Cell Motility and the Cytoskeleton 36:246–252 (1997) Physical Properties of Dystrophin Rod Domain Edith Kahana,1* G. Flood,2 and W.B. Gratzer1 1Medical Research Council Muscle and Cell Motility Unit, King’s College, London 2National Centre for Macromolecular Hydrodynamics, University of Leicester, England We have prepared two fragments of the human dystrophin rod domain, each containing eight spectrin-like repeating units, by expression in Escherichia coli. The first corresponds to the central portion of the rod, the other to three repeats from the N-terminal end, fused to five repeats from the C-terminal end. The latter makes up the entire mutant rod, found in a patient with mild (Becker-type) muscular dystrophy. Both fragments were found to possess an ordered, stable structure, and had the form of short rod-like particles in the electron microscope. Molecular weight determinations by sedimentation equilibrium revealed that both polypeptides were monomeric in solution, suggesting that the dystrophin rod domain is incapable of forming an antiparallel homodimer. This supports the inference from sequence analyses [Winder et al., 1995: FEBS Lett. 369:27–33, 1996: Biochem. Soc. Trans. 24:2805] that the dystrophin rod domain lacks the arrangement of sites required for lateral self-association, and that dystrophin, unlike the other known proteins of the spectrin superfamily, may thus exist as a monomer. Cell Motil. Cytoskeleton 36:246–252, 1997. r 1997 Wiley-Liss, Inc. Key words: dystrophin; spectrin repeats INTRODUCTION The manner in which dystrophin functions in stabilising the sarcolemma is unclear. It serves as a link between the transmembrane glycoprotein complex, which in turn connects the membrane to the extracellular matrix, and the cortical actin cytoskeleton [Ahn and Kunkel 1993]. The actin-binding site is in a globular N-terminal domain and the glycoprotein-binding site is in the Cterminal domain. The latter is preceded in the sequence by a cysteine-rich domain with unknown, but apparently essential function, and between this and the actin-binding domain is a rod domain, comprising 24 degenerate spectrin-like repeating units. Dystrophin is mainly located in specialised zones of the sarcolemma, the costameres, surrounding the Z-discs of the myofibrils [Minetti et al., 1992; Porter et al., 1992; Straub et al., 1992]. These probably represent the sites of mechanical coupling between the sarcolemma and the contractile system. The dystrophin is presumed to protect the membrane against the shearing stresses to which it is subjected, in much the same way as the spectrin lattice stabilises the red cell r 1997 Wiley-Liss, Inc. membrane. Spectrin is responsible for the elasticity of the membrane with which it associates, and the dystrophin rod domain, with its spectrin-related sequence [Koenig and Kunkel, 1990] and conformation [Kahana et al., 1994], must be supposed to exert a similar effect. Whereas total absence of dystrophin leads to severe (Duchenne) muscular dystrophy (DMD), many types of mutation results in much milder or intermediate forms of the disease, known as Becker muscular dystrophy (BMD). Thus, whereas deletions of the C-terminal or cysteinerich domains invariably cause DMD, loss of the actinbinding domain or of a large proportion of the rod are associated with BMD [Ahn and Kunkel, 1993; Ohlendieck, 1996]. It appears therefore that the membrane can survive and the muscle continue to function, even if in *Correspondence to: E. Kahana, Department of Hematology, Tel Aviv Sourasky Medical Center and Sackler Faculty of Medicine, 6 Weizmann Street, Tel-Aviv 64239, Israel. Received 2 August 1996; accepted 11 October 1996. Dystrophin Rod Domain impaired form, despite a presumed loss in membrane elasticity. Little is known about the organisation of the dystrophin-containing complex on the membrane, but most published models have explicitly or implicitly assumed that the dystrophin rod forms an antiparallel dimer, like all other known proteins of the spectrin class. This would allow a continuous structure to be generated at the membrane surface. The evidence for dimer formation is equivocal, however [Pons et al., 1990], and a recent analysis of the rod domain sequence concludes that dimerisation, as it occurs in the spectrins