- Email: [email protected]
Ionic interactions between proteins in nonequilibrium pH gradient electrophoresis: Histones affect the migration of high mobility group nonhistone chromatin proteins
ANALYTICAL
132,294-304
BIOCHEMISTRY
(1983)
Ionic Interactions between Proteins in Nonequilibrium pH Gradient Electrophoresis: Histones Affect the Migration of High Mobility Group Nonhistone Chromatin Proteins LISA WEN,* RODNEY K. TWETEN,? PAUL J. ISACKSON,* JOHN J. IANDOLO,~ AND GERALD R. REECK*~’ *Department of Biochemistry andtDivisionof Biology,KansasStateUniversity,Manhattan,Kansas66506 Received November 22, 1982 In two-dimensional gel electrophoresis of the high mobility group (HMG) proteins, it has proved necessary to use nonequilibrium pH gradient electrophoresis (NEPHGE) in the first dimension rather than isoelectric focusing, because of the basic character of most of the HMG proteins [D. Tyrell, P. J. Isa&son, and G. R. Reeck (1982)Anal. Biochem.119, 433-4391. In this paper it is reported that in samples that contain histories, the mobilities of HMG proteins (particularly HMG-I , HMG-2, and HMG-E) are severely distorted in NEPHGE. This presumably results from formation of complexes between histones and HMG proteins through ionic interactions. Analysis of HMG proteins by NEPHGE/sodium dodecyl sulfate-gel electrophoresis is thus precluded in samples containing histones. Our results raise the possibility of similar artifacts occurring in NEPHGE (or isoelectric focusing) analysis of other proteins with regions of high charge density. KEY WORDS: nonequilibrium pH gradient electrophoresis; two-dimensional gel electrophomsis; HMG proteins; histones; nitrocellulose blot.
basic HMG proteins to reach their steady states. Our previous two-dimensional electrophoresis was carried out with material extracted from chicken erythrocyte chromatin with 0.35 M NaCl. The same NEPHGE/sodium dodecyl sulfate (SDS) approach failed, however, to give suitable results with extracts obtained with higher NaCl concentrations (L. Clow and G. R. Reeck, unpublished observations). The results described in this paper explain why the technique fails with certain types of HMG protein preparations: The presence of histones (either the lysine-rich histones, Hl or H5, or the core histones) can severely distort the mobilities of the HMG proteins in NEPHGE.
The high mobility group (HMG)2 proteins are a small, intensively studied set of nonhistone chromatin proteins ( 1). Multiple forms of several of the HMG proteins are known to exist (2,3). High-resolution, two-dimensional electrophoretic analysis of the HMG proteins would be useful in assessing multiplicity and in studying the occurrence of the multiple forms in a variety of samples and circumstances. We (4) have recently reported good resolution of the HMG protein by two-dimensional electrophoresis in which the first dimension was nonequilibrium pH gradient electrophoresis (NEPHGE). That approach, instead of isoelectric focusing (to equilibrium), was used because alkaline pH gradients could not be maintained long enough to allow the
MATERIALS
’ To whom correspondence should be addressed. * Abbreviations used: HMG, high mobility group; NEPHGE, nonequilibrium pH gradient electrophoresis; SDS, sodium dodecyl sulfate; TCA, trichloroacetic acid; IEF, isoelectric focusing. 0003-2697183 Copyright All
rights
0
1983
of reproducl~on
Academic in any
Press. form
Inc. reserved.
METHODS
Preparation of HMG proteins. Chicken erythrocyte chromatin was prepared and extracted with 0.35 M NaCl as previously de294
$3.00 by
AND
NONEQUILIBRlUM
pH
GRADIENT
ELECTROPHORESIS
scribed (3,4). Such extracts contain the canonical HMG proteins (HMG-1, HMG-2, HMG-E, HMG-14, and HMG- 17), some histone H 1, and a very small amount of histone H5. The total protein concentration in an extract was estimated, after lyophihzation, by the biuret method, using bovine serum albumin as standard. Histones H 1 and H5 were removed from 0.35 M NaCl extracts as follows. The extract (20 ml) was applied directly to a Bio-Rex 70 column (2.5 X 15 cm) that had been equilibrated to 0.7 M NaCl/SO mM TrisHCl (pH 7.5). The same solvent was applied after the sample had been applied, resulting in elution of all of the HMG proteins. H 1 and H5 were at least largely retained by the BioRex column. They were subsequently eluted with 2 M NaCl. Preparation ofhistones. Chromatin that had been treated with 0.35 M NaCl was extracted with 5% perchloric acid to release histones H 1 and H5. The perchlo& acid extract was dialyzed against 0.4 M HzS04 for 6 h. After dialysis, 5 vol of cold ethanol was added to precipitate histones HI and H5. The precipitate, recovered by centrifugation and dissolved in distilled water, was also extensively dialyzed in 50 mM Tris-HCI (pH 7.5). The extract was then chromatographed on a BioRex 70 column with a NaCl gradient of 0.5 to 2 M in 50 mM Tris-HCl (pH 7.5). Hl and H5 eluted, respectively, at approximately 0.74 and 0.9 M NaCl. Core histones were extracted with 0.4 M HzS04 from pellets obtained by centrimgation of chromatin in 5% perchloric acid. The core histones were ethanol precipitated from such extracts. Histone concentrations were determined spectrophotometrically using the following extinction coefficients (A$!$): histone H 1I 0.10; histone H5, 0.23; core histones, 0.4. Isolation and purification of antibodies against HMG-1. Antibodies were elicited in rabbits against purified HMG-1, with an injection protocol described by Bustin et al. (5). Anti-HMG-1 antibodies were purified by affinity chromatography on a column contain-
OF
HMG
PROTEINS
295
ing immobilized HMG proteins. The proteins that were immobilized were in the 2% trichloroacetic acid (TCA)-soluble fraction of a 0.35 M NaCl extract of chicken erythrocyte chromatin. That fraction is composed almost exclusively of the HMG proteins (including HMG-14 and HMG-17). The proteins were linked covalently to cyanogen-bromide-activated Sepharose 2B and used as described elsewhere to purify antibodies (6). HMG- 1, HMG-2, and HMG-E are immunologically cross-reactive and unrelated immunochemitally to HMG-14 and HMG-17 (5,7). Thus, a column on which a mixture of HMG proteins is immobilized is suitable for purifying anti-HMG-1 antibodies from an antiserum obtained by injection of purified HMG- 1. Purified antibodies were iodinated by the chloramine-T method (8). Free iodine was removed from the iodination mixture by gel chromatography on a Sephadex G-75 column (24 X 1.2 cm) that was equilibrated and eluted with 0.015 M sodium borate (pH 8.4)/O. 15 M NaCl. Two-dimensional gel electrophoresis. We used NEPHGE/SDS electrophoresis as devised by O’Farrell et al. (9) and applied by us to 0.35 M NaCl extracts of chicken erythrocyte chromatin (4). The first-dimension gels contained a 1: 1 mixture of LKB ampholytes of pH ranges 5-7 and 3.5-10. First-dimension electrophoresis was carried out for 2550 V-h. Immunochemical detection after two-dimensional electrophoresis. The transfer of proteins from unstained two-dimensional gels to nitrocellulose membranes (10) was carried out essentially as described by Bowen et al. (11). The gel was soaked for 3 h in 50 mM NaC1/2 mM EDTA/4 M urea/l0 mM TrisHCl (pH 7.0) and sandwiched between a nitrocellulose membrane and Whatman 3MM paper with glass plates. The sandwich was immersed in transfer buffer [50 mM NaCl/2 mM EDTA/O.l% NaI/lO mM Tris-HCl (pH 7.4)] for 3 days. Nonspecific protein adsorption was eliminated by soaking the nitrocellulose in blocking buffer (3% bovine serum albumin/ 10% preimmune serum in transfer buffer) at
296
WEN ET AL.
