the particles (due to their natural origin) resulted in poor efficiency. Besides efficiency issues, limited availability and high costs prevented their spread in liquid chromatography. Thanks to continuous developments in diamond production technology over the last 40 years, high purity synthetic particles of controlled size and shape became easily accessible at relatively low costs [14]. The physical and chemical properties of these high quality particles renewed interest for chromatographic applications. Diamond particles possess excellent thermal conductivity [14], pH stability [15] and are biocompatible [16], which could make them useful for protein separations. The core of the commercial state-of-the-art FLARE®
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
material is a carbonized poly(divinylbenzene) particle with a diameter of ca. 3.4 m. Poly(allylamine)–nanodiamond hetero-layers are deposited onto the surface of the carbonized core by a modified layer-by-layer method [17]. The resulting core–shell is synthesized to a shell thickness of ca. 0.1 m and a finished particle size of 3.6 m with a surface area around 23 m2 /g. The particles are then functionalized, cross-linked and sieved. The surface area is somewhat lower than on silica-based or graphitic phases but provides sufficient retentivity. Studies in the past few years showed the applicability of carbon/nanodiamond/polymer based shell particles both in normal and reversed phase conditions [18]. Thanks to the improvement in the manufacturing technology of these particles and columns, a continuous increase of kinetic efficiency in the separation of low molecular weight solutes has been reported [14]. Latest results showed promising 100,000–120,000 N/m values for the separation of alkyl benzenes which corresponds to a reduced plate height (h) value of ∼2.3 [17]. These efficiencies are outstanding compared to earlier carbon/nanodiamond/polymer based shell particles, but are still below the efficiencies reported for silica based shell materials. The 3.6 m wide pore shell silica particles showed 150,000–200,000 N/m [9], while sub-3 m shell silica particles dedicated for macromolecular separations provided 200,000–260,000 N/m [8], measured by low molecular weight analytes. This clearly reflects that further improvement in the manufacturing technology of carbon/nanodiamond/polymer material is needed. On the other hand, thanks to its mixed-mode hydrophobicanion exchange surface, it could provide alternative selectivity and retention mechanism, compared to traditional reversed phases. To our knowledge, no data on the applicability of these materials for protein separations have been reported to date. The aim of this study was to evaluate the kinetic efficiency and selectivity of this novel stationary phase for protein characterization. Model proteins and real life samples of native, oxidative stressed and reduced therapeutic proteins were analyzed using the FLARE® column to demonstrate its possibilities and limitations. 2. Experimental 2.1. Instrumentation Chromatographic experiments were performed on a Waters Acquity I-Class UPLC system (Waters, Milford, MA, USA). The instrument was equipped with a binary solvent manager, autosampler, thermostated column compartment, and UV detector. The autosampler was equipped with a flow through needle injection system. The average extra-column peak variance of our system 2 ∼0.5–2.5 L2 . The UV detector operated was found to be around ec with a 500 nL flow cell, set to 280 nm and 40 Hz sampling rate. Data acquisition, data handling and instrument control were performed by Empower Pro 2 (Waters, Milford, MA, USA) software. 2.2. Chemicals and columns Acetonitrile (gradient grade), trifluoroacetic acid (TFA, puriss p.a.), hydrogen peroxide (30 wt.%, ACS reag. grade), dithiothreitol (DTT, ≥99.0%), methionine (≥98.0%) and protein standards such as transferrin (human, MW ∼ 77.0 kDa), cytochrome-c (from horse heart, MW ∼ 12.4 kDa) and albumin (BSA from bovine serum, MW ∼ 69.3 kDa) were purchased from Sigma–Aldrich (Buchs, Switzerland). Water was obtained from a MilliQ Purification System from Millipore (Bedford, MA, USA). Recombinant interferon alfa2A (MW ∼ 19.2 kDa, Roferon) was obtained from Roche Pharma (Switzerland). ˚ colFLARE® C18 mixed-mode (100 mm × 2.1 mm 3.6 m, 180 A) umn was purchased from Diamond Analytics (Orem, UT, USA), and
131
˚ column Aeris Widepore XB-C18 (100 mm × 2.1 mm, 3.6 m, 200 A) was purchased from Phenomenex (Torrance, CA, USA). 2.3. Methodology 2.3.1. Sample preparation For the kinetic evaluation, protein standard mixture of transferrin, BSA and cytochrome-c was prepared containing 0.3 mg/mL of each protein in water. For the selectivity studies Interferon alfa-2A was injected as received. Reduced and oxidized proteins have been prepared from 100 L of the Interferon solutions. A small amount of DTT was added to reduce the protein, then the sample was incubated at 30 ◦ C for 30 min in dark. Oxidation was carried out by adding 1 L of hydrogen-peroxide solution and incubation at 30 ◦ C for 60 min. Oxidation was quenched by adding a small amount of methionine to the sample. 2.3.2. Chromatographic methods For the gradient separation of the proteins, mobile phase “A” was prepared from water and mobile phase “B” was prepared from acetonitrile. Peak capacity was measured applying gradient elution of 20–45%B on the FLARE® column, and 30–55%B on the Aeris WP column (in order to keep the apparent retention factors in the same range). Gradient times were set to 4, 12, 20, 28, 36, and 44 min, followed by 1.5 min equilibration. Flow rate was set to 0.2 mL/min and 0.4 mL/min. Mobile phases contained 0.1, 0.2, 0.3 and 0.5% TFA. Columns were thermostated at 30 ◦ C and 50 ◦ C. 2 L of the samples were injected in full loop injection mode. Selectivity studies on Interferon variants included gradient runs of 25–50%B on the FLARE® column. Gradient times were set to 7 min, flow rate was fixed at 0.3 mL/min in all cases. Mobile phases contained 0.3 and 0.5% TFA, and temperature was set at 30, 50 and 90 ◦ C. Detection has been carried out using fluorescence detection (ex : 280 nm, em : 360 nm, 20 Hz sampling rate). 2 L of the samples were injected in full loop injection mode. 2.3.3. Determining peak capacity Peak capacity is a concept first described by Giddings [19] and soon put to good use by Horvath for gradient chromatography [20]. It is a measure of the separation power that includes the entire chromatographic space together with the variability of the peak width over the chromatogram. In this study, peak capacities were experimentally determined from the gradient time (tg ) and the average measured peak width at 50% height (W50% ). The following equation was used to estimate the peak capacity based on peak width at 4, corresponding to a resolution of Rs = 1 between consecutive peaks: P =1+
tg 1.7 · w50%
(1)
In order to avoid the imprecision associated with the measurement of peak widths at base for proteins often containing closely related variants (i.e. for a heterogeneous sample) the peak width at half height was preferred in this study. This way, the impurities present in the sample, and partially resolved from the main component, did not confuse the measurement. Peak capacity was plotted as a function of gradient time. 3. Results and discussion When high molecular weight solutes have to be separated, the column performance is mostly determined by the mass transfer resistance because of the low diffusivity of proteins. Recent work suggests that the improvement in mass transfer of the new generation of superficially porous particles can mostly be explained by the reduction in external film mass transfer resistance (cf term in
132
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
the Knox equation) and not by the reduction in transparticle mass transfer resistance (cp term in the Knox equation) [21]. Increasing the shell porosity accessible to proteins accelerates the mass transfer through the stationary film surrounding the superficially porous particles. According to a recent study of Gritti and Guiochon, the external film mass transfer resistance controls more than 90% of the solid–liquid mass transfer resistance of proteins [21]. The external film transfer becomes slower with decreasing access to the internal volume. Based on these theoretical considerations, it is clear that efficient protein separation requires fine shell type particles having large porosity, large average pore size and decreased shell thickness. However on particles possessing very narrow shell thickness, the retentivity and loadability are somewhat limited. In addition, compared to silica-based materials, detrimental secondary interactions between residual silanols and proteins – as the ion-exchange mechanism – does not exists on carbon and polymer based stationary phases. The surface of synthetic diamond has no functional groups [16] however it is also possible that some of the nanodiamond surfaces may not be completely coated and any oxygenated moieties on those heterogeneous surfaces could also contribute to tailing of more polar analytes [15]. It was shown that both the retention and peak broadening were strongly influenced by