Mutagenesis at the a–b interface impairs the cleavage of the dystroglycan precursor Francesca Sciandra 1, *, Manuela Bozzi 2, *, Simona Morlacchi 1,3 , Antonio Galtieri 4 , Bruno Giardina 1,2 and Andrea Brancaccio 1 1 Istituto di Chimica del Riconoscimento Molecolare (CNR), c ⁄ o Istituto di Biochimica e Biochimica Clinica, Universita ` Cattolica del Sacro Cuore, Rome, Italy 2 Istituto di Biochimica e Biochimica Clinica, Universita ` Cattolica del Sacro Cuore, Rome, Italy 3 Dipartimento di Biologia Animale ed Ecologia Marina, Universita ` degli Studi di Messina, Italy 4 Dipartimento di Chimica Organica e Biologica, Universita ` di Messina, Italy Introduction Dystroglycan (DG) is a ubiquitous membrane-span- ning protein complex that was originally identified and characterized in rabbit skeletal muscle [1–3]. DG is expressed in skeletal and cardiac muscle, in the central Keywords alanine scanning; dystroglycan; dystroglycan precursor; laminin binding; post-translational processing Correspondence A. Brancaccio, Istituto di Chimica del Riconoscimento Molecolare (CNR), c ⁄ o Istituto di Biochimica e Biochimica Clinica, Universita ` Cattolica del Sacro Cuore, L.go F. Vito 1, 00168 Rome, Italy Fax: +39 6 3053598 Tel: +39 6 3057612 E-mail: andrea.brancaccio@icrm.cnr.it *These two authors contributed equally to this work (Received 29 April 2009, revised 10 June 2009, accepted 3 July 2009) doi:10.1111/j.1742-4658.2009.07196.x The interaction between a-dystroglycan (a-DG) and b-dystroglycan (b-DG), the two constituent subunits of the adhesion complex dystroglycan, is crucial in maintaining the integrity of the dystrophin–glycoprotein complex. The importance of the a–b interface can be seen in the skeletal muscle of humans affected by severe conditions, such as Duchenne muscular dystrophy, where the a–b interaction can be secondarily weakened or completely lost, causing sarcolemmal instability and muscular necrosis. The reciprocal binding epi- topes of the two subunits reside within the C-terminus of a-DG and the ectodomain of b-DG. As no ultimate structural data are yet available on the a–b interface, site-directed mutagenesis was used to identify which specific amino acids are involved in the interaction. A previous alanine-scanning analysis of the recombinant b-DG ectodomain allowed the identification of two phenylalanines important for a-DG binding, namely F692 and F718. In this article, similar experiments performed on the a-DG C-terminal domain pinpointed two residues, G563 and P565, as possible binding counterparts of the two b-DG phenylalanines. In 293-Ebna cells, the introduction of ala- nine residues instead of F692, F718, G563 and P565 prevented the cleavage of the DG precursor that liberates a- and b-DG, generating a pre-DG of about 160 kDa. This uncleaved pre-DG tetramutant is properly targeted at the cell membrane, is partially glycosylated and still binds laminin in pull- down assays. These data reinforce the notion that DG processing and its membrane targeting are two independent processes, and shed new light on the molecular mechanism that drives the maturation of the DG precursor. Structured digital abstract l MINT-7214494: alpha DG (uniprotkb:Q62165) binds (MI:0407)tobeta DG (uni- protkb: Q62165)bysolid phase assay (MI:0892) l MINT-7214516: laminin (uniprotkb:P19137) binds (MI:0407)tobeta DG (uniprotkb:Q62165) by pull down ( MI:0096) Abbreviations DG, dystroglycan; DGC, dystrophin–glycoprotein complex; EGFP, enhanced green fluorescent protein; WGL, wheat germ lectin. FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS 4933 and peripheral nervous system and in several epithelial tissues [3,4]. Homozygous null mice for the DG gene dag-1 die early during embryogenesis, at day E6.5, as a result of defects in Reichert’s membrane, the first extra-embryonic basement membrane deposited during murine development [5]. Indeed, DG plays a crucial role in the assembly of several basement membranes, promoting the recruit- ment of laminins and other extracellular matrix mole- cules during morphogenesis, tissue remodelling, cell polarization and wound healing [6–10]. DG is also implicated in the myelinization of nerves and in the sta- bilization of the neuromuscular junction [11,12]. More- over, in skeletal muscle, together with sarcoglycans, dystrobrevins, syntrophins and sarcospan, DG forms the dystrophin–glycoprotein complex (DGC), which connects the extracellular matrix to the actin cytoskele- ton, and is thought to offer stabilization to the muscle fibres during the contraction–relaxation cycle [13]. Although no primary genetic alterations of DG have been linked to human diseases to date, mutations in other components of the DGC are associated with dis- tinct forms of muscular dystrophy. Primary mutations in dystrophin, laminin-2 and any of the sarcoglycans cause Duchenne muscular dystrophy, congenital mus- cular dystrophy and limb-girdle muscular dystrophy, respectively [2]. In these forms of muscular dystrophy, DG membrane targeting and stability can be strongly perturbed. DG is composed of two interacting subunits, a and b, which are translated from a single mRNA molecule, generating a precursor protein of 895 residues that is post-translationally cleaved into the two noncovalently associated subunits [1]. The cleavage site is highly con- served among vertebrates and lies between residues G653 and S654 [14,15]. The detailed mechanism and functional significance of the post-translational pro- cessing of the DG precursor are still largely unknown, but