room temperature for 3 h. After rinsing with water, the nitrocellulose membrane was soaked for 8 h at room temperature in blocking buffer containing 12%anti-HMG- 1 antibodies at 3 X IO5 cpm/ml. The membrane was washed five times with 0.1 M Tris-HCl (pH 7.0)/l% NaCl and thoroughly dried. It was then used to expose Kodak X-Omat AR film for about 12 h. RESULTS
Radioimmunochemical detection of HMG proteins a@er two-dimensional electrophoresis. Figure 1A is a stained electrophoretogram obtained, as described under Materials and Methods, on the proteins extracted from chicken erythrocyte chromatin with 0.35 M NaCl. It is thus comparable to Fig. 1 of Tyrell et al. (4). Figure 1A is included to allow direct comparison to the autoradiogram in Fig. 1B and to document changes in the electrophoretie pattern resulting from slightly modified conditions for NEPHGE. The changes are most dramatic for HMG-14. In our previous two-dimensional electrophoretograms (4) we observed a single elongated spot for HMG14. With the slightly altered conditions used in the present work, HMG-14 separates into several spots. All of the spots labeled as HMG14 in Fig. 1A are also observed in two-dimensional electrophoresis of a TCA-solubility fraction that contains only HMG-14 and HMG-17. [Results not shown. The fraction used was that on which the electrophoretogram in Fig. 4 of Tyrell et al. (4) was obtained.] More spots were also seen in the current work for HMG-2, a second form of which was observed between a more basic HMG-2 spot and the most basic form of HMG-E. Indicated in Fig. 1A by white X’s are the two forms of histone H 1. Each extends to the right of the X as a horizontal streak, as we observed previously (4). Immediately below the X’s is a major spot (labeled Y in Fig. 1A) that has the same mobility in the second dimension as HMG- 1. We suggested previously that this spot below the H 1‘s might be histone
H5 (4). Instead, spot Y is actually HMG-1, as we demonstrate below. Figure 1B is an autoradiogram produced by exposure of a nitrocellulose blot of an electrophoretogram, probed with ‘251-antiHMG-1 antibodies. The electrophoretogram had been obtained on the same sample and under the same electrophoretic conditions as the electrophoretogram of Fig. IA. The antibodies, which had been elicited in rabbits against purified chicken erythrocyte HMG- 1, react with the spot known from our previous work to be HMG-1. Also, as expected from the immunochemical cross-reactivity between HMG- 1 and HMG-E (7), the antibodies reacted with the several forms of HMG-E on the nitrocellulose blot. Curiously, the antibodies reacted poorly, if at all, with HMG-2, despite the close immunochemical relationship that is known to exist between HMG-1 and HMG-2 (5). We have found that the antibodies react with native HMG-2 in double immunodiffusion (L. Wen, unpublished observation), so failure to react with HMG-2 on the nitrocellulose blot is presumably due in part to the physical (probably denatured) state of HMG-2 on the blot. The same procedure as that which gave rise to Fig. 1B was also carried out with iz51-IgG (immunoglobulin G) from normal rabbit serum. No radioactive spots were detected on the nitrocellulose blot (not shown). From the standpoint of this paper, the most interesting aspect of Fig. 1B is the reaction of the anti-HMG- 1 antibodies with material that has a wide range of mobilities in the first dimension and the mobility of HMG-1 in the second dimension. Extending to the left of the bulge in Fig. 1B (which corresponds to the position of the HMG-1 spot in Fig. 1A) is a streak that terminates in a distorted spot (which corresponds to the position of spot Y in Fig. 1A). This demonstrates that the material in spot Y is immunochemically crossreactive with HMG-1 and has the same mobility in SDS electrophoresis as HMG- 1. (Of only technical interest is the streaked extension rightward from the HMG-1 spot to a small
-NEPHGE
B
BASIC
ACIDIC
FIG. 1. (A) Two-dimensional gel electrophoresis of proteins extracted from chicken erythrocyte chromatin with 0.35 M NaCl. White x’s are histones HI. Spot Y is shown in succeeding text and figures to be HMG-1. (B) Autoradiogram after two-dimensional electrophoresis of proteins extracted from chicken erythrocyte chromatin with 0.35 M NaCI. Electrophoresis conditions were the same as for (A). After electrophorcsis. proteins were transferred to nitrocellulose and exposed to ‘ZS1-anti-HMG-l antibodies. 291
298
WEN ET AL.
amount of reactive material that was at the origin of the NEPHGE gel.) The fact that antiHMG-1 antibodies react with a streak that connects HMG-1 with spot Y suggests either that many forms of HMG- I exist with a moreor-less continuous distribution of pi’s or that the streak and spot Y are artifactual extensions from the HMG-1 spot of what is actually one polypeptide (HMG-1). In what follows we demonstrate that the latter is the case and that the streaking is a result of the presence of HI and H5 in the extract. Electrophoresis of extracts depleted in histones HI and H.5. By chromatography on BioRex 70, as described under Materials and Methods, we removed essentially all of histones Hl and H5 from a 0.35 M NaCl extract of chicken erythrocyte chromatin. A twodimensional electrophoretogram on the Hl/H5depleted extract is shown in Fig. 2. This electrophoretogram is to be compared to that in Fig. 1A. The most obvious differences are the apparent absence of histones HI in Fig. 2 and the great diminution in Fig. 2 in the amount
FIG.
2. Two-dimensional
of material in the region corresponding to spot Y in Fig. IA. More subtle differences are the absence of material streaking to the basic side of HMG-2 in Fig. 2 and improved resolution of the multiple forms of HMG-14 in Fig. 2. One-dimensional SDS-gel electrophoresis of the proteins removed from the extract by the Bio-Rex column demonstrated the presence of histone Hl and a trace amount of histone HS, and the absence of material with the mobility of HMG-1 (not shown). We interpret these results to indicate that removal from the extract of histones HI and H5 resulted in the migration of HMG-1 as a single spot rather than as two spots joined by a streak. Electrophoresis of reconstituted mixtures of HMG proteins and histones. Figure 3 demonstrates the effect on the electrophoretic pattern of HMG proteins of the addition of purified histone Hl to extracts that had been depleted in Hl and H5. Judging from the intensities of the histone HI spots, we added to the sample for Fig. 3A an amount of Hl that corresponded approximately to the level of
gel electrophoresis of extract depleted in histones Hl and H5.
