10
Eur. J. Biochem. 67, 379-3388 (1976) Structural Studies of Chicken Erythrocyte Histone H 5 Colyn CRANE-ROBINSON, Shirley E. DANBY, E. Morton BRADBURY, Annie GAREL, Anne-Marie KOVACS, Madeleine CHAMPAGNE, and Michel DAUNE Biophysics Laboratory, Portsmouth Polytechnic, and Institut de Biologie Moleculaire et Cellulaire du Centre National de la Recherche Scientifique, Strasbourg (Received April 13/June 1, 1976) Spectroscopic studies (nuclear magnetic resonance, circular dichroism and infrared) have been carried out on chicken erythrocyte histone H5 and on three peptides cleaved therefrom: 1-31, 32-197 and 58-197. It is shown that at ionic strengths above 0.1M part of the H5 molecule takes up a globular conformation containing 14% a helix but no p sheet structure. Several details of the circular dichroism and nuclear magnetic resonance spectra indicate that the globular region is located in the N-terminal half of the molecule and this proposal is supported by the observation that the peptide 32- 197 is largely incapable of folding and the peptide 59- 197 is completely in- capable of folding. Structural similarities and differences between histone H 5 and histone H 1 are discussed. The nucleated erythrocytes of birds, reptiles, am- phibians and fish contain a very basic histone, H5, that has not been found in other tissues [l - 51. H5 largely, but not completely, replaces H 1 and the two histones are not dissimilar in amino acid composition and molecular weight [4,6 - 81. A significant difference between the two histones is that H 5 contains 11 % arginine whilst H1 contains only 2%. In this respect H 5 resembles the 4 1 histones of marine invertebrate sperm (in particular sea urchins) which totally replace the HI and also contain about 11% arginine in addition to roughly 25% lysine 191. Early observations of the occurrence and composition of H5 resulted in the suggestion that its function is the total suppression of the genome in the fully mature erythrocyte [lo]. Detailed studies on anemic chickens have subsequently shown, however, that H 5 can be detected even in erythroblasts (when a wide range of protein synthesis is taking place), and does not appear suddenly at the mature erythrocyte stage when protein synthesis is finally terminated [I 1 - 151. The function of H 5 in re- lation to genome suppression is, therefore, not as simple as originally thought. H 5 nevertheless is closely related to H I [I61 and a number of experi- ments have implicated H1 in the condensation of diffuse chromatin [17 - 191. In particular the state of chromatin is thought to depend on the degree of phosphorylation of H 1 [20,21] and recently it has been shown that H 5 is phosphorylated and dephosphorylat- ed in viuo [22]. There is at present no evidence to Abbreviations. NMR, nuclear magnetic resonance; CD, circular dichroism. suggest why H1 should be replaced in large part by H 5 and the present structural investigation of H 5 was carried out to point out the similarities and differences between H 5 and H 3. Related studies on H 1 have al- ready been published [23,24]. H 5 contains about 197 residues [25] and H 1 about 216 [26]. The sequence of H 5 has been determined up to residue 111 [27] and when this is compared to the sequence of rabbit thymus RTL 3 H1 [26] (and D. Cole, private communication) striking analogies are observed in the central and fairly hydrophobic sequences between H 1 Tyr-72 (58 in H5) and H 1 Phe-106 (93 in H5). Both molecules have C-terminal halves that are very rich in basic groups and alanine, but H 1 differs from H 5 in that the N-terminal 35 re- sidues are also very basic and contain a high propor- tion of proline; the N-terminal region of H5 is not particularly hydrophilic. Hydrodynamic studies have shown that H 5 does not aggregate at high ionic strength [8,28] and thereby resembles H1 and not the remaining four histone fractions. The monomeric state of H 1 has led to pro- posals that its function is quite separate from that of the other histones and this may also be true for H5. In chicken erythrocytes the stoichiometry of histones H 5 and H 1 together, with respect to the other four histones, appears to be similar to that of H1 alone in calf thymus [7,29]. A number of spectroscopic techniques are applied here to establish the presence and nature of secondary and tertiary structure in H 5 and preliminary attempts to determine the position of the structured region have

Structural Studies of Chicken Erythrocyte Histone H5

Embed Size (px)

Citation preview

Page 1: Structural Studies of Chicken Erythrocyte Histone H5

Eur. J . Biochem. 67, 379-3388 (1976)

Structural Studies of Chicken Erythrocyte Histone H 5 Colyn CRANE-ROBINSON, Shirley E. DANBY, E. Morton BRADBURY, Annie GAREL, Anne-Marie KOVACS, Madeleine CHAMPAGNE, and Michel D A U N E

Biophysics Laboratory, Portsmouth Polytechnic, and Institut de Biologie Moleculaire et Cellulaire du Centre National de la Recherche Scientifique, Strasbourg

(Received April 13/June 1, 1976)

Spectroscopic studies (nuclear magnetic resonance, circular dichroism and infrared) have been carried out on chicken erythrocyte histone H 5 and on three peptides cleaved therefrom: 1-31, 32-197 and 58-197. It is shown that at ionic strengths above 0.1M part of the H 5 molecule takes up a globular conformation containing 14% a helix but no p sheet structure. Several details of the circular dichroism and nuclear magnetic resonance spectra indicate that the globular region is located in the N-terminal half of the molecule and this proposal is supported by the observation that the peptide 32- 197 is largely incapable of folding and the peptide 59- 197 is completely in- capable of folding. Structural similarities and differences between histone H 5 and histone H 1 are discussed.