and a-actinin, is unlikely [Winder et al., 1995, 1996]. We have attempted here to address this problem by examining the properties of large fragments of the rod domain and also to analyse the conformational characteristics of the rod, about which little was previously known. MATERIALS AND METHODS Dystrophin rod fragments were prepared by expression in Escherichia coli. We selected a segment from the centre of the rod, comprising repeats 8–15, and the shortened rod domain from a patient with BMD [England et al., 1990]. In the latter, repeats 4 to 18 and half of repeat 19 are deleted, leaving a fusion of the first three and the last five repeats of the rod. The cDNA for both protein sequences was kindly provided by Professor K.E. Davies. We term the products RFM (rod fragment, middle) and RFE (rod fragment, extremities). RFM begins at met-1141 in the dystrophin sequence and ends with gln-1973. The following primers were synthesised: sense-strand nt 3629 to 3659, with an added phased NdeI restriction site; antisense-strand nt 26127 to 26107, with an added stop codon and a HindIII restriction site. The restriction enzymes did not cut the cloned fragment. DNA fragments were prepared by 15 cycles of PCR, using thermostable Vent DNA polymerase. The cloning vector was pMW172 [Way et al., 1990]. The dephosphorylated vector and the PCR products were purified by gel electrophoresis. The fragment was ligated into the cloning vector and transformed into a nonexpressing (TG1) strain of E. coli and recombinants were selected by PCR, using the T7 vector sense and antisense primers. The DNA was introduced into the expression strain (BL21/DE3) and the bacteria were grown to late log phase at 37°C in LB medium, containing 0.1 mg ml21 ampicillin. Sequences from the 38 and 58 ends of the recombinant DNA were determined to confirm its identity. RFE begins at asp-329 and ends at lys-3117, residues 665 to 2367 being deleted. Expression was based on DNA from the BMD patient. A vector, pT79, modified from pMW172 by Dr P.J. Marsh to embody additional 247 restriction sites, was employed. A sense primer, corresponding to nt 1187 to 1216, with an added NdeI restriction site, was synthesised, as well as an anti-sense primer, corresponding to nt 29559 to 29539, with an added stop codon and a NotI restriction site. Cloning and expression were performed as above. Both RFE and RFM were found in inclusion bodies, as well as in the soluble fraction of the bacteria. The inclusion bodies were collected, dispersed in 6 M guanidinium chloride, 5 mM dithiothreitol, and the protein was recovered by chromatography in the same solvent on a Sephacryl S-300 (Pharmacia, Piscataway, NJ) column of 2.5 3 100 cm. The product at this stage was contaminated with nucleotide material. This was eliminated by precipitating the protein with an equal volume of saturated ammonium sulphate. It was redissolved and dialysed against 50 mM sodium chloride, 20 mM sodium phosphate, pH 7.6, and rechromatographed on Superose 12 in a Pharmacia FPLC system. Fractions were screened for purity by gel electrophoresis in SDS. N-terminal sequencing was carried out on the products to establish their identities. (We are indebted for these determinations to A.C.M. Harris, National Institute for Medical Research, London.) Concentrations were measured spectrophotometrically, using calculated specific absorptivities [Perkins, 1986]. For chemical cross-linking, proteins at a concentration of 60 to 170 µg ml21 in 0.1 M triethanolamine, pH 8.5, on ice were treated with 1 mg ml21 dimethylsuberimidate [Davis and Stark, 1970] and examined by SDS-gel electrophoresis in 5% acrylamide gels. Red cell spectrin was used as a control. Molecular weight determinations by sedimentation equilibrium were performed in a Beckman (Palo Alto, CA) Optima XL-A analytical ultracentrifuge, using absorption optics at 280, 234, or 220 nm on solutions in 0.15 M sodium chloride, 50 mM sodium phosphate, pH 7.6. Sedimentation velocities were measured in the same instrument at 42,000 rpm. Partial specific volumes were calculated from the amino acid compositions [Perkins, 1986]. For electron microscopy (performed by Dr. A.P.R. Brain, King’s College London Electron Microscopy Unit) solutions of protein at 50–200 µg ml21 in 50 mM sodium chloride, 