NONEQUILIBRIUM
pH GRADIENT
ELECTROPHORESIS
OF HMG
PROTEINS
FIG;. 3. Two-dimensional gel eleetrophoresis of reconstituted mixture of HMG proteins and histone A) Purified histone H 1 (18 pg) was added to an extract that had been depleted in Hl and H5. ‘Il/HS-depleted extract was estimated to contain 250 pg of protein. (B) Purified histone Hl (74 rg) added to the same amount of extract used in (A).
299
300
WEiN
Hl in the original extract. The result was the appearance, below the Hl spots, of a substantial amount of material in the position of spot Y. Also, the intensity of the HMG- 1 spot (relative to the intensities of the HMG-2 and HMG-E spots) is less in Fig. 3A than in Fig. 2. These results are consistent, then, with the presence of histone H 1 resulting in much faster migration of some of the HMG-1 during NEPHGE than the migration of HMG-1 in the absence of H 1. In the sample for the electrophoretogram in Fig. 3B we added four times as much histone Hl as was added to the sample for Fig. 3A. The result was the virtual disappearance in the NEPHGE/SDS electrophoretogram of material at the position where HMG-1 migrates in the absence of histone HI and the appearance of still more material in the position of spot Y. In addition, note that in Fig. 3B there is marked streaking to the basic side of the HMG-2 spots. Electrophoresis of a sample to which we added twice as much H 1 as was added to the sample for Fig. 3B resulted in a shift of all of the HMG-2 and HMG-E to positions below that of spot Y, in addition to the shift of all of the HMG-1 to spot Y (result not shown). As might be expected, histone H5, a sequence homolog of HI (12), has an effect comparable to that of Hl when reconstituted with the H l/H5-depleted extract (Fig. 4). The core histones had a similar effect (Fig. 5). DISCUSSION
Histones, when added to samples containing HMG proteins, can severely distort the mobilities of HMG- 1, HMG-2, and HMG-E (Figs. 2-5). This presumably results from migration of the histones and the HMG proteins as complexes during NEPHGE. The addition of histone Hl at an appropriate level to an Hl/HS-depleted sample resulted in an electrophoretogram (Fig. 3A) similar to that observed by electrophoresis of 0.35 M NaCl extract (Fig. IA). Thus, it is likely that the occurrence of spot Y in Fig. 1A is due to the
ET
AL.
presence of histone Hl in the 0.35 M NaCl extract. That explanation is consistent with the reaction of anti-HMG-1 antibodies with material in the position of spot Y (Fig. IB), and also with the essential disappearance of spot Y when an extract is depleted of histones Hl and H5 (Fig. 2). The HMG proteins are completely dissociated from chromatin at 0.35 M NaCl (13), whereas histone Hl is only partially dissociated (14). Nonetheless, because H 1 occurs at a much higher level in chromatin than do the HMG proteins, the amount of Hl removed from chicken erythrocyte chromatin by 0.35 M NaCl is comparable to the amount of HMG proteins removed. It is fortuitous that the amount of Hl released is sufficient to distort the NEPHGE migration of only a portion of HMG- 1. Higher levels of H 1 preclude altogether the analysis by NEPHGE/ SDS-gel electrophoresis of not only HMG- 1, but HMG-2 and HMG-E as well. That is observed, for instance, in electrophoresis of material (principally H 1, H5, and HMG proteins) extracted from chicken erythrocy-te chromatin with 5% perchloric acid (L. Wen, unpublished observations). A comparable effect would be expected in electrophoresis of any HMG protein sample that contains histones at levels higher than the levels of HMG proteins: for example, sulfuric acid extracts, SDS extracts, and extracts obtained at NaCl concentrations greater than 0.35 M. In addition to the pronounced effects on the electrophoresis of the high-molecularweight HMG proteins, histones appear to reduce the resolution of multiple forms of HMG- 14. Reasonably good electrophoretograms of HMG- 14 (and HMG- 17) can nonetheless be obtained in samples with very high ratios of HI and H5 to HMG proteins (such as a 5% perchloric acid extract). That the several spots of HMG- 14 that we observe are indeed different forms of that protein and not electrophoretic artifacts is supported by our recent chromatographic resolution of several forms of HMG- 14 (L. Wen and G. R. Reeck, unpublished observations).
NONEQUILIBRIUM
pH GRADIENT
ELECTROPHORESIS
OF HMG
PROTEINS
FIG. 4. Two-dimensional gel electrophoresis of reconstituted mixture of HMG proteins and histone H5 (7 rg). The H l/HS-depleted extract was estimated to contain 250 pg of protein.
The O’Farrell approach to two-dimensional gel electrophoresis (9,15) uses solvents in both dimensions that are designed to promote solubility of proteins or, more accurately, of polypeptides. For essentially the same reasons that the solvents promote solubility, they disrupt protein/protein interactions. In the second dimension, SDS and 2-mercaptoethanol assure the essential absence of such interactions. The 9 M urea in the solvent for the first dimension would be expected to eliminate neady all protein/protein interactions by reducing the strength of hydrophobic forces ( 16) which are typically involved in such interactions (17), and by converting most proteins to random coils and thereby eliminating interactions that are dependent upon three-dimensional structure. Formation of complexes, in the solvent of the first dimension, between histones and HMG proteins is, however, the simplest means of explaining the results presented in this paper. We have not directly
proven the existence of such complexes, but they have been observed by other investigators under other, nondenaturing conditions. In particular, histone Hl has been shown to complex with HMG-1 (1819) and HMG-2 (18). The complexes are stable only at rather low ionic strength (19), which indicates that they result from electrostatic interactions. By necessity, the solvent for NEPHGE has a low ionic strength, thereby potentially allowing electrostatic interactions between urea-denatured polypeptides of high charge density. The C-terminal third of HMG-1 (or of HMG-2) is known to be extraordinarily acidic; the last 80 residues of HMG-I carry a net charge of -3 1 at neutral pH (20). We have recently postulated that that portion of the molecule is a physically distinct domain from the rest of the molecule, which itself has two domains (2 1). Complexes presumably form between the high-molecular-weight HMG proteins and histones through electrostatic at-
302
WEN ET AL.
FIG. 5. Two-dimensional gel electrophoresis of reconstituted mixture of HMG proteins and core histones (140 pg). The HI/HSdepleted extract was estimated to contain 250 fi of protein.