experimental evidence has demonstrated its impor- tance for the correct function of DG. Indeed, a trans- genic mouse overexpressing the uncleaved precursor developed muscular dystrophy, and the expression of the noncleavable DG protein in neuroepithelial cells reduced their proliferation and differentiation in neurons [16,17]. b-DG is a transmembrane protein whose cytoplas- mic domain binds actin via the interaction with dystro- phin, and may act as a scaffold platform for signalling proteins interacting with the adaptor protein Grb2, but also with ezrin and extracellular signal-regulated kinase [18]. a-DG, in turn, is a peripheral protein char- acterized by a dumbbell-like structure with two globu- lar domains at the N- and C-termini, separated by an elongated central and highly glycosylated mucin-like domain [19]. a-DG binds with high affinity a variety of extracellular matrix molecules, such as laminin, agrin and perlecan. The reduction of the glycosylated shell of DG is thought to perturb its binding affinity towards extracellular matrix molecules [20]. Indeed, several forms of congenital muscular dystrophy are caused by mutations in a number of known or putative glycosyltransferases, leading to hypoglycosylation of a-DG in both skeletal muscle and brain [21]. However, a-DG retains contact with the plasma membrane through binding with b-DG, and the inter- action is independent of glycosylation [22,23]. The interaction between the two subunits involves the C-terminal domain of a-DG and the extracellular domain of b-DG, which belongs to the increasingly populated family of natively unfolded proteins, charac- terized by high conformational plasticity [22,24]. The reciprocal binding epitopes have been mapped between amino acids 550 and 565 of the C-terminal domain of a-DG and in the region located between the amino acid positions 691 and 719 of b-DG [24,25]. Recently, detailed mutagenesis analysis of the interaction between the two DG subunits identified two phenylala- nine residues (F692 and F718), belonging to the b-DG ectodomain, that are essential for the binding to a-DG in vitro [26]. In this study, extending the molecular analysis to the C-terminal portion of the b-DG binding epitope of a-DG [25], we identified some new residues that are important for the stability of the a–b inter- face. Results Alanine scanning of the b-DG binding epitope within the C-terminal domain of a-DG We have previously demonstrated that a linear amino acid sequence of 15 residues between positions 550 and 565 of a-DG is sufficient to interact with b-DG in experiments carried out with recombinant proteins [25]. Following these preliminary data, alanine scan- ning was performed on three amino acid positions belonging to the N-terminal portion of this linear sequence, namely W551, F554 and N555, in order to evaluate the contribution of each amino acid side- chain to the stability of the a–b interface [26]. As none of these three mutations seem to significantly affect the interaction with b-DG, we extended our alanine-scan- ning approach to the C-terminus of the 550–565 linear sequence. We expressed and purified a series of recombinant proteins spanning the C-terminal domain of a-DG, Mutagenesis induces an uncleaved dystroglycan F. Sciandra et al. 4934 FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS a-DG(485–630), carrying the following point muta- tions: S556A, Q559A, M561A, Y562A, G563A, L564A and P565A (Fig. 1). The affinity of each mutant towards the soluble recombinant biotinylated b-DG ectodomain, b-DG(654–750), was measured by solid- phase binding assays. Although solid-phase binding assays were carried out in nonequilibrium conditions, they provide apparent dissociation constants that are fully comparable with those measured with more accu- rate techniques, such as surface plasmon resonance [26]. a-DG(485–630) and its mutants a-DG(485–630) S556A, a-DG(485–630)Q559A, a-DG(485–630)M561A, a-DG(485–630)Y562A, a-DG(485–630)G563A, a-DG (485–630)L564A and a–DG(485–630)P565A were coated onto a microtitre plate, whereas biotinylated b-DG(654– 750) was used as a soluble ligand at increasing concentra- tions (up to 20 lm). The mutants a-DG(485–630)S556A, a-DG(485–630) Y562A and a-DG(485–630)L564A bind b-DG(654– 750) with the same affinity as the wild-type (see Fig. 2A), whereas a-DG(485–630)Q559A, a-DG(485– 630)M561A, a-DG(485–630)G563A and a-DG(485– 630)P565A show a slightly reduced affinity for b-DG(654–750), suggesting that these latter mutations might destabilize the a–b interface (see Fig. 2B). In Table 1, it can be seen that the lowest affinities (corre- sponding to the highest apparent dissociation constants) refer to the mutants a-DG(485–630)G563A and a-DG(485–630)P565A. In order to further validate these results, we have produced the double mutant a-DG(485–630)G563A-P565A and measured its affin- ity towards b-DG(654–750). The double substitution Fig. 1. Panel of mutants of murine DG fused to GFP. The a- and b-subunits of mammalian DG contain several well-conserved domains: (a) the N-terminal domain, the mucin-like region and the C-terminal region of a-DG, the latter containing the b-DG binding epi- tope (amino acids 550–565); (b) the ectodomain, the transmembrane region (TM) and the cytosolic domain of b-DG. The b-DG binding epi- tope was mutated by alanine scanning to produce the following mutants: S556A, Q559A, M561A, Y562A, G563A, L564A and P565A. A mutant deleted of the whole b-DG binding epitope (DGD550–565) was also generated. All the mutations