5 mM imidazole, pH 7.6, containing 60% (v/v) glycerol, were sprayed onto mica, shadowed at an angle of 6° with platinum, and coated with carbon. The replicas were transferred to grids and examined in a Philips (Mahwah, NJ) 301G instrument. Circular dichroism spectra were measured in a Jobin-Yvon CD6 Dicrograph (Longjumeau, France), using cells of 0.5–2 mm path-length in a thermostated housing. For thermal denaturation profiles the temperature was varied in steps of 5°C. 248 Kahana et al. RESULTS The purified expressed dystrophin rod segments, RFM and RFE, were at least 95% pure, as judged by gel electrophoresis in SDS. The apparent molecular weights were about 97,000 and 125,000, respectively, in accordance with expectation for these fragments. Circular dichroism showed both polypeptides to be highly structured. The molar residue ellipticities at 222 nm, averaged in each case from data on several column fractions, were 222,200 and 223,900 degrees cm2 dmol21, respectively; these correspond, in the absence of other ordered structures, to about 62 and 67% a-helix [Greenfield and Fasman, 1969], or fitting the spectra with a basis-set of standard proteins by the CONTIN programme [Provencher and Glöckner, 1981], average a-helix contents of 77 and 75% are obtained, with 3 and 4% b-structure, respectively. To determine whether the polypeptide chains have entered their stable native fold, thermal denaturation profiles were determined by circular dichroism (Fig. 1). These show an extended plateau region, in which the structure remains essentially unchanged, followed by a sigmoidal transition to an unfolded state. The conformation is thus stable below the transition region. The transition itself is broader than those of a single repeating unit from the rod domain [Kahana et al., 1994; Kahana and Gratzer, 1995], indicating that the unfolding process is not fully cooperative along the rod in either case. Chemical cross-linking with dimethylsuberimidate gave no evidence of the formation of species of higher molecular weight than the monomer by either polypeptide. Spectrin dimers under the same conditions were totally cross-linked to the covalent dimer (Fig. 2). Sedimentation equilibrium distributions of RFM at concentrations of 40, 70, and 100 µg ml21, scanned at 234 and 220 nm, were well fitted by a monodisperse, ideal solute with molecular weight (based on a calculated partial specific volume of 0.740 ml g21 ) of 110,000 6 8,000 (n 5 5) (theoretical value, 97,000). RFE, measured at concentrations of 120, 170, and 230 mg ml21 and scanned at 280 and 234 nm, similarly gave an equilibrium distribution consistent with a monodisperse solute of molecular weight 130,000 6 5000 (n 5 6) (theoretical value, 126,300). Results are summarised in Table I. The precision of these measurements was probably limited by the rather low protein concentrations, and does not exclude fits that allow the presence of some aggregated material; it leaves no doubt, however, that both polypeptides are monomeric, rather than dimeric. Sedimentation velocity measurements gave corrected sedimentation coefficients (s020,w) of 4.9S for RFM and 4.6S for RFE. Corresponding frictional ratios are 1.15 and 1.25, respectively, implying a relatively low degree of asymmetry for both particles. That both are Fig. 1. Thermal denaturation profiles, measured by circular dichroism, of middle rod fragment, RFM (W) and rod with central deletion, RFE (M). significantly asymmetric and/or flexible is nevertheless clear, since globular proteins of the same molecular weight would sediment at about 6.0S and 7.2S, respectively [Halsall, 1967]. Electron microscopy of the shadowed preparations showed the presence of elongated particles in both polypeptides. The estimated lengths were about 33 6 7 nm (100 particles measured) for RFM and about 38 6 6 nm (105 particles) for RFE (Fig. 3). In the latter case there appears to be some tendency towards end-to-end association. As Figure 3 shows, however, globular particles were also distributed throughout the fields. DISCUSSION The expressed fragments were shown by conformational criteria to have entered the native protein fold. Both were monomeric in solution. By electron microscopy they had the form of short, flexible rods, although globular particles were