traction of the highly negatively charged C-terminal domains of the HMG proteins to the very basic histone molecules. Since addition of HI to Hl/HS-depleted extracts has its initial effect on the mobility of HMG-1, the Hl/HMG-1 interaction appears to be stronger (at least in the solvent for the first dimension) than the interactions of Hl with HMG-2 and HMG-E. Because the core histones have effects on HMG protein mobility similar to the effects of Hl and HS, we presume that complexes can also form between the core histones and HMG-1, HMG-2, and HMG-E. The effects of histones on the electrophoretic pattern of HMG-14 may be presumed, by a continuation of the same line of reasoning, to result from still weaker ionic interaction between histones and HMG- 14. The C-terminal portion of HMG-14 carries a net negative charge (22), although a much smaller one than that of the C-terminal third of the high-molecular-weight HMG proteins. An in-
teraction of HMG-14 with histones would most likely involve the acidic C-terminal portion of HMG-14. To our knowledge, there is no precedent for our proposal of electrostatic interactions between proteins in NEPHGE (or IEF), at least in the O’Farrell conditions, which include 9 M urea. Certain ionic interactions leading to artifactual effects have been observed in IEF (23); those interactions were between carrier ampholytes and polyelectrolytes, particularly polyanions such as heparin. Some interactions of that sort are apparently abolished in 8 M urea (24), though no mechanism has been proposed for that effect. Other ionic interactions between ampholytes and polyanions (e.g., those with polyphosphates) are apparently weakened but not eliminated by 8 M urea (24). The interactions that we are postulating, at least those that involve the high-molecularweight HMG proteins, differ from the binding
NONEQUILIBRIUM
pH GRADIENT
ELECTROPHORESIS
of (low-molecular-weight) ampholytes to polyanions in that we postulate interactions between two macromolecules (for instance, histone Hl and HMG-1). Each participating molecule thus carries many charges. That H 1 binds to HMG- 1 in the NEPHGE soIvent but carrier ampholytes would not be expected to do so (24) results from the large number of positive charges on the histone. [H 1 carries a net positive charge of 55 at neutral pH (25), whereas carrier ampholytes would have a net positive charge of 4 or 5 at the low pH at which interactions with polyanions have been observed (23).] Hence, the interaction of HI with HMG-1 would be expected to be much stronger than the interaction of ampholytes with HMG-1 or other polyanions. The H 1/HMG- 1 interaction persists in 9 M urea. The portion of HMG-1 that is presumably involved (domain C) appears by virtue of its peculiar amino acid composition to be unable to assume an ordered secondary or tertiary structure and, therefore, likely occurs as a random coil even under nondenatuting conditions (2 I ). Similarly, a substantiai portion of histone H1 appears to occur as a random coil in the absence of denaturants (26). Hence, urea may actually have very little effect on the interaction of H 1 with HMG- 1 and its homologs. Several procedures used to extract HMG proteins from chromatin also release histones in sufficiently large amounts to disallow use of NEPHGE/SDS electrophoresis for analyzing HMG proteins. Once histones are removed from such samples, NEPHGE/SDS electrophoresis can provide excellent resolution of the HMG proteins, including separation of multiple forms of the proteins. The technique should therefore be useful in studying the biochemical origin of those multiple forms and in investigating their biological significance. Whether artifacts in NEPHGE or IEF are produced by ionic interactions among proteins other than those we have studied remains to be seen. The histones and the HMG proteins are both peculiar in amino acid composition
OF HMG
PROTEINS
303
and their interaction in NEPHGE may be quite unique. On the other hand, in complex mixtures of proteins, such as are frequently analyzed by two-dimensional gel electrophoresis, pairs of equally peculiar but uncharacterized proteins may well occur. The apparent interaction in NEPHGE between histones and HMG proteins can therefore serve as an alert to a type of artifact that may be of some general significance. ACKNOWLEDGMENTS This work was supported by the Kansas Agricultural Experiment Station and NIH Grants CA- 17782 and GM29203 to G.R.R. and AI-17474 to J.J.1. G.R.R. is the recipient of NIH Research Career Development Award CA-00425 This is Publication 82-660-J of the Kansas Agricultural Experiment Station. Note added in proof For a similar sort of electrophoresis (but lacking urea in the first dimension), Nicolas and Goodwin have noted the requirement that histones HI and H5 be absent from samples of HMG proteins [Nicolas, R. H.. and Goodwin, G. H. (1982) in The HMG Chromosomal Proteins (Johns, E., ed.), p. 60, Academic Press, New York/London].
REFERENCES I. Goodwin, G. H., and Johns, E. W. (1973) Eur. .1. Biochem. 40, 214-219. 2. Mathew, C. G. P., Goodwin, G. H., Gooderham, K., Walker, 1. M., and Johns, E. W. (1979) Bi&em. Biophys. Rex Commun. 87, 1243-1251. 3. Isa&on, P. J., Debold, W., and Reeck, G. R. (I 980) FELLS L&f. 119, 337-342. 4. TyrelI, D., Isa&on, P. J., and Reeck. G. R. (1982) Anal. B&hem. 119,433-439. 5. Bustin, M., Hopkins, R. B., and Isenberg. 1. (1978) J. Biol. Chem. 253, 1694-1699. 6. Corfman, R. S., and Reeck, G. R. (1982) Biochim. Biophys. Acta 715, 170- 174. 7. Romani, M., Vidali, G., Tahourdin, C. S. M., and Bustin, M. (1980) J. Biol. Chem. 255, 468-474. 8. Krause, R. M., and McCarty, M. ( 1962) J. Exp. Med. 115,49-62. 9. O’Farrell, P. Z., Goodman, H. M., and O’Farrell, P. H. (1977) Cell 12, 1133-l 142. 10. Tweten, R. K., and Iandoio, J. J. ( 198 I ) fnfzcf. Immun. 34,900-907. Il. Bowen, B., Steinberg, J., Laemmli, U. K., and Weintraub, H. ( 1980) Nucl. Acid Res. 8, 1-2 1.
304
WEN ET AL,
12. Yaguchi, M., Roy, C., Dove, M., and Seligy, V. (1977) B&hem. Biophys. Res. Commun. 76, 100-106. 13. Isackson, P. J., Glow, L. G., and Reeck, G. R. (1981) FEBS Left. 125, 30-34. 14. Kumar, N. M., and Walker, I. 0. (1980) Nucl. Acid Res. 8, 3535-3551. 15. O’Farrell, P. H. (1975) .I Biol. Chem. 250, 40074021. 16. Wetlaufer, D. B., Mahk, S. K., Stroller, L., and Coffin, R. L. (1964) J. Amer. Chem. Sot. 86, 508-5 14. 17. Edelhoch, H., and Osborne, J. C., Jr. (1976) Advun. Protein Chem. 30, 183-250. 18. Smerdon, M. J., and Isenberg, I. (1976) Biochemistry 15,4242-4247. 19. Cary, P. S., Shooter, K. V., Goodwin, G. H., Johns, E. W., Olayemi, J. Y., Hartman, P. G., and Bradbury, E. M. (1979) B&hem. J. 183, 657-662.
20. Walker, J. M., Gooderham, K., Hastings, J. R. B., Mayes, E., and Johns, E. W. (1980) FEBS Lett. 122,264-270. 2 1, Reeck, G. R., Isackson, P. J., and Teller, D. C. Nature (London) 300,16-19. 22. Walker, J. M., Goodwin, G. H., and Johns, E. W. (1979) FEBS Lett. 100, 394-398. 23. Righetti, P. G., Gianazza, E., and Bosisio, A. B. (1980) in Recent Developments in Chromatography and Electrophoresis (Frigerio, A., and McCamish, eds.), Vol. 10, pp. 89-117, Elsevier, Amsterdam. 24. Gianazza, E., and Righetti, P. G. (1978) B&him. Biophys. Acta 540, 357-364. 25. Cole, R. D. (1977) in The Molecular Biology of the Mammalian Genetic Apparatus (Ts’o, P. 0. P., ed.), Vol. 1, pp. 93-104, Elsevier, Amsterdam. 26. Bradbury, E. M., Cary, P. D., Chapman, G. E., CraneRobinson, C., Danby, S. E., Rattle, H. W. E., Boublik, M., Palau, J., and Avilcs, F. J. (1975) Eur. J. Biochem. 52,605-6 13.
