Assembly defects of desmin disease mutants carrying deletions in the α-helical rod domain are rescued by wild type protein

Journal of Structural Biology 158 (2007) 107–115 www.elsevier.com/locate/yjsbi

Assembly defects of desmin disease mutants carrying deletions in the -helical rod domain are rescued by wild type protein Harald Bär a,¤, Norbert Mücke b, Hugo A. Katus a, Ueli Aebi c, Harald Herrmann d a

b

Department of Cardiology, University of Heidelberg, D-69120 Heidelberg, Germany Division of Biophysics of Macromolecules, German Cancer Research Center (DKFZ), D-69120 Heidelberg, Germany c Maurice E. Müller Institute for Structural Biology, Biozentrum, University of Basel, CH-4056 Basel, Switzerland d Department of Molecular Genetics, German Cancer Research Center (DKFZ), D-69120 Heidelberg, Germany Received 26 September 2006; accepted 30 October 2006 Available online 10 November 2006

Abstract Most mutations of desmin that cause severe autosomal dominant forms of myoWbrillar myopathy are point mutations and locate in the central -helical coiled-coil rod domain. Recently, two in-frame deletions of one and three amino acids, respectively, in the -helix have been described and discussed to drastically interfere with the architecture of the desmin dimer and possibly also the formation of tetramers and higher order complexes [Kaminska, A., Strelkov, S.V., Goudeau, B., Olive, M., Dagvadorj, A., Fidzianska, A., Simon-Casteras, M., Shatunov, A., Dalakas, M.C., Ferrer, I., Kwiecinski, H., Vicart, P., Goldfarb, L.G., 2004. Small deletions disturb desmin architecture leading to breakdown of muscle cells and development of skeletal or cardioskeletal myopathy. Hum. Genet. 114, 306–313.]. Therefore, it was proposed that they may poison intermediate Wlament (IF) assembly. We have now recombinantly synthesized both mutant proteins and subjected them to comprehensive in vitro assembly experiments. While exhibiting assembly defects when analyzed on their own, both one-to-one mixtures of the respective mutant protein with wild type desmin facilitated proper Wlament formation. Transient transfection studies complemented this fundamental Wnding by demonstrating that wild type desmin is also rescuing these assembly defects in vivo. In summary, our Wndings strongly question the previous hypothesis that it is assembly incompetence due to molecular rearrangements caused by the mutations, which triggers the development of disease. As an alternative, we propose that these mutations cause subtle agedependent structural alterations of desmin IFs that eventually lead to disease. © 2006 Elsevier Inc. All rights reserved. Keywords: MyoWbrillar myopathy; Desmin; Deletion mutation; Assembly; Intermediate Wlament disease

1. Introduction Intermediate Wlament (IF) proteins are major constituents of metazoan cells. Together with microtubules and microWlaments they constitute the cytoskeleton proper in combination with various cross-bridging proteins and celljunction complexes (Herrmann and Aebi, 2004). In man, more than 70 IF proteins are expressed in a cell-type speciWc way in parallel to distinct routes of embryonic development (Herrmann et al., 2003; Lazarides, 1982). Next to

*

Corresponding author. Fax: +49 6221 42 3519. E-mail address: [email protected] (H. Bär).

1047-8477/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.jsb.2006.10.029

their unique role in tissue architecture, IF proteins received much attention when it became apparent that mutations in IF proteins can cause tissue-speciWc disease. For instance, mutations in epidermal keratins cause blistering skin diseases, those in glial Wbrillary acidic protein (GFAP) lead to Alexander disease, a severe neurological disorder, and mutations in the nuclear IF proteins, the lamins, were found to cause at least 12 distinct diseases (for overview, see Omary et al., 2004). This complex set of diseases mirrors the fact that lamins, in stark contrast to cytoplasmic IF proteins are by and large not expressed in tissue-speciWc patterns. Mutations in the desmin gene can give rise to desminopathy, a subgroup of the genetically and clinically

