Eur. J. Biochem. 77, 487-493 (1977)
Sites of Histone/Histone Interaction in the H3 - H4 Complex Lothar BOHM Biochemistry Department, University of Cape Town Hiroaki HAYASHI, Peter D. CARY, Tom MOSS, Colyn CRANE-ROBINSON, and E. Morton BRADBURY Biophysics Laboratories, Portsmouth Polytechnic (Received March 18, 1977)
Sites of interaction between histones H3 and H4 have been probed by investigating complex formation, firstly between histone H4 and three peptides cleaved by chemical means from histone H3 (residues 1 - 90 and 1- 120 using cyanogen bromide and residues 42- 135 using N-bromosuccinimide), secondly between histone H3 and two peptides cleaved from histone H4 (residues 1- 84 using cyanogen bromide and residues 38 - 102 using chymotrypsin) and thirdly between the H4 peptide (residues 38 - 102) and the three H3 peptides (residues 1- 90, 1- 120 and 42- 135). The criterion for complex formation is the appearance of characteristic perturbed resonances in the aromatic region of the 270-MHz proton resonance spectrum of the peptide mixture. It is concluded that loss of 37 N-terminal residues from histone H4 and 41 N-terminal residues from histone H3 does not prevent complex formation, whilst the loss of 18 C-terminal residues from H4 and 45 C-terminal residues from H3 does prevent it; the last 15 C-terminal residues of H3 are, however, not required for forming a complex. The regions important for complex formation are therefore defined as residues 42- 120 in histone H3 and residues 38 - 102 in histone H4. In a physical study of the tetrameric state of the histones H3 and H4, it was shown [hat the complex between them is characterised by discrete perturbations in the proton magnetic resonance spectrum [l]. Such effects are a direct consequence of the tertiary structure of the complex and are always observed with globular proteins. The frictional ratio of the H3 . H4 complex (1.99 in 50 mM acetate/bisulphite pH 5, 1.7 in 25 mM phosphate pH 7 [l, and T. Moss, unpublished data]) is, however, much above that expected for a fully compact globular protein. It was concluded [l]that whilst the complex has elements of tertiary structure, it is not fully globular and has free unstructured sections of polypeptide chain, probably the N-terminal. If certain regions of the H4 and H3 chains are not included in the globular structure, then it should be possible to form the complex with incomplete polypeptide chains. This communication describes the results of NMR investigations of complex formation (a) between histone H3 and two H4 peptides H4(38 102) and H4(1-84), (b) between histone H4 and Abbreviation. NMR, nuclear magnetic resonance, Nomenclature. Histones are named according to international agreement, see Ciba Found. Symp. 28 (1975). Peptide fragments of the histones are indicated by the histone abbreviation followed by residue numbers in parentheses; e.g. H3(1-90) represents residues 1 - 90 of histone H3.
three H3 peptides H3(1- 90), H3(1- 120) and H3(42135) and (c) between peptide Hl(38- 102) and three peptides of H3, viz: (1 -90), (1 - 120) and (42- 135). From the results, regions of the two chains are defined that are not essential for complex formation and, conversely, other regions are found to be essential for complex formation. The criterion used for complex formation is the appearance of the characteristic perturbed peaks in the aromatic region of the NMR spectrum. This criterion has the advantage that it indicates the presence of the tertiary structure specific for the H3 . H4 complex, i.e. if complexes form with conformational characteristics other than those of the native H3 . H4, they will be detected as a different spectrum. It is important also that the method be able to select between cross-complex formation and self-aggregated histone. This is possible since the latter gives rise to very broad peaks that are essentially invisible in a highresolution NMR spectrum. MATERIALS AND METHODS Isolation and Purification of Chicken Erythrocyte Histone H3 Dimer
Whole chicken erythrocyte histone was isolated and fractionated as previously described [3]. The H3
488
Sites of Histone/Histone Interaction in the H3 . H4 Complex
Table 1. Amino acid composition of peptides f r o m chicken erythrocyte histone H3 and calf thymus histone H4 Amino acid
H3 (1 - 90) calc.
obs.
H3(1-120)
H3(42-135)
H4(38- 102)
H4(1-84)
calc.
obs.
calc.
obs.
calc.
obs.
calc.
obs.
12 2 14 4 10 6 13 5 6 16 1 6 2 5 11 3 4
9.1 2.5 13.8 4.6 9.0 4.4 13.5 5.7 6.0 15.9
5 1 13 5 5 4 13 3 3 10 1 5 2 7 11 2 4
5.4
5 1 8 3 6 1 5 0 8 5 0 8 1 3 5 4 2
3.9 1.3 8.7 3.3 5.4 2.2 4.7 0.2 7.6 5.5 7.6 1.0 3.4 5.1 3.7 1.4
10 2 12 4 6 2 5 1 13 6 0 7
8.75 1.8 12.6 4.0 5.0 2.0 5.0 1.5 12.4 6.6
residues/molecule Lysine Histidine Arginine Aspartic acid Threonine a Serine" Glutamic acid Proline Glycine A 1anine Cystine Valine Methionine Isoleucine Leucine Tyrosine" Phenylalanine
11 1 13 2 8 5 9 5 5 11 0 4 1 3 7 2 3
N-terminal amino acid
Alanine
Response to iodosobenzoate - ive to form S-S dimer
a
10.4 1.0 12.6 2.0 6.8 4.1 10.0 4.4 4.6 11.9 -
3.9 b
2.9 7.5 1.9 2.6
d
6.1 b 5.2 11.6 2.2 4.0
Alanine
+ ive
12.5 5.3 4.0 3.3 13.5 3.1 3.8 11.5 4.2 5.4 10.0 3.7
O
6 6 2 1
~
7.6 b
5.3 6.4 2.6 1.2
Arginine
~
ive'
No correction for hydrolytic losses. Homoserine lactone present. Oxidised by treatment with N-bromosuccinimide. Oxidised by CNBr treatment. Detected as spirolactone.
fraction was used directly for dimerization as given below. Alternatively, arginine-rich histones extracted from chicken erythrocyte deoxyribonucleoprotein with ethanol/HCl according to the procedure of Johns [4] were oxidized and the dimer obtained by fractionation. In a typical experiment 348 mg of a mixture of histones H3, H2A and H4 were dissolved in 12 ml 8 M urea pH 2 (HCl), gently stirred and disaggregated for 30 min. The pH was adjusted to 7.4 using NaOH and 10 mlO.5 M Tris. 5 mg iodosobenzoic acid (Pierce Chemical) neutralized with NaOH to give 1 ml were added and the solution was gently stirred for 12 h at 20 "C. The pH was readjusted to 7.4 and 7 mg iodosobenzoic acid prepared as above were added. The oxidation was continued for another 6 h. The dialyzed and freeze-dried residue (312 mg) was taken up in 8 M urea pH 2 (HCl) and fractionated on a 5 x 100-cm Sephadex G-100 column in 0.02 M HCl as previously described [5]. The HCl eluant from peak TI was freeze-dried twice, redissolved in water and freeze-dried to give 86 mg electrophoretically pure H3-S-S-H3 dimer. By the same procedure 60mg pure histone H3 prepared by pH-5.1 fractionation of whole histone [3] yielded 47 mg H3-S-S-H3 dimer.