the activity of residual silanols on silica-based materials (even when using TFA in the mobile phase) [7]. On silica-based stationary phases, there is a balance between the hydrophobic and ion exchange mechanisms. As an example, on Aeris WP C4 material, the strongest interactions (ion exchange and hydrogen bonding mechanism) was predominant over the weak hydrophobic interaction, even when working in very acidic conditions (0.1% TFA, pH 1.8) where the number of charged silanols was limited [7]. The stationary phase used in this study (FLARE® ) includes a significant concentration of protonated amino groups, due to the presence of poly(allylamine) within the shell structure, so the resulting retention mechanism must be considered as mixed mode including hydrophobic partitioning and anion-exchange [14–17]. Therefore, the carbon based FLARE® C18 mixed-mode phase operates in a reversed-phase and weak anion exchange mode, depending on the pH of the mobile phase. But the main difference compared to silica-based materials is that the ion-exchange affinity of the weak anion-exchange material (allylamine) is probably negligible at low pH. Moreover the additive used in this study possibly masks the ionic sites of the stationary phase as it is an ion-pairing reagent (TFA) and TFA also forms ion-pairs with the positively charged proteins. These all suggest that at low pH (pH < 2), the main driving force is mostly the hydrophobic interactions with the C18 alkylchains on this carbon-based material. However it also has to be considered that hydrogen-bonding interactions may occur between the ion-exchanger functional groups and proteins. Finally, this 3.6 m FLARE® C18 mixed mode material possesses 180 A˚ average pore size. This value is close to the pore size of the silica based Aeris 3.6 m material that is now a goldstandard in proteins’ RP separations. This pore size combined with the core–shell structure should suit for the separation of large molecules. All these features suggest that this new carbon-based material might be applicable for protein separations, even if the initial purpose of the provider was to introduce some unique selectivity for the separation of small molecules [22] and to further improve pH and temperature stability for RP-LC separations (the FLARE® material is stable up to 100 ◦ C and from pH 1 to 13). 3.1. Peak capacity: the effect of temperature, flow rate and additive concentration In the following section, a systematic study shows the achievable peak capacity of the new 3.6 m wide-pore carbon-based
Fig. 1. Effect of mobile phase temperature on peak capacity. Black signs represent results observed at 30 ◦ C, white signs show results at 50 ◦ C. Triangles: CYT-C, squares: TRFE, rhombuses: BSA. Flow rate: 0.2 mL/min, TFA concentration: 0.2%. Other conditions as specified in Section 2.3.2.
superficially porous material in representative gradient elution conditions. Three model proteins including transferrin (TRFE), cytochrome-c (CYT-C) and bovine serum albumin (BSA) were studied under generic gradient conditions (see Section 2.3.2). In RP-LC practice, proteins are separated in gradient elution mode at elevated temperature. Elevated temperature on silica-based phases is beneficial since it decreases the secondary interactions between residual silanols and positively charged biomolecules. Moreover, the use of high temperatures strongly enhances analyte diffusion [23,24]. Except elevated temperature, it is also important to add ionpairing reagents, such as TFA, to the mobile phase in order to increase the efficiency of protein separations [25,26]. To emulate real-life separations, we applied 30 and 50 ◦ C as column temperature, added 0.1–0.5% TFA into the mobile phase and used gradient spans between 4 and 44 min which are fairly common in the current practice. The influence of the gradient steepness on peak capacity was also investigated under several conditions. These experimental variables (temperature, TFA concentration and gradient steepness) are often used to adjust a proper separation in practical every day work. The gradient steepness strongly impacts resolution under gradient conditions as it affects the retention factor of the solute in the mobile phase composition upon elution. The peak capacity of 100 mm long narrow bore FLARE® C18 mixed-mode column was calculated and plotted against the gradient time when changing the other parameters. First, the impact of temperature was investigated. Several gradient spans were set and peak capacity was determined at 30 and 50 ◦ C when applying 0.2 mL/min flow rate and 0.2% TFA. Fig. 1A shows representative chromatograms of the test proteins obtained
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
with the 4 min long gradient at two temperatures. As shown, this fast separation is quite impressive (an average peak capacity of around P ∼ 55 in 5 min). All the three proteins eluted as sharp peaks. Similar peak widths can be observed with wide-pore fully porous silica materials of conventional size (3–5 m) [9]. This means that this carbon based material is indeed suitable for the separation of proteins. Fig. 1B shows the corresponding peak capacity plots and two important conclusions can be drawn. Firstly, the peak capacity does not change with temperature. That is quite uncommon since in most cases, peak capacity increases significantly at elevated temperature, through the reduction of secondary ion-exchange interaction strength. It proves that the ionic interactions are probably negligible on this special phase under the applied conditions. Moreover, it provides an important advantage compared to silica based phases, as there is no need to work at high temperature with this carbon based material. Proteins RPLC separations are generally performed at 50–80 ◦ C to obtain acceptable peak shape and peak width. However, the risk of thermal on-column degradation of large proteins may increase with temperature [27]. Higher temperature in the range of 70–80 ◦ C, were also tested on the FLARE® column, but a significant rise in baseline was observed and peak capacity was not modified between 30 and 80 ◦ C on the FLARE® C18 mixed-mode phase. Secondly, the peak capacity reaches its maximum at 15–20 min gradient spans and the efficiency cannot be improved further. It suggests that there is no need for very long gradient separations. By using silica-based materials with the same column dimension and similar analytes, a plateau is observed at around 50–80 min. With the 100 mm × 2.1 mm FLARE® column, peak capacity (P) ∼ 80–120 for cytochrome-C, P between 60 and 90 for transferrin and P between 40 and 50 for BSA were observed. In Fig. 2, the effect of flow rate on peak capacity is illustrated. Fig. 2A shows chromatograms observed at 0.2 and 0.4 mL/min, while Fig. 2B demonstrates the corresponding peak capacity plots for several gradient spans. Practically, the same efficiency was obtained at 0.2 and 0.4 mL/min. It suggests that the resistance to mass transfer does not change a lot in this flow rate range. With the 3.6 m silica-based core–shell type Aeris material, the observed peak capacity was significantly higher at lower flow rate (0.25 mL/min) compared to elevated flow rate (0.4 mL/min) [9]. Usually, with large molecules possessing low diffusivity, the resistance to the external and intra-particle mass transfer worsens drastically with the flow rate yielding band broadening (loss in peak capacity). In the studied flow rate range, this increase in the mass transfer resistance seems to be negligible on the FLARE® material for proteins with molar masses between 12 and 77 kDa. This fact entails the conclusion that high flow rates can be applied on the carbon-based material without significant efficiency loss. This material seems to be well suited for fast protein separations. In Fig. 3, the effect of mobile phase additive TFA concentration is shown. TFA at a concentration of 0.05–0.1% is commonly used for the analysis of peptides and proteins, as it provides excellent ion pairing and solvating characteristics. It reduces residual silanol interactions and therefore inhibits peak broadening and tailing [28]. TFA concentrations up to 0.5% can be useful for solubilizing large or hydrophobic proteins. As shown in Fig. 3, the peak capacity of the FLARE® column was drastically improved by increasing the TFA concentration for all three model proteins. As an example, for cytochrome-c, P = 38, 122, 140 and 167 were observed with 0.1, 0.2, 0.3 and 0.5% TFA when applying 44 min long gradients. Similar tendencies were observed with the other proteins. These results suggest that 0.1% TFA is not sufficient to perform efficient separation on this FLARE® material. At least 0.2% TFA in the mobile phase is required to utilize the possibilities of this carbon-based material for protein separations. With silica-based materials, 0.1% TFA provides sharp peaks, while an additional increase in TFA
133
Fig. 2. Effect of mobile phase flow rate on peak capacity. Black signs represent results observed at 0.4 mL/min, white signs show results at 0.2 mL/min. Triangles: CYT-C, squares: TRFE, rhombuses: BSA. Temperature: 30 ◦ C, TFA concentration: 0.2%. Other conditions as specified in Section 2.3.2.