108

H. Bär et al. / Journal of Structural Biology 158 (2007) 107–115

heterogeneous group of myoWbrillar myopathies. After the Wrst mutations in this gene were discovered to give rise to a rare but devastating illness, which can aVect both skeletal and cardiac muscle (Goldfarb et al., 1998; Munoz-Marmol et al., 1998), the number of identiWed cases increased signiWcantly (Arbustini et al., 2006; Bär et al., 2005b; Fidzianska et al., 2005; Goldfarb et al., 2004; Pruszczyk et al., 2006). As a rule, desminopathies follow an autosomal dominant pattern of inheritance although some cases occurred de novo and also single autosomal recessive cases have been reported (Goldfarb et al., 2004). Most pathogenic mutations discovered to date are missense mutations which lead to the exchange of single amino acids in the intermediate Wlament (IF) protein desmin (Bär et al., 2004; Goldfarb et al., 2004; Paulin et al., 2004). Among these, the majority is located in highly conserved amino acid sequence stretches in the second half of the central -helical coiled-coil rod domain, termed coil 2B. We recently showed that in contrast to prior assumptions about 40% of these missense mutations are assembly-competent when examined both in vitro and in vivo (Bär et al., 2005a, 2006a,b) although they all lead to desmin aggregation in myoWbers of aVected patients, the pathognomonic feature of this disease entity (Goldfarb et al., 2004). Moreover, we demonstrated that some of the proline mutations residing in the coiled-coil forming part of the -helix are tolerated for in vitro and in vivo assembly (Bär et al., 2005a, 2006a,b). This latter type of mutation introduces a “potent helix breaker” into the -helical rod domain and has therefore been assumed to be particularly harmful (Paulin et al., 2004). Apart from these missense mutations, some more complex desmin mutations were discovered, which represent an even more severe insult to the architecture of the desmin molecule (Munoz-Marmol et al., 1998; Park et al., 2000; Schröder et al., 2003). Among them are two mutations which lead to an in-frame deletion of one (DesN366) and three (DesE359_S361) amino acids, respectively (Kaminska et al., 2004). By combining three-dimensional molecular modeling with transfection studies it was hypothesized that these two mutations lead to a severe modiWcation of the molecular geometry of the desmin coiled coil. It was furthermore speculated that this change in dimer architecture would result in “fatal damage to desmin Wlament assembly” which would lead to aggregate formation and, ultimately, development of severe myopathy (Kaminska et al., 2004). In the present study, we investigated the assembly competence of these two desmin deletion mutant variants by in vitro assembly studies of the recombinant proteins and “time-lapse” electron microscopy. Furthermore, we tried to gain insight into the network forming capability of equimolar mixtures of the respective mutant protein with wild type desmin analogous to the situation in the heterozygous patients, where mutant and wild type protein is coexpressed. We demonstrate that in a copolymeric situation with wild type protein both deletion mutants are able to