The nucleated erythrocytes of birds, reptiles, am- phibians and fish contain a very basic histone, H5, that has not been found in other tissues [l - 51. H 5 largely, but not completely, replaces H 1 and the two histones are not dissimilar in amino acid composition and molecular weight [4,6 - 81. A significant difference between the two histones is that H 5 contains 11 % arginine whilst H1 contains only 2%. In this respect H 5 resembles the 4 1 histones of marine invertebrate sperm (in particular sea urchins) which totally replace the H I and also contain about 11% arginine in addition to roughly 25% lysine 191. Early observations of the occurrence and composition of H 5 resulted in the suggestion that its function is the total suppression of the genome in the fully mature erythrocyte [lo]. Detailed studies on anemic chickens have subsequently shown, however, that H 5 can be detected even in erythroblasts (when a wide range of protein synthesis is taking place), and does not appear suddenly at the mature erythrocyte stage when protein synthesis is finally terminated [I 1 - 151. The function of H 5 in re- lation to genome suppression is, therefore, not as simple as originally thought. H 5 nevertheless is closely related to H I [I61 and a number of experi- ments have implicated H1 in the condensation of diffuse chromatin [17 - 191. In particular the state of chromatin is thought to depend on the degree of phosphorylation of H 1 [20,21] and recently it has been shown that H 5 is phosphorylated and dephosphorylat- ed in viuo [22]. There is at present no evidence to Abbreviations. NMR, nuclear magnetic resonance; CD, circular dichroism.

suggest why H1 should be replaced in large part by H 5 and the present structural investigation of H 5 was carried out to point out the similarities and differences between H 5 and H 3 . Related studies on H 1 have al- ready been published [23,24].

H 5 contains about 197 residues [25] and H 1 about 216 [26]. The sequence of H 5 has been determined up to residue 11 1 [27] and when this is compared to the sequence of rabbit thymus RTL 3 H1 [26] (and D. Cole, private communication) striking analogies are observed in the central and fairly hydrophobic sequences between H 1 Tyr-72 (58 in H5) and H 1 Phe-106 (93 in H5). Both molecules have C-terminal halves that are very rich in basic groups and alanine, but H 1 differs from H 5 in that the N-terminal 35 re- sidues are also very basic and contain a high propor- tion of proline; the N-terminal region of H 5 is not particularly hydrophilic.

Hydrodynamic studies have shown that H 5 does not aggregate at high ionic strength [8,28] and thereby resembles H1 and not the remaining four histone fractions. The monomeric state of H 1 has led to pro- posals that its function is quite separate from that of the other histones and this may also be true for H5. In chicken erythrocytes the stoichiometry of histones H 5 and H 1 together, with respect to the other four histones, appears to be similar to that of H1 alone in calf thymus [7,29].

A number of spectroscopic techniques are applied here to establish the presence and nature of secondary and tertiary structure in H 5 and preliminary attempts to determine the position of the structured region have

Page 2: Structural Studies of Chicken Erythrocyte Histone H5

380 Erythrocyte Histone H 5

been made by studying peptides cleaved from the H 5 molecule.

MATERIALS AND METHODS

Histone H 5 Preparation and Cleatluge

Chicken erythrocyte histone H 5, the two cyanogen bromide peptides 1 - 31 and 32 - 197 and the N-bro- mosuccinimide peptide 59 - 197 were prepared as described earlier [25]. Their purity was checked using acrylamide gel electrophoresis according to Leboy et al. [30] and by amino acid analysis.

Spectroscopic Measurements

NMR spectra were obtained on a Bruker WH 270-MHz Fourier transform spectrometer. Samples were dissolved in 99.8% 'H20 in 5-mm tubes. Reso- lution enhancement by convolution difference was ob- tained by the methods of Campbell et al. [31].

Circular dichroism spectra were obtained using a Jobin Yvon Mark 111 dicrograph for all measure- ments except those for the heat denaturation study which was carried out on a Roussel Jouan CD 185 dicrograph (see [32]).

RESULTS AND DISCUSSION

The Folding of' Histone H5 : Secondary Structure

In dilute solution (0.1 mg/ml) at pH 5.7, H 5 hi- stone is essentially in the form of a flexible random coil. This is seen, in part, from the CD spectrum (Fig.1) in which the ellipticities at the two negative extrema are [Q]222 = - 16000 and [8],98 = - 18000". The conformation of the H5 histone is found to be very sensitive indeed to the ionic strength of the so- lution ; thus even an increase in protein concentration from 0.1 mg/ml to 5 mg/ml results in an increase of the negative ellipticity at 222 nm from - 1600" to - 24000 due to the rise in counter-ion concentration to approximately 20 mM. Fig. 2 shows a complete plot of the change in secondary structure (measured from [O],,,) as the ionic strength is increased at pH 6-7 using potassium fluoride. Comparison of the order- ing effects of the chlorides and fluorides of sodium and potassium showed close similarities between the effects of different uni-univalent salts. A correlation was found, however, between the ordering effect and the nature of the anion, chloride being more effective than fluoride. This accords with their positions in the Hoffmeister series. The maximum value of the [Q]222

extremum at high ionic strength in NaCl is about - 5200". On the basis of a coil ellipticity of - 1000" (average for several histones and histone fragments in 8 M urea solution) and a helical ellipticity of

Wavelength (nrn) 190 200 210 2x1 230 240 250 260

Fig. 1. CD spectru of histone H 5 . (1 ) ~ H,O, 0.1 mg/m!, pH 5.7. (2) . . . ' . . H,O, 5 mg/ml, pH 5.2. ( 3 ) ---- H,0/1 M KF, 0.1 mg/ml, pH 7.4

- 30000" [33], this represents 14.5% helicity or 29 re- sidues. The ellipticities presented here are in agreement with those of Garel et al. [32] but not with those of Williams and Seligy 1341.