Cyanogen Bromide Cleavage of Histone H3 Cyanogen bromide cleavage was essentially according to the method of Givol and Porter [6] and was performed on the H3-S-S-H3 dimer. In a typical experiment 106 mg dimer in 20 ml 70 % (v/v) formic acid and 70 mg CNBr were kept at room temperature under gentle agitation in a closed counting vial with a minimum of air space. After 5 and 20 h 70 mg CNBr were again added giving a total of 210 mg CNBr (147-fold molar excess). After 30 h 10 vol. of water were added. The freeze-dried residue was taken up in 8 M urea pH 2 (HCl) and fractionated on a 5 x 100-cm Sephadex G-100 column to give 34 mg, 9 mg and 6 mg peptide in peaks 3, 4 and 5 respectively (Fig. 1A solid line). These peptides correspond to the fragments H3(1-90), H3(91- 121)2 dimer and H3(121-135) previously identified by sequence analysis [7]. On refractionation on Sephadex G-100, 34 mg crude peptide (peak 3, Fig. 1A) yielded 25 mg H3(1-90) peptide which was electrophoretically homogeneous on polyacrylamide gels [8], by dansylation [9] gave alanine as the N-terminal amino acid which had an amino acid composition as indicated (Table 1).
L. Bohm, H. Hayashi, P. D. Cary, T. Moss, C. Crane-Robinson, and E. M. Bradbury
489
G-100 column to give intact H3 monomer followed by an H3 peptide (peaks 1 and 2 respectively, Fig. 1B). By dansylation the N-terminal amino acid of this peptide preparation was found to be alanine. Its electrophoretic mobility on acetic acid/urea/polyacrylamide gels [8] corresponded with the expected molecular weight (120 residues) and was higher than the mobility of histone H2A (129 residues) but lower than the mobility of histone H4 (102 residues). After oxidation with iodosobenzoate the electrophoretic mobility of this peptide was found to be considerably lower than in the presence of mercaptoethanol and was similar to the mobility of histone H1. From the specificity of cleavage, elution from Sephadex, electrophoretic mobility, response to oxidation, end-group analysis and amino acid composition (Table 1) this peptide was identified as the H3(1- 120) fragment. N-Bromosuccinimide Cleavage of Histone H3
Eluate volume (ml)
Fig. 1. Elution projilox of' SepIzu~Io.~ G I 0 0 chromatography of ( A ) CNBr total cleavage product, (B) rechromatography of peaks I + 2 of ( A ) and ( C ) N-bromosuccinimide totul cleavage product. Column sizes: (A) 5 x 100 cm, (B) 2.5 x 90 cm, (C) 5 x 100 cm
From the chromatogram (Fig. 1A, solid line) and the peptide yields it was obvious that CNBr cleavage was incomplete and that the fraction eluting in peak 2 (Fig. 1A) may represent an H3 fragment exceeding 90 residues. The apparent size on Sephadex and the specificity of cleavage suggested that this may be the H3(1 - 120) fragment arising from preferential cleavage cleavage at methionine-120. To favour incomplete cleavage, the amount of CNBr was reduced to a total of 100 mg (70-fold molar excess on 106 mg H3-S-S-H3 dimer) which were added in two equal portions at 0 and 5 h. The cleavage was terminated after 10 h and the freeze-dried residue was fractionated as described above (Fig. 1A, broken line). The eluant from peaks 1 and 2 (Fig. 1A) was pooled, dialyzed, freeze-dried, reduced in 8 M urea/l% 2-mercaptoethanol and refractionated on a 2.5 x 90-cm Sephadex
N-Bromosuccinimide cleavage was done according to the method of Ramachandran and Witkop [lo]. H3-S-S-H3 dimer was used as the substrate. In a typical experiment 140 mg dimer were dissolved in 50 ml 50% (vjv) acetic acid and titrated with 200-pl aliquots of N-bromosuccinimide reagent (96 mg Nbromosuccinimide recrystallized from glacial acetic acid dissolved in 5 ml 50 % vjv acetic acid). After each addition of reagent the ultraviolet spectrum was recorded against a blank containing identical aliquots of reagent in 50 ml 50 % (vjv) acetic acid. The total consumption was 2.3 ml. Cleavage was accompanied by a blue shift of the spectrum from 280-260 nm. The end point was taken when a 100-p1 aliquot of reagent caused no further increase of the absorbance at 260 nm and the solution remained yellow. After addition of 5 vol. of water the solution was freezedried. The residue was taken up in 6 M urea pH 2 (HC1) and fractionated on a 5 x 100-cm Sephadex G-100 column in 0.02 M HCl. Repeated freeze-drying of the HCl eluant yielded 8,13,88 and 9 mg of peptide from peaks 1, 2, 3 and 4 respectively (Fig. 1C). Peak 1 after refractionation on Sephadex G-300 in 0.02 M HCl was electrophoretically homogeneous in acetic acid/urea/polyacrylamide gels [8] and in sodium dodecylsulfate/polyacrylamide gels [l 11. In both systems the electrophoretic mobility of the peptide was higher than that of intact histones H3 and H2A but lower than that of histone H4. By dansylation arginine was detected as the N-terminal amino acid, glutamine was absent. Sephadex elution and electrophoretic mobility in relation to intact histone H3, specificity of cleavage and end-group analysis suggested a large peptide lacking part of the N-terminal chain. By these criteria and from the amino acid composition (Table 1) the peptide was identified as the H3(42- 135) fragment.
490
Sites of Histone/Histone Interaction in the H 3 . H4 Complex
Calf Thymus Histone H4 This was prepared according to the method of Johns [12] and purified by exclusion chromatography on Biogel P10, eluting with HCl, pH 2.