132,294-304
BIOCHEMISTRY
(1983)
Ionic Interactions between Proteins in Nonequilibrium pH Gradient Electrophoresis: Histones Affect the Migration of High Mobility Group Nonhistone Chromatin Proteins LISA WEN,* RODNEY K. TWETEN,? PAUL J. ISACKSON,* JOHN J. IANDOLO,~ AND GERALD R. REECK*~’ *Department of Biochemistry andtDivisionof Biology,KansasStateUniversity,Manhattan,Kansas66506 Received November 22, 1982 In two-dimensional gel electrophoresis of the high mobility group (HMG) proteins, it has proved necessary to use nonequilibrium pH gradient electrophoresis (NEPHGE) in the first dimension rather than isoelectric focusing, because of the basic character of most of the HMG proteins [D. Tyrell, P. J. Isa&son, and G. R. Reeck (1982)Anal. Biochem.119, 433-4391. In this paper it is reported that in samples that contain histories, the mobilities of HMG proteins (particularly HMG-I , HMG-2, and HMG-E) are severely distorted in NEPHGE. This presumably results from formation of complexes between histones and HMG proteins through ionic interactions. Analysis of HMG proteins by NEPHGE/sodium dodecyl sulfate-gel electrophoresis is thus precluded in samples containing histones. Our results raise the possibility of similar artifacts occurring in NEPHGE (or isoelectric focusing) analysis of other proteins with regions of high charge density. KEY WORDS: nonequilibrium pH gradient electrophoresis; two-dimensional gel electrophomsis; HMG proteins; histones; nitrocellulose blot.
basic HMG proteins to reach their steady states. Our previous two-dimensional electrophoresis was carried out with material extracted from chicken erythrocyte chromatin with 0.35 M NaCl. The same NEPHGE/sodium dodecyl sulfate (SDS) approach failed, however, to give suitable results with extracts obtained with higher NaCl concentrations (L. Clow and G. R. Reeck, unpublished observations). The results described in this paper explain why the technique fails with certain types of HMG protein preparations: The presence of histones (either the lysine-rich histones, Hl or H5, or the core histones) can severely distort the mobilities of the HMG proteins in NEPHGE.
The high mobility group (HMG)2 proteins are a small, intensively studied set of nonhistone chromatin proteins ( 1). Multiple forms of several of the HMG proteins are known to exist (2,3). High-resolution, two-dimensional electrophoretic analysis of the HMG proteins would be useful in assessing multiplicity and in studying the occurrence of the multiple forms in a variety of samples and circumstances. We (4) have recently reported good resolution of the HMG protein by two-dimensional electrophoresis in which the first dimension was nonequilibrium pH gradient electrophoresis (NEPHGE). That approach, instead of isoelectric focusing (to equilibrium), was used because alkaline pH gradients could not be maintained long enough to allow the
MATERIALS
’ To whom correspondence should be addressed. * Abbreviations used: HMG, high mobility group; NEPHGE, nonequilibrium pH gradient electrophoresis; SDS, sodium dodecyl sulfate; TCA, trichloroacetic acid; IEF, isoelectric focusing. 0003-2697183 Copyright All
rights
0
1983
of reproducl~on
Academic in any
Press. form
Inc. reserved.
METHODS
Preparation of HMG proteins. Chicken erythrocyte chromatin was prepared and extracted with 0.35 M NaCl as previously de294
$3.00 by
AND
NONEQUILIBRlUM
pH
GRADIENT
ELECTROPHORESIS
scribed (3,4). Such extracts contain the canonical HMG proteins (HMG-1, HMG-2, HMG-E, HMG-14, and HMG- 17), some histone H 1, and a very small amount of histone H5. The total protein concentration in an extract was estimated, after lyophihzation, by the biuret method, using bovine serum albumin as standard. Histones H 1 and H5 were removed from 0.35 M NaCl extracts as follows. The extract (20 ml) was applied directly to a Bio-Rex 70 column (2.5 X 15 cm) that had been equilibrated to 0.7 M NaCl/SO mM TrisHCl (pH 7.5). The same solvent was applied after the sample had been applied, resulting in elution of all of the HMG proteins. H 1 and H5 were at least largely retained by the BioRex column. They were subsequently eluted with 2 M NaCl. Preparation ofhistones. Chromatin that had been treated with 0.35 M NaCl was extracted with 5% perchloric acid to release histones H 1 and H5. The perchlo& acid extract was dialyzed against 0.4 M HzS04 for 6 h. After dialysis, 5 vol of cold ethanol was added to precipitate histones HI and H5. The precipitate, recovered by centrifugation and dissolved in distilled water, was also extensively dialyzed in 50 mM Tris-HCI (pH 7.5). The extract was then chromatographed on a BioRex 70 column with a NaCl gradient of 0.5 to 2 M in 50 mM Tris-HCl (pH 7.5). Hl and H5 eluted, respectively, at approximately 0.74 and 0.9 M NaCl. Core histones were extracted with 0.4 M HzS04 from pellets obtained by centrimgation of chromatin in 5% perchloric acid. The core histones were ethanol precipitated from such extracts. Histone concentrations were determined spectrophotometrically using the following extinction coefficients (A$!$): histone H 1I 0.10; histone H5, 0.23; core histones, 0.4. Isolation and purification of antibodies against HMG-1. Antibodies were elicited in rabbits against purified HMG-1, with an injection protocol described by Bustin et al. (5). Anti-HMG-1 antibodies were purified by affinity chromatography on a column contain-
OF
HMG
PROTEINS
295
ing immobilized HMG proteins. The proteins that were immobilized were in the 2% trichloroacetic acid (TCA)-soluble fraction of a 0.35 M NaCl extract of chicken erythrocyte chromatin. That fraction is composed almost exclusively of the HMG proteins (including HMG-14 and HMG-17). The proteins were linked covalently to cyanogen-bromide-activated Sepharose 2B and used as described elsewhere to purify antibodies (6). HMG- 1, HMG-2, and HMG-E are immunologically cross-reactive and unrelated immunochemitally to HMG-14 and HMG-17 (5,7). Thus, a column on which a mixture of HMG proteins is immobilized is suitable for purifying anti-HMG-1 antibodies from an antiserum obtained by injection of purified HMG- 1. Purified antibodies were iodinated by the chloramine-T method (8). Free iodine was removed from the iodination mixture by gel chromatography on a Sephadex G-75 column (24 X 1.2 cm) that was equilibrated and eluted with 0.015 M sodium borate (pH 8.4)/O. 15 M NaCl. Two-dimensional gel electrophoresis. We used NEPHGE/SDS electrophoresis as devised by O’Farrell et al. (9) and applied by us to 0.35 M NaCl extracts of chicken erythrocyte chromatin (4). The first-dimension gels contained a 1: 1 mixture of LKB ampholytes of pH ranges 5-7 and 3.5-10. First-dimension electrophoresis was carried out for 2550 V-h. Immunochemical detection after two-dimensional electrophoresis. The transfer of proteins from unstained two-dimensional gels to nitrocellulose membranes (10) was carried out essentially as described by Bowen et al. (11). The gel was soaked for 3 h in 50 mM NaC1/2 mM EDTA/4 M urea/l0 mM TrisHCl (pH 7.0) and sandwiched between a nitrocellulose membrane and Whatman 3MM paper with glass plates. The sandwich was immersed in transfer buffer [50 mM NaCl/2 mM EDTA/O.l% NaI/lO mM Tris-HCl (pH 7.4)] for 3 days. Nonspecific protein adsorption was eliminated by soaking the nitrocellulose in blocking buffer (3% bovine serum albumin/ 10% preimmune serum in transfer buffer) at
296
WEN ET AL.