form Wlaments in vitro and Wlamentous networks in vivo. This behaviour is very surprising, since the deletions are located in a very conserved part of the -helical coiled-coil domain and have therefore been expected to interfere with the functional architecture of the dimer and to “poison” IF assembly. 2. Material and methods 2.1. Cloning and mutagenesis Mutations were introduced by site-directed mutagenesis (Quickchange®, Stratagene, Germany) into the full-length clone of the human desmin WT cDNA and were veriWed by sequencing. For protein expression, desmin wild type (DesWT) or the mutant cDNAs were subcloned into the prokaryotic expression vector pDS5 as described previously (Bär et al., 2005a; Herrmann et al., 1999). For transfection studies, the full-length clones were inserted into the unique EcoRI site of the eukaryotic expression vector p163/ 7, which drives expression with a MHC promoter (Niehrs et al., 1992), and the bicistronic vector pIRES2-EGFP (BD Clontech, Germany). 2.2. Protein chemical methods and viscometry The Escherichia coli strain TG1 (Amersham, Germany) was transformed with DesWT and mutant desmin plasmids, respectively. Recombinant desmin proteins were puriWed from inclusion bodies as described (Bär et al., 2005b; Herrmann et al., 1992, 2004; Hofmann et al., 1991). For in vitro reconstitution of puriWed recombinant protein, 0.5–1.0 mg of protein was dialyzed at a concentration of 0.5–1.0 mg/ml overnight into a buVer containing 5 mM Tris–HCl (pH 8.4), 1 mM EDTA, 0.1 mM EGTA and 1 mM DTT (“Tris-buVer”) using regenerated cellulose dialysis tubing (Spectra/Por®, MWCO 50.000; Roth, Germany). Assembly was started by addition of equal amounts of “assembly buVer” (45 mM Tris–HCl, pH 7.0, 100 mM NaCl). Viscosity measurements were routinely performed at a protein concentration of 0.3 mg/ml in an Ostwald viscometer (Cannon-Nanning, Semi-Viscometer, Zematra BV, The Netherlands) at 50 mM NaCl, 25 mM Tris–HCl (pH 7.5), 37 °C, as described (Hofmann et al., 1991). Time-lapse assembly experiments and negative staining were performed as described (Herrmann et al., 1999). For mixing experiments, equal amounts of mutant and WT desmin were combined in 9.5 M urea prior to dialysis into “Tris-buVer” in order to allow heterodimer formation. 2.3. Analytical ultracentrifugation Analytical ultracentrifugation experiments were carried out in “Tris-buVer”, using an Optima XLA Beckman analytical ultracentrifuge. Data analysis was performed as described (Mücke et al., 2004).

H. Bär et al. / Journal of Structural Biology 158 (2007) 107–115

2.4. Oxidative crosslinking The Destail (truncated at Ile 416) and DesE359_S361 or DesN366 protein were mixed 1:1 in 9.5 M urea prior to dialysis into “Tris-buVer”. DTT was initially added in order to reduce cysteine residues, but was then removed by extensive dialysis prior to initiation of the crosslinking reaction. Crosslinking was carried out in “Tris-buVer” using 250 M H2O2 and 25 M CuCl2 for 1 h at 37 °C (Rogers et al., 1996). The reaction products were analysed by electrophoresis on 10% polyacrylamide gels in the absence of a reducing agent (Laemmli, 1970). 2.5. Electron microscopy Filaments were absorbed to glow-discharged carboncoated copper grids and negatively stained by a 15 s treatment with 0.1% uranyl acetate. The specimens were inspected and micrographs recorded in a Zeiss model 900 transmission-electron microscope (Carl Zeiss, Oberkochen, Germany) as described (Herrmann et al., 1996). Electron microscopic images were scanned and processed using the free software ImageJ 1.32j, which has been developed at the National Institutes of Health, Bethesda, USA (http:// rsb.info.nih.gov/ij). For Wlament width measurements, at least 100 Wlament diameters were determined. 2.6. Cell culture and microscopic procedures For transfection studies, we used human adrenocortical carcinoma cells completely devoid of cytoplasmic intermediate Wlaments (SW13), murine 3T3 Wbroblastderived cells and murine C2.7 myoblasts (Franke et al., 1978; Hedberg and Chen, 1986; Pinset et al., 1988). Cells were plated on glass coverslips, placed in six-well plates for 1–2 days, grown to »30% conXuency and then transiently transfected with 5 g plasmid DNA per plate using Fugene 6® according to the manufacturer’s protocol