Cyanogen Bromide Cleavage of Histone H4 H 3 ( 1- 135)
The peptide H4( 1 - 84) was prepared by essentially the same method as used for the two H3 peptides. The details of preparation and purification are given elsewhere 1131. The product was electrophoretically pure on urea/acetic acid gels 181. Table 1 gives the amino acid composition of the sample used.
t i 4 ( 1 -102)
H 3 ( 1- 120)
H 4 ( 1-102)
Chymotrypsin Cleavage of Histone H4 The bond between leucine-37 and alanine-38 is a primary cleavage point for chymotrypsin in the H4 molecule. From a limited chymotryptic digest, therefore, it is possible to obtain the peptide H4(38 - 102). The details of preparation and purification are given elsewhere 1131. The product was electrophoretically pure on urea/acetic acid gels [8]. Table 1 gives the amino acid composition of the peptide sample used.
Preparation of Complexes Stock solutions of all the peptides and intact histones were made up in H 2 0 , pH 3 at about 10 mg/ ml. The precise concentration was obtained by measuring the absorbance of the solutions at 280 nm using an absorption coefficient of 1400 cm-l (mol tyrosine)-'. The concentration of the solution containing the peptide H3(42- 135) was obtained by measuring the peptide absorbance and using the methods described elsewhere [14,15]. The solutions were then made 8M in urea by adding solid urea and the final volume noted. Dithiothreitol was added next, to a concentration of 1 mg/ml. Equimolar mixtures of the peptides and intact histones (and the three peptide mixtures) were then made up and allowed to stand at 4 "C overnight. Afterwards the mixtures were dialysed against 'HzO, pH 6.5, 0.1 mg/ml dithiothreitol, with several changes of dialysate for at least 48 h. The final dialysate contained no dithiothreitol. The absorbance of the solutions at 280 nm was again measured to check that no considerable loss of peptide or histone had occurred, making a correction for volume increase during dialysis.
Nuclear Magnetic Resonance Spectra of the peptide mixtures were obtained in 2 H 2 0 at pH z 6 in 5-mm tubes using a Bruker 270-MHz Fourier transform spectrometer. The free induction decay of 5 x lo3 or lo4 pulses per spectrum was multiplied by a negative exponential (equivalent
Fig.2. 270-MHz N M R spectra of the aromatic resonances of ( A ) a control H3 . H4 complex and ( B - D ) histone H4 mixed with three peptides cleaved from histone H3. Peaks 1 and 2 are perturbed resonances characteristic of the native complex. Peak 0 is at the shift of unperturbed tyrosine resonance (two protons ortho to -OH)
to 2-Hz line broadening) to improve the signal-tonoise ratio. RESULTS AND DISCUSSION Fig.2 shows the aromatic spectrum of a control sample of intact calf thymus histone H4 mixed in equimolar ratio with intact chicken erythrocyte histone H3 and with three peptides cleaved from the H3 molecule. The upfield spectra are not shown because the ring-current shifted peaks observed in the native H3 . H4 complex are very broad and difficult to observe and therefore not easily used as a criterion of complex formation. The perturbed aromatic resonances characteristic of the complex (see [l]) are indicated by vertical lines in Fig.2-4 and can probably be assigned to tyrosine residues. Fig.2 shows that peptides H3(42- 135) and H3(1- 120) both interact with intact histone H4 to give a spectrum
491
L. Bohm, H. Hayashi, P. D. Cary, T. Moss, C. Crane-Robinson, and E. M. Bradbury
01 2
11 I
H3( 1-135) H4(38-102)
I I I I
1 9
I 8
I 7
H3( 1- 135) H4(1-84)
I 6
I
I
5
9
I 8
1 7
1 6
I 5
6 (PPm) Fig.3. 270-MHz N M R spectra oJ the aromatic resonances of ( A ) a control H 3 . H4 complex and ( B , C ) histone H3 mixed with two peptides cleaved f r o m histone H4
6 (PPm) Fig.4. 270-MHz N M R spectra of the aromatic resonances of ( A ) a control H3 . H4 complex and ( B - D ) mixtures of peptide H4(38- 102) with three peptides cleaved from histone H3
like that of the control. Peptide H3( 1- 90), however, shows no spectral perturbations when mixed with intact histone H4 and gives the sharp simple spectrum expected of a random coil protein mixture. [The very weak resonance at peak positions 1 and 2 is probably due to impurities of intact H3 or H3(1- 120)remaining in the sample of H3(1- 90).] This type of spectrum is the result of equivalence among all residues of a given type in the chain i.e. tyrosine, phenylalanine and histidine, in this part of the spectrum. No specific interaction takes place therefore between histone H4 and peptide H3(1-90). When the mixtures H4/H3(42135), H4/H3(1- 120) and the control H4/H3 mixture were made up to 3 mM phosphate, pH 7, there was little effect on the spectrum. The addition of phosphate to the H4/H3(1- 90) mixture resulted in considerable line broadening and a reduction of apparant peak area by more than 70%. This is a clear indication
of the formation of high-molecular-weight aggregates in this mixture, a characteristic of the separate histones, and is further evidence that the peptide H3(190) does not interact specifically with histone H4. Fig.3 shows the aromatic spectrum of equimolar mixtures of intact histone H3 with the two peptides cleaved from H4, H4(1- 84) and H4(38 - 102), together with a control H3 . H4 complex. The spectra show that the C-terminal peptide H4(38 - 102) gives rise to the perturbed resonances of the control whilst the N-terminal peptide H4(1- 84) does not. The spectrum of mixture H3/H4(1- 84) does show some broad resonance in the region 6.5-6.7 ppm and a very weak component at the position of peak 1. This is probably due to self-aggregation and to a small amount of uncleaved histone H4 remaining in the sample of peptide H4(1- 84). It is clear, however, that peptide H4(1- 84) does not form a complex with histone H3
492
Sites of Histone/Histone Interaction in the H3 . I14 Complex 33
1
H3
I
(INTACT)
1
H4
. 1
H3
t
‘1
/
H4
PiPTIDE
CROSS-INTNIUCTIONS
135
H3 (IFTACT )
1
1
I
1
H4
120
1
I
I
H3
102
38
102
102
135
1 102
42
38 I
CLOAVACES
H4
N-TSRMI NAL
1
I
38
H4
I 1 2
CLSAVA CES
120
Hj
I
13 5
H3
t
C-TERMINAL
I J
7
I
H4
135
I
102
N and C-ThRMIVAL CLBAVAGB
Fig. 5 . Cross-interactions betweenpeptides of histones H3 andH4. (-)