The CD spectrum of histone H 5 in 1 M KF (Fig. 1) does not show an increased negative ellipticity around 218 nm that would indicate the presence of ,b' sheet structures. The absence of structure has been con- firmed from the infrared spectra of H 5 histone at 50 mg/ml concentration in 1 M NaC1/2H20. The amide I band centred at 1640 cm-I showed no evi- dence of a component at 1610- 1620 cm-', character- istic of ,b' sheets.

Tertiary Structure offfistone H 5

The formation of secondary structure is accompa- nied by tertiary folding and this we have followed in two ways: by NMR and from the extrinsic Cotton effects of the aromatic residues at around 280 nm in the CD spectrum. Fig. 3 shows this region of the CD spectrum for a 5-mg/ml solution of H 5 under solution conditions of essentially maximal secondary structure and under two denaturing conditions. In 1 M KF, pH 7.4, strong positive tyrosine Cotton effects can be seen at 280 and 290 nm and a set of weak bands are detected between 260 and 270 nm that can be assigned

Page 3: Structural Studies of Chicken Erythrocyte Histone H5

C. Crane-Robinson, S. E. Danby, E. M. Bradbury, A. Garel, A:M. Kovdcs, M. Champagne, and M. Daune 381

0.1 1 .o 10 I (mM)

Pig. 2. Variation of'ell@ticity at 222 und 228 nrn with ionic strength usingpotassiumfluoride. Protein concentration 5 mg/ml (0, A) and 0.1 mg/ml (0). The contribution of the protein to the ionic strength was taken as 38 m M at 5 mg/ml and 0.36 mM at 0.1 mg/ml, pH 7.4. Path lengths: 0.1 mm for (0. 0) 1 cm for [O],,, (A)

1 I \

; ', 1 1 1 1 I ! I

Fig. 3. CD spectra of histone H S in the region of the aromatic residue extrinsic Cotton <fJects. Protein concentration 5 mg/ml, pH 7.4. Path length 1 cm. (I) ---- 1 M KF, 20 "C. (2) ...... 50 mM KF, 20 c. (3) 4 M guanidinium hydrochloride, 20°C

I M KF, 75',C and in water 20°C. (4)

to phenykdlanine. Since all these bands largely dis- appear in 4 M guanidinium hydrochloride or on heating, it is concluded that tertiary folding takes place in 1 M KF, pH 7.4 and that both tyrosine and phenyl- alanine residues are included in the fold. The three

tyrosines are a t positions 28, 53 and 58 and the single phenykdianine residue is at position 93 and these ob- servations therefore suggest that at least the N-ter- minal half of the molecule undergoes folding. Fig.2 includes a plot of the variation of [Q],,, with ionic strength using the same solutions as used for obtdin- ing [812,, data. It can be seen from Fig.2 that forma- tion of secondary structure ([Q],,,) begins at somewhat lower ionic strengths than does the formation of tertiary structure [Q],,, . This indicates rather low co- operativity in the protein folding. This point is in- vestigated further (see below) by denaturation studies.

High-resolution NMR can also indicate the for- mation of a tertiary fold by the presence of well- defined'resonances displaced from their normal che- mical shift by the anisotropy of a nearby residue such as an aromatic ring. Fig.4 shows the complete spec- trum of H5 and the presence of ring-current shifted methyl peaks between 0 and 1 ppm due to valine, leu- cine and isoleucine is immediately apparent. Likewise, the non-equivalence of the three tyrosine (N, 0 and P) and the three histidine residues (A, B and D) is ap- parent in the aromatic spectrum between 6 and 9 ppm. Despite these perturbations, which indicate tertiary folding, the majority of the resonances above 4 ppm (largely from lysine, argine, serine and alanine) show no such effects and the upfield spectrum looks like that of a random coil chain and not like that of a globular protein. It follows that a large section of the H 5 chain that is rich in these residues is not globular. Amino acid sequence and composition data suggest this to be the C-terminal half of the molecule and the situation is similar to that of H 1 in this respect. The aromatic residues are, in contrast, situated entirely in the N-terminal half of the molecule. Fig.5 shows the onset of tertiary structure with increasing ionic strength as revealed by changes in the aromatic spec- trum. There are three histidine residues, at positions

Page 4: Structural Studies of Chicken Erythrocyte Histone H5

382 Erythrocyte Histone H 5

Gain x 64

H ~ H O ju I . ! , , I I , I , , I ,

9 8 7 6 5 4 3 2 1 0 I 1

U 8 ' I

6 (PPm) Fig. 4. 270-MHz Fourier transform spectrum of H 5 , 50 mglrnl, p H 4, 1 M NaCI. 6000 pulses. Protein fully structured

