Articles Anal. Chem. 1995,67,2517-2523
Mixed-Mode Retention of Peptides Phosphate-Modified Polybutadiene-Coated Zirconia Liiang Sunt and Peter W. Cam*,*
Department of Chemistly and Institute for Advanced Studies in Bioprocess Technology, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455
Zirconia HPLC packing materials were found to be potentially advantageous for large-scale protein separations due to their excellent pH stability and mechanical stability. However, Lewis acid sites on zirconia’s surface cause irreversible adsorption of proteins due to their interactions with hard Lewis bases such as the carboxyl groups in proteins. Although the Lewis acid sites can be effectively blocked by adsorbing phosphate ions onto zirconia’s surface, proteins and peptides m o t be eluted using a typical reversed-phase mobile phase. In this work, we found that the separation of peptides on a phosphate-modifiedpolybutadiene-coatedzirconia (PBDZrOz) can be brought about by using a mobile phase containing both an organic modifier and a high concentration of sodium perchlorate. The salt is needed to cancel the Coulombic interactionsbetween the negatively charged stationary phase and the positively charged proteins. To understand the retention mechanism of proteins and peptides on phosphate-mded PBD-Zr02, this work was aimed at the study of the surface characteristics of the phosphate-modifiedPBD-Zr02. We found that the phosphate-modified PBD-Zr02 phase has both reversed-phase and cation-exchange characteristics under the acidic mobile-phase conditions used for protein and peptide separations. The PBD coating provides hydrophobic moieties, and the phosphate ions adsorbed on zirconia’s surface provide cation-exchange sites. Reversed-phase separation of a peptide standard mixture and cationexchange separation of a cationic peptide standard mixture on the same phosphate-modifiedPBD-ZrO2 column shows excellent column resolution in both modes. Although mixed-mode stationary phases provide unique selectivity, the secondary equilibrium on phosphatemodified PBD-Zr02 can cause peak broadening. Applications of the phosphate-modified PBD-ZrOz to peptide separations are demonstrated here. Monoclinic zirconia is stable from pH 1 to pH 14.1-3 In addition, zirconia has excellent solvent and mechanical stability. +
Department of Chemistry.
* Department of Chemistry and Institute for Advanced Studies in Bioprocess Technology. (1) Nawrocki, J.: Carr, P. W.; Rigney, M. P.; McCormick, A V. /. Chromafogr. 1993,657,229-282.
0003-2700/95/0367-2517$9.00/0 0 1995 American Chemical Society
For largescale bioseparations, pH-stable packing materials have the advantage that they can be cleaned and sterilized using hot acids and bases to remove irreversibly adsorbed proteins and perhaps pyrogen~.4*~ The limited pH stability of silica-based packing materials and the limited solvent stability (shrink and swell) of polymeric packing has motivated studies of zirconia-based HPLC packings for protein separations. A number of zirconia-based stationary phases have been developed. These include the following: siloxane bonded-phase z i r ~ o n i a ,polyb ~,~ utadiene-coated zirconia (PBD-Zr02),8Jo and carbonclad zirconia,11J2which all behave as reversed-phase packing materials; phosphatemodified z i r ~ o n i a , a~ cationexchange ~J~ packing mate rial; fluoridemodified z i r ~ o n i a , ~ an ~ - lanion-exchange ~ packing material; and protein-modified zirconia, used as affinity chromatographic packing material4 Among these packing materials, phosphatemodified and fluoridemodifiedzirconiaI6have been successfully used for protein separations. Although PBDZrO2 has excellent pH stability and efficiency for the separation of small molecules, the separation of proteins on PBD-Zr02 has not been successful due to irreversible adsorption of proteins.6It was also found that proteins cannot be eluted from a phosphate modified PBD-Zr02 using typical reversed-phase mobile phases. The separation of peptides on PBD-Zr02 has not been previously studied. (2) Nawrocki, J.; Dunlap, C. J.; Carr, P. W.; Blackwell, J. A Biofechnol. Prog. 1994,10,561-573. (3) Rigney, M. P.; Funkenbush, E. F.; Carr, P. W. 1.Chromatogr. 1990,499, 291-304. (4) Glavanovich, M. H. Ph.D. Thesis, University of Minnesota, Minneapolis, MN, 1994. (5) Weary, M.; Pearson, F. Biochem. Phamacol. 1988,1 22-29. (6)Glajch, J. L;Kirkland J. J.; Kohler, J. J. Chromafogr. 