Journal of Chromatography, 359 (1986) 11 l-1 19
`
`Elsevier Science Publishers B.V.. Amsterdam - Printed in The Netherlands CHROMSYMP. 782 REVERSED-PHASE CHROMATOGRAPHY OF PROTEINS ON RESIN- BASED WIDE-PORE PACKINGS KATHLEEN A. TWEETEN and THOMAS N. TWEETEN*
`
`Polymer Laboratories Inc., 160 Old Farm Road, Amherst, MA 01002 (U.S.A.)
`
`SUMMARY PLRP-S, a macroporous poly(styrene-divinylbenzene) HPLC packing mate- rial, was evaluated for analysis of proteins. Materials with a particle size of 8 pm and pore sizes of 100 A, 300 A, and 1000 A were tested under similar chromatographic conditions to determine the effect of pore size and surface area on retention, selec- tivity, and efficiency. This resin-based material has a highly homogeneous hydro- phobic surface, based on phenyl moieties contributed by the copolymers of the pack- ing. Thus, unlike silica, PLRP-S has no bonded alkyl groups or residual silanol sites. In addition, the PLRP-S material is stable over a wide range of solvents, pH, and ionic strengths. The elution order of a series of proteins on PLRP-S 300 A was similar to that observed on a variety of alkyl-bonded silica-based reversed-phase columns. The percentage of acetonitrile in the mobile phase necessary to elute three of these proteins (ribonuclease A, cytochrome c, and ovalbumin) was determined. Protein desorption occurred over a very narrow range of 2-4% increase in concentration of the organic modifier. No apparent change in protein selectivity or peak area occurred with repetitive injections of a mixture of proteins, indicating little relative loss of protein on the column. Pore size appeared to have little effect on the selectivity or retentivity of the test proteins. It appeared that the loss of total surface area due to increased pore size was offset by the increased availability of pore surface area for protein-packing interaction. The resin-based PLRP-S 300 A was used for analysis of wheat proteins, bovine pancreatic enzymes, high-molecular-weight proteins, and whey proteins. Over the broad range of molecular weights and hydrophobicities in these applications, proteins were eluted as sharp, symmetrical peaks. Thus, PLRP-S offers an effective alternative to reversed-phase silica for the analysis of proteins. INTRODUCTION In recent years reversed-phase liquid chromatography has been adopted for the analysis of peptides and proteins. The utility of a variety of macroporous sili- ca-based packings for reversed-phase separation of proteins has been widely recog- nizedlp5. Because proteins generally carry a net charge under the mobile phase con- * Address for correspondence: The Pillsbury Co., 311 2nd St. S.E., Minneapolis, MN 55414, U.S.A. 0021-9673/86/$03.50 0 1986 Elsevier Science Publishers B.V.
`
`1
`
`MTX1025
`
`