109

(Roche, Germany). 24, 48 and 72 h after transfection, cells were processed for immunocytochemistry. BrieXy, cells were Wxed in methanol for 5 min followed by permeabilisation in acetone for 3 min at ¡20°C. After rehydration, specimens were blocked in 10% donkey serum in phosphate-buVered saline (PBS) for 30 min. The coverslips were incubated with the polyclonal rabbit anti-desmin serum (dilution 1:100; Progen, Germany) together with the monoclonal anti-vimentin antibody Vim3B4 (undiluted; Progen, Germany) for 60 min at room temperature. After thoroughly rinsing in PBS, a Cy-3 labelled donkeyanti-mouse antibody (dilution 1:300; dianova, Germany) and an Alexa 488 labelled donkey-anti-rabbit antibody (dilution 1:100; Invitrogen, Germany) were applied simultaneously for 30 min together with DAPI (dilution 1:1000; 4,6-diamidino-2-phenylindole; Roche Diagnostics, Germany) for nuclear staining. The coverslips were Wnally mounted on glass slides in Fluoromount G (Southern Biotechnology Associates, USA). Cells were viewed by confocal laser scanning Xuorescence microscopy (DMIRE 2, Leica, Germany). 3. Results 3.1. Analytical ultracentrifugation Following denaturation with 9.5 M urea, we Wrst reconstituted the recombinant desmin variants by dialysis into a low ionic strength buVer containing 5 mM Tris–HCl (pH 8.4), 1 mM EDTA, 0.1 mM EGTA and 1 mM DTT (“TrisbuVer”; Herrmann et al., 2004). Next, complex formation of both mutant desmins was investigated relative to recombinant DesWT by sedimentation velocity ultracentrifugation. Accordingly, DesWT and both mutant variants were found to sediment as homogenous species with s-values of 5.4 S (DesWT), 5.8 S (DesE359_S361) and 5.5 S (DesN366) (Fig. 1A), indicating tetramer formation (Bär et al., 2006b; Herrmann et al., 1996).

Fig. 1. Characterization of the soluble assembly precursors of (A) DesWT (black squares), DesE359_S361 (open circles), DesN366 (open rhombus) and (B) equimolar mixtures of DesWT with the respective mutant protein (DesE359_S361, open circle, and DesN366, open rhombus) by analytical ultracentrifugation. Sedimentation velocity analysis was performed in “Tris-buVer”, the curves presented are representative g(s¤) plots.

110

H. Bär et al. / Journal of Structural Biology 158 (2007) 107–115

In order to analyze whether the respective mutant desmin variants and DesWT associate into homogenous complexes, we mixed equimolar amounts of DesWT with either one of the two mutant desmin variants in 9.5 M urea prior to dialysis. We then performed sedimentation velocity ultracentrifugation as described above to analyze the resulting complexes (Fig. 1B). Both mixtures yielded sedimentation coeYcients of 5.7 S (DesWT/DesE359_S361) and 5.4 S (DesWT/DesN366), respectively. These values are in the same range as those determined for DesWT and the two mutant variants individually, indicating that they were present as tetrameric complexes and that they did not lead to any kind of aggregation due to a potential incompatibility of the wild type and the mutant protein at the dimeric and/or tetrameric stage (Bär et al., 2005a, 2006b; Herrmann et al., 1996; Mücke et al., 2004).

Next, we performed oxidative crosslinking of the endogenous cysteine residues (C333) in order to verify that both mutants are in fact able to form heterodimeric coiled coils with DesWT. To distinguish the mutant from the wild type desmin variant by gel electrophoresis, we employed a tailless wild type desmin variant (Destail; truncated at Ile 416 (Bär et al., 2006b)). “Native” gel electrophoresis documented that indeed cross-linked heterodimers were present in the mixture of soluble complexes (data not shown). 3.2. In vitro assembly studies Under standard assembly conditions, DesWT assembled into smooth, extended Wlaments upon addition of “assembly buVer” (Fig. 2A). In contrast, the recombinant

Fig. 2. Electron microscopic analysis of negatively stained assembly products of (A) DesWT, (B) the mutants DesE359_S361 and DesN366, and (C) equimolar mixtures of DesWT with either one of the mutants. Assembly was stopped at 10 s, 5, 10 and 60 min, respectively, by addition of 0.1% glutaraldehyde. Note the enhanced adhesiveness and formation of Wlamentous aggregates for the mutant Des366 (group III according to Bär et al., 2005a). Bar, 100 nm.