under these conditions. As in the first experiment the two peptide-containing mixtures and the control were then made 3 mM in phosphate, pH 7. The control H3/H4 and the H3/H4(38 - 102) mixtures showed essentially no change, but very considerable line broadening and apparant loss of peak area was observed with the H3/H4(1- 84) mixture. We therefore conclude that the peptide H4(38 - 102), but not H4(1- 84) interacts specifically with histone H3. To substantiate the results obtained on the first two sets of complexes, a further three equimolar peptide mixtures were studied, each containing the peptide H4(38 - 102) (previously found to interact with histone H3) together with the three peptides from histone H3, viz: H3(1-90), H3(42- 135) and H3(1- 120). Fig.4 shows the aromatic region of the spectrum and it can be seen that the characteristic perturbed peaks are observed in the two mixtures containing the peptides H3(42- 135) and H3(1- 120), but not in that containing H3(1- 90). The spectra obtained with complexes formed by the peptide H3(42 - 135) are of particular interest since this N-bromosuccinimide peptide does not contain the aromatic side chains of the remaining two tyrosine residues (positions 54 and 100) which were destroyed in the cleavage reaction. (This was checked by noting the presence of only phenylalanine resonances in the spectrum of the peptide in the random coil form at pH 3.) Fig. 2 shows the spectrum of peptide H3(42 135) interacting with intact histone H4 and Fig.4
Complex formation observed; (-----)complex formation not observed
that with peptide H4(38 - 102). The spectrum between 6 and 7 ppm is similar for the two complexes and it follows that all of the peaks in this region are due to the four tyrosines of histone H4 (provided, of course, that no phenylalanine resonance is perturbed that far upfield). The appearance of the spectrum suggests that peaks 2 and 1 are each due to a single tyrosine (two protons each) and peak 0 to two unperturbed tyrosines (four protons). Although it is not possible at present to make assignments of peaks 0, 1 and 2 to particular residues in histone H4, the use of modification procedures such as nitration or iodination should permit this to be done in the future. The data on complex formation are summarised in Fig. 5 in which a solid line indicates cross-complex formation. The major conclusion that follows from the data is that it is the C-terminal rather than the N-terminal portions of the H3 and H4 molecules that are the most important for complex formation. Firstly, it is noted that intact histone H3, and also its two peptides H3(1- 120) and H3(42- 135), interacts not only with intact histone H4 but also with its peptide H4(38102). It is clear therefore that residues 1- 37 of histone H4 (36 % of the molecule) are not essential for complex formation. We have previously proposed that the N-terminal region of histone H4 is not involved in the tertiary structure of the complex [l] and have also shown that there is no secondary structure in this portion of the H4 chain [13]. In a parallel way, the
L. Bohm, H. Hayashi, P. D. Cary, T. Moss, C. Crane-Robinson, and E. M. Bradbury
fact that intact histone H4 interacts with the peptide H3 (42- 135), shows that the N-terminal 41 residues of histone H3 (30% of the molecule) are likewise not essential for complex formation. The importance of the C-terminal portions of both chains is seen from the observation that neither H3(1- 90) (67 % of the H3 molecule) nor H4(1- 84) (82 % of the H4 molecule) are able to form cross-complexes with their intact histone partners. This demonstrates the critical importance of the C-terminal 19 residues of histone H4 and 45 residues of histone H3. In the case of histone H3, however, it is found that its peptide H3(1- 120) does form a complex with both histone H4 and peptide H4(48 - 102), indicating that for histone H3 the most critical region for interaction does not include the final 15 residues. It is concluded therefore, on the basis of the peptides studied here, that the regions of the two chains that are critical for complex formation are residues 42 - 120 for histone H3 and residues 38 - 102 for histone H4. These may not be the absolute minimum chain lengths required for complex formation but other peptides would be required to define this even more precisely. It is perhaps surprising that folding and complex formation can occur between molecules that have lost 36 % (H4) and 30 % (H3) of their N-terminal amino acids. This must be due both to the extremely high association constant between histones H3 and H4 (estimated at z 1021 [16]) and also to the fact that a good proportion of their residues are conformationally external to the tertiary complex. The present data demonstrate that the external residues are mainly N-terminal. Is it possible to define precisely the boundary between the two structural domains? Without a more detailed probe of the structure of the whole of the complex, one cannot be certain that the N-terminal sections missing in the minimal complex H4(38 - 102) . H3(42- 135) are all
external to the tertiary structure. The NMR spectral perturbations are a reflection of the integrity of the hydrophobic cores of the tertiary structure and are probably not dependent on the presence or absence of several residues on the perifery of the globular structure. A very high association constant between the folded C-terminal portions of histones H3 and H4 could result in complex formation when several residues are missing and thus 40 or so residues is probably an overestimate of the size of the free random coil Nterminal chains in the native complex.
REFERENCES 1. Moss, T., Cary, P. D., Crane-Robinson, C. & Bradbury, E. M. (1976) Biochemistry, 15, 2261 -2267. 2. Reference deleted. 3. van der Westhuyzen, D. R., Bohm, E. L. & von Holt, C. (1974) Biochim. Biophys. Acta, 359, 341 - 345. 4. Johns, E. W. (1964) Biochem. J . 92, 55-59. 5. Brandt, W. F. & von Holt, C. (1971) FEBS Lett. 14, 338-342. 6. Givol, D. & Porter, R. R. (1965) Biochem. J. 97, 32C- 34C. 7. Brandt, W. F. & von Holt, C. (1974) Eur. J . Biochem. 46,419429. 8. Panyim, S. & Chalkley, R. (1969) Arch. Biochem. Biophys. 130, 337 - 346. 9. Gray, W. R. (1972) Methods Enzymol. 25B, 121- 143. 10. Ramachandran, L. K. & Witkop, B. (1967) Methods Enzymol. 11,283-299. 11. Maizel, J. V. (1971) in Methods in Virology (Maramorosch, K. & Koprowski, H., eds) vol. 5 , pp. 179-246, Academic Press, New York. 12. Johns, E. W. (1967) Biochem. J . 105, 611-615. 13. Crane-Robinson, C., Hayashi, H., Cary, P. D. (1977) Eur. J . Biochem. in the press. 14. Wadell, W. J. (1956) J . Lab. Clin. Med. 48, 311-314. 15. Tombs, M. P., Soutar, F. & Maclagan, N. F. (1959) Biochem. J . 73.167-171. 16. D’Anna, J. A., Jr & Isenberg, I. (1974) Biochemistry, 13,4993 4997.
L. Bohm, Department of Biochemistry, University of Cape Town, Private Bag, Rondebosch, Cape Town, South Africa H. Hayashi, P. D. Cary, T. Moss, C. Crane-Robinson *, and E. M. Bradbury, Biophysics Laboratories, Portsmouth Polytechnic, St Michael’s Building, White Swan Road, Portsmouth, Hampshire, Great Britain, PO1 2DT
*
To whom correspondence should be addressed.