`

`112 K. A. TWEETEN, T. N. TWEETEN ditions used for reversed-phase chromatography, of major concern in the use of silica-based packings is the presence of residual and accessible surface silanols. In- teraction of proteins with the silanol groups leads to mixed-mode separation with resultant poor chromatographic resolution due to peak broadening and tailing6. A new high-resolution resin-based packing (PLRP-S 300 A) has been intro- duced for reversed-phase analysis of proteins. While offering the advantage of com- patibility with a wide range of solvents, pH, and ionic strengths, the highly cross- linked poly(styrenedivinylbenzene) PLRP-S can also be operated at normal flow- rates and column pressures without loss of column performance. The resin-based material has a highly homogeneous, non-polar surface due to phenyl moieties contributed by the copolymers of the packing material. Thus, unlike silica-based material, PLRP-S has no bonded alkyl groups or residual silanol groups. In the present study, PLRP-S material of different pore sizes was evaluated for its utility in the chromatography of proteins. The efficiency and selectivity of the material for the chromatography of proteins having a wide range of hydrophobicities and molecular weights was determined. EXPERIMENTAL
`
`Apparatus
`
`High-performance liquid chromatography (HPLC) was performed on PLRP- S 100 A, 300 A, and 1000 A columns (Polymer Labs., Church Stretton, U.K.) using two Knauer Type 64 pumps (Polymer Labs.), a Knauer Type 50B HPLC gradient programmer (Polymer Labs.), a Knauer dynamic mixing chamber (Sonntek, Wood- cliff Lake, NJ, U.S.A.) and a Rheodyne Model 7125 loop injector (Rheodyne, Cotati, CA, U.S.A.). All columns were 150 x 4.6 mm I.D. and contained 8-pm packing. A Knauer Type 87 variable-wavelength detector (Polymer Labs.) and a Trilab 2000 chromatography data system (Trivector, West Chester, PA, U.S.A.) were used for detection, data collection, and analysis.
`
`Reagents
`
`All proteins were obtained from Sigma (St. Louis, MO, U.S.A.). Acetonitrile and water were of HPLC quality and were purchased from J. T. Baker (Phillipsburg, NJ, U.S.A.) as was the trifluoroacetic acid (TFA).
`
`Chromatographic conditions
`
`All chromatographic separations were carried out at ambient temperature. Mobile phases were vacuum-degassed and sonicated. Flow-rates generally were 1.0 ml/min, and proteins were injected in lo-p1 volumes. Unless otherwise stated for gradient elution, solvent A was 0.1% TFA in acetonitrile-water (1:99) and solvent B was 0.1% TFA in acetonitrile-water (95:5). Proteins were detected at 220 nm. RESULTS AND DISCUSSION
`
`Efect of organic mod@er concentration on protein retentivity
`
`The percentage of acetonitrile in the mobile phase necessary to elute ribonu- clease A, cytochrome c, and ovalbumin was determined. PLRP-S 300 A columns
`
`2
`
`

`

`RPC OF PROTEINS ON WIDE-PORE PACKINGS 113
`
`were
`
`equilibrated with a mobile phase containing increasing (1% increments) ace- tonitrile concentrations, individual proteins were injected, and retention times were determined. The elution characteristics of the proteins on PLRP-S were similar to those reported for reversed-phase silica - 7 l l. Following injection of the proteins into the column, there appeared to be little or no desorption until a critical concentration of acetonitrile was reached (Fig. 1). At this point, the retention of the various proteins greatly decreased. With a further 24% increase in organic modifier concentration, the proteins were no longer retained. Preliminary results (data not shown) suggested that when the organic modifier concentration was increased to high levels (greater than 70% acetonitrile for ribonuclease A and cytochrome c) proteins were once again adsorbed on the PLRP-S packing. Increased protein retention at high modifier con- centrations has been reported for reversed-phase bonded silicagql 2. The percentages of acetonitrile required for initial desorption of ribonuclease A and cytochrome c were 26% and 31%, respectively. The desorption of ovalbumin was observed to be more complex. Initial desorption was detected at 48% acetonitrile but multiple, distorted peaks eluted. This may be due to the presence of ovalbumin species in varying states of unfoldingi3, column deglycosylation14, or the partial separation of the heterogeneous glycosylated forms of ovalbumin15-17. Upon increas- ing the organic modifier concentration to 50%, the
`
`k’
`
`for ovalbumin decreased five-fold, and a single species of protein was eluted. * I
`
`40--
`
`?
`.‘-I
`
`. .
`
`E
`(Y 30--
`E
`;r
`
`,,
`
`6
`s 20--
`
`5
`.lJ
`
`lz
`
`”
`
`lO--
`
`0+ 20
`
`I
`I
`I
`
`:
`I
`I
`I
`I
`I
`I
`I
`I
`
`P
`
`h
`
`&*A.-
`
`____________
`
`8_,
`
`30
`
`40
`Percent Acetonitrile
`
`SO
`
`i
`60
`
`Fig. 1. Effect of organic modifier concentration on protein retentivity. Proteins were injected into a PLRP-S 300 A column equilibrated with increasing (1% increments) of acetonitrile. Retention times were determined for ribonuclease A (O-O), cytochrome c (B---m), and ovalbumin (U-0). The asterisks indicate the modifier concentrations below which no protein desorption was detected.
`
`3
`
`