were intro- duced into the wild-type murine DG cDNA sequence and cloned into a pEGFP vector for cell transfection experiments, or introduced into a plasmid, allowing quantitative expression of recombinant C-termi- nal a-DG peptides in E. coli cells (see Experimental procedures). A B C Fig. 2. Solid-phase binding assays. a-DG(485–630) (black) and its mutants, a-DG(485–630)S556A (red), a-DG(485–630)Y562A (green), a-DG(485–630)L564A (blue) (A), a-DG(485–630)Q559A (red), a-DG (485–630)M561A (green), a-DG(485–630)G563A (blue), a-DG(485– 630)P565A (magenta) (B) and a-DG(485–630)G563A-P565A (red) (C), were coated onto a microtitre plate, whereas biotinylated b-DG(654–750) was used as a soluble ligand at increasing concen- trations. Each continuous line corresponds to a representative experiment (from a set of at least three experiments with similar results), and was obtained by fitting experimental data to a single class of equivalent binding sites equation (see Experimental procedures). F. Sciandra et al. Mutagenesis induces an uncleaved dystroglycan FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS 4935 of both G563 and P565 with alanine completely inhib- ited the interaction between a-DG(485–630) and b-DG(654–750), at least in the ligand concentration range explored (Fig. 2C). Interestingly, G563 and P565 are also fully conserved in DGs from phylogenetically distant species (Fig. S1, see Supporting information). These results indicate that, together, G563 and P565 might significantly contribute to the a–b interface and to the stability of the whole DG complex. Transfection of 293-Ebna cells with mutated DGs: western blot and fluorescence microscopy In order to analyse in eukaryotic cells the effects of the point mutations that impair the interaction between a- and b-DG, the same mutations tested in solid-phase binding assays were introduced within the entire mur- ine DG cDNA, which was cloned into the pEGFP vec- tor and used to transiently transfect 293-Ebna cells. Enhanced green fluorescent protein (EGFP) was fused at the C-terminal region of b-DG to increase its molec- ular mass by 25 kDa; the presence of GFP allows endogenous b-DG to be distinguished unambiguously from exogenous b-DG-EGFP in western blot analysis. Western blot of total protein extracts of cells overex- pressing DG-EGFP constructs carrying the single point mutations S556A, Q559A, M561A, G563A, L564A and P565A confirmed the presence of the expected 68 kDa band corresponding to exogenous b-DG-EGFP when the samples were probed with both anti-b-DG and anti-EGFP IgG (Fig. 3A,B). However, G563A displayed an additional faint band of about 100 ⁄ 200 kDa (Fig. 3A,B). Interestingly, the same band was also detectable in the two double mutants, G563A ⁄ P565A and F692A ⁄ F718A (Fig. 3C– E). The latter mutant hits the two phenylalanines belonging to the b-DG ectodomain, F692A and F718A, that have been shown previously to be key residues for binding with a-DG in vitro [26]. This higher band is likely to correspond to the unprocessed DG precursor (hereafter pre-DG), as the mutation S654A, located at the physiological a ⁄ b maturation cleavage site G653– S654, produces a single band with a molecular weight estimated at 160 kDa that has the same electrophoretic mobility as displayed by the double mutants G563A ⁄ P565A and F692A ⁄ F718A (Fig. 3C,D) [16,27,28]. On the basis of these results, we hypothesized that perturbation of the network of interactions that is likely to stabilize the a–b interface within the DG com- plex may interfere with the cleavage of the DG precur- sor. To further validate this hypothesis, we generated two additional constructs, one carrying the four mutations G563A, P565A, F692A and F718A, DGG563A_P565A_F692A_F718A, and the second with deletion of the whole b-DG binding epitope between amino acids 550 and 565 within the C-termi- nal domain of a-DG, DGD550–565 [25]. As expected, the products of both constructs appeared on SDS- PAGE as a single 160 kDa band, albeit less intense than that observed for the mutant S654A, indicating an instability and a major susceptibility to degradation of the former mutant pre-DGs (Fig. 3C–E). A possible scale in the amounts of pre-DG is as fol- lows: DGS654A > DGG563A_P565A_F692A_F718A > DG(D550–565) > DGF692AF718A ‡ DGG563A P565A > DGG563A (Fig. 3E). Fluorescence microscopy analysis showed that the DG precursors are likely to be properly targeted at the plasma membrane, as cells expressing the un- cleavable DG mutants are indistinguishable from those expressing wild-type DG (Fig. 4). In addition, the quadruple mutation G563A ⁄ P565A ⁄ - F692A ⁄ F718A and the deletion of the 550–565 region did not significantly affect the trafficking or membrane targeting of pre-DG (Fig. 4). The diffused and punctuated label throughout the cytoplasm and around the plasma membrane, featured by cells transfected with both wild-type and mutated DG, was probably a result of overexpression of exogenous EGFP-tagged proteins. Wheat germ lectin (WGL)-driven enrichment of mutant pre-DGs The DG gene encodes a unique polypeptide precur- sor consisting of 895 amino acids with a calculated Table 1. Apparent equilibrium dissociation constants (K D ) calcu- lated by solid-phase binding assays. Mean apparent K D values and relative standard deviations, calculated for the interaction between a-DG(485–630) and its mutants and b-DG(654–750) in solid-phase binding assays. The values were averaged over a number of inde- pendent experiments indicated in parentheses. For the