room temperature for 3 h. After rinsing with water, the nitrocellulose membrane was soaked for 8 h at room temperature in blocking buffer containing 12%anti-HMG- 1 antibodies at 3 X IO5 cpm/ml. The membrane was washed five times with 0.1 M Tris-HCl (pH 7.0)/l% NaCl and thoroughly dried. It was then used to expose Kodak X-Omat AR film for about 12 h. RESULTS
Radioimmunochemical detection of HMG proteins a@er two-dimensional electrophoresis. Figure 1A is a stained electrophoretogram obtained, as described under Materials and Methods, on the proteins extracted from chicken erythrocyte chromatin with 0.35 M NaCl. It is thus comparable to Fig. 1 of Tyrell et al. (4). Figure 1A is included to allow direct comparison to the autoradiogram in Fig. 1B and to document changes in the electrophoretie pattern resulting from slightly modified conditions for NEPHGE. The changes are most dramatic for HMG-14. In our previous two-dimensional electrophoretograms (4) we observed a single elongated spot for HMG14. With the slightly altered conditions used in the present work, HMG-14 separates into several spots. All of the spots labeled as HMG14 in Fig. 1A are also observed in two-dimensional electrophoresis of a TCA-solubility fraction that contains only HMG-14 and HMG-17. [Results not shown. The fraction used was that on which the electrophoretogram in Fig. 4 of Tyrell et al. (4) was obtained.] More spots were also seen in the current work for HMG-2, a second form of which was observed between a more basic HMG-2 spot and the most basic form of HMG-E. Indicated in Fig. 1A by white X’s are the two forms of histone H 1. Each extends to the right of the X as a horizontal streak, as we observed previously (4). Immediately below the X’s is a major spot (labeled Y in Fig. 1A) that has the same mobility in the second dimension as HMG- 1. We suggested previously that this spot below the H 1‘s might be histone
H5 (4). Instead, spot Y is actually HMG-1, as we demonstrate below. Figure 1B is an autoradiogram produced by exposure of a nitrocellulose blot of an electrophoretogram, probed with ‘251-antiHMG-1 antibodies. The electrophoretogram had been obtained on the same sample and under the same electrophoretic conditions as the electrophoretogram of Fig. IA. The antibodies, which had been elicited in rabbits against purified chicken erythrocyte HMG- 1, react with the spot known from our previous work to be HMG-1. Also, as expected from the immunochemical cross-reactivity between HMG- 1 and HMG-E (7), the antibodies reacted with the several forms of HMG-E on the nitrocellulose blot. Curiously, the antibodies reacted poorly, if at all, with HMG-2, despite the close immunochemical relationship that is known to exist between HMG-1 and HMG-2 (5). We have found that the antibodies react with native HMG-2 in double immunodiffusion (L. Wen, unpublished observation), so failure to react with HMG-2 on the nitrocellulose blot is presumably due in part to the physical (probably denatured) state of HMG-2 on the blot. The same procedure as that which gave rise to Fig. 1B was also carried out with iz51-IgG (immunoglobulin G) from normal rabbit serum. No radioactive spots were detected on the nitrocellulose blot (not shown). From the standpoint of this paper, the most interesting aspect of Fig. 1B is the reaction of the anti-HMG- 1 antibodies with material that has a wide range of mobilities in the first dimension and the mobility of HMG-1 in the second dimension. Extending to the left of the bulge in Fig. 1B (which corresponds to the position of the HMG-1 spot in Fig. 1A) is a streak that terminates in a distorted spot (which corresponds to the position of spot Y in Fig. 1A). This demonstrates that the material in spot Y is immunochemically crossreactive with HMG-1 and has the same mobility in SDS electrophoresis as HMG- 1. (Of only technical interest is the streaked extension rightward from the HMG-1 spot to a small
-NEPHGE
B
BASIC
ACIDIC
FIG. 1. (A) Two-dimensional gel electrophoresis of proteins extracted from chicken erythrocyte chromatin with 0.35 M NaCl. White x’s are histones HI. Spot Y is shown in succeeding text and figures to be HMG-1. (B) Autoradiogram after two-dimensional electrophoresis of proteins extracted from chicken erythrocyte chromatin with 0.35 M NaCI. Electrophoresis conditions were the same as for (A). After electrophorcsis. proteins were transferred to nitrocellulose and exposed to ‘ZS1-anti-HMG-l antibodies. 291
298
WEN ET AL.
amount of reactive material that was at the origin of the NEPHGE gel.) The fact that antiHMG-1 antibodies react with a streak that connects HMG-1 with spot Y suggests either that many forms of HMG- I exist with a moreor-less continuous distribution of pi’s or that the streak and spot Y are artifactual extensions from the HMG-1 spot of what is actually one polypeptide (HMG-1). In what follows we demonstrate that the latter is the case and that the streaking is a result of the presence of HI and H5 in the extract. Electrophoresis of extracts depleted in histones HI and H.5. By chromatography on BioRex 70, as described under Materials and Methods, we removed essentially all of histones Hl and H5 from a 0.35 M NaCl extract of chicken erythrocyte chromatin. A twodimensional electrophoretogram on the Hl/H5depleted extract is shown in Fig. 2. This electrophoretogram is to be compared to that in Fig. 1A. The most obvious differences are the apparent absence of histones HI in Fig. 2 and the great diminution in Fig. 2 in the amount
FIG.
2. Two-dimensional
of material in the region corresponding to spot Y in Fig. IA. More subtle differences are the absence of material streaking to the basic side of HMG-2 in Fig. 2 and improved resolution of the multiple forms of HMG-14 in Fig. 2. One-dimensional SDS-gel electrophoresis of the proteins removed from the extract by the Bio-Rex column demonstrated the presence of histone Hl and a trace amount of histone HS, and the absence of material with the mobility of HMG-1 (not shown). We interpret these results to indicate that removal from the extract of histones HI and H5 resulted in the migration of HMG-1 as a single spot rather than as two spots joined by a streak. Electrophoresis of reconstituted mixtures of HMG proteins and histones. Figure 3 demonstrates the effect on the electrophoretic pattern of HMG proteins of the addition of purified histone Hl to extracts that had been depleted in Hl and H5. Judging from the intensities of the histone HI spots, we added to the sample for Fig. 3A an amount of Hl that corresponded approximately to the level of
gel electrophoresis of extract depleted in histones Hl and H5.
NONEQUILIBRIUM
pH GRADIENT
ELECTROPHORESIS
OF HMG
PROTEINS
FIG;. 3. Two-dimensional gel eleetrophoresis of reconstituted mixture of HMG proteins and histone A) Purified histone H 1 (18 pg) was added to an extract that had been depleted in Hl and H5. ‘Il/HS-depleted extract was estimated to contain 250 pg of protein. (B) Purified histone Hl (74 rg) added to the same amount of extract used in (A).