concentration does not significantly improve efficiency. This is illustrated in Fig. 4, which shows that the peak capacity plots observed with the silica-based Aeris® material were almost similar when adding 0.1 and 0.3% TFA in the mobile phase. In addition, significant increase in retention was observed on the FLARE® column when increasing the TFA concentration. This is most likely due to a decrease in overall polarity (increase in hydrophobicity) of the proteins as the TFA concentration increased and also due to the decrease in polarity of the stationary phase through ion pairing effects of TFA. Since TFA has a strong impact on efficiency, probably hydrogen bonding interactions significantly contribute to the overall retention mechanism. On the other hand, high TFA concentration may decrease the compatibility and sensitivity when applying mass spectrometric detection. 3.2. Real-life example, separation of closely related proteins (interferon ˛-2A variants) Characterization of a bio-pharmaceutical product, performed with appropriate analytical techniques, provides useful information on expression efficiency, product purity and protein stability in its formulation. Reversed-phase liquid chromatography is commonly employed to separate the closely related impurities or degradants of a protein, such as oxidized-, reduced- or deamidated forms. 3.2.1. Improving efficiency The resolution power of this new phase was evaluated for the separation of interferon ␣-2A variants. Interferon ␣-2A has a broad
134
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
Fig. 3. Effect of mobile phase TFA concentration on peak capacity. From white to black signs results are shown at 0.1%, 0.2%, 0.3% and 0.5% TFA concentration, respectively. Triangles: CYT-C, squares: TRFE, rhombuses: BSA. Temperature: 30 ◦ C, flow rate: 0.2 mL/min. Other conditions as specified in Section 2.3.2.
spectrum of antiviral, cytotoxic, and immunomodulating activity and has been successfully used in the treatment of several diseases, such as Hepatitis C. The chromatographic conditions were optimized by a systematic variation of the gradient span and steepness, and of the column temperature and flow rate. It was found that a 7 min long gradient at a flow rate of 0.3 mL/min was appropriate to separate the related oxidized and reduced forms of interferon. During the method optimization, it seemed that TFA concentration significantly impacts the resolution between interferon related peaks. As above discussed, the TFA concentration may have a huge
Fig. 4. Peak capacity plots observed on a 100 mm × 2.1 mm Aeris® C18 column with 0.1 and 0.3% TFA respectively. White and black signs results correspond to 0.1 and 0.3% TFA. Triangles: CYT-C, squares: TRFE, rhombuses: BSA. Temperature: 30 ◦ C, flow rate: 0.2 mL/min. Other conditions as specified in Section 2.3.2.
Fig. 5. Representative chromatograms of Interferon degradants (oxidized and reduced forms) observed on FLARE® C18 mixed mode 100 mm × 2.1 mm column with 0.3% (A) and 0.5% (B) TFA. Other conditions as specified in Section 2.3.2.
impact on peak widths and retention times on the carbon-based FLARE® material. Fig. 5 shows the corresponding chromatograms, obtained with 0.3 and 0.5% TFA in the mobile phase. As illustrated, the selectivity was not drastically modified with TFA concentration but an absolute shift in retention was observed. However, due to the drastic improvement of peak widths with increased TFA concentration, the resolution between the critical peak pair (intact and reduced-1 forms) improved considerably. The quality of the separation in Fig. 5B was similar to the ones achieved using stateof-the-art silica-based materials [7]. To conclude on efficiency and TFA concentration, higher amount of TFA is required (0.2–0.5%) to perform efficient separations than typically used for silica-based stationary phases (0.05–0.1%). 3.2.2. Tuning selectivity Beside efficiency, selectivity can also improve resolution between critical peak pairs. Therefore, the impact of temperature and TFA concentration on selectivity was studied. The effect of temperature was evaluated at three levels (30, 50 and 90 ◦ C), while TFA concentration was tested at two levels (0.3 and 0.5%). Fig. 6A shows the change in selectivity between (1) the oxidized2 and oxidized1 forms, (2) the oxidized1 and intact, and (3) the intact and reduced forms. Fig. 6B and C illustrates selectivity maps between oxidized1 and intact forms and between intact and reduced2 forms. In Fig. 6A, it can be seen that all three selectivities change systematically with the temperature. Selectivity increases with temperature for oxidized2–oxidized1 and oxidized1–intact forms while it decreases for intact–reduced2 forms. For better understanding the changes in selectivity, surface plots were created. Fig. 6B shows the case when selectivity increases significantly with temperature, while Fig. 6C illustrates the opposite behavior. On the other hand, temperature does not affect drastically the peak
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
135
Fig. 6. The impact of mobile phase temperature and TFA concentration on the selectivity between interferon related peaks (A) and predicted surface plots for selectivity between oxidized-1 and intact form (B) and reduced-2 and intact form (C). Column: FLARE® C18 mixed mode 100 mm × 2.1 mm. Conditions as specified in Section 2.3.2.
capacity (efficiency). There are only minor effects in the temperature range of 30–90 ◦ C. To conclude on TFA concentration and temperature, TFA concentration strongly affects the efficiency and slightly impacts selectivity, while temperature mostly influences selectivity.
4. Conclusion The possibilities of a new superficially porous material possessing a carbon core and nanodiamond–polymer shell (FLARE® C18 mixed mode) were evaluated for the separation of large proteins. This stationary phase includes a significant amount of protonated amino groups, due to the presence of poly(allylamine) within the shell structure, the operative retention mechanism is
considered a mix between reversed phase and anion-exchange. However, under the applied conditions (0.1–0.5% TFA in the mobile phase), it seemed that the main retention mechanism for proteins was hydrophobic interactions with the C18 alkylchains on this carbon based material. Based on the results presented in this study, the particle structure (superficially ˚ porous), stationary phase chemistry and pore size (180 A) makes this new material suitable for proteins reversed phase separations. The main advantage of this carbon-based phase compared to silica-based materials is that there is no need for working at high temperature. Peak capacity was not changed significantly between 30 and 80 ◦ C, proving that the ionic interactions are probably negligible on this phase. The lack of silanol groups and related interactions make it possible to perform protein separations at
136
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
ambient temperature with reasonable efficiency. This way, the often mentioned risk of thermal on-column degradation of large proteins is avoided. Another important difference compared to silica-based RPphases is that this carbon-based material requires higher amount of TFA. A proportion of TFA comprised between 0.2 and 0.5% was found to be sufficient to provide narrow peaks. Peak capacity was continuously improved with TFA concentration in the range of 0.1–0.5%. The maximum performance was reached with relatively fast gradients (15–20 min). By using silica-based materials with the same column dimension, normally the peak capacity saturates at around 50–80 min long gradients. This feature makes this carbonbased material useful for fast protein separations. It seemed that selectivity between closely related proteins could be changed mostly through mobile phase temperature. Finally, this superficially porous carbon-based material was successfully applied to real life protein separations. Interferon alfa2A variants (oxidized, native and reduced forms) were separated within 6 min. The quality of interferon separation was comparable to that obtained on current state-of-the-art silica-based phases. Based on our results, the new 3.6 m carbon-based superficially porous RP material seems to be promising for fast protein separations. Additional measurements are planned to be performed on a new prototype deactivated C18 phase. References [1] A. Staub, D. Guillarme, J. Schappler, J.-L. Veuthey, S. Rudaz, Intact protein analysis in the biopharmaceutical field, J. Pharm. Biomed. Anal. 55 (2011) 810–822. [2] S. Fekete, J.-L. Veuthey, D. Guillarme, New trends in reversed-phase liquid chromatographic separations of therapeutic peptides and proteins: theory and applications, J. Pharm. Biomed. Anal. 69 (2012) 9–27. [3] S. Fekete, A.L. Gassner, S. Rudaz, J. Schappler, D. Guillarme, Analytical strategies for the characterization of therapeutic monoclonal antibodies, Trends Anal. Chem. 42 (2013) 74–83. [4] M.E. Swartz, U.P.L.CTM : an introduction and review, J. Liq. Chromatogr. Relat. Technol. 28 (2005) 1253–1263. [5] S. Fekete, I. Kohler, S. Rudaz, D. Guillarme, Importance of instrumentation for fast liquid chromatography in pharmaceutical analysis, J. Pharm. Biomed. Anal. 87 (2014) 105–119. [6] S. Fekete, A. Grand-Guillaume Perrenoud, D. Guillarme, Evolution and Current trends in liquid and supercritical fluid chromatography, Curr. Chromatogr. 1 (2014) 15–40. [7] S. Fekete, R. Berky, J. Fekete, J.-L. Veuthey, D. Guillarme, Evaluation of recent very efficient wide-pore stationary phases for the reversed-phase separation of proteins, J. Chromatogr. A 1252 (2012) 90–103. [8] S. Fekete, E. Oláh, J. Fekete, Fast liquid chromatography: the domination of core–shell and very fine particles, J. Chromatogr. A 1228 (2012) 57–71.