a-DG(485– 630) double mutant, showing a strongly reduced affinity towards b-DG(654–750), the K D value could not be calculated (n.d.; see Experimental procedures). Immobilized protein ⁄ biotinylated protein K D,app (lM) a-DG wt ⁄ b-DG wt 3.3 ± 1.0 (6) a-DG(S556A) ⁄ b-DG wt 3.2 ± 1.2 (4) a-DG(Q559A) ⁄ b-DG wt 4.3 ± 0.3 (3) a-DG(M561A) ⁄ b-DG wt 4.5 ± 1.4 (3) a-DG(Y562A) ⁄ b-DG wt 3.1 ± 0.8 (5) a-DG(G563A) ⁄ b-DG wt 4.7 ± 0.6 (3) a-DG(L564A) ⁄ b-DG wt 3.5 ± 1.4 (3) a-DG(P565A) ⁄ b-DG wt 5.5 ± 1.2 (4) a-DG(G563A–P565A) ⁄ b-DG wt n.d. (3) Mutagenesis induces an uncleaved dystroglycan F. Sciandra et al. 4936 FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS molecular mass of about 98 kDa. Based on their apparent mobility on SDS-PAGE, pre-DGG 563A_P565A, pre-DGG563A_P565A_F692A_F718A, pre-DG(D550–565) and pre-DGS654A should be highly, or at least partially, glycosylated. In order to further clarify this aspect, total protein extracts obtained from 293-Ebna cells transfected with the uncleavable DG mutants were incubated with aga- rose-immobilized WGL that specifically binds N-acet- ylglucosamine residues (Fig. 5A). All the pre-DG mutants were pulled down and enriched by this pro- cedure, suggesting the presence of N-acetylglucos- amine moieties within the uncleaved precursors (Fig. 5B). Densitometric analysis confirmed the minor stability of preDGG563A_P565A, preDGG563A_ P565A_F692A_F718A and preDG(D550–565) when compared with pre-DGS654A (Fig. 5C). Laminin binding properties of mutant pre-DGs DG serves as a receptor for a variety of extracellular ligands, such as laminin, agrin and perlecan. Full chemi- cal deglycosylation of a-DG in vitro is known to disrupt its ability to bind other extracellular matrix proteins [29,30]. Therefore, laminin conjugated to Sepharose beads was used to test the capacity of the mutant pre-DGs (pre-DGG563A_P565A_F692A_F718A, pre- DG D550–565 and pre-DGS654A) to interact with com- mercial mouse laminin-1 (Fig. 6A). Mutant pre-DGs remained bound to laminin even after several washing steps (Fig. 6B). This interaction was inhibited using EDTA, suggesting that the binding between laminin and the mutant pre-DGs is reversible and dependent on divalent calcium cations, as expected for the laminin– DG interaction (data not shown) [29,31]. A B C D E Fig. 3. Western blot of total protein extracts. 293-Ebna cells were transfected with DG mutants and their protein extracts were probed with anti-b-DG (anti-43-DAG) (A and C) or anti-GFP (B and D). The DG mutants carrying the point mutations S556A, Q559A, M561A, G563A, L564, P565A display a single band corresponding to the cleaved b-DG-GFP (A and B). The double DG mutants, G563A ⁄ P565A and F692A ⁄ F718A, show the presence of an additional higher band at 160 kDa that is likely to correspond to the unprocessed pre-DG (C and D). Pre-DG is also expressed in the presence of the mutation S654A, the quadruple mutation G563A ⁄ P565A ⁄ F692A ⁄ F718A and the deletion of the entire b-DG binding epitope between amino acids 550–565 (C and D). A lower band, at about 50 kDa, probably originates from further pro- teolysis of b-DG-GFP. The black boxes indi- cate pre-DG-GFP and b-DG-GFP. The amounts of cleaved DG-GFP (open bars) and unprocessed pre-DG (filled bars) were quan- tified by densitometry, averaging the values of the band intensities obtained from five independent experiments (E). Such a quanti- tative analysis shows how the differences in band intensities between some of the mutants are not significant. F. Sciandra et al. Mutagenesis induces an uncleaved dystroglycan FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS 4937 Discussion Towards the identification of single amino acids within the C-terminal region of a-DG that are crucial for the interaction with b-DG In this study, we focused alanine scanning on the a-DG amino acid linear sequence 550–565 in an attempt to identify which residues were responsible for binding with b-DG. A series of point mutations, S556A, Q559A, M561A, Y562A, G563A, L564A and P565A, was introduced into the recombinant protein a-DG(485–630), and their affinities towards recombi- nant biotinylated b-DG(654–750) were measured. The K D values reported in Table 1 show that only the point mutations hitting odd positions (i.e. G563A and P565A) elicit some slight effect on the interaction with b-DG, whereas those at even positions (i.e. S556A, Y562A and L564A) do not produce any effect (see Fig. 2A,B and Table 1). Moreover, the double muta- tion G563A ⁄ P565A completely inhibits the interaction between a- and b-DG (Fig. 2C). The interaction between a- and b-DG seems to induce some local secondary structures. Indeed, our results may suggest that the a-DG linear sequence QLMYGLP assumes a b-strand conformation, with the amino acids Q559, M561, G563 and P565 pointing towards the b-DG ectodomain and interacting with it. Retrospectively, our previous NMR experiments, car- ried out by exploiting the synthetic peptide a-DG(550– 585) in free and b-DG-bound fashion, suggested a greater involvement of Q559, M561 and G563 than of S558, L560, Y562 and L564 in binding the recombi- nant b-DG ectodomain; the alternate fashion of these side-chain contributions could indeed be reminiscent of a b-strand conformation (see Fig. 4 of [25]). Further- more, the hypothesis that the QLMYGLP amino acid stretch would assume a b-strand conformation is corroborated by a model of the a-DG C-terminal domain based on sequence homology with a member of the cadherin family [32]. However, further