299
300
WEiN
Hl in the original extract. The result was the appearance, below the Hl spots, of a substantial amount of material in the position of spot Y. Also, the intensity of the HMG- 1 spot (relative to the intensities of the HMG-2 and HMG-E spots) is less in Fig. 3A than in Fig. 2. These results are consistent, then, with the presence of histone H 1 resulting in much faster migration of some of the HMG-1 during NEPHGE than the migration of HMG-1 in the absence of H 1. In the sample for the electrophoretogram in Fig. 3B we added four times as much histone Hl as was added to the sample for Fig. 3A. The result was the virtual disappearance in the NEPHGE/SDS electrophoretogram of material at the position where HMG-1 migrates in the absence of histone HI and the appearance of still more material in the position of spot Y. In addition, note that in Fig. 3B there is marked streaking to the basic side of the HMG-2 spots. Electrophoresis of a sample to which we added twice as much H 1 as was added to the sample for Fig. 3B resulted in a shift of all of the HMG-2 and HMG-E to positions below that of spot Y, in addition to the shift of all of the HMG-1 to spot Y (result not shown). As might be expected, histone H5, a sequence homolog of HI (12), has an effect comparable to that of Hl when reconstituted with the H l/H5-depleted extract (Fig. 4). The core histones had a similar effect (Fig. 5). DISCUSSION
Histones, when added to samples containing HMG proteins, can severely distort the mobilities of HMG- 1, HMG-2, and HMG-E (Figs. 2-5). This presumably results from migration of the histones and the HMG proteins as complexes during NEPHGE. The addition of histone Hl at an appropriate level to an Hl/HS-depleted sample resulted in an electrophoretogram (Fig. 3A) similar to that observed by electrophoresis of 0.35 M NaCl extract (Fig. IA). Thus, it is likely that the occurrence of spot Y in Fig. 1A is due to the
ET
AL.
presence of histone Hl in the 0.35 M NaCl extract. That explanation is consistent with the reaction of anti-HMG-1 antibodies with material in the position of spot Y (Fig. IB), and also with the essential disappearance of spot Y when an extract is depleted of histones Hl and H5 (Fig. 2). The HMG proteins are completely dissociated from chromatin at 0.35 M NaCl (13), whereas histone Hl is only partially dissociated (14). Nonetheless, because H 1 occurs at a much higher level in chromatin than do the HMG proteins, the amount of Hl removed from chicken erythrocyte chromatin by 0.35 M NaCl is comparable to the amount of HMG proteins removed. It is fortuitous that the amount of Hl released is sufficient to distort the NEPHGE migration of only a portion of HMG- 1. Higher levels of H 1 preclude altogether the analysis by NEPHGE/ SDS-gel electrophoresis of not only HMG- 1, but HMG-2 and HMG-E as well. That is observed, for instance, in electrophoresis of material (principally H 1, H5, and HMG proteins) extracted from chicken erythrocy-te chromatin with 5% perchloric acid (L. Wen, unpublished observations). A comparable effect would be expected in electrophoresis of any HMG protein sample that contains histones at levels higher than the levels of HMG proteins: for example, sulfuric acid extracts, SDS extracts, and extracts obtained at NaCl concentrations greater than 0.35 M. In addition to the pronounced effects on the electrophoresis of the high-molecularweight HMG proteins, histones appear to reduce the resolution of multiple forms of HMG- 14. Reasonably good electrophoretograms of HMG- 14 (and HMG- 17) can nonetheless be obtained in samples with very high ratios of HI and H5 to HMG proteins (such as a 5% perchloric acid extract). That the several spots of HMG- 14 that we observe are indeed different forms of that protein and not electrophoretic artifacts is supported by our recent chromatographic resolution of several forms of HMG- 14 (L. Wen and G. R. Reeck, unpublished observations).
NONEQUILIBRIUM
pH GRADIENT
ELECTROPHORESIS
OF HMG
PROTEINS
FIG. 4. Two-dimensional gel electrophoresis of reconstituted mixture of HMG proteins and histone H5 (7 rg). The H l/HS-depleted extract was estimated to contain 250 pg of protein.
The O’Farrell approach to two-dimensional gel electrophoresis (9,15) uses solvents in both dimensions that are designed to promote solubility of proteins or, more accurately, of polypeptides. For essentially the same reasons that the solvents promote solubility, they disrupt protein/protein interactions. In the second dimension, SDS and 2-mercaptoethanol assure the essential absence of such interactions. The 9 M urea in the solvent for the first dimension would be expected to eliminate neady all protein/protein interactions by reducing the strength of hydrophobic forces ( 16) which are typically involved in such interactions (17), and by converting most proteins to random coils and thereby eliminating interactions that are dependent upon three-dimensional structure. Formation of complexes, in the solvent of the first dimension, between histones and HMG proteins is, however, the simplest means of explaining the results presented in this paper. We have not directly
proven the existence of such complexes, but they have been observed by other investigators under other, nondenaturing conditions. In particular, histone Hl has been shown to complex with HMG-1 (1819) and HMG-2 (18). The complexes are stable only at rather low ionic strength (19), which indicates that they result from electrostatic interactions. By necessity, the solvent for NEPHGE has a low ionic strength, thereby potentially allowing electrostatic interactions between urea-denatured polypeptides of high charge density. The C-terminal third of HMG-1 (or of HMG-2) is known to be extraordinarily acidic; the last 80 residues of HMG-I carry a net charge of -3 1 at neutral pH (20). We have recently postulated that that portion of the molecule is a physically distinct domain from the rest of the molecule, which itself has two domains (2 1). Complexes presumably form between the high-molecular-weight HMG proteins and histones through electrostatic at-
302
WEN ET AL.
FIG. 5. Two-dimensional gel electrophoresis of reconstituted mixture of HMG proteins and core histones (140 pg). The HI/HSdepleted extract was estimated to contain 250 fi of protein.