493
Sites of Histone/Histone Interaction in the H3 - H4 Complex Lothar BOHM Biochemistry Department, University of Cape Town Hiroaki HAYASHI, Peter D. CARY, Tom MOSS, Colyn CRANE-ROBINSON, and E. Morton BRADBURY Biophysics Laboratories, Portsmouth Polytechnic (Received March 18, 1977)
Sites of interaction between histones H3 and H4 have been probed by investigating complex formation, firstly between histone H4 and three peptides cleaved by chemical means from histone H3 (residues 1 - 90 and 1- 120 using cyanogen bromide and residues 42- 135 using N-bromosuccinimide), secondly between histone H3 and two peptides cleaved from histone H4 (residues 1- 84 using cyanogen bromide and residues 38 - 102 using chymotrypsin) and thirdly between the H4 peptide (residues 38 - 102) and the three H3 peptides (residues 1- 90, 1- 120 and 42- 135). The criterion for complex formation is the appearance of characteristic perturbed resonances in the aromatic region of the 270-MHz proton resonance spectrum of the peptide mixture. It is concluded that loss of 37 N-terminal residues from histone H4 and 41 N-terminal residues from histone H3 does not prevent complex formation, whilst the loss of 18 C-terminal residues from H4 and 45 C-terminal residues from H3 does prevent it; the last 15 C-terminal residues of H3 are, however, not required for forming a complex. The regions important for complex formation are therefore defined as residues 42- 120 in histone H3 and residues 38 - 102 in histone H4. In a physical study of the tetrameric state of the histones H3 and H4, it was shown [hat the complex between them is characterised by discrete perturbations in the proton magnetic resonance spectrum [l]. Such effects are a direct consequence of the tertiary structure of the complex and are always observed with globular proteins. The frictional ratio of the H3 . H4 complex (1.99 in 50 mM acetate/bisulphite pH 5, 1.7 in 25 mM phosphate pH 7 [l, and T. Moss, unpublished data]) is, however, much above that expected for a fully compact globular protein. It was concluded [l]that whilst the complex has elements of tertiary structure, it is not fully globular and has free unstructured sections of polypeptide chain, probably the N-terminal. If certain regions of the H4 and H3 chains are not included in the globular structure, then it should be possible to form the complex with incomplete polypeptide chains. This communication describes the results of NMR investigations of complex formation (a) between histone H3 and two H4 peptides H4(38 102) and H4(1-84), (b) between histone H4 and Abbreviation. NMR, nuclear magnetic resonance, Nomenclature. Histones are named according to international agreement, see Ciba Found. Symp. 28 (1975). Peptide fragments of the histones are indicated by the histone abbreviation followed by residue numbers in parentheses; e.g. H3(1-90) represents residues 1 - 90 of histone H3.
three H3 peptides H3(1- 90), H3(1- 120) and H3(42135) and (c) between peptide Hl(38- 102) and three peptides of H3, viz: (1 -90), (1 - 120) and (42- 135). From the results, regions of the two chains are defined that are not essential for complex formation and, conversely, other regions are found to be essential for complex formation. The criterion used for complex formation is the appearance of the characteristic perturbed peaks in the aromatic region of the NMR spectrum. This criterion has the advantage that it indicates the presence of the tertiary structure specific for the H3 . H4 complex, i.e. if complexes form with conformational characteristics other than those of the native H3 . H4, they will be detected as a different spectrum. It is important also that the method be able to select between cross-complex formation and self-aggregated histone. This is possible since the latter gives rise to very broad peaks that are essentially invisible in a highresolution NMR spectrum. MATERIALS AND METHODS Isolation and Purification of Chicken Erythrocyte Histone H3 Dimer
Whole chicken erythrocyte histone was isolated and fractionated as previously described [3]. The H3
488
Sites of Histone/Histone Interaction in the H3 . H4 Complex
Table 1. Amino acid composition of peptides f r o m chicken erythrocyte histone H3 and calf thymus histone H4 Amino acid
H3 (1 - 90) calc.
obs.
H3(1-120)
H3(42-135)
H4(38- 102)
H4(1-84)
calc.
obs.
calc.
obs.
calc.
obs.
calc.
obs.
12 2 14 4 10 6 13 5 6 16 1 6 2 5 11 3 4
9.1 2.5 13.8 4.6 9.0 4.4 13.5 5.7 6.0 15.9
5 1 13 5 5 4 13 3 3 10 1 5 2 7 11 2 4
5.4
5 1 8 3 6 1 5 0 8 5 0 8 1 3 5 4 2
3.9 1.3 8.7 3.3 5.4 2.2 4.7 0.2 7.6 5.5 7.6 1.0 3.4 5.1 3.7 1.4
10 2 12 4 6 2 5 1 13 6 0 7
8.75 1.8 12.6 4.0 5.0 2.0 5.0 1.5 12.4 6.6
residues/molecule Lysine Histidine Arginine Aspartic acid Threonine a Serine" Glutamic acid Proline Glycine A 1anine Cystine Valine Methionine Isoleucine Leucine Tyrosine" Phenylalanine
11 1 13 2 8 5 9 5 5 11 0 4 1 3 7 2 3
N-terminal amino acid
Alanine
Response to iodosobenzoate - ive to form S-S dimer
a
10.4 1.0 12.6 2.0 6.8 4.1 10.0 4.4 4.6 11.9 -
3.9 b
2.9 7.5 1.9 2.6
d
6.1 b 5.2 11.6 2.2 4.0
Alanine
+ ive
12.5 5.3 4.0 3.3 13.5 3.1 3.8 11.5 4.2 5.4 10.0 3.7
O
6 6 2 1
~
7.6 b
5.3 6.4 2.6 1.2
Arginine
~
ive'
No correction for hydrolytic losses. Homoserine lactone present. Oxidised by treatment with N-bromosuccinimide. Oxidised by CNBr treatment. Detected as spirolactone.
fraction was used directly for dimerization as given below. Alternatively, arginine-rich histones extracted from chicken erythrocyte deoxyribonucleoprotein with ethanol/HCl according to the procedure of Johns [4] were oxidized and the dimer obtained by fractionation. In a typical experiment 348 mg of a mixture of histones H3, H2A and H4 were dissolved in 12 ml 8 M urea pH 2 (HCl), gently stirred and disaggregated for 30 min. The pH was adjusted to 7.4 using NaOH and 10 mlO.5 M Tris. 5 mg iodosobenzoic acid (Pierce Chemical) neutralized with NaOH to give 1 ml were added and the solution was gently stirred for 12 h at 20 "C. The pH was readjusted to 7.4 and 7 mg iodosobenzoic acid prepared as above were added. The oxidation was continued for another 6 h. The dialyzed and freeze-dried residue (312 mg) was taken up in 8 M urea pH 2 (HCl) and fractionated on a 5 x 100-cm Sephadex G-100 column in 0.02 M HCl as previously described [5]. The HCl eluant from peak TI was freeze-dried twice, redissolved in water and freeze-dried to give 86 mg electrophoretically pure H3-S-S-H3 dimer. By the same procedure 60mg pure histone H3 prepared by pH-5.1 fractionation of whole histone [3] yielded 47 mg H3-S-S-H3 dimer.