H. Bär et al. / Journal of Structural Biology 158 (2007) 107–115

mutant desmin variants exhibited distinct assembly defects (Fig. 2B). DesE359_S361 formed extended Wlamentous networks (“group I” according to Bär et al., 2005a). However, the Wlamentous structures were thicker and exhibited a more irregular diameter (»13.5 § 2.0 nm) than those depicted for DesWT (»12.4 § 1.4 nm). In contrast, DesN366, which gave rise to highly irregular Wlamentous precursors after 10 s of assembly, exhibited subsequent deterioration of assembly leading to the formation of Wlamentous aggregates (“group III” according to Bär et al., 2005a). In order to investigate the inXuence of the mutant proteins on wild type desmin assembly, we mixed equimolar amounts of the respective mutant with DesWT in 9.5 M urea prior to dialysis into “Tris-buVer” (Fig. 2C). Upon initiation of assembly of the mixture of DesWT and DesE359_S361, long, Xexible and smooth Wlaments of »12.4 § 1.6 nm diameter were formed. Similarly, the mixture of DesN366 and DesWT was able to form extended Wlamentous networks. However, these Wlaments exhibited a somewhat more irregular diameter (» 11.5 § 3.1 nm).

111

3.4. Transfection studies Last but not least, we analyzed the in vitro capability of both mutants to form Wlamentous networks upon transient transfection with expression constructs. Within the observed time course, DesE359_S361 formed only short Wlamentous structures, which often aggregated into huge proteinaceous masses in the perinuclear region in human SW13 cells, which are completely devoid of cytoplasmic IF proteins (Fig. 4). In contrast, DesN366 formed small aggregates, which were distributed evenly throughout the cytoplasm (Fig. 4). When the mutants were transfected into murine 3T3 Wbroblasts, a cell line expressing endogenous vimentin, both DesE359_S361 and DesN366 were able to integrate into the pre-existing IF network. Hence, the available vimentin network was able to rescue the assembly defect of both mutant proteins that inhibited the formation of extended IF networks in SW13 cells. In murine C2.7 myoblast cells which express an elaborate endogenous desmin cytoskeleton, both mutants fully integrated into the pre-existing IF network (data not shown). 4. Discussion

3.3. Viscometry In order to follow desmin assembly and to monitor potential inXuences of the various mutations on DesWT network formation, we performed viscometric analyses as described previously (Bär et al., 2005a,b). Compared to DesWT, the drastic increase in relative viscosity in DesE359_S361 indicates the formation of a highly viscous gel (Fig. 3). For DesN366, the viscosity proWle revealed only a minor increase after 5 min of assembly, and the drop that took place thereafter was indicative of aggregation. This is conWrmed by the Wlamentous aggregates depicted by electron microscopy (see Fig. 2). The addition of equimolar amounts of DesWT to both mutant variants resulted in normalization of the viscosity proWles (Fig. 3B).

Mutations in various genes have been found to give rise to diverse cardiac and skeletal myopathies (Ahmad et al., 2005; Morita et al., 2005; Seidman and Seidman, 2001). Since the Wrst report that mutations in the desmin gene are responsible for a subgroup of myoWbrillar myopathy (Goldfarb et al., 1998), the number of known disease-causing mutations has risen rapidly (Arbustini et al., 2006; Bär et al., 2005b; Fidzianska et al., 2005; Goldfarb et al., 2004; Muntoni et al., 2006; Pruszczyk et al., 2006). In order to understand the functional implications of desmin mutations for the development of myopathy, we set out to investigate the impact of various missense mutations in the desmin gene on the protein’s propensity to selfassemble in vitro into IFs (Bär et al., 2005a, 2006b, 2005b). Systematic analysis of up to now 21 mutant desmin

Fig. 3. Viscometric analysis of (A) wild type desmin, the two mutant desmins, as well as (B) equimolar mixtures of wild type desmin with either one of the respective mutants. Assembly was monitored by viscometry at a protein concentration of 0.3 mg/ml. Black closed squares: DesWT; black rhombus: DesE359_S361; black triangle: DesN366; open rhombus: equimolar mixture (in 9.5 M urea) of DesE359_S361 and wild type desmin; open triangle: equimolar mixture of DesN366 and DesWT. Abscissa, time (min); ordinate, relative viscosity.