`

`114 K. A. TWEETEN, T. N. TWEETEN
`
`Protein recovery
`
`Selectivity and column performance of PLRP-S in protein chromatography
`
`To determine protein recovery from PLRP-S 300 A, increasing amounts of ribonuclease A, cytochrome c, and ovalbumin were individually injected in a series (2-fold increments) from 1.5 pg to 100 pg back to 1.5 pg. Mobile phases containing acetonitrile concentrations giving k’ values of about 4 were used (0.1% TFA in acetonitrile-water (27:73), acetonitrile-water (33:67), and acetonitrile-water (51:49) for ribonuclease A, cytochrome c, and ovalbumin, respectively). After each injection, protein retention time, peak area, and peak height were determined at 220 nm. For all three proteins, there was a very good linear response between the amount of protein injected and both detector peak area and peak height response up to 100 pg. The peak area resulting from the initial injection of 1.5 pg of each protein was compared to that of the final injection of 1.5 pg at completion of the series. An increase of about 8% in peak area for the final injection relative to the initial injection was detected for the various proteins and may indicate some protein loss on the packing during the initial injection. A certain amount of column “conditioning” may be required, especially when small quantities of protein are injected into a new col- umn.
`To further investigate protein recovery on PLRP-S 300 A and to examine column selectivity and efficiency, repetitive injections of a mixture of ribonuclease A and ovalbumin were made over a two-day period. Proteins were eluted with a 20- 60% gradient of solvent B over 15 min. Retention times, peak areas, and peak heights were determined, and the results are summarized in Table I. Throughout the series of injections, the column demonstrated good selectivity, as peak areas and retention times for both proteins remained relatively constant. Whereas ribonuclease A showed no decrease in peak height, some reduction in peak height for the more hydrophobic ovalbumin was observed with time. These results indicate that PLRP-S has repro- ducible selectivity over a broad range of protein hydrophobicities. Because of the homogeneous surface characteristics of the PLRP-S, it is expected that selectivity variations between batches of the packing would be minimized.
`PLRP-S packings of two different pores sizes, 300 A and 1000 A, were com- pared for chromatography of proteins over a molecular weight range of 12 OOO- TABLE I EFFECT OF REPETITIVE INJECTIONS OF PROTEIN ON PLRP-S 300 8, SELECTIVITY AND COLUMN PERFORMANCE E = the gradient elution capacity factor, x = the mean value of 32 injections and s = the standard deviation.
`
`Effect of pore size on protein retentivity, pore permeation, and column performance
`
`Protein
`
`R
`
`Retention time (s)
`
`Peak height
`
`Peak area
`
`x
`
`s
`
`x
`
`s
`
`2
`
`s
`
`Ribonuclease A 2.9 405 4.5 381 15 4320 91 Ovalbumin 7.7 899 8.6 156 35 3918 117
`
`4
`
`

