Molecular symmetry and arrangement of subunits in extracellular hemoglobin from the nematode Ascaris mum

Transcription

1 Eur. J. Biochem. 201, FEBS V Molecular symmetry and arrangement of subunits in extracellular hemoglobin from the nematode Ascaris mum Saleh DARAWSHE and Ezra DANlEL Department of Biochemistry, George S. Wise Faculty of Life Sciences, Tel-Aviv University, Israel (Received January 8, 1991) - EJB The arrangement of subunits and molecular symmetry of extracellular hemoglobin from the nematode Ascaris suum, an 11.7s molecule of molecular mass 332 kda and composed of eight identical subunits, was studied. Dissociation of the molecule at alkaline and acid ph yielded 4.6s and 2.7s components, identified as polypeptide-chain dimers and monomers, respectively. Cross-linking with glutardialdehyde followed by SDS/PAGE resulted in a maximum number of eight bands identified in order of decreasing mobility as monomeric and 2-8 cross-linked-polypeptide-chain species. Comparison with values predicted from theory shows that the distribution of protein among the various cross-linked species, obtained after different extents of exposure to cross-linker, is consistent with a two-layered arrangement of subunits involving one type of interaction between subunits from different layers and another between subunits within the same layer. Electron micrographs of the molecule showed two profiles, a square and a rectangle. We propose a model for the molecule which is eight subunits arranged in two layers, stacked in an eclipsed orientation. The proposed model is consistent with the results from sedimentation, cross-linking and electron microscopy. Taken together, our findings indicate D4 symmetry for Ascaris hemoglobin. In the nematodes, hemoglobin has been found in many animal-parasitic and in a few free-living species [l, 21. Among nematode hemoglobins, extracellular hemoglobin from Ascuris suum, an intestinal parasite of the pig, has been the most extensively studied. Interest in this hemoglobin has been aroused by the finding that it is associated with exceptionally high oxygen-binding affinity [3]. Structural investigations have hitherto been directed towards characterization of the molecule with respect to its molecular mass and that of the constituent polypeptide chains, the heme content and the amino acid composition. These studies have established that Ascaris hemoglobin is composed of eight identical single-polypeptide-chain subunits [4-61. Each subunit has been found to carry two binding sites for heme [6]. Incomplete realization of the full binding capacity for heme has afforded an explanation for the relatively low heme content reported for this hemoglobin. The purpose of the present study is to determine the symmetry and arrangement of subunits in Ascaris hemoglobin. MATERIALS AND METHODS Materials Extracellular hemoglobin was prepared from the nematode A. suum as described elsewhere [6]. Glutardialdehyde (25% by vol.) was a Merck product. Coomassie brilliant blue R was obtained from Sigma. Buffers were prepared from Correspondence to E. Daniel, Department of Biochemistry, Tel- Aviv University, 1L Tel-Aviv, Israel Nofe. This paper is dedicated to Professor Abel Schejter on the occasion of his 60th birthday. analytical-grade reagents. Double-distilled water was used throughout. Sedimentation Sedimentation velocity centrifugation was performed in a Beckman model E ultracentrifuge using Schlieren-phase plate optics. Sedimentation coefficients were corrected to S20,w in the usual way [7]. Cr osslinking Crosslinking of Ascaris hemoglobin with glutardialdehyde was carried out at 20 C in 0.01 M sodium phosphate, ph 7.0. The protein concentration was 2 mg/ml and the glutardialdehyde concentration was 0.09Oh (by vol.). At various times, aliquots were withdrawn from the reaction mixture and cross-linking stopped by the addition of SDS. The cross-linked species were analyzed by SDS/PAGE. SDSjPAGE Electrophoresis was performed on polyacrylamide gels in tubes (0.5 cm x 10 cm) in the presence of 1% (rnassivol.) SDS and 1% (by vol.) 2-mercaptoethanol as described by Weber et al. [8]. Samples were heated for 2 min at 10O"C, then applied to the gel. Staining was carried out with Coomassie brilliant blue R. The gels were thoroughly destained and photographed. The photographic films were scanned at 550 nm in a scan-recording densitometer. Evaluation of densitometric scans Preliminary examination of the densitometric scans showed that the shape of a well-separated electrophoretic