/I His C4H is C f l His C

I 1 I I

9 8 7 6 6 (PPm)

Fig, 5. Sult-induced folding of histone H 5 . Aromatic region of 270- MHz spectrum. Protein concentration 10 mg/ml, pH 3.20000pulses per spectrum

25, 57 and 62 and in 2H20, pH 3, the C2H spectrum (at z 8.7 ppm) shows that even in the random coil state one residue is not precisely equivalent to the other two. All three tyrosine residues are essentially equiva- lent under these conditions (peaks at 6.8 and 7.1 ppm) and the phenylalanine protons are centred at 7.3 ppm. On salt addition all three histidines become non- equivalent: the C2H proton peaks are labelled A, B and D and the corresponding C4H proton peaks H, I and M. Non-equivalence is also seen for the three

tyrosine residues : the peaks from protons ortho to the OH group are labelled N, 0 and P in the 1 M NaCl spectrum. The remaining tyrosine protons are largely in peak K and peak J contains the five phenylalanine protons. It follows from these observations that the aromatic residues are largely included in the tertiary fold of H5, although it cannot be concluded unam- biguously that all are included since one histidine residue (B) and one tyrosine residue (N) show little perturbation of chemical shift. Lack of perturbation can mean that a residue remains essentially in its flexible coil state, or that it has no anisotropic neigh- bours in the fold. A changing chemical shift on fold- ing means that a residue changes its chemical (i.e. magnetic) environment. The presence of tertiary fold- ing in a protein can also be seen from the observation of ring-current-shifted peaks upfield of 1 ppm. H5 shows such perturbed peaks and they can be seen clearly in Fig. 4. Their presence likewise inplicates the aromatic residues in the tertiary folded region.

The titration of histidine residues can be followed in the NMR spectrum of a protein and provide infor- mation on the inclusion of such residues into a fold. Fig.6a shows a set of spectra obtained over the pH range 3 to 9.4 with an H 5 histone solution of 50 mg/ml in 1 M NaC1, i.e. with the protein in its folded con- formation. The upfield displacement with increasing pH of the C2H peaks A, B and D and the C4H peaks H, I and M can readily be seen and the pK values of the three residues determined to be 6.5, 6.5 and 7.6 respectively. Two residues (peaks A and B) therefore, have essentially normal pK values, whereas the histi- dine (D) that exhibits the greatest chemical shift per- turbation, also has an increased pK: this could be a consequence of the histidine residue taking part in an ionic linkage. This titration study also led to the conclusion that one of the ring-current-shifted peaks in the 0 - 1 ppm region is perturbed by a histidine residue. Fig. 6 b shows the changes with pH of this region of the spectrum: the peak labelled 5 is seen to move upfield with increasing pH and to be at its mid-

Page 5: Structural Studies of Chicken Erythrocyte Histone H5

C. Crane-Robinson, S. E. Danby, E. M. Bradbury, A. Garel, A.-M. Kovacs, M. Champagne, and M. Daune 383

b a

9:0 8.5 8.0 7.5 7.0 6.5 5 (PPm)

PH 9.4

8.9

7.9

7.65

7.1

6.7

6.6

6.4

6.2 5.6

5.1

3 .O

Y PH 8.9

7.9

7.1

6.7

6.6

Fig.6a. Low-field spectrum of histone H 5 (50 mglrnl) showing titration of the three histidine residues. Histidine C2H peaks A, B and D correspond to C,H peaks H, I and M. Peaks C, E and G are due to hard-to-exchange hydrogens that slowly disappear during the course of the experiment

Fig. 6b. Ring-current-shifted region of histone H5 showing the effects of p H change. All spectra have resolution enhancement by convolution difference

way position at about pH 6.6. This peak, the area of which corresponds to that of a single methyl group is therefore perturbed by one of the histidines that have been labelled A and B. None of the other ring-current- shifted peaks 1 to 6 is perturbed by a histidine residue and thus tyrosine or phenylalanine must be the per- turber. A change in the region of peak 5 is also noted between pH 3 and 5 and it follows that a carboxyl group is situated in the fold close to the apolar residue giving rise to peak 5.

Structure Prediction for Histone H 5

From the foregoing data it is clear that under physiological conditions of pH and ionic strength a large part of the H 5 molecule is in the form of a glo- bular fold that contains some helical structure. As

stated above, the general appearance of the upfield NMR spectrum suggests that not all the residues are included in this globular region. The sequence has so far been established up to residue 11 1 [35,25,16,27]. The composition and sequence of the N-terminal half of the molecule is quite compatible with the formation of globular structure, whereas the C-terminal half does not have the composition expected of a globular protein; in fact the protein appears to change its char- acter at about residue 100. On the N-terminal side of residue 100 the basic to acidic ratio is about 2, there are 37% hydrophobic residues (including alanine) and 4% proline. On the C-terminal side of residue 100, half of the residues are basic, there is only one acidic residue and of the 19 hydrophobic residues, 16 are alanine: furthermore 9% of the residues are proline. Thus the C-terminal half of H 5 is similar in composi-

Page 6: Structural Studies of Chicken Erythrocyte Histone H5

384 Erythrocyte Histone H 5

4-0- - - - - _ _ _ _ - - _ 20- .' T g K ~ E ~ A " R ~ ? ~ ~ A > ~ M ~ E t s BrCN SucNBr _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ 80