`

`RPC OF PROTEINS ON WIDE-PORE PACKINGS 115 335 000. Equal amounts of protein were injected into each column and were eluted using a 20-60% gradient of solvent B over 15 min (ribonuclease A, cytochrome c, ovalbumin) or 30-60% solvent B over 12 min (fl-amylase, apoferritin, thyroglobulin). Peak heights, retention times, and areas were determined. Elution times for each of the proteins were similar on both columns. An in- crease in pore size from 300 8, to 1000 A appeared to have no significant effect on peak heights over the molecular weight range investigated (Table II). The loss of surface area which accompanied the increase in pore size from 300 A to 1000 8, appeared to be offset by the increase in the region of the pore surface available for interaction of the proteins with the packing material. Preliminary results (data not shown) indicated a significant increase in peak height when high-molecular-weight fibrous proteins (collagen and fibrinogen) were chromatographed on PLRP-S 1000 A compared to PLRP-S 300 A. Ribonuclease A, cytochrome c, apoferritin, ovalbumin, p-amylase, and thy- roglobulin were chromatographed on PLRP-S with pore sizes of 100 A, 300 A, and 1000 A. The proteins were chromatographed under conditions where they were not retained (70% acetonitrile). The elution times of the proteins were compared to that of thiourea, which was used as a marker for the total permeation volume of the columns. The elution times for thiourea on the 100-A, 300-A, and 1000-A columns were 87, 98 and 110 s, respectively. The proteins all eluted prior to the thiourea and their order of elution suggests some selective permeation of pores in all three column types (Fig. 2). The time required for the proteins to be eluted from the columns was observed to decrease with increasing molecular weight. On the 300-A and 1000-k, columns all the proteins appeared to be able to enter the packing pores. On the 100-8, column, selective permeation was restricted to proteins with a molecular weight of less than 200 000. The 300-A pore type, based on these results and those of the column performance studies would appear to offer the most adequate ratio between surface area and accessible pore volume for highest-efficiency separations of globular proteins. The 1000-8, column, on the other hand, demonstrated potential for more efficient separations of fibrous proteins.
`
`Applications of PLRP-S 300 2
`
`A variety of protein samples were chromatographed on PLRP-S 300 A. Analy- sis of a mixture of standard proteins demonstrated that the order of elution of these TABLE II EFFECT OF PLRP-S PORE SIZE ON PROTEIN PEAK HEIGHT
`
`Protein
`
`Molecular
`weight (daltons)
`
`Peak area
`
`Peak height
`
`300 A
`
`1000 A
`
`300 ;i
`
`1000 A
`
`12 400
`
`3680
`
`3665
`
`305
`
`300
`
`Ribonuclease A 13 600 Apoferritin 22 150* Ovalbumin 45 000 fl-Amylase 200 000 Thyroglobulin 335 ooo* * Subunit molecular weight.
`
`4066 4045 376 395 10 649 10 996 353 355 4873 4215 277 276 5846 5872 242 253 3216 3184 49 50
`
`5
`
`Cytochrome c
`

`

`K. A. TWEETEN. T. N. TWEETEN 4+ , SO 99 70 90 90 Retent i on T i me (seconds) Fig. 2. Effect of pore size on selective permeation of packing pores by proteins. Ribonuclease A, cyto- chrome c, ovalbumin, apoferritin, j?-amylase, and thyroglobulin were chromatographed on PLRP-S 100 A (m-m), 300 A (o--O), and 1000 8, (0-e) with 0.1% TFA in acetonitrile-water (70:30) and reten- tion times were determined. , 5
`
`20
`
`JO
`
`Retention time (mid
`
`Fig. 3. Analysis of a mixture of standard proteins on PLRP-S 300 A. Proteins were chromatographed on a 250 x 4.6 mm I.D. column with a linear gradient of 2660% solvent B over 22 min at a flow-rate of 1.5 ml/min. Peaks: 1 = ribonuclease A, 2 = insulin, 3 = cytochrome c, 4 = lysozyme, 5 = bovine serum albumin, 6 = myoglobin and 7 = ovalbumin. (Courtesy of L. L. Lloyd, Polymer Labs.).
`
`6
`
`15
`

`

`Retention
`
`160
`
`time (mid
`
`RPC OF PROTEINS ON WIDE-PORE PACKINGS 117 t i0 is
`Fig. 4. Analysis of gliadin proteins from Scout wheat flour on PLRP-S 300 A. The ethanol-extracted wheat proteins were chromatographed on a 150 x 4.6 mm I.D. column. Solvent A was 0.1% TFA in acetonitrile-water (1585) and solvent B was 0.1% TFA in acetonitrile-water (80:20). A gradient of 20- 50% B over 120 min at 0.5 ml/min was used. (Courtesy of D. L. Wetzel, Kansas State University). 1 5
`
`10
`
`15 20 25 30 Retention time (min) Fig. 5. Analysis of representative whey proteins on PLRP-S 300 A. The proteins were chromatographed on a 150 x 4.6 mm I.D. column with a gradient of 3648% solvent B over 24 min. Peaks: 1 = a- lactalbumin, 2 = j?-lactoglobulin (B chain), and 3 = fi-lactoglobulin (A chain).
`
`7
`
`