also present and probably represented polypeptides that had undergone conformational collapse under the shearing stresses encountered during drying. Globular objects can similarly be discerned in electron micrographs of intact dystrophin [Pons et al., 1990]. The lengths of the rods were in the range of 30–40 nm. Judged by the extended length of erythroid spectrin, as measured in the electron microscope by the same procedure [Shotton et al., 1979], a length of about 30 nm would be expected for both fragments. On the other hand, the observed length of the very stable and rigid a-actinin rod, comprising four spectrin-like repeats, was 25 nm [Imamura et al., 1988] or 22.5 nm [Flood et al., 1995]. Dystrophin Rod Domain Fig. 2. Cross-linking of rod fragments with dimethylsuberimidate. Analysis of products by gel electrophoresis in SDS. A: RFM (middle rod). Lanes a, unreacted rod fragment; b, protein after treatment with cross-linking reagent at protein concentrations of 170 µg ml21; c, protein reacted at 90 µg ml21; d, untreated erythroid spectrin; e, spectrin after treatment with reagent at protein concentration of 170 µg ml21; f, spectrin treated at 90 µg ml21. B: RFE (rod fragment with central deletion). Lanes a, untreated protein; b, protein treated with cross-linking reagent; c, untreated erythroid spectrin; d, spectrin treated with cross-linking reagent. Left-hand tracks in both panels contain molecular weight markers. 249 250 Kahana et al. Fig. 3. Shadowing electron microscopy of rod fragments, RFM (A) and RFE (B). Bar 5 50 nm. TABLE I. Sedimentation Equilibrium of Dystrophin Fragments, RFM and RFE Fragment Middle rod (RFM) Concentration (µg ml21 ) Wavelength (nm) Molecular weighta 40 220 234 220 234 234 114,500 98,000 107,700 108,000 118,600 234 280 234 280 234 280 127,200 126,700 129,200 137,000 125,600 136,600 70 100 Fused rod extremities (RFE) 120 170 230 aBest fit to monodisperse ideal solute. This value would lead to a predicted length of 45–50 nm for RFE and RFM. The length of a single Drosophila spectrin repeat in the crystal is close to 5 nm [Yan et al., 1993], and on this basis a length of approximately 40 nm would be expected for the fragments. Considering especially the variability of sequence in the dystrophin rod repeats, a difference in contour length and stiffness can certainly be envisaged. The asymmetry of the hydrodynamic particle in each case is relatively low, implying that both chains are far from rigid. The rigidity might be expected to be lower for a monomer than for a dimer of laterally associated subunits; nevertheless, sedimentation coefficients of spectrin and fodrin tetramer, spectrin dimer, a-actinin rods of four and three repeats and of the dystrophin fragments RFE and RFM, all adhere quite closely to the same sedimentation coefficient-molecular weight relation, namely s ~ M 0.44. The expressed dystrophin rod fragment, RFM, corresponds to the central portion of the rod domain, beginning 7 residues before the start of repeat 8 and terminating at the end of repeat 15. The fragment, RFE, derived from a natural dystrophin mutant [England et al., 1990], contains repeats 1 to 3 and the 11 amino acid residues preceding, fused to a segment which starts in the middle (centre of the B-helix) of repeat 19 and terminates 77 residues beyond the end of the rod domain; it thus contains 8 repeats, like RFM. The identification of the repeat boundaries is based on the alignment of Winder et al.  (which is almost identical to that of Koenig and Kunkel ). In the other proteins of the same superfamily, spectrin and a-actinin, hetero- and homodimers, respectively, are formed by pairwise association between the repeats of antiparallel subunits, the first (N-terminal repeat) with the last (C-terminal repeat), the second with the last-but-one, and so on. Thus, if this simple scheme applies to dystrophin, RFM has the possibility of self-associating through three antiparallel pairs of repeats and RFE through 6 such pairs. In the case of a-actinin, the four repeats of the isolated rod domain are sufficient to form an extremely tight dimer [Imamura et al., 1988; Kahana and Gratzer, 1991], but in spectrin the situation is more complex. For erythroid spectrin Speicher and his colleagues [Speicher et al., 1992; Ursitti et al., 