[9] S. Fekete, R. Berky, J. Fekete, J.-L. Veuthey, D. Guillarme, Evaluation of a new wide pore core–shell material (AerisTM WIDEPORE) and comparison with other existing stationary phases for the analysis of intact proteins, J. Chromatogr. A 1236 (2012) 177–188. [10] S. Fekete, J. Fekete, K. Ganzler, Characterization of new types of stationary phases for fast liquid chromatographic applications, J. Pharm. Biomed. Anal. 50 (2009) 703–709. [11] R. Hayes, A. Ahmed, T. Edge, H. Zhang, Core–shell particles: preparation, fundamentals and applications in high performance liquid chromatography, J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.05.010. [12] L.R. Snyder, J.J. Kirkland, J.L. Glajch, Practical HPLC Method Development, 2nd ed., John Wiley, New York, 1997. [13] M.J. Telepchak, The mechanism of reverse phase liquid–solid chromatography, Chromatographia 6 (1973) 234–236. [14] A.A. Peristyy, O.N. Fedyanina, B. Paull, P.N. Nesterenko, Diamond based adsorbents and their application in chromatography, J. Chromatogr. A 1357 (2014) 68–86. [15] L.A. Wiest, D.S. Jensen, C.-H. Hung, R.E. Olsen, R.C. Davis, M.A. Vail, A.E. Dadson, P.N. Nesterenko, M.R. Linford, Pellicular particles with spherical carbon cores and porous nanodiamond/polymer shells for reversed-phase HPLC, Anal. Chem. 83 (2011) 5488–5501. [16] P. Nesterenko, P. Haddad, Diamond-related materials as potential new media in separation science, Anal. Bioanal. Chem. 396 (2010) 205–211. [17] C.-H. Hung, L.A. Wiest, B. Singh, A. Diwan, M.J.C. Valentim, J.M. Christensen, R.C. Davis, A.J. Miles, D.S. Jensen, M.A. Vail, A.E. Dadson, M.R. Linford, Improved efficiency of reversed-phase carbon/nanodiamond/polymer core–shell particles for HPLC using carbonized poly(divinylbenzene) microspheres as the core materials, J. Sep. Sci. 36 (2013) 3821–3829. [18] G. Saini, D.S. Jensen, L.A. Wiest, M.A. Vail, A. Dadson, M.L. Lee, V. Shutthanandan, M.R. Linford, Core–Shell diamond as a support for solid-phase extraction and high-performance liquid chromatography, Anal. Chem. 82 (2010) 4448–4456. [19] J.C. Giddings, Maximum number of components resolvable by gel filtration and other elution chromatographic methods, Anal. Chem. 39 (1967) 1027–1028. [20] Cs. Horvath, S.R. Lipsky, Peak capacity in chromatography, Anal. Chem. 39 (1967) 1893-1893. [21] F. Gritti, G. Guiochon, The mass transfer kinetics in columns packed with HaloES shell particles, J. Chromatogr. A 1218 (2011) 907–921. [22] http://diamond-analytics.com/articles/detecting the active ingredients in cannabis/ [23] H. Chen, C.S. Horvath, High speed high-performance liquid chromatography of peptides and proteins, J. Chromatogr. A 705 (1995) 3–20. [24] S. Heinisch, J.L. Rocca, Sense and nonsense of high temperature liquid chromatography, J. Chromatogr. A 1216 (2009) 642–658. [25] R.A. Everley, T.R. Croley, Ultra performance liquid chromatography/mass spectrometry of intact proteins, J. Chromatogr. A 1192 (2008) 239–247. [26] A. Shibue, C.T. Mant, R.S. Hodges, Effect of anionic ion-pairing reagent concentration (1–60 mM) on reversed-phase liquid chromatography elution behaviour of peptides, J. Chromatogr. A 1080 (2005) 58–67. [27] S. Fekete, S. Rudaz, J.L. Veuthey, D. Guillarme, Impact of mobile phase temperature on recovery and stability of monoclonal antibodies using recent reversed-phase stationary phases, J. Sep. Sci. 35 (2012) 3113–3123. [28] A. Apffel, S. Fisher, G. Goldberg, P.C. Goodley, F.E. Kuhlmann, Enhanced sensitivity for peptide mapping with electrospray liquid chromatography–mass spectrometry in the presence of signal suppression due to trifluoroacetic acidcontaining mobile phases, J. Chromatogr. A 712 (1995) 177–190.
Contents lists available at ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba
Evaluation of new superficially porous particles with carbon core and nanodiamond–polymer shell for proteins characterization Balázs Bobály a , Davy Guillarme b , Szabolcs Fekete b,∗ a b
Budapest University of Technology and Economics, Department of Inorganic and Analytical Chemistry, Szt. Gellért tér 4, 1111 Budapest, Hungary School of Pharmaceutical Sciences, University of Geneva, University of Lausanne, Bd d’Yvoy 20, 1211 Geneva 4, Switzerland
a r t i c l e
i n f o
Article history: Received 16 September 2014 Received in revised form 18 November 2014 Accepted 20 November 2014 Available online 27 November 2014 Keywords: Superficially porous Peak capacity FLARE® Aeris® Protein separation
a b s t r a c t A new superficially porous material possessing a carbon core and nanodiamond–polymer shell and pore size of 180 A˚ was evaluated for the analysis of large proteins. Because the stationary phase on this new support contains a certain amount of protonated amino groups within the shell structure, the resulting retention mechanism is most probably a mix between reversed phase and anion exchange. However, under the applied conditions (0.1–0.5% TFA in the mobile phase), it seemed that the main retention mechanism for proteins was hydrophobic interaction with the C18 alkylchains on this carbon based material. In this study, we demonstrated that there was no need to increase mobile phase temperature, as the peak capacity was not modified considerably between 30 and 80 ◦ C for model proteins. Thus, the risk of thermal on-column degradation or denaturation of large proteins is not relevant. Another important difference compared to silica-based materials is that this carbon-based column requires larger amount of TFA, comprised between 0.2 and 0.5%. Finally, it is important to mention that selectivity between closely related proteins (oxidized, native and reduced forms of Interferon ␣-2A variants) could be changed mostly through mobile phase temperature. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Thanks to its high resolving power, reversed phase liquid chromatography (RPLC) is an important and promising technique in the separation of biological macromolecules [1–3]. Due to the development of instruments [4,5] and column technology [6–10], the past decade served significant improvements in the separation of proteins under reversed phase conditions. Sub-2 m fully porous particles and core–shell materials possessing large pore size in the range of 170–300 A˚ provide outstanding performance when analyzing proteins larger than 10–15 kDa [9]. This is explained by the excellent kinetic performance produced by small fully porous particles and by the so-called core–shell advantages [11] when using superficially porous materials. Except for kinetic performance improvements, recently packed column developments focused on the introduction of various stationary phase chemistries (e.g. C3 C12, pentafluoro-phenyl, phenyl-hexyl, biphenyl, etc.), assumed to serve different selectivity from the conventional C18 phases.