60 s S R Q s I Q K Y I K s H Y K v G H N ~ D J ~ Q J K,L s I R R L ~ L ~ A ~ A ~ G ~ ~ _ ~ ~ K _ Q T K G v

110 100

t t SucNEr SucNBr

... K A K K K G P S R K A K D S K A L R F S

Fig. 7. Amino acid sequence of diickcn ervrhroc.vte histone H S herwen residues I and I 1 1 1271 .rhowing krliccrl structure predicted hx the Chou and Fasmunprocedure i- - -) und by Ptitsyn etal. (-----). Cleavage points also indicated: CNBr = cyanogen bromide, SucNBr = N-bromosuccinimide. Three B-bends predicted by the Chou and Fasman procedure are also included

Wavelength (rim)

I I I

Fig.3. CD spectra ojhistone H S peptides, all at % 0.1 mglml concentrution. (A) H 5 (1 -31) in H,O, pH 3.5. and also in H,Oil M KF, pH7.4.(B)H5(32-197)inH20,pH 3.5(-)andinH20/1 MKF,pH7 ,4 ( . . . . . . ) . In t ac tH5 in 1 MKF,pH7.4(-----) .(C)H5(59-197) in H,O, pH 3.5 (-) and in H,0]1 M KF, pH 7.4 (......)

tion to the last 100 residues of H1, which are now known not to form a globular structure [36]. On the basis of composition and sequence data the most likely position of the globular region is therefore in the N-terminal half of the molecule. The NMR evidence so far presented supports the view that part of the molecule is globular, whilst part is present as a flexible random coil.

Fig. 7 shows the H 5 sequence up to residue 3 11 and also includes predictions of a helix and p turns by the procedures of Chou and Fasinan [37] and by the methods of Ptitsyn, Lim and Finkelstein (private com- munication). Predictions of p sheet structure have not been included since the infrared spectrum demon- strates its absence. However, regions predicted as p sheet by the Chou and Fasman procedure (having ( P o ) > (P,)) may nevertheless have a (P,) sufficient to qualify for a helix prediction if p sheet is excluded. This occurs at two points in histone H5. Residues 66-72 have ( P o ) = 1.19 > (P,) = 1.10, however, the sequence 64 - 70 has (P,) = 1.19 and so is shown

predicted as a. Residues 80-84 have (Pp) = 1.21 > (P,) = 1.11, but the sequence 75 - 83 has (P,) = 1.20 and is likewise shown predicted as a helix. The total number of a predicted residues is 34 according to Chou and Fasman and 27 according to Ptitsyn etal. The ob- served number is about 29 a residues and so experi- ment and prediction are in reasonable agreement de- spite half of the molecule not having been considered. This suggests that all the helix is indeed within the first 100 residues of histone H 5.

Studies on Peptides from Histone H5

An attempt has been made to test the proposition that the globular region is in the N-terminal half by the study of several peptides cleaved from histone H 5. Cyanogen bromide treatment cleaves the molecule at the single methionine-31, yielding the two peptides H 5 (1 - 31) and H 5 (32 - 197). N-Bromosuccinimide cleaves the molecule at the three tyrosine residues and the peptide H 5 (59 - 197) has been recovered. Fig. 8

Page 7: Structural Studies of Chicken Erythrocyte Histone H5

C. Crane-Robinson, S. E. Danby, E. M. Bradbury, A. Garel, A.-M. Kovacs, M. Champagne, and M. Daune 385

I

8 7

Fig.9. NMR spectra of hislone H 5 peptide (59- 197) in

I) 4 3 2 I 0

fi ( p m )

' H Z O and in 2 H 2 0 j l M NaCl, p H 7.4

shows the CD spectra of these three peptides in water at acid pH (denaturing conditions) and at neutral pH in 1 M potassium floride. The spectrum of fully structured intact H5 in 1 M KF, pH 7.4 is shown for comparison in Fig. 8 b. The spectra of all three pep- tides are essentially the same in salt-free solution at acid pH and very similar to that of intact H5 under the same conditions (see Fig. 1) : they are clearly ran- dom coil. Addition of I M KF at pH 7.4 leads to no change in the spectrum of H 5 (1 - 31) and only a very small change in that of H 5 (59- 197). Essentially no secondary structure is induced in these two peptides at high ionic strength. The peptide H5 (32-197) shows evidence (Fig. 8 b) for the formation of a small amount of secondary structure on salt addition : there is a marked reduction in the negative ellipticity at 197 nm and a small increase in the negative ellipticity at 222 nm to -2300". This represents of the order of 5% helix i. e. very much less than the 14.5% estimated for the intact molecule. Since the sum of the elliptici- ties for the two cyanogen bromide peptides does not equal that of the intact molecule it follows that residue 31 (the point of cleavage) is located within a region of H5 that is not in a random coil state in 1 M KF. I t might be argued from this observation that residue 31 is within a helical segment; however, reduced ab- ility of both cyanogen bromide peptides to form sec- ondary structure is almost certainly due to their in- ability to form the characteristic tertiary structure as a consequence of cleavage.