1996] found that lateral, antiparallel association of the a- and b-chains was a function of the four repeats nearest to the N-terminus of the b-chain and a complementary set at the C-terminus of the a-chain. These were envisaged to nucleate the interaction, bringing the remaining, more weakly interacting, pairs of repeats into register. The first of the b-chain repeats was found to be necessary but insufficient for association and the smallest b-chain element that bound detectably to the a-chain comprised repeats 1 and 2; addition of repeats 3 and 4 increased the association constant by successive factors of 5 [Ursitti et al., 1996]. A feature of the strongly interacting repeats is an N-terminal insert of eight amino acids, also present in all a-actinin repeats. Dystrophin Rod Domain Requirements for inter-chain binding were also defined by Viel and Branton  for Drosophila spectrin. Two repeats in each chain were found to be essential, each of them containing the eight-residue insert. In addition some residues in the adjoining nonhomologous, terminal sequences were found to contribute to the association, although this was not the case in human spectrin [Ursitti et al., 1996]. A search of the dystrophin sequence revealed no eight-residue stretches homologous to the inserts in the a-actinin and terminal spectrin repeats. Allowing up to 3 amino acid differences, four related eight-residue sequences were found, but none was close to the Nterminus of a repeat in the internal alignment. The only direct experimental data that bear on the association state of native, intact dystrophin stem from electron microscopy. Pons et al.  concluded that their preparations contained mainly monomers with some dimers. Later data [Sato et al., 1992] should probably be disregarded because of the authors’ subsequent doubts concerning the identity of the material [Kake et al., 1995]. Winder et al.  suggest that the poor sequence conservation of surface residues in the dystrophin rod repeats and the presence of loops between coiled-coil elements argues strongly against lateral association. Moreover, the presumptive self-association sites, present in all erythroid spectrin repeats, occur in only a few dystrophin repeats, and of these all but two are out of register in an antiparallel alignment [Winder et al., 1996]. Our results demonstrate that neither the middle portion of the rod nor the ends, which would interact by pairs in an antiparallel dimer, are able to self-associate measurably. The rod domain is thus unlike those of erythroid spectrin and a-actinin, for example, in its behaviour. We have not proved that dystrophin is incapable of self-associating in other, unrelated modes, for example end-to-end. Madhavan and Jarrett  have generated a fragment of the C-terminal domain, which on immunoblots exhibits binding to intact dystrophin, but there are no data on the affinity of this interaction. To establish definitively whether dystrophin has any capacity to self-associate in a specific manner is likely to require studies on the intact native protein. ACKNOWLEDGMENTS This work was made possible through the support of the Muscular Dystrophy Group of Great Britain. We are grateful to P.J. Marsh, Randall Institute, King’s College, for help and advice on expression of the proteins, to A.P.R. Brain of the Electron Microscopy Unit of King’s College for electron microscopy, and to A.C.M. Harris, National Institute for Medical Research, London, for N-terminal sequencing. 251 REFERENCES Ahn, A.H., and Kunkel, L.M. (1993): The structural and functional diversity of dystrophin. Nature Genet. 3:283–291. Davis, G.E., and Stark, G.R. (1970): Use of dimethylsuberimidate, a cross-linking reagent, in studying the subunit structure of oligomeric proteins. Proc. Natl. Acad. Sci. U.S.A. 66:651–657. England, S.B., Nicholson, L.V.B., Johnson, M.A., Forrest, S.M., Love, D.R., Zubrzycka-Gaarn, E.E., Bulman, D.E., Harris, J.B., and Davies, K.E. (1990): Very mild muscular dystrophy associated with the deletion of 46% of dystrophin. Nature 343:180–182. Flood, G., Kahana, E., Gilmore, A.P., Rowe, A.J., Gratzer, W.B., and Critchley, D.R. (1995): Association of structural repeats in the a-actinin rod domain. Alignment of inter-subunit interactions. J. Mol. Biol. 252:227–234. Greenfield, N., and