112

H. Bär et al. / Journal of Structural Biology 158 (2007) 107–115

Fig. 4. Functional analyses of both mutant desmin variants, DesE359_S361 and DesN366, (A) in cells devoid of cytoplasmic IFs, i.e., SW13, and (B) in 3T3 Wbroblasts expressing endogenous vimentin. Cells depicted were processed for immunoXuorescence staining 48 h after transfection. Nevertheless , they are representative for the entire time course followed. Note the formation of short Wlamentous structures in SW13 cells when transfected with DesE359_S361. In contrast DesN366 forms small aggregates which are distributed throughout the cytoplasm. In contrast, both mutant proteins are able to integrate into the endogenous vimentin cytoskeleton in 3T3 Wbroblasts. SW13 cells were tested for spontaneous re-expression of endogenous vimentin by double-immunoXuorescence staining (data not shown). Co-localization of desmin and vimentin staining in 3T3 cells is indicated by the yellow color in the merge in the right column. Left column, transfected desmin variants (green); middle column, endogenous vimentin (red); right column, merge. Blue, DAPI. Bar, 10 m.

variants led to the discovery that these mutations interfere with the assembly process at distinct stages, thereby allowing us to group desmin mutations into four distinct assembly defect categories (Bär et al., 2005a). Unexpectedly, 12 out of the 21 mutant desmin proteins retained their ability to form bona Wde IFs in vitro (Bär et al., 2005a), although some mutations signiWcantly altered Wlament architecture as revealed by the measurement of the mass-per-length of the corresponding Wlaments (Bär et al., 2006b). Some mutants locating in the tail domain of desmin, while still being able to form Wlaments on their own, were found to poison wild type desmin assembly when equimolar mixtures of wild type and mutant protein were analyzed (Bär et al., unpublished data). Taken together, the assembly defects caused by single amino acid replacements are currently not predictable by the type or the location of the corresponding mutation.

We have now investigated the functional consequences for the desmin molecule as imposed by the in-frame deletion of a single (DesN366) and three (DesE359_S361) amino acids, respectively. Both deletions are located in close vicinity to a naturally occurring “stutter” motif (F356-E359) in desmin’s central -helical rod domain, i.e., a region where a local unwinding of the coiled coil occurs (Parry, 2005; Strelkov et al., 2002). Interference with this highly conserved incomplete heptad repeat by recombinant insertion of three amino acid codons caused the mutant protein to cease assembly at the unit-length Wlament state (Herrmann and Aebi, 1999). This suggests that the “stutter” has evolved into a critical structural feature for IF formation. The deletion of three amino acids as found in two Polish families suVering from progressive skeletal myopathy (DesE359_S361) was modeled to create an additional stutter, thus resulting in increased local

H. Bär et al. / Journal of Structural Biology 158 (2007) 107–115

unwinding of the coiled coil (Kaminska et al., 2004). The other mutant (DesN366) was found in a family from Spain suVering from skeletal as well as cardiac myopathy. Theoretically, it converts the stutter into a so-called “stammer” motif, thereby resulting in overwinding of the coiled coil (Kaminska et al., 2004). By three-dimensional molecular modeling, it was furthermore hypothesized that these changes taking place in coil 2B will result in a re-orientation of the entire C-terminal non--helical “tail” domain relative to the axis of the central –helical coiledcoil domain (Kaminska et al., 2004). In our experiments we now demonstrated that both mutants are able to form regular tetrameric complexes. In order to simulate the heterozygous situation prevailing in patient myocytes, where both wild type and mutant alleles are co-expressed, we performed mixing experiments with both proteins being present in equimolar ratios. The fact that we observed a rescue of Wlament assembly in this copolymeric situation is perhaps the most surprising Wnding of our analyses. Mixtures of mutant and wild type protein are capable of forming heterodimers and heterotetrameric structures that are seemingly identical to wild type as revealed by analytical ultracentrifugation, although both mutants have been suggested to severely alter the three-dimensional structure of the respective dimer. Furthermore, we have demonstrated by electron microscopy that these copolymeric complexes assemble into smooth Wlaments. By capillary viscometry, a method very sensitive to detect aggregation of proteins, we revealed that the mixture of mutant and wild type protein formed normal networks, too. Even in transfected cells, both mutants were found to readily co-assemble with either pre-existing vimentin (3T3 Wbroblasts) or desmin (C2.7 myoblasts) and to integrate into the respective Wlamentous networks. The formation of seemingly normal IFs by combining the mutant with wild type protein is in no case a common phenomenon, as we have previously provided evidence for quite the opposite: assemblyincompetent mutant protein can “poison” wild type IF formation even when present only in small amounts (Bär et al., 2005b). Furthermore, all tested mutants exhibiting severe assembly defects in vitro also disrupted the endogenous vimentin cytoskeleton in transfected Wbroblast cells (Bär et al., 2005a, 2006a). In contrast, the assemblyincompetent mutant DesN366 is now found to readily integrate into extensive Wlament networks. The molecular modeling of alterations of the threedimensional structure of the desmin molecule caused by a distinct mutation as done by Kaminska and coworkers (Kaminska et al., 2004) is, according to our data, at present still in its infancy and does not directly allow to predict the assembly properties of a mutant IF protein. Furthermore, the fact that theoretically severe architectural alteration of the coiled-coil rod domain is obviously tolerated for IF assembly indicates that IF assembly is somewhat “relaxed” towards individual changes of the amino acid sequence, at least in some parts of the mole-