1987,384,81-90. (7) Tanaka N.; Araki, M. Adu. Chromafogr. 1989,30,81-123. (8) Rigney, M. P. Ph.D. Thesis, University of Minnesota, Minneapolis, MN, 1988. (9) Yu J.; Rassi, Z. E. J. Chromatogr. 1993,631,91-106. (10) Rgney, M. P.; Weber T. P.; Carr, P. W. J. Chromatogr. 1989,484,273291. (11) Weber, T. P.; Carr, P. W. Anal. Chem. 1990,62,2620-2625. (12) Weber, T. P.; Carr, P. W.J Chromatogr. 1990,319,31-52. (13) Schafer, W. A; Carr, P. W. J. Chromatog?. 1991, 587,137-147. (14) Schafer, W. A; Carr, P. W.J Chromafogr, 1991,587,149-160. (15) Blackwell, J. A; Carr, P. W. J. Chromafogr. 1991,549,43-57. (16) Blackwell, J. A; Carr, P. W. J Chromafogr. 1991,549,59-75. (17) Blackwell, J. A; Carr, P. W. Anal. Chem. 1992,64,863-873. (18) Blackwell, J. A; Carr, P. W. J. Liq. Chromafogr. 1991,14,2875-2889. (19) Blackwell, J. A; Carr, P. W. Anal. Chem. 1992,64,853-862.
Analytical Chemistry, Vol. 67,No. 15,August 1, 1995 2517
Our recent studies20.21 show that the irreversible adsorption of proteins on PBD-Zr02 is due to strong, hard Lewis acid-base interactions between the carboxyl groups on proteins and the Lewis acid sites on zirconia. To reduce such surface interactions, the Lewis acid sites on zirconia were blocked by adsorbing a hard Lewis base such as phosphate.13J4 This work showed that the adsorption of negatively charged phosphate on PBD-Zr02 results in mixed-mode interactions between proteins and the stationary phase. We found that a number of proteins could be eluted from phosphate-modified PBD-Zr02 using a reversed-phase mobile phase containing both an organic modifier and a high concentration of sodium perchlorate.20 Mixed-mode stationary phase^^^'^^ can provide better selectivity for protein and peptide separations. However, the use of mixedmode stationary phases frequently involves the disadvantage of peak broadening and low recovery due to secondary interactions between the solutes and the stationary phase.22 One well-known example of this problem is the effect of silanol groups on the retention and peak shape of amines when silica-based bonded phases are ~ s e d . The 2 ~ ~secondary ~~ interactions of amines with the surface silanol groups often result in broad and asymmetric peaks. On some mixed-mode stationary phases, irreversible adsorption of proteins has been observed.26 The goal of this work is to characterize the surface properties of the phosphate-modified PBD-2-02 as modified by dynamic adsorption under acidic mobile-phase conditions. This will lead to a better understanding of the retention mechanism of proteins and peptides and therefore enable us to manipulate the surface propertiesz0in order to optimize the separation conditions. Small organic molecules and peptides were used to probe the surface properties of the phosphate-modified PBD-Zr02. The application of this phase to peptide separations was explored. Examples of separations of reversed-phase and cation-exchange peptide standard mixtures on phosphate-modified PBD-Zr02 show excellent resolution in both the reversed-phase mode and the cationexchange mode. An application of this stationary phase to tryptic mapping is also shown in this work. EXPERIMENTAL SECTION
Chemicals. All solvents used in this work were HPLC grade and were obtained from Fisher Scientific (Fairlawn, NJ). HPLC water was prepared by passing house deionized water through a Barnstead Nanopure deionizing system with an organic-free cartridge and a 0.2 pm filter. Trifluoroacetic acid, phosphoric acid, sodium perchlorate monohydrate, and sodium hydroxide were from Aldrich (Milwaukee, WI). Aniline, benzylamine, m-xylenediamine, 1-naphthylamine, dibenzylamine, and N,N-dimethyl-1naphthylamine were obtained from Aldrich. Solutions (0.1%w/v) of pure amines were typically prepared in 5050 acetonitrile-water. Carbobenzoxyarginine was obtained from Peptide International (Louisville, KY). The peptides listed in Table 1 were obtained ~~
_____~
~
~~
~~
(20) Sun, L.; Carr, P. W. Submitted. (21) Sun, L.; McCormick. A. V.; Carr, P. W. /. Chromatogr. 1994,658, 465473. (22) Nau, D. R. In HPLC of Proteins Peptides and Polynucleotides;M. T. W., H e m . Ed.; VCH Publishers, Inc.: New York, 1991; pp 331-395. (23) Zhu, B. Y.; Mant, C. T.; Hodges, R S. J. Chromatogr. 1992,594, 75-86. (24) Snyder, L. R In High Pelfomance Liquid ChromatographyAdvances and Perspectives; Horvath, Cs., Ed.; Academic Press: New York, 1983; Vol. 3, p 1.