`

`Retention time (mid
`
`Fig. 6. Analysis of bovine pancreatic enzymes on PLRP-S 300 A. A gradient of 3740% solvent B over 15 min, followed by 40-45% solvent B over 10 min, was used for chromatography of the enzymes on a 150 x 4.6 mm I.D. column. Peaks: 1 = trypsinogen, 2 = trypsin, 3 = chymotrypsin, and 4 = chymo- trypsinogen. proteins from PLRP-S (Fig. 3) is similar to that seen on a variety of alkyl-bonded silica columns. A separation of gliadin proteins from Scout wheat flour is shown in Fig. 4. Reversed-phase chromatography has been shown to be an effective method for finger-printing of wheat varieties 18. PLRP-S provided high resolution of the pro- teins in the flour extract. Also of commercial importance is the ability to compare proteins in various whey preparations 19. A separation of the major proteins present in whey is shown in Fig. 5. Near-baseline resolution was obtained for the A and B chains of P-lactoglobulin. The PLRP-S was able to provide excellent separation of chymotrypsin and its proenzyme, chymotrypsinogen (Fig. 6). Both of these proteins were readily separated from the bovine pancreatic enzyme trypsin and its precursor form, trypsinogen. Over the broad range of molecular weights and hydrophobicities illustrated by these applications, proteins were eluted as sharp peaks with no ob- servable tailing. Thus, the PLRP-S packing with its homogeneous, non-polar surface provided for high resolution separations of proteins. ACKNOWLEDGEMENTS We thank L. L. Lloyd and F. P. Warner (Polymer Labs.) for advice and M. F. Barnett and Z. Dryzek (Polymer Labs.) for supplying experimental PLRP-S col- umns .
`
`8
`
`K. A. TWEETEN, T. N. TWEETEN
`

`

`RPC OF PROTEINS ON WIDE-PORE PACKINGS 119 REFERENCES 1 J. Rivier, J. Liq. Chromatogr., 1 (1978) 389. 2 R. V. Lewis, A. Fallon, S. Stein, K. D. Gibson and S. Undenfriend, Anal. Biochem., 104 (1980) 153. 3 H. P. J. Bennett, J. Chromatogr., 266 (1983) 501. 4 J. D. Pearson, W. C. Mahoney, M. A. Hermodson and F. E. Regnier, J. Chromatogr., 207 (1981) 325. 5 F. E. Regnier, Methods Enzymol., 91 (1983) 137. 6 H. Engelhardt and H. Muller, Chromatographia, 19 (1984) 19. 7 M. Rubenstein, Anal. Biochem., 98 (1978) 1. 8 E. C. Nice, M. W. Capp, N. H. C. Cooke and M. J. O’Hare, J. Chromatogr., 218 (1981) 569. 9 M. T. W. Hearn and B. Grego, J. Chromatogr., 218 (1981) 497. 10 G. Xindu and P. W. Carr, J. Chromatogr., 269 (1983) 96. 11 N. H. C. Cooke, B. G. Archer, M. J. O’Hare, E. C. Nice and M. W. Capp, J. Chromatogr., 255 (1981) 115. 12 M. T. W. Hearn and B. Grego, J. Chromatogr., 255 (1983) 125. 13 K. Benedek, S. Dong and B. L. Karger, J. Chromatogr., 317 (1984) 227. 14 M. W. Berchtold, K. J. Wilson and C. W. Heinzmann, Biochemistry, 21 (1982) 6552. 15 G. Vavacek and F. E. Regnier, Anal. Biochem., 109 (1980) 345. 16 M. T. W. Hearn and B. Grego, J. Chromatogr., 296 (1984) 61. 17 A. R. Kerlavage, C. J. Weitzmann, T. Hasan and B. S. Cooperman, J. Chromatogr., 266 (1983) 2 18 J. A. Bietz, Cereal Chem., 62 (1985) 201. 19 R. J. Pearce, Aus. J. Dairy Tech., 38 (1983) 114. .25.
`
`9
`
`