∗ Corresponding author. Tel.: +41 36 30 395 6657; fax: +41 22 379 6808. E-mail address: [email protected] (S. Fekete). http://dx.doi.org/10.1016/j.jpba.2014.11.039 0731-7085/© 2014 Elsevier B.V. All rights reserved.
Surface modifications have been executed on the traditional silica or hybrid-based materials, which are the workhorses of liquid chromatography [12]. Recently, an alternative, carbon–nanodiamond based C18 superficially porous material has been introduced to the market under the trademark FLARE® from Diamond Analytics. This new column technology could be particularly interesting in the separation of large analytes, since it possesses an average pore size of 180 A˚ and offers the benefits of core–shell particle technology. Natural diamond microparticles have been first reported as chromatographic packing in 1973 [13]. However, the low surface area and inhomogeneous adsorption properties of the particles (due to their natural origin) resulted in poor efficiency. Besides efficiency issues, limited availability and high costs prevented their spread in liquid chromatography. Thanks to continuous developments in diamond production technology over the last 40 years, high purity synthetic particles of controlled size and shape became easily accessible at relatively low costs [14]. The physical and chemical properties of these high quality particles renewed interest for chromatographic applications. Diamond particles possess excellent thermal conductivity [14], pH stability [15] and are biocompatible [16], which could make them useful for protein separations. The core of the commercial state-of-the-art FLARE®
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
material is a carbonized poly(divinylbenzene) particle with a diameter of ca. 3.4 m. Poly(allylamine)–nanodiamond hetero-layers are deposited onto the surface of the carbonized core by a modified layer-by-layer method [17]. The resulting core–shell is synthesized to a shell thickness of ca. 0.1 m and a finished particle size of 3.6 m with a surface area around 23 m2 /g. The particles are then functionalized, cross-linked and sieved. The surface area is somewhat lower than on silica-based or graphitic phases but provides sufficient retentivity. Studies in the past few years showed the applicability of carbon/nanodiamond/polymer based shell particles both in normal and reversed phase conditions [18]. Thanks to the improvement in the manufacturing technology of these particles and columns, a continuous increase of kinetic efficiency in the separation of low molecular weight solutes has been reported [14]. Latest results showed promising 100,000–120,000 N/m values for the separation of alkyl benzenes which corresponds to a reduced plate height (h) value of ∼2.3 [17]. These efficiencies are outstanding compared to earlier carbon/nanodiamond/polymer based shell particles, but are still below the efficiencies reported for silica based shell materials. The 3.6 m wide pore shell silica particles showed 150,000–200,000 N/m [9], while sub-3 m shell silica particles dedicated for macromolecular separations provided 200,000–260,000 N/m [8], measured by low molecular weight analytes. This clearly reflects that further improvement in the manufacturing technology of carbon/nanodiamond/polymer material is needed. On the other hand, thanks to its mixed-mode hydrophobicanion exchange surface, it could provide alternative selectivity and retention mechanism, compared to traditional reversed phases. To our knowledge, no data on the applicability of these materials for protein separations have been reported to date. The aim of this study was to evaluate the kinetic efficiency and selectivity of this novel stationary phase for protein characterization. Model proteins and real life samples of native, oxidative stressed and reduced therapeutic proteins were analyzed using the FLARE® column to demonstrate its possibilities and limitations. 2. Experimental 2.1. Instrumentation Chromatographic experiments were performed on a Waters Acquity I-Class UPLC system (Waters, Milford, MA, USA). The instrument was equipped with a binary solvent manager, autosampler, thermostated column compartment, and UV detector. The autosampler was equipped with a flow through needle injection system. The average extra-column peak variance of our system 2 ∼0.5–2.5 L2 . The UV detector operated was found to be around ec with a 500 nL flow cell, set to 280 nm and 40 Hz sampling rate. Data acquisition, data handling and instrument control were performed by Empower Pro 2 (Waters, Milford, MA, USA) software. 2.2. Chemicals and columns Acetonitrile (gradient grade), trifluoroacetic acid (TFA, puriss p.a.), hydrogen peroxide (30 wt.%, ACS reag. grade), dithiothreitol (DTT, ≥99.0%), methionine (≥98.0%) and protein standards such as transferrin (human, MW ∼ 77.0 kDa), cytochrome-c (from horse heart, MW ∼ 12.4 kDa) and albumin (BSA from bovine serum, MW ∼ 69.3 kDa) were purchased from Sigma–Aldrich (Buchs, Switzerland). Water was obtained from a MilliQ Purification System from Millipore (Bedford, MA, USA). Recombinant interferon alfa2A (MW ∼ 19.2 kDa, Roferon) was obtained from Roche Pharma (Switzerland). ˚ colFLARE® C18 mixed-mode (100 mm × 2.1 mm 3.6 m, 180 A) umn was purchased from Diamond Analytics (Orem, UT, USA), and
131
˚ column Aeris Widepore XB-C18 (100 mm × 2.1 mm, 3.6 m, 200 A) was purchased from Phenomenex (Torrance, CA, USA). 2.3. Methodology 2.3.1. Sample preparation For the kinetic evaluation, protein standard mixture of transferrin, BSA and cytochrome-c was prepared containing 0.3 mg/mL of each protein in water. For the selectivity studies Interferon alfa-2A was injected as received. Reduced and oxidized proteins have been prepared from 100 L of the Interferon solutions. A small amount of DTT was added to reduce the protein, then the sample was incubated at 30 ◦ C for 30 min in dark. Oxidation was carried out by adding 1 L of hydrogen-peroxide solution and incubation at 30 ◦ C for 60 min. Oxidation was quenched by adding a small amount of methionine to the sample. 2.3.2. Chromatographic methods For the gradient separation of the proteins, mobile phase “A” was prepared from water and mobile phase “B” was prepared from acetonitrile. Peak capacity was measured applying gradient elution of 20–45%B on the FLARE® column, and 30–55%B on the Aeris WP column (in order to keep the apparent retention factors in the same range). Gradient times were set to 4, 12, 20, 28, 36, and 44 min, followed by 1.5 min equilibration. Flow rate was set to 0.2 mL/min and 0.4 mL/min. Mobile phases contained 0.1, 0.2, 0.3 and 0.5% TFA. Columns were thermostated at 30 ◦ C and 50 ◦ C. 2 L of the samples were injected in full loop injection mode. Selectivity studies on Interferon variants included gradient runs of 25–50%B on the FLARE® column. Gradient times were set to 7 min, flow rate was fixed at 0.3 mL/min in all cases. Mobile phases contained 0.3 and 0.5% TFA, and temperature was set at 30, 50 and 90 ◦ C. Detection has been carried out using fluorescence detection (ex : 280 nm, em : 360 nm, 20 Hz sampling rate). 2 L of the samples were injected in full loop injection mode. 2.3.3. Determining peak capacity Peak capacity is a concept first described by Giddings [19] and soon put to good use by Horvath for gradient chromatography [20]. It is a measure of the separation power that includes the entire chromatographic space together with the variability of the peak width over the chromatogram. In this study, peak capacities were experimentally determined from the gradient time (tg ) and the average measured peak width at 50% height (W50% ). The following equation was used to estimate the peak capacity based on peak width at 4, corresponding to a resolution of Rs = 1 between consecutive peaks: P =1+
tg 1.7 · w50%
(1)
In order to avoid the imprecision associated with the measurement of peak widths at base for proteins often containing closely related variants (i.e. for a heterogeneous sample) the peak width at half height was preferred in this study. This way, the impurities present in the sample, and partially resolved from the main component, did not confuse the measurement. Peak capacity was plotted as a function of gradient time. 3. Results and discussion When high molecular weight solutes have to be separated, the column performance is mostly determined by the mass transfer resistance because of the low diffusivity of proteins. Recent work suggests that the improvement in mass transfer of the new generation of superficially porous particles can mostly be explained by the reduction in external film mass transfer resistance (cf term in
132
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