(25) Sadek, P. C.; Carr, P. W. J. Chromatogr. Sci. 1983,21, 314. (26) Berkowitz, S. A.; Baker, J. T. Adu. Chromatogr. 1989,29, 175-219.
2518 Analytical Chemistty, Vol. 67, No. 15,August 1, 1995
from Bachem California (Torrance, CA). Peptide samples were prepared as 0.1%(w/v) aqueous solutions. The synthetic peptide standards listed in Table 2 were obtained from Synthetic Peptides Inc. (Edmonton, Alberta, Canada). These peptides were dissolved in water according to the instructions in the catalog. TPCK-treated trypsin, bovine cytochrome c, horse cytochrome c, and dog cytochrome c were obtained from Sigma (St. Louis, MO) . Tryptic Digestion. Bovine cytochrome c, horse cytochrome c, and dog cytochrome c were dissolved in a pH 8.0, 0.1 M ammonium bicarbonate buffer at a concentration of 2 mg/mL. Trypsin was dissolved in the same buffer at a concentration of 0.2 mg/mL. Trypsin solution was added to each of the cytochrome c solutions at a 1:l volume ratio. The mixtures are then incubated at 37 "C for 24 h. The digestion was terminated by deactivating trypsin by acidifying the digests with 1%TFA, and the digests were kept frozen until used. Chromatography Columns. Zirconia particles (batch no. Coac-14) used in this study were prepared by the "polymerinduced colloid aggregation" (PICA) m e t h ~ d . The ~ ~ , main ~~ advantage of this method is that the resulting particles have a narrow size di~tribution.2~ The average diameter of these particles is 6 pm. The surface area of this packing is 22 m2/g, and the average pore diameter (4 times the ratio of pore volume to surface area) is 170 A by nitrogen sorptometry. To ensure reproducible surface properties after sintering, zirconia particles were washed with 0.5 M hydrochloric acid and then 0.5 M sodium hydroxide at room temperature. The zirconia particles described above were used for the preparation of PBD-Zr02. PBD-Zr02 was prepared according to a modified procedure based on the work of Rigney et d.1° Zirconia particles were dried in a vacuum oven at 110 "C for 8 h and cooled over phosphorus pentoxide immediately before the coating preparation. The particles were suspended in a 0.1%(w/v) PBD in hexane solution, and the slurry was equilibrated for 8 h. No precautions were taken to dry the hexane. A PBD loading of 0.6 mg of PBD/m2 of total surface area was used. Next, 2.5% (w/w of dicumyl peroxide-PBD) dicumyl peroxide was added to the slurry and then the hexane was evaporated at 35 "C under vacuum in a rotary evaporator. After cross-linking PBD at 160 "C under vacuum, the coated material was extensively cleaned in an extracting device,29 refluxing with 110 "C toluene for 4 h. Compared to the sequential solvent washing procedure,1° the hot extraction procedure more efficiently removes excess PBD. Finally, the particles were rinsed with hexane to remove residual toluene and were then air-dried. The carbon loading of the PBDcoated ZrOz was 0.45 mg/m2. PBD-coated ZrO? was typically packed into 5 x 0.46 cm and 15 x 0.40 columns. Approximately 2 g of zirconia-based packing was used to pack a 5 x 0.46 cm column. Parker-Hannifin 316 stainless steel column end fittings were used with 2 pm stainless steel screens (Alltech, Deerfield, IL). Columns were packed by a magnetically stirred upward slurry technique in which the particles are suspended in a solvent and forced into a column at 350 bar with pure isopropyl alcohol. The slurry solvent used for PBD-Zr02 was 9O:lO isopropyl alcohol-hexane. (27) Sun, L.; Annen, M. J.; Lorenzano-Porras, F.; Carr, P. W.; McCormick, A. V. J. Colloid Intelface Sci. 1994,163, 464-473. (28) h n e n , M. J.; Lorenzano-Porras, F.; Carr, P. W.; McCormick, A. V.; Flickinger, M. C. J. Colloid Intelface Sci. 1995,170, 299-307. (29) Aue, W. A; Daniewski, M. M.; Muller, J.; Laba. J. P. Anal. Chem. 1977, 49, 146551466,