2 170 I I I I 0'; PH Fig. 1. Dependence of the sedimentation coefficient of Ascaris hemoglobin on ph. The protein concentration was approximately 1.2 mg/ ml. Buffers (approximately 0.1 M) were used in the ph ranges indicated: HCljKCl (below ph 2.2); glycine/hci ( ); sodium acetate ( ); sodium phosphate ( ); glycine/naoh ( ); NaOH/Na2HP04 (above 10.6). (---) Sedimentation coefficients for whole molecules taken as 11.7S, and for two-subunit and one-subunit species of Ascaris hemoglobin predicted by the Kirkwood theory [18] on the basis of the model presented in Fig. 7. (0) Observed values; (0) values determined by Okazaki et al. [5] for single-subunit and two-subunit succinylated products of Ascaris hemoglobin band can be fitted to a Gaussian function. Resolution of the scan into Gaussian-component bands was therefore carried out. The fraction of protein associated with a given band was obtained from the area under the corresponding Gaussian function, relative to the total area associated with all the bands on the same gel. The reliability of the procedure rests on the assumption of proportionality between the Coomassie-stain incorporation and protein content in the gel. This assumption is reasonable in the case of Ascuris hemoglobin, since the polypeptide chains involved in cross-linking seem to be identical [6]. Calculations The theoretical distribution of protein among cross-linked ohgomers as a function of extent of reaction was calculated assuming possible models for an assembly of eight identical subunits. The calculation involved determination of the probability of occurrence of a given set of cross-links, carried out according to Sculley et al. [9], and evaluation of the fraction of cross-linked oligomers obtained upon disruption of the non-covalent interactions by SDS, carried out according to Hajdu et al. [lo]. Electron microscopy Negative staining was performed using 1 % (massivol.) uranyl acetate. Observation and photography were with a Jeol-Jem 1200 Ex electron microscope. Contrast enhancement was achieved using Markham rotation [Ill and superimposition of images in the same orientation [12]. Fig. 2. Electrophoresis of cross-linked Ascaris hemoglobin. Hemoglobin was reacted with glutardialdehyde and then electrnphoresed on 3.3% (massjvol.) polyacrylamide gels in the presence of SDS and 2- mercaptoethanol. Exposure times to cross-linker, in min, are indicated along the x-axis ficient of about 0.5s observed between approximate ph values of 8 and 9 (the ph at which the decrease occurs is not exactly known) is probably due to a conformational transition. Outside the ph range , dissociation of the molecule occurs. On the alkaline side, components with Sedimentation coefficients of 4.6 & 0.3s and, at still higher ph, 2.7 -t 0.2s are observed. On the acidic side, as at alkaline ph values, dissociation involes 4.6s and 2.7s components. The dissociation to 4.6s under both alkaline and acidic conditions is steep. The transition 4.6s to 2.7s at acidic ph is gradual and involves intermediates with sedimentation coefficients between the two limiting values. Cross-linking with glutardialdehyde On SDS/PAGE, Ascaris hemoglobin yields essentially a single band with a mobility corresponding to 43 kda. Exposure to glutardialdehyde results in the appearance of bands of lower mobilities (Fig. 2). A maximum number of eight bands are observed. As described elsewhere [6], these bands can be identified in order of decreasing mobilities, as monomeric and 2-8 cross-linked polypeptide chains. Exposure to cross-linker for relatively long times brought about an increase in the intensity of low-mobility bands and a loss in their ability to be resolved. Under the conditions used for cross-linking, no bands beyond those identified could be detected. RESULTS Dependence of the sedimentation coeflicient on ph Fig. 1 shows the effect of ph on the sedimentation coefficient of Ascaris hemoglobin. It is seen that the sedimentation coefficient has a constant value of _ 0.35 in the ph range 3.9 to approximately 8 and of 11.2 f 0.4s in the range approximately 9 to The small decrease in sedimentation coef- Analysis of data obtained from cross-linking Fig. 3 presents a densitometric scan of cross-linked products obtained by exposure of Ascaris hemoglobin to glutardialdehyde?or 9 min. The scan was resolved into Gaussian distributions and, from the relative area under each Gaussian curve, the fraction of corresponding cross-linked species was determined. Resolution of scans, corresponding to different exposure times to cross-linker, allows a follow-up

3 I I VlllVII VI v IV Ill II I Distance from origin lcml Fig. 3. Resolution qf'densitometric scan. The electrophoretic pattern of Ascaris hemoglobin exposed to glutardialdehyde for 9 min was scanned and the scan was resolved into Gaussian bands (---). (0) Observed values; (--) sum of resolved component bands. Resolved bands arc designated by Roman numerals corresponding to the number of cross-linked polypeptide chains ' Fraction of monomer Fig. 5. Variation with extent of reaction of protein distribution among cross-linkedspecies. Fractions of cross-linked oligomers obtained after various exposure times of hemoglobin to glutardialdehyde were plotted against corresponding fractions of remaining monomer. Observed values: (0) dimer; (A) trimer; (W) tetramer; (*) sum of pentamer, hexamer, heptamer and octamer. The solid lines represent theoretical curves, calculated for a dihedral two-layered model with k,/k, = 2.0 (see text). 