traction of the highly negatively charged C-terminal domains of the HMG proteins to the very basic histone molecules. Since addition of HI to Hl/HS-depleted extracts has its initial effect on the mobility of HMG-1, the Hl/HMG-1 interaction appears to be stronger (at least in the solvent for the first dimension) than the interactions of Hl with HMG-2 and HMG-E. Because the core histones have effects on HMG protein mobility similar to the effects of Hl and HS, we presume that complexes can also form between the core histones and HMG-1, HMG-2, and HMG-E. The effects of histones on the electrophoretic pattern of HMG-14 may be presumed, by a continuation of the same line of reasoning, to result from still weaker ionic interaction between histones and HMG- 14. The C-terminal portion of HMG-14 carries a net negative charge (22), although a much smaller one than that of the C-terminal third of the high-molecular-weight HMG proteins. An in-
teraction of HMG-14 with histones would most likely involve the acidic C-terminal portion of HMG-14. To our knowledge, there is no precedent for our proposal of electrostatic interactions between proteins in NEPHGE (or IEF), at least in the O’Farrell conditions, which include 9 M urea. Certain ionic interactions leading to artifactual effects have been observed in IEF (23); those interactions were between carrier ampholytes and polyelectrolytes, particularly polyanions such as heparin. Some interactions of that sort are apparently abolished in 8 M urea (24), though no mechanism has been proposed for that effect. Other ionic interactions between ampholytes and polyanions (e.g., those with polyphosphates) are apparently weakened but not eliminated by 8 M urea (24). The interactions that we are postulating, at least those that involve the high-molecularweight HMG proteins, differ from the binding
NONEQUILIBRIUM
pH GRADIENT
ELECTROPHORESIS
of (low-molecular-weight) ampholytes to polyanions in that we postulate interactions between two macromolecules (for instance, histone Hl and HMG-1). Each participating molecule thus carries many charges. That H 1 binds to HMG- 1 in the NEPHGE soIvent but carrier ampholytes would not be expected to do so (24) results from the large number of positive charges on the histone. [H 1 carries a net positive charge of 55 at neutral pH (25), whereas carrier ampholytes would have a net positive charge of 4 or 5 at the low pH at which interactions with polyanions have been observed (23).] Hence, the interaction of HI with HMG-1 would be expected to be much stronger than the interaction of ampholytes with HMG-1 or other polyanions. The H 1/HMG- 1 interaction persists in 9 M urea. The portion of HMG-1 that is presumably involved (domain C) appears by virtue of its peculiar amino acid composition to be unable to assume an ordered secondary or tertiary structure and, therefore, likely occurs as a random coil even under nondenatuting conditions (2 I ). Similarly, a substantiai portion of histone H1 appears to occur as a random coil in the absence of denaturants (26). Hence, urea may actually have very little effect on the interaction of H 1 with HMG- 1 and its homologs. Several procedures used to extract HMG proteins from chromatin also release histones in sufficiently large amounts to disallow use of NEPHGE/SDS electrophoresis for analyzing HMG proteins. Once histones are removed from such samples, NEPHGE/SDS electrophoresis can provide excellent resolution of the HMG proteins, including separation of multiple forms of the proteins. The technique should therefore be useful in studying the biochemical origin of those multiple forms and in investigating their biological significance. Whether artifacts in NEPHGE or IEF are produced by ionic interactions among proteins other than those we have studied remains to be seen. The histones and the HMG proteins are both peculiar in amino acid composition
OF HMG
PROTEINS
303
and their interaction in NEPHGE may be quite unique. On the other hand, in complex mixtures of proteins, such as are frequently analyzed by two-dimensional gel electrophoresis, pairs of equally peculiar but uncharacterized proteins may well occur. The apparent interaction in NEPHGE between histones and HMG proteins can therefore serve as an alert to a type of artifact that may be of some general significance. ACKNOWLEDGMENTS This work was supported by the Kansas Agricultural Experiment Station and NIH Grants CA- 17782 and GM29203 to G.R.R. and AI-17474 to J.J.1. G.R.R. is the recipient of NIH Research Career Development Award CA-00425 This is Publication 82-660-J of the Kansas Agricultural Experiment Station. Note added in proof For a similar sort of electrophoresis (but lacking urea in the first dimension), Nicolas and Goodwin have noted the requirement that histones HI and H5 be absent from samples of HMG proteins [Nicolas, R. H.. and Goodwin, G. H. (1982) in The HMG Chromosomal Proteins (Johns, E., ed.), p. 60, Academic Press, New York/London].
REFERENCES I. Goodwin, G. H., and Johns, E. W. (1973) Eur. .1. Biochem. 40, 214-219. 2. Mathew, C. G. P., Goodwin, G. H., Gooderham, K., Walker, 1. M., and Johns, E. W. (1979) Bi&em. Biophys. Rex Commun. 87, 1243-1251. 3. Isa&on, P. J., Debold, W., and Reeck, G. R. (I 980) FELLS L&f. 119, 337-342. 4. TyrelI, D., Isa&on, P. J., and Reeck. G. R. (1982) Anal. B&hem. 119,433-439. 5. Bustin, M., Hopkins, R. B., and Isenberg. 1. (1978) J. Biol. Chem. 253, 1694-1699. 6. Corfman, R. S., and Reeck, G. R. (1982) Biochim. Biophys. Acta 715, 170- 174. 7. Romani, M., Vidali, G., Tahourdin, C. S. M., and Bustin, M. (1980) J. Biol. Chem. 255, 468-474. 8. Krause, R. M., and McCarty, M. ( 1962) J. Exp. Med. 115,49-62. 9. O’Farrell, P. Z., Goodman, H. M., and O’Farrell, P. H. (1977) Cell 12, 1133-l 142. 10. Tweten, R. K., and Iandoio, J. J. ( 198 I ) fnfzcf. Immun. 34,900-907. Il. Bowen, B., Steinberg, J., Laemmli, U. K., and Weintraub, H. ( 1980) Nucl. Acid Res. 8, 1-2 1.
304
WEN ET AL,
12. Yaguchi, M., Roy, C., Dove, M., and Seligy, V. (1977) B&hem. Biophys. Res. Commun. 76, 100-106. 13. Isackson, P. J., Glow, L. G., and Reeck, G. R. (1981) FEBS Left. 125, 30-34. 14. Kumar, N. M., and Walker, I. 0. (1980) Nucl. Acid Res. 8, 3535-3551. 15. O’Farrell, P. H. (1975) .I Biol. Chem. 250, 40074021. 16. Wetlaufer, D. B., Mahk, S. K., Stroller, L., and Coffin, R. L. (1964) J. Amer. Chem. Sot. 86, 508-5 14. 17. Edelhoch, H., and Osborne, J. C., Jr. (1976) Advun. Protein Chem. 30, 183-250. 18. Smerdon, M. J., and Isenberg, I. (1976) Biochemistry 15,4242-4247. 19. Cary, P. S., Shooter, K. V., Goodwin, G. H., Johns, E. W., Olayemi, J. Y., Hartman, P. G., and Bradbury, E. M. (1979) B&hem. J. 183, 657-662.
20. Walker, J. M., Gooderham, K., Hastings, J. R. B., Mayes, E., and Johns, E. W. (1980) FEBS Lett. 122,264-270. 2 1, Reeck, G. R., Isackson, P. J., and Teller, D. C. Nature (London) 300,16-19. 22. Walker, J. M., Goodwin, G. H., and Johns, E. W. (1979) FEBS Lett. 100, 394-398. 23. Righetti, P. G., Gianazza, E., and Bosisio, A. B. (1980) in Recent Developments in Chromatography and Electrophoresis (Frigerio, A., and McCamish, eds.), Vol. 10, pp. 89-117, Elsevier, Amsterdam. 24. Gianazza, E., and Righetti, P. G. (1978) B&him. Biophys. Acta 540, 357-364. 25. Cole, R. D. (1977) in The Molecular Biology of the Mammalian Genetic Apparatus (Ts’o, P. 0. P., ed.), Vol. 1, pp. 93-104, Elsevier, Amsterdam. 26. Bradbury, E. M., Cary, P. D., Chapman, G. E., CraneRobinson, C., Danby, S. E., Rattle, H. W. E., Boublik, M., Palau, J., and Avilcs, F. J. (1975) Eur. J. Biochem. 52,605-6 13.