I
10.0
6-
54-
k ’ 321 -
0 1
I
I
I
I
0.1 0.062
0.125
0.250
0.500
1.000
[NaCIO,] M
Figure I. Effect of ionic strength on the retention of amines on phosphate-modified PED-Zr02. Both axes are plotted on a logarithm scale: e, N,N-dimethyl-1-naphthylamine; 0 , dibenzylamine; v, 1-naphthylamine; A, m-xylenediamine; ,. benzylamine; 0 , aniline. Mobile phase is an aqueous solution of sodium perchlorate and 0.1 M phosphoric acid at pH 1.6; 1 mUmin, 254 nm detection, 30 “C. Table 1. Peptides for Studying the Effect of Secondary Adsorption
Chromatographic Experiments. Chromatographic experiments were performed on a Perkin-Elmer 3B Series chromatograph at room temperature and on a HP 1090 chromatograph at 30 “C. UV detection at 254 nm was typically used. Typical injection volumes were 10 pL. The mobile phase was a water and acetonitrile system with 0.1%TFA, 0.08 M phosphoric acid, and 1M sodium perchlorate. Details are described in the captions of the figures. RESULTS AND DISCUSSION
Cation-&change Properties of Phosphate-ModifiedPBD2102. To characterize the negative surface charge of the phosphatemodifled PBD-Zr02,small amines were used as probe solutes to study the dependence of their retention on ionic strength and pH of the mobile phase. AU of the amines studied have pKas much higher than the pHs used in this study. Therefore, these amines are protonated and bear one or more positive charges that can interact with the surface negative charges. Figure 1 shows that, at pH 1.6, the retention of small amines decreases upon increasing the concentration of sodium perchlorate in the mobile phase. The retention of the only diamine used, m-xylenediamine, decreased faster upon increase in ionic strength than the singlecharged amines. These results are typical for cation exchangers and are consistent with theoretical models for ionexchange chr~matography.~~ We also note that there is considerable evidence for a hydrophobic contribution to retention for these species. The retention order is in good agreement with their relative size. Furthermore, even at a rather high ionic strength the k’ values do not become zero. An increase in pH from pH 1 to pH 3 leads to an increase in the surface negative charge on phosphatemodifled PBD-Zr02 (30) Poole, C . F.; Poole, S. K In Chromatografihy Today; Elsevier Science Publishers B. V.: Amsterdam, 1991; pp 422-438. (31) Stumm, W.; Kummert, R; Sigg, L. Croat. Chem. Actu 1980, 53, 291. (32) Hingston, F. J.; Atkinson, R J.; Posner, A M.; Quirk, J. P. Nature 1967, 215,1459.
peptide” MW Nb
1 2 3
4
structureC,d
1060 2 HzN-ArgPro-Pro-Gly-Phe-Ser-Pro-Phe-&-OH 1298 3 HzN-Arg-ArgHypHypGly-PheSer-Phe-ArgOH 1394 4 HzN-Lys-Lys-Arg-Pro-HypGly-Thi-Ser-Phe-TiAripOH 1647 6 HzN-Ty-Gly-Gly-Phe-Leu-ArgArgAgAg-ProL~s-L~u-L~s-OH
1, bradykinin; 2, [~Arp, H7p2z3, ~-Phe’]-bradykinin; 3, Lys-Lys[Hyp3,p-(2-thienyl)-Ala5