The conformation of the two C-terminal peptides has also been studied using NMR. Fig.9 shows the sDectrum of H5 (59-197) in water and 1 M NaC1.

pH 7.4. The peptide should contain only two aromatic residues, one phenylalanine and one histidine. How- ever, N-bromosuccinimide treatment is known to break the imidazole ring and so the low-field spectrum comes only from phenylalanine-93. Salt addition does not cause the formation of any perturbed resonances (as are observed in intact H 5) and it is concluded that no tertiary structure is induced by salt. A slight broadening of the phenylalanine resonance at 7.3 ppm and the apolar CH3 resonance at 0.9 ppm is noticed however, at high ionic strength: this is probably due to some non-specific aggregation taking place. Fig. 10 shows the low-field spectrum of the cyanogen bromide peptide H 5 (32- 197) in water and in 50 and 100 mM NaCI. At higher ionic strengths the solution becomes somewhat cloudy at 10 mg/ml protein and there was marked broadening of the NMR spectrum: this must be due to non-specific aggregation occurring. This peptide contains two histidines (57 and 62), one phe- nylalanine (93) and two tyrosines (53 and 58). Even in the random coil state the two histidine C2H re- sonances at 8.6 and 8.7 ppm (B and D, Fig. 10) are not equivalent; however, Fig. 5 shows clearly that one histidine C2H peak (D, chemical shift 8.6ppm) is shifted upfield from the other two even in the un- structured state of intact H5. The histidine C2H re-

. sonances labelled D in Fig. 5 and 10 are therefore probably from the same residue: the corresponding C4H peak is labelled M in both figures. The same labelling system is used in Fig. 6A and 11. The two tyrosine residues of the peptide are essentially equi- valent in water. On addition of 100 mM NaCl there is a slight upfield displacement of histidine resonance

Page 8: Structural Studies of Chicken Erythrocyte Histone H5

386

Phe Tvr

Erythrocyte Histone H 5

Phe Tyr

52OC

4OoC

35OC

I I I I

9 0 7 6 6 (ppm)

Fig. 10. Low-field spectra of H 5 peptide (32- 197) in 2 H 2 0 j M ~ C l p H 3. Histidine peaks B, D and M and tyrosine P probably cor- respond to similarly labelled peaks in Fig. 4,5,6 and I 1

D and of one tyrosine resonance. The displacement of histidine D in the peptide is in the same direction as that of histidine D in intact H5. Likewise the up- field displacement of a single tyrosine resonance in H 5 (32-197) is in the same direction, but also of lower magnitude, than that of the tyrosine P of intact H 5 (see Fig.1 and 11). The NMR spectra of H 5 (32 - 197) therefore suggest that this peptide retains the native capacity for folding to a finite but only very low degree. In fact peaks D and P of Fig. 10 both move 15% of the total displacement of peaks D and P when intact histone H 5 folds. It is therefore possible that the peptide H 5 (32 - 197) assumes the native confor- mation for 15% of the time. The small but significant changes in the CD spectra are consistent with this interpretation. The results obtained for H 5 (32 - 197) imply that at least some if not all of residues 1 - 31 are essential for the maintainance of the H5 fold. The complete lack of secondary or tertiary structure in H 5 (59- 197) is in agreement with this proposal.

The overall result of these studies of cleaved pep- tides is that the region of H 5 that takes up a globular structure is in the N-terminal region of the chain and starts before residue 31. The sequence between 1 and 31 suggests that all of these residues could be glob- ular: they have a net charge of only + 3 and contain 9 hydrophobic residues. The Chou and Fasman pro- cedure in fact predicts a helix between residues 1 and 7. The C-terminal end of the globular structure is not defined experimentally at present : the sequence and

33°C

24OC

1 8 O C

A B C D E F G 1- No H

I I I

9 8 7 6 6 (ppm)

Fig. 11. Heat denaturutivn of a solution of 50 mglml histone H 5 in ' H Z O , pH 3. No added salt. Low-field 270-MHz spectrum. Peaks C, E, F and G are due to exchangeable protons

composition data suggest a limit around residue 100, but this needs to be confirmed by further cleavage studies.

Denaturation Studies

A folded region consisting of about 100 residues without disulphide bridges is not large. We have in- vestigated the co-operativity and kinetics of H 5 fold- ing by following the terminal denaturation using NMR and CD at both 288 and 222 nm. Fig. 11 illustrates the changes in the NMR spectrum of the aromatic residues. Four peaks (A and D due to histidine C2H, 0 and P due to the protons ortho to OH in two tyrosine residues) are readily seen to move monotonically from their perturbed positions at 18 "C to their random coil position at 52 "C. The spectrum in general and these four peaks in particular show that at all tem-