experi- ments are needed to validate this hypothesis. Interestingly, the crucial importance of G563 and P565 could be deduced from the analysis of a multiple alignment of DG sequences from species phylogeneti- cally distant from humans or mouse, including lower vertebrate and several invertebrate species (Fig. S1, see Supporting information), where these amino acids are always conserved despite a very low overall sequence homology. Our new data on G563 and P565, together with the results of our previous study, in which two phenylalanines belonging to the b-DG ectodomain, F692 and F718, were recognized as key residues for the interaction with a-DG, point towards the identifi- cation of the major molecular cornerstones of the a–b interface. Fig. 4. Immunofluorescence of 293-Ebna cells transfected with the pEGFP vector, empty or carrying wild-type or mutated DGs. All the uncleavable mutants are expressed and targeted to the plasma membrane (open arrowheads), showing a fluorescence pattern similar to that of wild-type DG (WT). GFP was expressed throughout the cytoplasm. Mutagenesis induces an uncleaved dystroglycan F. Sciandra et al. 4938 FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS The a–b interface is essential for the correct cleavage of the DG precursor A heterologous cell expression system was used to verify whether the mutations analysed in vitro might also influence the expression and stability of DG in cells. 293-Ebna cells were transfected with the entire DG gene carrying the single mutations, S556A, Q559A, M561A, G563A, L564A and P565A, and cloned into a pEGFP vector. As demonstrated by western blot of total cell extracts, the single point mutations do not drastically alter the stability of DG, which is correctly processed into the two subunits (Fig. 3A,B). Only the mutant G563A showed an addi- tional faint band at about 160 kDa, probably caused by a small amount of the uncleaved DG precursor, pre-DG, which spans both the a- and b-subunits of DG (Fig. 3A,B,E). Interestingly, the two DG A B C Fig. 5. WGL enrichment of total protein extracts of untransfected cells (NT) and cells transfected with wild-type or mutated DGs. (A) Assay rationale: WGL specifically binds to the N-acetylglucosamine moieties covalently linked to the core protein of a-DG. Therefore, b-DG-GFP, which is noncovalently associated with a-DG (or directly pre-DG), can be retained from the immobilized WGL molecules. (B) Western blot carried out with the anti-b-DG IgG clearly shows that both wild-type (b-DG-GFP) and mutant (b-DG-GFP and mainly pre-DG) DG proteins can be specifically eluted by WGL beads. Only the eluted fractions, collected upon extensive washing, were loaded onto the gel; the wash fractions did not contain any rele- vant signal (data not shown). (C) The amounts of cleaved DG-GFP and unprocessed pre-DG were quantified by densitometry, averag- ing the values of the band intensities obtained from three indepen- dent experiments. A B Fig. 6. Laminin-Sepharose pull-down of mutant pre-DGs. (A) Assay rationale: laminin, covalently bound to CNBr-Sepharose, interacts with a-DG. b-DG-GFP is retained by laminin-Sepharose beads through the interaction with the a-subunit. (B) Pull-down of wild- type (WT) DG or DG carrying the mutations indicated on the spe- cific panels. Western blot carried out with the anti-b-DG IgG clearly shows that both wild-type and mutant DG proteins specifically bind laminin (lane E: elution fraction) after extensive washing (lanes W1 and W5); FT (flow-through). The black boxes indicate b-DG-GFP and pre-DG-GFP. F. Sciandra et al. Mutagenesis induces an uncleaved dystroglycan FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS 4939 constructs carrying the double mutations, G563A ⁄ P565A and F692A ⁄ F718A, which, in solid- phase assays completely inhibit the binding between a- and b-DG (Fig. 2C) [26], display significant amounts of pre-DG, with respect to the correctly cleaved b-DG-EGFP (Fig. 3C–E). The correct cleavage is completely inhibited in the DG construct carrying the four mutations G563A, P565A, F692A and F718A, (Fig. 3C–E), suggesting that interaction between the reciprocal binding epi- topes of the two subunits forming the mature DG complex is necessary for correct processing of the DG precursor. Consistent with this hypothesis, the entire deletion (knockin) of the b-DG binding epitope within the a-DG subunit (positions 550–565) abolishes the cleavage, producing the 160 kDa pre-DG (Fig. 3C–E). The mutants pre-DGG563A_P565A_F692A_F718A, pre-DG D550–565 and pre-DGS654A specifically bind WGL, which indicates that they are at least partially glycosylated (Fig. 5B); furthermore, laminin pull-down experiments show that mutated DG precursors har- bour some laminin binding epitopes (Fig. 6B), clearly indicating a residual functionality of hypoglycosylated and unprocessed pre-DGs. Depicting a possible model for DG precursor processing The mechanism and functional significance of DG pre- cursor processing still remain largely elusive. In several human and murine cell lines and tissues, DG was always detected as a heteromeric complex, suggesting that precursor cleavage is a very early post-transla- tional event along the route of DG maturation. Muta- tions in the amino acid sites crucial for the interaction between a- and b-DG, namely G563 and P565 within the C-terminal domain of a-DG and their counterparts F692 and F718 within the b-DG ectodomain, ‘freeze’ the DG precursor as a relatively stable and partially glycosylated monomeric intermediate. Our