Cyanogen Bromide Cleavage of Histone H3 Cyanogen bromide cleavage was essentially according to the method of Givol and Porter [6] and was performed on the H3-S-S-H3 dimer. In a typical experiment 106 mg dimer in 20 ml 70 % (v/v) formic acid and 70 mg CNBr were kept at room temperature under gentle agitation in a closed counting vial with a minimum of air space. After 5 and 20 h 70 mg CNBr were again added giving a total of 210 mg CNBr (147-fold molar excess). After 30 h 10 vol. of water were added. The freeze-dried residue was taken up in 8 M urea pH 2 (HCl) and fractionated on a 5 x 100-cm Sephadex G-100 column to give 34 mg, 9 mg and 6 mg peptide in peaks 3, 4 and 5 respectively (Fig. 1A solid line). These peptides correspond to the fragments H3(1-90), H3(91- 121)2 dimer and H3(121-135) previously identified by sequence analysis [7]. On refractionation on Sephadex G-100, 34 mg crude peptide (peak 3, Fig. 1A) yielded 25 mg H3(1-90) peptide which was electrophoretically homogeneous on polyacrylamide gels [8], by dansylation [9] gave alanine as the N-terminal amino acid which had an amino acid composition as indicated (Table 1).
L. Bohm, H. Hayashi, P. D. Cary, T. Moss, C. Crane-Robinson, and E. M. Bradbury
489
G-100 column to give intact H3 monomer followed by an H3 peptide (peaks 1 and 2 respectively, Fig. 1B). By dansylation the N-terminal amino acid of this peptide preparation was found to be alanine. Its electrophoretic mobility on acetic acid/urea/polyacrylamide gels [8] corresponded with the expected molecular weight (120 residues) and was higher than the mobility of histone H2A (129 residues) but lower than the mobility of histone H4 (102 residues). After oxidation with iodosobenzoate the electrophoretic mobility of this peptide was found to be considerably lower than in the presence of mercaptoethanol and was similar to the mobility of histone H1. From the specificity of cleavage, elution from Sephadex, electrophoretic mobility, response to oxidation, end-group analysis and amino acid composition (Table 1) this peptide was identified as the H3(1- 120) fragment. N-Bromosuccinimide Cleavage of Histone H3
Eluate volume (ml)
Fig. 1. Elution projilox of' SepIzu~Io.~ G I 0 0 chromatography of ( A ) CNBr total cleavage product, (B) rechromatography of peaks I + 2 of ( A ) and ( C ) N-bromosuccinimide totul cleavage product. Column sizes: (A) 5 x 100 cm, (B) 2.5 x 90 cm, (C) 5 x 100 cm
From the chromatogram (Fig. 1A, solid line) and the peptide yields it was obvious that CNBr cleavage was incomplete and that the fraction eluting in peak 2 (Fig. 1A) may represent an H3 fragment exceeding 90 residues. The apparent size on Sephadex and the specificity of cleavage suggested that this may be the H3(1 - 120) fragment arising from preferential cleavage cleavage at methionine-120. To favour incomplete cleavage, the amount of CNBr was reduced to a total of 100 mg (70-fold molar excess on 106 mg H3-S-S-H3 dimer) which were added in two equal portions at 0 and 5 h. The cleavage was terminated after 10 h and the freeze-dried residue was fractionated as described above (Fig. 1A, broken line). The eluant from peaks 1 and 2 (Fig. 1A) was pooled, dialyzed, freeze-dried, reduced in 8 M urea/l% 2-mercaptoethanol and refractionated on a 2.5 x 90-cm Sephadex
N-Bromosuccinimide cleavage was done according to the method of Ramachandran and Witkop [lo]. H3-S-S-H3 dimer was used as the substrate. In a typical experiment 140 mg dimer were dissolved in 50 ml 50% (vjv) acetic acid and titrated with 200-pl aliquots of N-bromosuccinimide reagent (96 mg Nbromosuccinimide recrystallized from glacial acetic acid dissolved in 5 ml 50 % vjv acetic acid). After each addition of reagent the ultraviolet spectrum was recorded against a blank containing identical aliquots of reagent in 50 ml 50 % (vjv) acetic acid. The total consumption was 2.3 ml. Cleavage was accompanied by a blue shift of the spectrum from 280-260 nm. The end point was taken when a 100-p1 aliquot of reagent caused no further increase of the absorbance at 260 nm and the solution remained yellow. After addition of 5 vol. of water the solution was freezedried. The residue was taken up in 6 M urea pH 2 (HC1) and fractionated on a 5 x 100-cm Sephadex G-100 column in 0.02 M HCl. Repeated freeze-drying of the HCl eluant yielded 8,13,88 and 9 mg of peptide from peaks 1, 2, 3 and 4 respectively (Fig. 1C). Peak 1 after refractionation on Sephadex G-300 in 0.02 M HCl was electrophoretically homogeneous in acetic acid/urea/polyacrylamide gels [8] and in sodium dodecylsulfate/polyacrylamide gels [l 11. In both systems the electrophoretic mobility of the peptide was higher than that of intact histones H3 and H2A but lower than that of histone H4. By dansylation arginine was detected as the N-terminal amino acid, glutamine was absent. Sephadex elution and electrophoretic mobility in relation to intact histone H3, specificity of cleavage and end-group analysis suggested a large peptide lacking part of the N-terminal chain. By these criteria and from the amino acid composition (Table 1) the peptide was identified as the H3(42- 135) fragment.
490
Sites of Histone/Histone Interaction in the H 3 . H4 Complex
Calf Thymus Histone H4 This was prepared according to the method of Johns [12] and purified by exclusion chromatography on Biogel P10, eluting with HCl, pH 2.