Page 9: Structural Studies of Chicken Erythrocyte Histone H5

C. Crane-Robinson, S. E. Danby, E. M. Bradbury, A. Garel, A,-M. Kovacs, M. Champagne, and M. Dame 387

peratures during denaturation the protein is in rapid equilibrium between the folded and denatured states at a rate greater than lo's-' i.e. life times in both states less than s. The spectra of Fig. 5, showing the reverse process of folding at room temperature, differ somewhat from those of Fig. 11 in that the equi- librium between the folded (1 M NaCl) and denatur- ed ('HZO) states is not as rapid as under the above conditions. Thus on the addition of 14 mM NaCl at room temperature the perturbed peaks A, D, 0 and P are immediately apparent showing the formation of a small amount of folded H 5 ; however, the chemical shift of P in 14 mM NaCl is not quite that in the fully structured state (1 M NaCl). This 'intermediate' ex- change rate has been estimated from a simulation programme for the dependence of line shapes and positions on exchange life times. The result obtajted for the spectrum (Fig.5) in 50 mM NaCl at room temperature was a life time of 5 ms. These results show that the folding of H 5 is very much more rapid than that of the usual globular proteins and similar to that of H 1 [23]. The fast kinetics are presumably a direct result of H 5 having a relatively small globular segment containing no disulphide bonds. Several ex- periments have suggested that H 1 is involved in the considerable changes of chromatin structure that occur at mitosis [20] and the ability of H 5 and H 1 to undergo rapid conformational change could be related to this function in vivo. Another difference between the globular structure of H5 and that of the well-studied globular proteins such as lysozyme or ribonuclease lies in the apparently low cooperativity of H 5 denaturation. This has been studied by follow- ing the thermal denaturation of H 5 at a concentration of 5 mg/ml in 1 M KF, pH 7.4 by the three different methods : (a) noting the temperature dependence of the chemical shift of the most displaced tyrosine re- sidue P as an indicator of tertiary structure (Fig. 11 shows a typical set of spectra obtained on temperature denaturation of H 5, (b) following the extrinsic tyro- sine ellipticity at 288 nm as another indicator of ter- tiary structure (Fig.3 shows that this CD band dis- appears on denaturation) and (c) following the ellip- ticity at 222 nm as an indicator of secondary structure. The data are shown in Fig. 12 and it is apparent that the temperature of half denaturation t l /Z is much lower for the CD measurements at 288 nm ( t l j 2 = 46°C) than for the CD measurements at 222 nm or the NMR data (f1/2 = 59°C + 62°C). At 46°C for example, there is no loss of secondary structure and little change in the NMR spectrum whilst according to the ellipticity at 288 nm, the protein is half denatured. These differen- ces reflect the fact that the three parameters are dif- ferent indicators of the precise structural state of the molecule. The overall picture however, is that the de- naturation is not highly cooperative and the unfolding cannot be described as a two-state process. It should

I

-3.0' ' 1 -10 0 10 20 30 40 50 60 To 80 90

Temperature ("C)

Fig. 12. Melting profile of a solution of' 5 iifg/!~71 hiatone H 5 in I M KF, p H 7.4 obtained by three methods.Curve (1) ellipticity at 222 nm. Curve (2) chemical shift of tyrosine resonance P relative to the shift in fully structured H 5. Curve (3) ellipticity at 288 nm

be noted that it is only the folding and denaturation processes which show these clear differences from globular proteins. In particular there is no suggestion that the tertiary structure of the fully folded state of histone H 5 is in any way less specific than that of, say, lysozyme. The highly specific and absolutely re- producible NMR spectrum is a direct indicator of that. The enthalpy difference between the denatured and the native state (obtained from a plot of 1nK against 1 / T using the three sets of data in Fig. 12) is evaluated as 22 kcal mol-' or 92 kJ (from [O],,,), as 38 kcal rno1-I or 159 kJ rno1-I (from the NMR data) and as 53 kcal mol-' or 222 kJ mol-' (from [O],,,). Since the data obtained from the tyrosine ex- trinsic Cotton effect at 288nm shows the least co- operativity, the value of AH,,, = 22 kcal mol-' (92 kJ mol-') must be the closest to the true value.

CONCLUSIONS

The spectroscopic evidence presented here shows that histone H 5 contains both secondary and tertiary structure. Both CD and NMR spectra show that the aromatic residues including phenylakanine-93, are in- volved in the tertiary structure. Since these residues are entirely within the N-terminal half of the molecule and furthermore the composition of the C-terminal half is so basic (50 as to effectively prelude globular- ity, it is proposed that only the N-terminal half takes up a globular structure. This is a close parallel to the conformation of calf thymus H 1. An important dif- ference between H 5 and H 1 may, however, be that in H 5 the globular structure extends right up to the

Page 10: Structural Studies of Chicken Erythrocyte Histone H5

388 C. Crane-Robinson, S. E. Danby, E. M. Bradbury, A. Garel, A.-M. Kovacs, M. Champagne, and M. Daune: Erythrocyte Histone H 5

N terminus: cleavage at residue 31 is shown to effec- tively destroy tertiary folding. In H 1, by contrast, the 40 N-terminal residues are extremely basic, very variable and probably not globular. These N-terminal differences between H 1 and H 5 could be functionally important.

This work was partially supported by grant 7570187 of the Delegation Generale a la Recherche Scienrifique et Technique. of France and the Science Research Council of Great Britain. We are grateful to Dr G. E. Chapman for helpful discussion of this work.

REFERENCES

1.

2. 3.

4.

5.

6. 7.

8.

9.

10.

11. 12.

13.

14.

15.

Neelin, J. M., Callahan, P. X., Lamb, D. C. & Murray, K.

Knilica, L. S. (1964) Experientia (Basel) 20, 13-16. Vendrely, R., Genty, N. & Coirault, Y. (1965) Bull. Soc. Chin?.

Champagne, M., Mazen, A. & Wilhelm, X. (1968) Bull. Soc.