results strongly suggest that the network of interactions important for the build up of the a–b interface on pre- cursor cleavage is already established within the unc- leaved DG precursor and is strictly necessary for processing into the two subunits. The impairment of the correct formation of the a–b interface may destabi- lize pre-DGs; indeed, both pre-DGG563A_P565A_ F692A_F718A and pre-DG(D550–565) display lower expression levels compared with pre-DGS654A, in which most of the interactions underlying the a–b interface are still likely to take place (Fig. 3E). Such a network of interactions may also influence the glyco- sylation pattern of the DG precursor. This could be inferred from the different electrophoretic behaviour displayed by the uncleavable pre-DGs. Indeed, pre- DG S654A displays in western blot as a broader band (which could imply the presence of more carbohydrate groups) with respect to pre-DGG563A_P565A_ F692A_F718A and pre-DG(D550–565), where most of the a–b interactions cannot be established (Fig. 3C,D). The correct folding of the DG precursor may there- fore be important for the recognition by glyco- syltransferases, which should primarily take place at the level of the N-terminal portion of a-DG [33,34]. In particular, O-glycosyltransferases are thought to be crucial, especially for the extensive sugar decoration of the DG central mucin-like domain [19]. Apart from the correct folding of what could be defined as the ‘pre-a–b interface’, a few other factors have been proposed to play an important role in DG precursor processing: for example, the disulfide bridge between C669 and C713, within the b-DG ectodomain [35], and N-glycosylation [27]. The formation of this disulfide bridge may also contribute to the stabilization of the correct folding of the DG precursor necessary for spe- cific cleavage. As far as N-glycosylation is concerned, it has been shown by others that alanine substitution of N662, a putative N-glycosylation site in the b-DG ectodomain, prevents the cleavage of the precursor and strongly reduces its expression [27]. However, whether N-glycosylation really influences DG precursor cleav- age is still a matter of debate: other studies have shown that blocking N-glycosylation does not prevent cleavage [36]. On the basis of our data and other evidence from the literature, we propose the following scenario for DG maturation (shown in Fig. 7): immediately after translation, the DG core protein is translocated into the endoplasmic reticulum, where it is likely to adopt a stable three-dimensional conformation prior to any post-translational modifications. At this stage, an essential contribution for achieving a confor- mation that will allow subsequent cleavage is pro- vided by a network of interactions (in which G563, P565, F692 and F718 play a crucial role) that are likely to stabilize the mature a–b interface also on cleavage. It is still unclear whether cleavage is carried out by an unidentified protease or whether it occurs via an autocatalytic mechanism [28]. However, our data clearly show that precursor cleavage is dispensable for correct trafficking and membrane targeting of DG, as all our novel uncleavable mutants can be detected at the plasma membrane, and their localization is indis- tinguishable from that characterizing wild-type DG (Fig. 4); furthermore, they are still capable of binding Mutagenesis induces an uncleaved dystroglycan F. Sciandra et al. 4940 FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS laminin, fulfilling one of the most important functions of DG (Fig. 6). Conclusions Based on the available evidence, during evolution, there was a ‘free choice’ for the liberation, or not, of the two DG subunits. For example, in Caenorhabditis elegans, in which not only DG but a whole DGC orthologue has been identified and functionally charac- terized [37], the maturation of the DG complex into two subunits has not been observed and, accordingly, the motif Gly-Ser (653–654) at the cleavage site has not been conserved [38]. Clearly, further work is needed in order to fully understand the biological sig- nificance of why the two DG subunits are liberated [39]. How an abnormal a–b interface would affect human DG function is not yet known, as no primary muta- tions of the dag1 gene have been identified so far. However, in principle, it should be possible to find spe- cific mutations, or more likely polymorphisms, which, in mammals, would interfere with DG processing with- out grossly impairing DG function and displaying very mild phenotypic signs in virtually asymptomatic carri- ers. This is suggested by recent papers from other laboratories showing that DG does not take part in the later stages of embryonic development, or that hy- poglycosylated DG can be partially functional [40–42]. However, it will also be important to rule out the pos- sibility that the presence of an uncleaved DG precur- sor may, instead, be linked to severe neuromuscular pathologies. A concerted effort of biochemical, genetic and clini- cal studies is needed in order to finally address these points. At the present stage, our identification of mul- tiple point mutations that inhibit or affect the DG maturation pathway may provide a useful tool to investigate and shed light on the molecular details of such an important and mysterious process. Experimental procedures DNA manipulation The full-length cDNA encoding for murine DG was used as a template to generate, by PCR, two DNA constructs, one corresponding to the N-terminal