the Knox equation) and not by the reduction in transparticle mass transfer resistance (cp term in the Knox equation) [21]. Increasing the shell porosity accessible to proteins accelerates the mass transfer through the stationary film surrounding the superficially porous particles. According to a recent study of Gritti and Guiochon, the external film mass transfer resistance controls more than 90% of the solid–liquid mass transfer resistance of proteins [21]. The external film transfer becomes slower with decreasing access to the internal volume. Based on these theoretical considerations, it is clear that efficient protein separation requires fine shell type particles having large porosity, large average pore size and decreased shell thickness. However on particles possessing very narrow shell thickness, the retentivity and loadability are somewhat limited. In addition, compared to silica-based materials, detrimental secondary interactions between residual silanols and proteins – as the ion-exchange mechanism – does not exists on carbon and polymer based stationary phases. The surface of synthetic diamond has no functional groups [16] however it is also possible that some of the nanodiamond surfaces may not be completely coated and any oxygenated moieties on those heterogeneous surfaces could also contribute to tailing of more polar analytes [15]. It was shown that both the retention and peak broadening were strongly influenced by the activity of residual silanols on silica-based materials (even when using TFA in the mobile phase) [7]. On silica-based stationary phases, there is a balance between the hydrophobic and ion exchange mechanisms. As an example, on Aeris WP C4 material, the strongest interactions (ion exchange and hydrogen bonding mechanism) was predominant over the weak hydrophobic interaction, even when working in very acidic conditions (0.1% TFA, pH 1.8) where the number of charged silanols was limited [7]. The stationary phase used in this study (FLARE® ) includes a significant concentration of protonated amino groups, due to the presence of poly(allylamine) within the shell structure, so the resulting retention mechanism must be considered as mixed mode including hydrophobic partitioning and anion-exchange [14–17]. Therefore, the carbon based FLARE® C18 mixed-mode phase operates in a reversed-phase and weak anion exchange mode, depending on the pH of the mobile phase. But the main difference compared to silica-based materials is that the ion-exchange affinity of the weak anion-exchange material (allylamine) is probably negligible at low pH. Moreover the additive used in this study possibly masks the ionic sites of the stationary phase as it is an ion-pairing reagent (TFA) and TFA also forms ion-pairs with the positively charged proteins. These all suggest that at low pH (pH < 2), the main driving force is mostly the hydrophobic interactions with the C18 alkylchains on this carbon-based material. However it also has to be considered that hydrogen-bonding interactions may occur between the ion-exchanger functional groups and proteins. Finally, this 3.6 m FLARE® C18 mixed mode material possesses 180 A˚ average pore size. This value is close to the pore size of the silica based Aeris 3.6 m material that is now a goldstandard in proteins’ RP separations. This pore size combined with the core–shell structure should suit for the separation of large molecules. All these features suggest that this new carbon-based material might be applicable for protein separations, even if the initial purpose of the provider was to introduce some unique selectivity for the separation of small molecules [22] and to further improve pH and temperature stability for RP-LC separations (the FLARE® material is stable up to 100 ◦ C and from pH 1 to 13). 3.1. Peak capacity: the effect of temperature, flow rate and additive concentration In the following section, a systematic study shows the achievable peak capacity of the new 3.6 m wide-pore carbon-based
Fig. 1. Effect of mobile phase temperature on peak capacity. Black signs represent results observed at 30 ◦ C, white signs show results at 50 ◦ C. Triangles: CYT-C, squares: TRFE, rhombuses: BSA. Flow rate: 0.2 mL/min, TFA concentration: 0.2%. Other conditions as specified in Section 2.3.2.
superficially porous material in representative gradient elution conditions. Three model proteins including transferrin (TRFE), cytochrome-c (CYT-C) and bovine serum albumin (BSA) were studied under generic gradient conditions (see Section 2.3.2). In RP-LC practice, proteins are separated in gradient elution mode at elevated temperature. Elevated temperature on silica-based phases is beneficial since it decreases the secondary interactions between residual silanols and positively charged biomolecules. Moreover, the use of high temperatures strongly enhances analyte diffusion [23,24]. Except elevated temperature, it is also important to add ionpairing reagents, such as TFA, to the mobile phase in order to increase the efficiency of protein separations [25,26]. To emulate real-life separations, we applied 30 and 50 ◦ C as column temperature, added 0.1–0.5% TFA into the mobile phase and used gradient spans between 4 and 44 min which are fairly common in the current practice. The influence of the gradient steepness on peak capacity was also investigated under several conditions. These experimental variables (temperature, TFA concentration and gradient steepness) are often used to adjust a proper separation in practical every day work. The gradient steepness strongly impacts resolution under gradient conditions as it affects the retention factor of the solute in the mobile phase composition upon elution. The peak capacity of 100 mm long narrow bore FLARE® C18 mixed-mode column was calculated and plotted against the gradient time when changing the other parameters. First, the impact of temperature was investigated. Several gradient spans were set and peak capacity was determined at 30 and 50 ◦ C when applying 0.2 mL/min flow rate and 0.2% TFA. Fig. 1A shows representative chromatograms of the test proteins obtained
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
with the 4 min long gradient at two temperatures. As shown, this fast separation is quite impressive (an average peak capacity of around P ∼ 55 in 5 min). All the three proteins eluted as sharp peaks. Similar peak widths can be observed with wide-pore fully porous silica materials of conventional size (3–5 m) [9]. This means that this carbon based material is indeed suitable for the separation of proteins. Fig. 1B shows the corresponding peak capacity plots and two important conclusions can be drawn. Firstly, the peak capacity does not change with temperature. That is quite uncommon since in most cases, peak capacity increases significantly at elevated temperature, through the reduction of secondary ion-exchange interaction strength. It proves that the ionic interactions are probably negligible on this special phase under the applied conditions. Moreover, it provides an important advantage compared to silica based phases, as there is no need to work at high temperature with this carbon based material. Proteins RPLC separations are generally performed at 50–80 ◦ C to obtain acceptable peak shape and peak width. However, the risk of thermal on-column degradation of large proteins may increase with temperature [27]. Higher temperature in the range of 70–80 ◦ C, were also tested on the FLARE® column, but a significant rise in baseline was observed and peak capacity was not modified between 30 and 80 ◦ C on the FLARE® C18 mixed-mode phase. Secondly, the peak capacity reaches its maximum at 15–20 min gradient spans and the efficiency cannot be improved further. It suggests that there is no need for very long gradient separations. By using silica-based materials with the same column dimension and similar analytes, a plateau is observed at around 50–80 min. With the 100 mm × 2.1 mm FLARE® column, peak capacity (P) ∼ 80–120 for cytochrome-C, P between 60 and 90 for transferrin and P between 40 and 50 for BSA were observed. In Fig. 2, the effect of flow rate on peak capacity is illustrated. Fig. 2A shows chromatograms observed at 0.2 and 0.4 mL/min, while Fig. 2B demonstrates the corresponding peak capacity plots for several gradient spans. Practically, the same efficiency was obtained at 0.2 and 0.4 mL/min. It suggests that the resistance to mass transfer does not change a lot in this flow rate range. With the 3.6 m silica-based core–shell type Aeris material, the observed peak capacity was significantly higher at lower flow rate (0.25 mL/min) compared to elevated flow rate (0.4 mL/min) [9]. Usually, with large molecules possessing low diffusivity, the resistance to the external and intra-particle mass transfer worsens drastically with the flow rate yielding band broadening (loss in peak capacity). In the studied flow rate range, this increase in the mass transfer resistance seems to be negligible on the FLARE® material for proteins with molar masses between 12 and 77 kDa. This fact entails the conclusion that high flow rates can be applied on the carbon-based material without significant efficiency loss. This material seems to be well suited for fast protein separations. In Fig. 3, the effect of mobile phase additive TFA concentration is shown. TFA at a concentration of 0.05–0.1% is commonly used for the analysis of peptides and proteins, as it provides excellent ion pairing and solvating characteristics. It reduces residual silanol interactions and therefore inhibits peak broadening and tailing [28]. TFA concentrations up to 0.5% can be useful for solubilizing large or hydrophobic proteins. As shown in Fig. 3, the peak capacity of the FLARE® column was drastically improved by increasing the TFA concentration for all three model proteins. As an example, for cytochrome-c, P = 38, 122, 140 and 167 were observed with 0.1, 0.2, 0.3 and 0.5% TFA when applying 44 min long gradients. Similar tendencies were observed with the other proteins. These results suggest that 0.1% TFA is not sufficient to perform efficient separation on this FLARE® material. At least 0.2% TFA in the mobile phase is required to utilize the possibilities of this carbon-based material for protein separations. With silica-based materials, 0.1% TFA provides sharp peaks, while an additional increase in TFA
133
Fig. 2. Effect of mobile phase flow rate on peak capacity. Black signs represent results observed at 0.4 mL/min, white signs show results at 0.2 mL/min. Triangles: CYT-C, squares: TRFE, rhombuses: BSA. Temperature: 30 ◦ C, TFA concentration: 0.2%. Other conditions as specified in Section 2.3.2.