11, dimer; 111, trimer; IV, tetramer; V-VIII, sum of pentamer, hexamer, heptamer and octamer --.:/ a Number of polypeptide chains Fig. 4. Distvibution qf'protein among cross-linked species. The fraction of' protein in a given cross-linked oligomeric species was plotted versus the number of polypeptide chains in the oligomer. The data correspond to (a) 3, (b) 5, (c) 7 and (d) 9 min exposure to glutardialdehyde. (0) Observed values. Theoretical values for an assembly of eight identical subunits, calculated for integral values (1-8) of the abscissa, were plotted and joined by straight lines for better vizualization: (---) cyclic model; ( ) dihedral two-layered model with rate constants k,/k, = 2.0 (see text) of the variation in the intensity associated with a particular band with the extent of the reaction (Fig. 4). Inspection of the resolved gel scans revealed a tendency for alternation in the intensities of the bands. The fraction of tetramer was found to be slightly higher than either trimer or pentamer, the fraction of hexamer higher than either pentamer or heptamer. Fig. 5 shows the distribution of protein among the various cross-linked species plotted against remaining monomer fraction, a parameter chosen here and in Fig. 4 to represent the extent of reaction. Electron microscopy Fig. 6 is an electron micrograph of a negatively stained preparation of Ascaris hemoglobin. The field shows a large number of what appears to be molecular profiles. Difficulties were encountered, however, in selecting profiles suitable for structural analysis. Attempts to obtain micrographs of a better Fig. 6. Electron micrograph of negatively stained Ascaris hernoglobin. (0) Rectangular profile; 0 square profile quality, by varying the conditions of the experiment, were not successful. Close inspection of detailed features of individual profiles shows, nevertheless, that they fall into two types: a polygon with a central hole, and a rectangle divided into two halves by a longitudinal gap. Application of the technique of contrast enhancement further shows the following: (a) the polygonal profile is a square; (b) the length/breadth ratio for the rectangular profile is about 2: 1 ; (c) the side of the square and the breadth of the rectangle are almost equal (Fig. 7). DISCUSSION A.~caris hemoglobin is an 11.7s molecule of molecular mass 332 kda composed of eight identical subunits. Accord-

4 172 A 8 fm 0 n A Fig. 7. Proposed model for Ascaris hemoglobin. Schematic representation of a model composed of eight subunits arranged in two layers, stacked in an eclipsed orientation, the four subunits in each layer occupying the vertices of a square. Each subunit has been approximated by a cylinder with a height/diameter ratio equal to two. Projections of the model: al, top; bl, side. Electron-microscopic profiles of the molecule obtained by superposition of four rotational photographs: a2, square (rotation steps, 4 x 90"); bz, rectangle (rotation steps, 2 x 180") C ing to theory [13, 141, eight identical subunits can assemble to give closed structures in two types of symmetry, cyclic (C,) and dihedral (D4). Two types of intersubunit bonds are involved in the assembly : isologous, where each subunit contributes an identical binding domain, and heterologous, where each subunit contributes a different binding domain [15]. In the cyclic structure, the intersubunit bonds are heterologous and all of one type. In contrast, the bonds linking the subunits together in the dihedral structures cannot be all of the same type. Thus, there are two types of isologous bonds in an alternating ring, one type of heterologous and one type of isologous bond in a two-layered eclipsed structure, and one type of heterologous and two types of isologous bond in a two-layered staggered structure (Fig. 8). A relatively recent addition for the determination of the quaternary structure of a protein oligomer is that of crosslinking with a bifunctional reagent [9, 10, 16, 171. The method is based on the premise that the intersubunit cross-links formed upon exposure to a bifunctional reagent are governed by the symmetry of the molecule, in a manner similar to the bonds connecting the subunits in the native structure. Theoretical treatments designed for four-subunit assemblies have been derived and successfully applied to a number of tetrameric proteins [lo, 161. The approach has been extended to proteins composed of more than four subunits [9]. In the application of the cross-linking method, the aim is to determine which of the allowed structures is able to describe the experimental data. The suitability of a given model is examined by assigning values for the rate constants of crosslinking associated with the model, and comparing the observed and predicted distributions of protein among the crosslinked species. Such a comparison showed that neither the cyclic nor the