Cyanogen Bromide Cleavage of Histone H4 H 3 ( 1- 135)
The peptide H4( 1 - 84) was prepared by essentially the same method as used for the two H3 peptides. The details of preparation and purification are given elsewhere 1131. The product was electrophoretically pure on urea/acetic acid gels 181. Table 1 gives the amino acid composition of the sample used.
t i 4 ( 1 -102)
H 3 ( 1- 120)
H 4 ( 1-102)
Chymotrypsin Cleavage of Histone H4 The bond between leucine-37 and alanine-38 is a primary cleavage point for chymotrypsin in the H4 molecule. From a limited chymotryptic digest, therefore, it is possible to obtain the peptide H4(38 - 102). The details of preparation and purification are given elsewhere 1131. The product was electrophoretically pure on urea/acetic acid gels [8]. Table 1 gives the amino acid composition of the peptide sample used.
Preparation of Complexes Stock solutions of all the peptides and intact histones were made up in H 2 0 , pH 3 at about 10 mg/ ml. The precise concentration was obtained by measuring the absorbance of the solutions at 280 nm using an absorption coefficient of 1400 cm-l (mol tyrosine)-'. The concentration of the solution containing the peptide H3(42- 135) was obtained by measuring the peptide absorbance and using the methods described elsewhere [14,15]. The solutions were then made 8M in urea by adding solid urea and the final volume noted. Dithiothreitol was added next, to a concentration of 1 mg/ml. Equimolar mixtures of the peptides and intact histones (and the three peptide mixtures) were then made up and allowed to stand at 4 "C overnight. Afterwards the mixtures were dialysed against 'HzO, pH 6.5, 0.1 mg/ml dithiothreitol, with several changes of dialysate for at least 48 h. The final dialysate contained no dithiothreitol. The absorbance of the solutions at 280 nm was again measured to check that no considerable loss of peptide or histone had occurred, making a correction for volume increase during dialysis.
Nuclear Magnetic Resonance Spectra of the peptide mixtures were obtained in 2 H 2 0 at pH z 6 in 5-mm tubes using a Bruker 270-MHz Fourier transform spectrometer. The free induction decay of 5 x lo3 or lo4 pulses per spectrum was multiplied by a negative exponential (equivalent
Fig.2. 270-MHz N M R spectra of the aromatic resonances of ( A ) a control H3 . H4 complex and ( B - D ) histone H4 mixed with three peptides cleaved from histone H3. Peaks 1 and 2 are perturbed resonances characteristic of the native complex. Peak 0 is at the shift of unperturbed tyrosine resonance (two protons ortho to -OH)
to 2-Hz line broadening) to improve the signal-tonoise ratio. RESULTS AND DISCUSSION Fig.2 shows the aromatic spectrum of a control sample of intact calf thymus histone H4 mixed in equimolar ratio with intact chicken erythrocyte histone H3 and with three peptides cleaved from the H3 molecule. The upfield spectra are not shown because the ring-current shifted peaks observed in the native H3 . H4 complex are very broad and difficult to observe and therefore not easily used as a criterion of complex formation. The perturbed aromatic resonances characteristic of the complex (see [l]) are indicated by vertical lines in Fig.2-4 and can probably be assigned to tyrosine residues. Fig.2 shows that peptides H3(42- 135) and H3(1- 120) both interact with intact histone H4 to give a spectrum
491
L. Bohm, H. Hayashi, P. D. Cary, T. Moss, C. Crane-Robinson, and E. M. Bradbury
01 2
11 I
H3( 1-135) H4(38-102)
I I I I
1 9
I 8
I 7
H3( 1- 135) H4(1-84)
I 6
I
I
5
9
I 8
1 7
1 6
I 5
6 (PPm) Fig.3. 270-MHz N M R spectra oJ the aromatic resonances of ( A ) a control H 3 . H4 complex and ( B , C ) histone H3 mixed with two peptides cleaved f r o m histone H4
6 (PPm) Fig.4. 270-MHz N M R spectra of the aromatic resonances of ( A ) a control H3 . H4 complex and ( B - D ) mixtures of peptide H4(38- 102) with three peptides cleaved from histone H3
like that of the control. Peptide H3( 1- 90), however, shows no spectral perturbations when mixed with intact histone H4 and gives the sharp simple spectrum expected of a random coil protein mixture. [The very weak resonance at peak positions 1 and 2 is probably due to impurities of intact H3 or H3(1- 120)remaining in the sample of H3(1- 90).] This type of spectrum is the result of equivalence among all residues of a given type in the chain i.e. tyrosine, phenylalanine and histidine, in this part of the spectrum. No specific interaction takes place therefore between histone H4 and peptide H3(1-90). When the mixtures H4/H3(42135), H4/H3(1- 120) and the control H4/H3 mixture were made up to 3 mM phosphate, pH 7, there was little effect on the spectrum. The addition of phosphate to the H4/H3(1- 90) mixture resulted in considerable line broadening and a reduction of apparant peak area by more than 70%. This is a clear indication
of the formation of high-molecular-weight aggregates in this mixture, a characteristic of the separate histones, and is further evidence that the peptide H3(190) does not interact specifically with histone H4. Fig.3 shows the aromatic spectrum of equimolar mixtures of intact histone H3 with the two peptides cleaved from H4, H4(1- 84) and H4(38 - 102), together with a control H3 . H4 complex. The spectra show that the C-terminal peptide H4(38 - 102) gives rise to the perturbed resonances of the control whilst the N-terminal peptide H4(1- 84) does not. The spectrum of mixture H3/H4(1- 84) does show some broad resonance in the region 6.5-6.7 ppm and a very weak component at the position of peak 1. This is probably due to self-aggregation and to a small amount of uncleaved histone H4 remaining in the sample of peptide H4(1- 84). It is clear, however, that peptide H4(1- 84) does not form a complex with histone H3
492
Sites of Histone/Histone Interaction in the H3 . I14 Complex 33
1
H3
I
(INTACT)
1
H4
. 1
H3
t
‘1
/
H4
PiPTIDE
CROSS-INTNIUCTIONS
135
H3 (IFTACT )
1
1
I
1
H4
120
1
I
I
H3
102
38
102
102
135
1 102
42
38 I
CLOAVACES
H4
N-TSRMI NAL
1
I
38
H4
I 1 2
CLSAVA CES
120
Hj
I
13 5
H3
t
C-TERMINAL
I J
7
I
H4
135
I
102
N and C-ThRMIVAL CLBAVAGB
Fig. 5 . Cross-interactions betweenpeptides of histones H3 andH4. (-)