concentration does not significantly improve efficiency. This is illustrated in Fig. 4, which shows that the peak capacity plots observed with the silica-based Aeris® material were almost similar when adding 0.1 and 0.3% TFA in the mobile phase. In addition, significant increase in retention was observed on the FLARE® column when increasing the TFA concentration. This is most likely due to a decrease in overall polarity (increase in hydrophobicity) of the proteins as the TFA concentration increased and also due to the decrease in polarity of the stationary phase through ion pairing effects of TFA. Since TFA has a strong impact on efficiency, probably hydrogen bonding interactions significantly contribute to the overall retention mechanism. On the other hand, high TFA concentration may decrease the compatibility and sensitivity when applying mass spectrometric detection. 3.2. Real-life example, separation of closely related proteins (interferon ˛-2A variants) Characterization of a bio-pharmaceutical product, performed with appropriate analytical techniques, provides useful information on expression efficiency, product purity and protein stability in its formulation. Reversed-phase liquid chromatography is commonly employed to separate the closely related impurities or degradants of a protein, such as oxidized-, reduced- or deamidated forms. 3.2.1. Improving efficiency The resolution power of this new phase was evaluated for the separation of interferon ␣-2A variants. Interferon ␣-2A has a broad
134
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
Fig. 3. Effect of mobile phase TFA concentration on peak capacity. From white to black signs results are shown at 0.1%, 0.2%, 0.3% and 0.5% TFA concentration, respectively. Triangles: CYT-C, squares: TRFE, rhombuses: BSA. Temperature: 30 ◦ C, flow rate: 0.2 mL/min. Other conditions as specified in Section 2.3.2.
spectrum of antiviral, cytotoxic, and immunomodulating activity and has been successfully used in the treatment of several diseases, such as Hepatitis C. The chromatographic conditions were optimized by a systematic variation of the gradient span and steepness, and of the column temperature and flow rate. It was found that a 7 min long gradient at a flow rate of 0.3 mL/min was appropriate to separate the related oxidized and reduced forms of interferon. During the method optimization, it seemed that TFA concentration significantly impacts the resolution between interferon related peaks. As above discussed, the TFA concentration may have a huge
Fig. 4. Peak capacity plots observed on a 100 mm × 2.1 mm Aeris® C18 column with 0.1 and 0.3% TFA respectively. White and black signs results correspond to 0.1 and 0.3% TFA. Triangles: CYT-C, squares: TRFE, rhombuses: BSA. Temperature: 30 ◦ C, flow rate: 0.2 mL/min. Other conditions as specified in Section 2.3.2.
Fig. 5. Representative chromatograms of Interferon degradants (oxidized and reduced forms) observed on FLARE® C18 mixed mode 100 mm × 2.1 mm column with 0.3% (A) and 0.5% (B) TFA. Other conditions as specified in Section 2.3.2.
impact on peak widths and retention times on the carbon-based FLARE® material. Fig. 5 shows the corresponding chromatograms, obtained with 0.3 and 0.5% TFA in the mobile phase. As illustrated, the selectivity was not drastically modified with TFA concentration but an absolute shift in retention was observed. However, due to the drastic improvement of peak widths with increased TFA concentration, the resolution between the critical peak pair (intact and reduced-1 forms) improved considerably. The quality of the separation in Fig. 5B was similar to the ones achieved using stateof-the-art silica-based materials [7]. To conclude on efficiency and TFA concentration, higher amount of TFA is required (0.2–0.5%) to perform efficient separations than typically used for silica-based stationary phases (0.05–0.1%). 3.2.2. Tuning selectivity Beside efficiency, selectivity can also improve resolution between critical peak pairs. Therefore, the impact of temperature and TFA concentration on selectivity was studied. The effect of temperature was evaluated at three levels (30, 50 and 90 ◦ C), while TFA concentration was tested at two levels (0.3 and 0.5%). Fig. 6A shows the change in selectivity between (1) the oxidized2 and oxidized1 forms, (2) the oxidized1 and intact, and (3) the intact and reduced forms. Fig. 6B and C illustrates selectivity maps between oxidized1 and intact forms and between intact and reduced2 forms. In Fig. 6A, it can be seen that all three selectivities change systematically with the temperature. Selectivity increases with temperature for oxidized2–oxidized1 and oxidized1–intact forms while it decreases for intact–reduced2 forms. For better understanding the changes in selectivity, surface plots were created. Fig. 6B shows the case when selectivity increases significantly with temperature, while Fig. 6C illustrates the opposite behavior. On the other hand, temperature does not affect drastically the peak
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
135
Fig. 6. The impact of mobile phase temperature and TFA concentration on the selectivity between interferon related peaks (A) and predicted surface plots for selectivity between oxidized-1 and intact form (B) and reduced-2 and intact form (C). Column: FLARE® C18 mixed mode 100 mm × 2.1 mm. Conditions as specified in Section 2.3.2.
capacity (efficiency). There are only minor effects in the temperature range of 30–90 ◦ C. To conclude on TFA concentration and temperature, TFA concentration strongly affects the efficiency and slightly impacts selectivity, while temperature mostly influences selectivity.