alternating model fits with the cross-linking results of Ascaris hemoglobin by glutardialdehyde. The incompatibility of the cyclic model with the data is illustrated in Fig. 4 for extents of reaction with cross-linker corresponding to residual monomer fractions of A two-layered Fig. 8. Possible models for Ascaris hemoglobin and their modes of dissociation. A, cyclic ring; B, alternating ring; C, two-layered eclipsed; D, two-layered staggered. The symmetry is cyclic (point group 8 or C,) in A, and dihedral (point group 422 or D4) in B - D. Binding regions on a subunit, represented for simplicity by a sphere, are indicated by a - d model with a ratio k,/k, = 2.0, where k, and k, are the rate constants of cross-linking between subunits in different layers and within a layer, respectively, was found to provide a satisfactory fit with the data obtained for the various extents of reaction examined (Figs 4 and 5). The fact that the crosslinking data can be satisfactorily accounted for by a twolayered model involving only two cross-linking rate constants, renders a staggered arrangement unlikely and favors an arrangement of subunits of the eclipsed type. On the basis of the cross-linking results and the evidence from electron microscopy, we propose a model for Ascaris hemoglobin. The model consists of eight subunits arranged into two layers, stacked in an eclipsed orientation. The four subunits of each layer occupy the vertices of a square (Fig. 7). The symmetry underlying the arrangement of the subunits is D4. The projections of the model are consistent with the square and rectangular profiles seen under the electron microscopy and analysed by contrast enhancement technique. Okazaki et al. [5] studied the dissociation of Ascaris hemoglobin brought about by reaction with succinic anhydride. They reported the isolation of succinylated products with sedimentation coefficients of 4.4s and 2.9s and their identification, on the basis of molecular-mass determinations, as dimers and monomers of a single polypeptide-chain subunit. The 4.6s and 2.7s components obtained in the present study as first and ultimate products in the ph dissociation of

5 Ascaris hemoglobin may therefore be identified as polypeptide-chain dimers and monomers, respectively. According to this interpretation, values of sedimentation coefficients between 4.6s and 2.7s may be attributed to a rapid dimer/ monomer equilibrium. The observation of a dimeric subunit as an intermediate in the dissociation of Ascaris hemoglobin is consistent with one of the possible modes of dissociation of the two-layered eclipsed model. As illustrated in Fig. 8, nonidentity of the intersubunit bonds in this model opens the possibility of obtaining intermediates, subunit dimers in one case and half-molecules in the other, provided that the dissociating agent is sufficiently discriminative. Furthermore, the sedimentation-coefficient values calculated for two-subunit and single-subunit species of Ascaris hemoglobin by the Kirkwood theory [18], are in fair agreement with the corresponding values observed experimentally (Fig. 1). The pattern of dissociation of the molecule, as reflected in the ph dependence of the sedimentation coefficient, is thus seen to provide support for the model proposed in this study for the arrangement of the subunits in Ascaris hemoglobin. We thank Dr M. Sculley for sending us his computer program for a six-subunit molccule, Dr Y. Tsfadia for performing some of the computations, Y. Delarea for the electron microscopy and A. Shoob for the photographic work Tsfadia, Y., Shaked, I. & Daniel, E. (1990) Eur. J. Biochem. ly3, REFERENCES Ellenby, C. &Smith, L. (1966) Comp. Biochem. Physiol. 19, Van Holde, K. E. (1975) in Theproteins, 3rd edn (Neurath, H. & 877. Hill, R., eds) vol. 1, pp , Academic Press, London 2. Atkinson, H. J. (1975) J. Eq. Biol. 62, 1-9. and New York Okazaki, T. & Wittenberg, J. B. (1965) Biochim. Biophys. Acra 111, Wittenberg, B. A,, Okazaki, T. & Wittenberg, J. B. (1965) Biochim. Biophys. Acta 111, Okazaki, T., Briehl, R. W., Wittenberg, J. B. & Wittenberg, B. A. (1965) Biochim. Biophys. Actu 111, Darawshe, S., Tsafadyah, Y. & Daniel, E. (1987) Biochem. J. 242, Svedberg, T. & Pedersen, K. 0. (1940) The ultracentrifuge, pp , Clarendon Press, Oxford. 8. Webel; K., Pringle, J. R. & Osborn, M. (1972) Methuds Enzymol. 26, Sculley, M. J., Treacy, G. B. & Jeffrey, P. D. (1984) Biophys. Chem. 19, Hajdu, J., Bartha, F. & Friedrich, P. (1976) Eur. J. Biochem. 68, Markham, R., Frey, S. & Hills, G. J. (1963) Virology 20, Valentine, R. C., Shapiro, B. M. & Stadtman, E. R. (1968) Biochemistry 7, Klotz, I. M., Darnall, D. W. & Langerman, N. R. (1975) in The proteins, 3rd edn (Neurath, H. & Hill, R., eds) vol. 1, pp , Academic Press, London and New York. 14. Matthews, B. W. & Bernhard, S. A. (1973) Annu. Rev. Biophp. Bioeng. 2, Monod, J., Wyman, J. & Changeux, J.-P. (1965) J. Mol. Biol. 12, Hucho, F., Miillner, H. & Sund, H. (1975) Eur. J. Biochem. 59,