Miki, B. L. A. &Neelin, J. M. (1975) Can. J . Biochem. 53,

Mazen, A. & Champagne, M. (1969) FEBS Lett. 2, 248-250. Mazen, A. &Champagne, M. (1972) Biochimie (Paris) 54,

Diggle, J. H., McVittie, J. D. & Peacocke, A. R. (1975) Eur. J . Biochem. 56, 173- 182.

Palau, J., Ruiz-Carillo, A. & Subirana, J . A. (1969) Eur. J . Biochem. 7,209-213.

Seligy, V. L., Adams, G. H. M. & Neelin, J. M. (1973) Bioche- mistry of Gene Expression in Higher Organisms (Pollak, J . & Lee, J. W., eds) Australia and New Zealand Book Company, TTY Ltd.

Adams, G. H. M. &Neelin, J. M. (1972) Cytohios6, 133-142. Appels, R., Wells, J. R. E. & Williams, A. F. (1971) J . Cell.

Appels, R. & Wells, J. R. E. (1972) J . Mol. Biol. 70, 425-

Billet, M. R. & Hindley, J. (1972) Eur. J . Biochen7. 2K, 451 -

Sotirov, N. & Johns, E. W. (1973) Exp. Cell. Res. I , 13-16.

(1964) Can. J . Biochm. 42, 1743 - 1752.

Biol. 47,2233 - 2240.

Chim. Biol. 50, 1261 - 1272.

1158 - 1169.

1273 - 1279.

Sci. I0,47-49.

434.

462.

16. Sautiere, P., Kmiecik, D., Loy, O., Briand, G., Biserte, G., Garel, A. & Champagne, M. (1975) FEBS Lett. 50, 200- 203.

17. Littau, V. C., Burdick, C. J., Allfrey, V. G. & Mirsky, A. E. (1965) Proc. Nail Acad. Sci. U.S.A. 54, 1240-1243.

18. Mirsky, A. E., Burdick, C. J., Davidson, E. N. & Littau, V. C. (1968) Proc. Natl Acud. Sci. U.S.A. 61, 592-597.

19. Bradbury, E. M., Carpenter, B. G. & Rattle, H. W. E. (1973) Nuture (Lond.) 241,123 - 126.

20. Bradbury, E. M., Inglis, R. J., Matthews, H. R. & Sarner, N. (1973) Eur. .I. Biochem. 33, 131 -139.

21. Lake, R. S. (1973) Nut. New Biol. 242, 145-146. 22. Tobin, R. S. & Seligy, V. L. (1975) J . Biol. Chem. 250, 358-

364. 23. Bradbury, E. M., Cary, P. D., Chapman, G. E., Crane-Robin-

son, C., Danby, S. E., Rattle, H: W. E., Boublik, M., Palau, J. &Aviles, F. J. (1975) Eur. J . Biochem. 52, 605-613.

24. Bradbury, E. M., Chapman, G. E., Danby, S. E., Hartman, P. G. & Riches, P. L. (1975) Eur. J . Biochem. 57, 521 -528.

25. Garel, A,, Mazen, A., Champagne, M., Sautiere, P., Kmiecik, D., Loy, 0. & Biserte, G. (1975) FEBS Lett. 50, 195-199.

26. Jones, G. M. T., Rall, S. C. & Cole, R. D. (1974) J. Biol. Chem. 249, 2548-2553.

27. Sautiere, P., Briand, G., Kmiecik, D., Loy, O., Biserte, G., Ga- rel, A. & Champagne, M. (1976) FEBS Lett. 63, 164- 166.

28. Champagne, M., Pouyet, J., Ouellet, L. & Garel, A. (1970) Bull. Soc. Chim. Biol. 52, 377 -389.

29. Wright, E. B., Olins, D. E. (1975) Biochem. Biophys. Res. Com- mun. 63,642- 650.

30. Leboy,O. S., Cox,E. C. & Flaks, J . G. (1964) Proc. Nut1

31. Campbell, 1. D., Dobson, C. M., Williams, R. J. P. & Xavier,

32. Garel, A,, Kovacs, A. M., Champagne, M. & Daune, M. (1975)

33. Chen, Y. H., Yang, J . T. & Chau, K. H. (1974) Biochemistry,

34. Williams, R. E. & Seligy, V. L. (1973) FEBSLett. 33, 197-200. 35. Greenaway, P. J. & Murray, K. (1971) Nut. New Biol. 229,

36. Chapman, G. E., Hartman, P. G., Bradbury, E. M. (1976) Eur.

37. Chou, P. Y. & Fasman, G. D. (1974) Biochemistry, 13, 211 -

Acad. Sci. U.S.A. 52, 1367-1374.

A. V. (1973) J. Magn. Res. [ I , 172-181.

Biochim. Biophys. Acta, 395, 5-15.

13,3350-3359.

233 -238.

J . Biochem. 61, 69-75.

222 and 222 -245.

C. Crane-Robinson, S. E. Danby and E. M Bradbury, Biophysics Laboratory, Department of Physics, Portsmouth Polytechnic, Gun House, Hampshire Terrace. Portsmouth, Great Britain, PO1 2DZ

A. Garel, A.-M. Kovacs, M. Champagne and M. Daune, Institute de Biologic Molkculaire et Cellulaire du C.N.R.S., 15 Rue Rene-Descartes, Esplanade, F-67000 Strasbourg, France