under these conditions. As in the first experiment the two peptide-containing mixtures and the control were then made 3 mM in phosphate, pH 7. The control H3/H4 and the H3/H4(38 - 102) mixtures showed essentially no change, but very considerable line broadening and apparant loss of peak area was observed with the H3/H4(1- 84) mixture. We therefore conclude that the peptide H4(38 - 102), but not H4(1- 84) interacts specifically with histone H3. To substantiate the results obtained on the first two sets of complexes, a further three equimolar peptide mixtures were studied, each containing the peptide H4(38 - 102) (previously found to interact with histone H3) together with the three peptides from histone H3, viz: H3(1-90), H3(42- 135) and H3(1- 120). Fig.4 shows the aromatic region of the spectrum and it can be seen that the characteristic perturbed peaks are observed in the two mixtures containing the peptides H3(42- 135) and H3(1- 120), but not in that containing H3(1- 90). The spectra obtained with complexes formed by the peptide H3(42 - 135) are of particular interest since this N-bromosuccinimide peptide does not contain the aromatic side chains of the remaining two tyrosine residues (positions 54 and 100) which were destroyed in the cleavage reaction. (This was checked by noting the presence of only phenylalanine resonances in the spectrum of the peptide in the random coil form at pH 3.) Fig. 2 shows the spectrum of peptide H3(42 135) interacting with intact histone H4 and Fig.4
Complex formation observed; (-----)complex formation not observed
that with peptide H4(38 - 102). The spectrum between 6 and 7 ppm is similar for the two complexes and it follows that all of the peaks in this region are due to the four tyrosines of histone H4 (provided, of course, that no phenylalanine resonance is perturbed that far upfield). The appearance of the spectrum suggests that peaks 2 and 1 are each due to a single tyrosine (two protons each) and peak 0 to two unperturbed tyrosines (four protons). Although it is not possible at present to make assignments of peaks 0, 1 and 2 to particular residues in histone H4, the use of modification procedures such as nitration or iodination should permit this to be done in the future. The data on complex formation are summarised in Fig. 5 in which a solid line indicates cross-complex formation. The major conclusion that follows from the data is that it is the C-terminal rather than the N-terminal portions of the H3 and H4 molecules that are the most important for complex formation. Firstly, it is noted that intact histone H3, and also its two peptides H3(1- 120) and H3(42- 135), interacts not only with intact histone H4 but also with its peptide H4(38102). It is clear therefore that residues 1- 37 of histone H4 (36 % of the molecule) are not essential for complex formation. We have previously proposed that the N-terminal region of histone H4 is not involved in the tertiary structure of the complex [l] and have also shown that there is no secondary structure in this portion of the H4 chain [13]. In a parallel way, the
L. Bohm, H. Hayashi, P. D. Cary, T. Moss, C. Crane-Robinson, and E. M. Bradbury
fact that intact histone H4 interacts with the peptide H3 (42- 135), shows that the N-terminal 41 residues of histone H3 (30% of the molecule) are likewise not essential for complex formation. The importance of the C-terminal portions of both chains is seen from the observation that neither H3(1- 90) (67 % of the H3 molecule) nor H4(1- 84) (82 % of the H4 molecule) are able to form cross-complexes with their intact histone partners. This demonstrates the critical importance of the C-terminal 19 residues of histone H4 and 45 residues of histone H3. In the case of histone H3, however, it is found that its peptide H3(1- 120) does form a complex with both histone H4 and peptide H4(48 - 102), indicating that for histone H3 the most critical region for interaction does not include the final 15 residues. It is concluded therefore, on the basis of the peptides studied here, that the regions of the two chains that are critical for complex formation are residues 42 - 120 for histone H3 and residues 38 - 102 for histone H4. These may not be the absolute minimum chain lengths required for complex formation but other peptides would be required to define this even more precisely. It is perhaps surprising that folding and complex formation can occur between molecules that have lost 36 % (H4) and 30 % (H3) of their N-terminal amino acids. This must be due both to the extremely high association constant between histones H3 and H4 (estimated at z 1021 [16]) and also to the fact that a good proportion of their residues are conformationally external to the tertiary complex. The present data demonstrate that the external residues are mainly N-terminal. Is it possible to define precisely the boundary between the two structural domains? Without a more detailed probe of the structure of the whole of the complex, one cannot be certain that the N-terminal sections missing in the minimal complex H4(38 - 102) . H3(42- 135) are all
external to the tertiary structure. The NMR spectral perturbations are a reflection of the integrity of the hydrophobic cores of the tertiary structure and are probably not dependent on the presence or absence of several residues on the perifery of the globular structure. A very high association constant between the folded C-terminal portions of histones H3 and H4 could result in complex formation when several residues are missing and thus 40 or so residues is probably an overestimate of the size of the free random coil Nterminal chains in the native complex.
REFERENCES 1. Moss, T., Cary, P. D., Crane-Robinson, C. & Bradbury, E. M. (1976) Biochemistry, 15, 2261 -2267. 2. Reference deleted. 3. van der Westhuyzen, D. R., Bohm, E. L. & von Holt, C. (1974) Biochim. Biophys. Acta, 359, 341 - 345. 4. Johns, E. W. (1964) Biochem. J . 92, 55-59. 5. Brandt, W. F. & von Holt, C. (1971) FEBS Lett. 14, 338-342. 6. Givol, D. & Porter, R. R. (1965) Biochem. J. 97, 32C- 34C. 7. Brandt, W. F. & von Holt, C. (1974) Eur. J . Biochem. 46,419429. 8. Panyim, S. & Chalkley, R. (1969) Arch. Biochem. Biophys. 130, 337 - 346. 9. Gray, W. R. (1972) Methods Enzymol. 25B, 121- 143. 10. Ramachandran, L. K. & Witkop, B. (1967) Methods Enzymol. 11,283-299. 11. Maizel, J. V. (1971) in Methods in Virology (Maramorosch, K. & Koprowski, H., eds) vol. 5 , pp. 179-246, Academic Press, New York. 12. Johns, E. W. (1967) Biochem. J . 105, 611-615. 13. Crane-Robinson, C., Hayashi, H., Cary, P. D. (1977) Eur. J . Biochem. in the press. 14. Wadell, W. J. (1956) J . Lab. Clin. Med. 48, 311-314. 15. Tombs, M. P., Soutar, F. & Maclagan, N. F. (1959) Biochem. J . 73.167-171. 16. D’Anna, J. A., Jr & Isenberg, I. (1974) Biochemistry, 13,4993 4997.
L. Bohm, Department of Biochemistry, University of Cape Town, Private Bag, Rondebosch, Cape Town, South Africa H. Hayashi, P. D. Cary, T. Moss, C. Crane-Robinson *, and E. M. Bradbury, Biophysics Laboratories, Portsmouth Polytechnic, St Michael’s Building, White Swan Road, Portsmouth, Hampshire, Great Britain, PO1 2DT
*
To whom correspondence should be addressed.
493