Fasman, G.D. (1969): Computed circular dichroism spectra for the evaluation of protein conformation. Biochemistry 8:4108–4116. Halsall, H.B. (1967): Atassi-Gaudlin sedimentation coefficient and molecular weight relationship. Nature 215:880–882. Imamura, M., Endo, T., Kuroda, M., Tanaka, T., and Masaki, T. (1988): Substructure and higher structure of chicken smooth muscle a-actinin molecule. J. Biol. Chem. 263:7800–7805. Kahana, E., and Gratzer, W.B. (1991): Properties of the spectrin-like element of smooth muscle a-actinin. Cell Motil. Cytoskeleton 20:242–248. Kahana, E., and Gratzer, W.B. (1995): Minimum folding unit of dystrophin rod domain. Biochemistry 34:8110–8114. Kahana, E., Marsh, P.J., Henry, A.J., Way, M., and Gratzer, W.B. (1994): Conformation and phasing of dystrophin structural repeats. J. Mol. Biol. 235:1271–1277. Kake, T., Spissinger, T., Sato, O., Kimura, S., and Maruyama, K. (1995): Molecular shape of dystrophin with special reference to type-VI collagen. Proc. Jpn. Acad. Sci. Ser. B, Phys. Biol. Sci. 71:24–26. Koenig, M., and Kunkel, L.M. (1990): Detailed analysis of the repeat domain of dystrophin reveals four potential hinge segments that may confer flexibility. J. Biol. Chem. 265:4560–4566. Madhavan, R., and Jarrett, H.W. (1995): Interactions between dystrophin glycoprotein complex proteins. Biochemistry 34:12204– 12209. Minetti, C., Beltrame, F., Marcenaro, G., and Bonilla, E. (1992): Dystrophin at the plasma membrane of human muscle fibers shows a costameric localization. Neuromusc. Disord. 2:99–109. Ohlendieck, K. (1996): Towards an understanding of the dystrophinglycoprotein complex: Linkage between the extracellular matrix and the membrane cytoskeleton in muscle fibers. Eur. J. Cell Biol. 69:1–10. Perkins, S.J. (1986): Protein volumes and hydration effects. The calculation of partial specific volumes, neutron scattering matchpoints and 280-nm absorption coefficients for proteins and glycoproteins from amino acid sequences. Eur. J. Biochem. 158:169–180. Pons, F., Augier, N., Heilig, R., Léger, J., Mornet, D.E., and Léger, J-J. (1990): Isolated dystrophin molecules as seen by electron microscopy. Proc. Natl. Acad. Sci. U.S.A. 87:7851–7855. Porter, G.A., Dmytrenko, G.M., Winkelmann, J.C., and Bloch, R.J. (1992): Dystrophin colocalizes with b-spectrin in distinct subsarcolemmal domains in mammalian skeletal muscle. J. Cell. Biol. 117:997–1005. Provencher, S.W., and Glöckner, J. (1981): Estimation of globular protein secondary structure from circular dichroism. Biochemistry 20:33–37. Sato, O., Nonomura, Y., Kimura, S., and Maruyama, K. (1992): Molecular shape of dystrophin. J. Biochem. (Tokyo) 112:631–636. 252 Kahana et al. Shotton, D.M., Burke, B.E., and Branton, D. (1979): The molecular structure of human erythrocyte spectrin. Biophysical and electron microscopic studies. J. Mol. Biol. 131:303–329. Speicher, D.W., Weglarz, L., and DeSilva, T.M. (1992): Properties of human red cell spectrin heterodimer (side-to-side) assembly and identification of an essential nucleation site. J. Biol. Chem. 267:14775–14782. Straub, V., Bittner, R.E., Léger, J.J., and Voit, T. (1992): Direct visualization of the dystrophin network on skeletal muscle fiber membrane. J. Cell Biol. 119:1183–1191. Ursitti, J.A., Kotula, L., DeSilva, T.M., Curtis, P.J., and Speicher, D.W. (1966): Mapping the human erythrocyte b-spectrin dimer initiation site using recombinant peptides and correlation of its phasing with the a-actinin dimer site. J. Biol. Chem. 271:6636– 6644. Viel, A., and Branton, D. (1994): Interchain binding at the tail end of the Drosophila spectrin molecule. Proc. Natl. Acad. Sci. U.S.A. 91:10839–10843. Way, M., Pope, B., Gooch, J., Hawkins, M., and Weeds, A.G. (1990): Identification of a region in segment 1 of gelsolin critical for actin binding. EMBO J. 9:4103–4109. Winder, S.J., Gibson, T.J., and Kendrick-Jones, J. (1995): Dystrophin and utrophin: The missing link. FEBS Lett. 369:27–33. Winder, S.J., Gibson, T.J., and Kendrick-Jones, J. (1996): Low probability of dystrophin and utrophin coiled coil regions forming dimers. Biochem. Soc. Trans. 24:280S. Yan, Y., Winograd, E., Viel, A., Cronin, T., Harrison, S.C., and Branton, D. (1993): Crystal structure of the repetitive segments of spectrin. Science 262:2027–2030.