region of b-DG, b-DG(654–750), and the other to the C-terminal region of a-DG, a-DG(485–630) [22]. Appropriate primers were used to amplify the DNA sequences of interest. For b-DG(654– 750): forward, 5¢-CCCGGATCCTCTATCGTGGTGG AATGGACCAACA-3¢; reverse, 5¢-CCCGAATTCTTAG TAAACATCGTCCTCACTGCTCTCTTC-3¢ (BamHI and EcoRI restriction sites are given in italic type). For a-DG(485–630): forward, 5¢-CCCGTCGACAGTGGAGTG CCCCGTGGGGGAGAAC-3¢; reverse, 5¢-CCCGAATTC TTATACCAAAGCAATTTTTCTTGTGAATG-3¢ (SalI and EcoRI restriction sites are given in italic type). Single point mutations were introduced into the murine DG gene, cloned into the pEGFP vector, using the QuikChange site- PreDG AB C PreDG Fig. 7. Schematic model showing the influ- ence of the a–b interface on pre-DG cleav- age. (A) In the wild-type pre-DG, the correct interaction between the a- and b-domains stabilizes pre-DG in a conformation that can be proteolytically processed at its G ⁄ S cleavage site, liberating the a- and b-subun- its. The black double-headed arrow indicates the pre-a–b interface. (B) When S654, part of the cleavage site, is mutated, pre-DG is not proteolytically processed. It is possible that the interactions within the a–b interface are formed even in the uncleaved precursor, ensuring a certain stability of pre-DG. The black double-headed arrow indicates the pre-a–b interface. (C) When the a–b inter- face is impaired by specific mutations hitting the amino acids G563, P565 (within a-DG), F692 and F718 (within b-DG), pre-DG does not reach a conformation suitable for proteo- lytic cleavage. F. Sciandra et al. Mutagenesis induces an uncleaved dystroglycan FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS 4941 directed mutagenesis kit (StratageneÒ, Cedar Creek, TX, USA); all constructs were verified by automated sequenc- ing. The primers used for mutagenesis are reported below with the mutated codons in italic: S556A forward: 5¢-TGGGTTCAGTTTAACGCCAACA GCCAGCTCATG-3¢ S556A reverse: 5¢-CATGAGCTGGCTGTTGGCGTTA AACTGAACCCA-3¢ Q559A forward: 5¢-TTTAACAGCAACAGCGCGCTC ATGTATGGCCTG-3¢ Q559A reverse: 5¢-CAGGCCATACATGAGCGCGCT GTTGCTGTTAAA-3¢ M561A forward: 5¢-AGCAACAGCCAGCTCGCGTAT GGCCTGCCTGAC-3¢ M561A reverse: 5¢-GTCAGGCAGGCCATACGCGAG CTGGCTGTTGCT-3¢ Y562A forward: 5¢-AACAGCCAGCTCATGGCT GGCCTGCCTGACAGC-3¢ Y562A reverse: 5¢-GCTGTCAGGCAGGCCAGCCAT GAGCTGGCTGTT-3¢ G563A forward: 5¢-AGCCAGCTCATGTATGCCCTG CCTGACAGCAGC-3¢ G653A reverse: 5¢-GCTGCTGTCAGGCAGGGCATA CATGAGCTGGCT-3¢ L564A forward: 5¢-CAGCTCATGTATGGCGCGCCTG ACAGCAGCCAT-3¢ L564A reverse: 5¢-ATGGCTGCTGTCAGGCGCGCC ATACATGAGCTG-3¢ P565A forward: 5 ¢-CTCATGTATGGCCTGGCTGAC AGCAGCCATGTG-3¢ P565A reverse: 5¢-CACATGGCTGCTGTCAGCCAG GCCATACATGAG-3¢ S654A forward: 5¢-CAGAACATCACTCGGGGCGC TATCGTGGTGGAATGGACC-3¢ S654A reverse: 5¢-GGTCCATTCCACCACGATAGCGC CCCGAGTGATGTTCTG-3¢ G563AP565A forward:5¢-AGCCAGCTCATGTATG CCCTGGCTGACAGCAGC-3¢ G563AP565A reverse: 5¢-GCTGCTGTCAGCCAGGG CATACATGAGCTGGCT-3¢ The full-length DNA constructs carrying the point muta- tions were also used as templates to generate, by PCR, the DNA constructs for the expression of the a-DG(485–630) mutants in the Escherichia coli recombinant system (see below), employing the same primers as used to amplify the wild-type a-DG(485–630) sequence. For the production of the DG(D550–565) deletion mutant, the knocked-in DNA construct was generated by the overlap extension method [43] using 5¢-CCCGAAT TCATGTCTGTGGACAACTGGCTACTG-3¢ and 5¢- TTTCTCACCTACTAACTGCTGCTCT-3¢ as forward and reverse primers, respectively, for the first PCR, and 5¢- CAGTTAGTAGGTGAGAAAGACAGCAGCCATGTG-3¢ and 5¢-CCCGAATTCGGCTAGGGGGAACATACGGAG GGGG-3¢ for the second PCR. Protein expression, purification and biotinylation The DNA constructs were cloned into a bacterial vector that was appropriate for the expression of the protein as a thioredoxin fusion product, also containing an N-terminal 6His tag and a thrombin cleavage site [44]. The recombi- nant fusion proteins were expressed in E. coli BL21(DE3) Codon Plus RIL strain and purified using nickel affinity chromatography. The fragments of interest were obtained on thrombin cleavage. Tricine ⁄ SDS-PAGE was used to check the purity of the recombinant proteins under analy- sis. For solid-phase binding assays, recombinant b-DG(654–750) was biotinylated in 5 mm sodium phos- phate buffer at pH 7.4, with 0.5 mgÆmL )1 sulfo-N-hydroxyl- succinimido-biotin (S-NHS-biotin, PierceÒ, Rockford, IL, USA). The reaction was carried out for 30 min on ice and in the dark, and dialysed overnight against 10 mm Tris ⁄ HCl, 150 mm NaCl, pH 7.4. The optimal dilution for signal detection was determined by dot blot analysis and revealed by enhanced chemiluminescence (PierceÒ). Solid-phase binding assays To assess the binding properties of recombinant a-DG(485– 630) and its mutants with respect to biotinylated recombi- nant b-DG(654–750), solid-phase assays were performed as follows: approximately 0.5 lgofa-DG(485–630), its mutants and BSA were immobilized on microtitre plates in coating buffer (50 mm NaHCO 3 , pH 9.6) overnight at 4 °C. After washing with NaCl ⁄ P i buffer (2.5 mm KCl, 2 mm KH 2 PO 4 ,2mm Na 2 HPO 4 , 140 mm NaCl, pH 7.4) contain- ing 0.05% (v ⁄ v) Tween-20, 1.25 mm CaCl 2 and 1 mm MgCl 2 , wells were incubated with decreasing concentrations of recombinant biotinylated b-DG(654–750) in NaCl ⁄ P i containing 0.05% (v ⁄ v) Tween-20, 3% (w ⁄ v) BSA, 1.25 mm CaCl 2 and 1 mm MgCl 2 for 3 h at room temperature. After washing, the biotinylated b-DG(654–750) bound fraction was detected with