4. Conclusion The possibilities of a new superficially porous material possessing a carbon core and nanodiamond–polymer shell (FLARE® C18 mixed mode) were evaluated for the separation of large proteins. This stationary phase includes a significant amount of protonated amino groups, due to the presence of poly(allylamine) within the shell structure, the operative retention mechanism is
considered a mix between reversed phase and anion-exchange. However, under the applied conditions (0.1–0.5% TFA in the mobile phase), it seemed that the main retention mechanism for proteins was hydrophobic interactions with the C18 alkylchains on this carbon based material. Based on the results presented in this study, the particle structure (superficially ˚ porous), stationary phase chemistry and pore size (180 A) makes this new material suitable for proteins reversed phase separations. The main advantage of this carbon-based phase compared to silica-based materials is that there is no need for working at high temperature. Peak capacity was not changed significantly between 30 and 80 ◦ C, proving that the ionic interactions are probably negligible on this phase. The lack of silanol groups and related interactions make it possible to perform protein separations at
136
B. Bobály et al. / Journal of Pharmaceutical and Biomedical Analysis 104 (2015) 130–136
ambient temperature with reasonable efficiency. This way, the often mentioned risk of thermal on-column degradation of large proteins is avoided. Another important difference compared to silica-based RPphases is that this carbon-based material requires higher amount of TFA. A proportion of TFA comprised between 0.2 and 0.5% was found to be sufficient to provide narrow peaks. Peak capacity was continuously improved with TFA concentration in the range of 0.1–0.5%. The maximum performance was reached with relatively fast gradients (15–20 min). By using silica-based materials with the same column dimension, normally the peak capacity saturates at around 50–80 min long gradients. This feature makes this carbonbased material useful for fast protein separations. It seemed that selectivity between closely related proteins could be changed mostly through mobile phase temperature. Finally, this superficially porous carbon-based material was successfully applied to real life protein separations. Interferon alfa2A variants (oxidized, native and reduced forms) were separated within 6 min. The quality of interferon separation was comparable to that obtained on current state-of-the-art silica-based phases. Based on our results, the new 3.6 m carbon-based superficially porous RP material seems to be promising for fast protein separations. Additional measurements are planned to be performed on a new prototype deactivated C18 phase. References [1] A. Staub, D. Guillarme, J. Schappler, J.-L. Veuthey, S. Rudaz, Intact protein analysis in the biopharmaceutical field, J. Pharm. Biomed. Anal. 55 (2011) 810–822. [2] S. Fekete, J.-L. Veuthey, D. Guillarme, New trends in reversed-phase liquid chromatographic separations of therapeutic peptides and proteins: theory and applications, J. Pharm. Biomed. Anal. 69 (2012) 9–27. [3] S. Fekete, A.L. Gassner, S. Rudaz, J. Schappler, D. Guillarme, Analytical strategies for the characterization of therapeutic monoclonal antibodies, Trends Anal. Chem. 42 (2013) 74–83. [4] M.E. Swartz, U.P.L.CTM : an introduction and review, J. Liq. Chromatogr. Relat. Technol. 28 (2005) 1253–1263. [5] S. Fekete, I. Kohler, S. Rudaz, D. Guillarme, Importance of instrumentation for fast liquid chromatography in pharmaceutical analysis, J. Pharm. Biomed. Anal. 87 (2014) 105–119. [6] S. Fekete, A. Grand-Guillaume Perrenoud, D. Guillarme, Evolution and Current trends in liquid and supercritical fluid chromatography, Curr. Chromatogr. 1 (2014) 15–40. [7] S. Fekete, R. Berky, J. Fekete, J.-L. Veuthey, D. Guillarme, Evaluation of recent very efficient wide-pore stationary phases for the reversed-phase separation of proteins, J. Chromatogr. A 1252 (2012) 90–103. [8] S. Fekete, E. Oláh, J. Fekete, Fast liquid chromatography: the domination of core–shell and very fine particles, J. Chromatogr. A 1228 (2012) 57–71.
[9] S. Fekete, R. Berky, J. Fekete, J.-L. Veuthey, D. Guillarme, Evaluation of a new wide pore core–shell material (AerisTM WIDEPORE) and comparison with other existing stationary phases for the analysis of intact proteins, J. Chromatogr. A 1236 (2012) 177–188. [10] S. Fekete, J. Fekete, K. Ganzler, Characterization of new types of stationary phases for fast liquid chromatographic applications, J. Pharm. Biomed. Anal. 50 (2009) 703–709. [11] R. Hayes, A. Ahmed, T. Edge, H. Zhang, Core–shell particles: preparation, fundamentals and applications in high performance liquid chromatography, J. Chromatogr. A (2014), http://dx.doi.org/10.1016/j.chroma.2014.05.010. [12] L.R. Snyder, J.J. Kirkland, J.L. Glajch, Practical HPLC Method Development, 2nd ed., John Wiley, New York, 1997. [13] M.J. Telepchak, The mechanism of reverse phase liquid–solid chromatography, Chromatographia 6 (1973) 234–236. [14] A.A. Peristyy, O.N. Fedyanina, B. Paull, P.N. Nesterenko, Diamond based adsorbents and their application in chromatography, J. Chromatogr. A 1357 (2014) 68–86. [15] L.A. Wiest, D.S. Jensen, C.-H. Hung, R.E. Olsen, R.C. Davis, M.A. Vail, A.E. Dadson, P.N. Nesterenko, M.R. Linford, Pellicular particles with spherical carbon cores and porous nanodiamond/polymer shells for reversed-phase HPLC, Anal. Chem. 83 (2011) 5488–5501. [16] P. Nesterenko, P. Haddad, Diamond-related materials as potential new media in separation science, Anal. Bioanal. Chem. 396 (2010) 205–211. [17] C.-H. Hung, L.A. Wiest, B. Singh, A. Diwan, M.J.C. Valentim, J.M. Christensen, R.C. Davis, A.J. Miles, D.S. Jensen, M.A. Vail, A.E. Dadson, M.R. Linford, Improved efficiency of reversed-phase carbon/nanodiamond/polymer core–shell particles for HPLC using carbonized poly(divinylbenzene) microspheres as the core materials, J. Sep. Sci. 36 (2013) 3821–3829. [18] G. Saini, D.S. Jensen, L.A. Wiest, M.A. Vail, A. Dadson, M.L. Lee, V. Shutthanandan, M.R. Linford, Core–Shell diamond as a support for solid-phase extraction and high-performance liquid chromatography, Anal. Chem. 82 (2010) 4448–4456. [19] J.C. Giddings, Maximum number of components resolvable by gel filtration and other elution chromatographic methods, Anal. Chem. 39 (1967) 1027–1028. [20] Cs. Horvath, S.R. Lipsky, Peak capacity in chromatography, Anal. Chem. 39 (1967) 1893-1893. [21] F. Gritti, G. Guiochon, The mass transfer kinetics in columns packed with HaloES shell particles, J. Chromatogr. A 1218 (2011) 907–921. [22] http://diamond-analytics.com/articles/detecting the active ingredients in cannabis/ [23] H. Chen, C.S. Horvath, High speed high-performance liquid chromatography of peptides and proteins, J. Chromatogr. A 705 (1995) 3–20. [24] S. Heinisch, J.L. Rocca, Sense and nonsense of high temperature liquid chromatography, J. Chromatogr. A 1216 (2009) 642–658. [25] R.A. Everley, T.R. Croley, Ultra performance liquid chromatography/mass spectrometry of intact proteins, J. Chromatogr. A 1192 (2008) 239–247. [26] A. Shibue, C.T. Mant, R.S. Hodges, Effect of anionic ion-pairing reagent concentration (1–60 mM) on reversed-phase liquid chromatography elution behaviour of peptides, J. Chromatogr. A 1080 (2005) 58–67. [27] S. Fekete, S. Rudaz, J.L. Veuthey, D. Guillarme, Impact of mobile phase temperature on recovery and stability of monoclonal antibodies using recent reversed-phase stationary phases, J. Sep. Sci. 35 (2012) 3113–3123. [28] A. Apffel, S. Fisher, G. Goldberg, P.C. Goodley, F.E. Kuhlmann, Enhanced sensitivity for peptide mapping with electrospray liquid chromatography–mass spectrometry in the presence of signal suppression due to trifluoroacetic acidcontaining mobile phases, J. Chromatogr. A 712 (1995) 177–190.