alkaline phosphatase Vectastain AB Complex (Vector LaboratoriesÒ, Burlingame, CA, USA). Five milligrams of p-nitrophenyl phosphate dissolved in 10 mL of 10 mm diethanolaminine and 0.5 m MgCl 2 were added to every well containing 100 lL of this solution, and used as a substrate for the reaction with alkaline phospha- tase; the absorbance values were recorded at 405 nm. For each b-DG(654–750) concentration, the absorbance value (A i ) originating from coated BSA was subtracted from the values obtained with the coated wild-type or mutated a-DG samples under analysis. The data were fitted using a single class of equivalent binding sites equation, A i = A- sat [c ⁄ (K D + c)+A 0 ], where A i represents the absorbance measured at increasing concentrations of ligand, K D is the dissociation constant, c is the concentration of ligand, bioti- nylated b-DG(654–750), and A sat and A 0 are the absor- bances at saturation and in the absence of ligand, respectively. Data were normalized and reported as the Mutagenesis induces an uncleaved dystroglycan F. Sciandra et al. 4942 FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS [...]... (2008) Comparative study of a -dystroglycan glycosylation in dystroglycanopathies suggests that the hypoglycosylation of a -dystroglycan does not consistently correlate with clinical severity Brain Pathol doi:10.1111/j.1750-3639.2008.00198 42 Puckett R, Moore SA, Winder TL, Willer T, Romansky SG, Covault KK, Campbell KP & Abdenur JE (2009) Further evidence of Fukutin mutations as a cause of childhood... boundary · pixel area’ and volume units are calculated as ‘intensity units · mm2’ The volumes of at least three independent experiments were normalized with respect to the endogenous DG volume and used to calculate the mean values Data were exported to MicrosoftÔ Excel to generate the plot Preparation of laminin-Sepharose Preactivated CNBr Sepharose (Vector Laboratories) was suspended in 1 mm HCl for 15 min... characterization by NMR of the natively unfolded extracellular domain of b -dystroglycan: toward the identification of FEBS Journal 276 (2009) 4933–4945 ª 2009 The Authors Journal compilation ª 2009 FEBS F Sciandra et al 25 26 27 28 29 30 31 32 33 34 35 36 37 the binding epitope for a -dystroglycan Biochemistry 42, 13717–13724 Bozzi M, Veglia G, Paci M, Sciandra F, Giardina B & Brancaccio A (2001) A synthetic...F Sciandra et al fractional saturation (%): 100· [(Ai ) A0) ⁄ (Asat ) A0)] For the double mutant a-DG(485–630)G563A-P565A, which displayed a strong reduction in binding affinity, the data could not be fitted according to the equation above, and were simply normalized, setting the maximal binding of wild-type a-DG(485–630), extrapolated by the fitting, as 100% The KD values reported in Table 1 were... Biology of the striated muscle dystrophin–glycoprotein complex Int Rev Cytol 265, 191–225 Smalheiser NR & Kim E (1995) Purification of cranin, a laminin binding membrane protein Identity with dystroglycan and reassessment of its carbohydrate moieties J Biol Chem 270, 15425–15433 Deyst KA, Bowe MA, Leszyk JD & Fallon JR (1995) The a -dystroglycan b -dystroglycan complex Membrane organization and relationship... structure of dystrophin-associated glycoproteins linking dystrophin to the extracellular matrix Nature 355, 696–702 2 Barresi R & Campbell KP (2006) Dystroglycan: from biosynthesis to pathogenesis of human disease J Cell Sci 119, 199–207 3 Cohn RD (2005) Dystroglycan: important player in skeletal muscle and beyond Neuromuscul Disord 15, 207–217 4 Durbeej M & Campbell KP (1999) Biochemical characterization of. .. N-terminal region of murine skeletal muscle a -dystroglycan discloses a modular architecture J Biol Chem 279, 44812–44816 Watanabe N, Sasaoka T, Noguchi S, Nishino I & Tanaka T (2007) Cys669–Cys713 disulfide bridge formation is a key to dystroglycan cleavage and subunit association Genes Cells 12, 75–88 Holt KH, Crosbie RH, Venzke DP & Campbell KP (2000) Biosynthesis of dystroglycan: processing of a precursor. .. Martin PT (2003) Inhibition of dystroglycan cleavage causes muscular dystrophy in transgenic mice Neuromuscul Disord 13, 365–375 Schroder JE, Tegeler MR, Grobhans U, Porten E, ¨ Blank M, Lee J, Esapa C, Blake DJ & Kroger S (2007) ¨ Dystroglycan regulates structure, proliferation and differentiation of neuroepithelial cells in the developing vertebrate CNS Dev Biol 307, 62–78 Batchelor CL, Higginson JR,... Densitometric analyses of different films were performed using a GS-800 imaging densitometer (Bio-Rad, Hercules, CA, USA) and analysed using Bio-Rad Quantity-One software The signal intensities of cleaved b-DG and unprocessed pre-DG bands were determined using volume analysis with object average background correction applied The volume is defined as the ‘sum of the intensities of the pixels within the volume boundary... h), the cells were fixed with 4% (v ⁄ v) paraformaldehyde at room temperature for 30 min and observed under a fluorescence microscope (NikonÒ, Tokyo, Japan) About 20 lg of wild-type or mutated pEGFPDG were also transfected using the calcium phosphate method: briefly, DNA was mixed with 125 mm CaCl2 and 50 mm Bes (Bes-buffered saline) The DNA–calcium phosphate complex was added to the cells After 24 h, the . [16,27,28]. On the basis of these results, we hypothesized that perturbation of the network of interactions that is likely to stabilize the a–b interface within the. Mutagenesis at the a–b interface impairs the cleavage of the dystroglycan precursor Francesca Sciandra 1, *, Manuela