113

cule. In order to more rationally understand the complexities of molecular interactions taking place in IF proteins during Wlament assembly and the impact of a given mutation on IF assembly and function, a detailed analysis of the atomic structure of coil 2 of desmin and eventually that of higher-order complexes is needed. At the present stage our data suggest that desminopathy is not necessarily caused by an impairment of Wlament assembly due to architectural alterations in the desmin dimer, but that most mutations instead result in more subtle changes of the desmin IF protein, leading to either altered biophysical properties of single Wlaments or a modiWed binding of IF-associated proteins (IFAPs) due to changes in the three-dimensional shape of the Wlament or modiWed surface charges (Fig. 5). Hence, desminopathy can be envisaged to be caused by interference of the mutated protein with several physiological parameters. A similar situation is encountered with the nuclear IF protein lamin A. Laminopathies are caused by point mutations in lamin A and they have been proposed to result from mutated proteins interfering with diVerent cellular activities. To explain how the over 200 known mutations lead to at least 12 particular partially overlapping disease entities, three models, which are not mutually exclusive, have been suggested (Gotzmann and Foisner, 2006; Mattout et al., 2006; Worman and Courvalin, 2005): (1) a mechanical stress model, implying that the mutant Wlaments are not functional and cause disturbances of cytoskeletal order; (2) a gene expression model, which implies that the mutant protein interferes with transcriptional activities; and (3) a cell fate model, assuming that the diVerentiation of stem cells is inXuenced, e.g., by alteration in cellular signaling pathways and/or cell cycle regulation. Notably, also mutations in lamin A can lead to muscular dystrophy and cardiomyopathy. It will be interesting to Wnd out, if and at what stage both lamin A and desmin mutations interfere with the same regulatory mechanisms. 4.1. Summary and perspective Both point mutations in the rod or tail domain and inframe deletions in the rod domain slowly but deWnitely cause desmin aggregation and severe myopathy. Despite the fact that mutant desmin variants in vitro form various kinds of structures ranging from seemingly normal IFs to catastrophically super-aggregated proteinaceous masses, all mutations ultimately lead to the deposition of aggregated desmin and mechanical dysfunction of cardiac and skeletal myocytes. Hence, at present we cannot present a unifying “mechanism of disease”, but we can depict some of the physiological parameters that are most probably aVected (Fig. 5). Future work will have to reveal how the desmin IF system is integrated into mechanotransduction and cellular signaling. Only this will eventually enable us to interfere with the development of desminopathy.

114

H. Bär et al. / Journal of Structural Biology 158 (2007) 107–115

Fig. 5. Schematic representation of how desmin mutations may lead to desminopathy (Bär et al., 2005a; Patterson, 2006; Paulin et al., 2004). Mutations in the desmin gene either allow Wlament formation or they obstruct proper assembly into desmin Wlaments. However, as patients are usually heterozygous for a given mutation, either assembly defects might be rescued by the presence of the wild type protein or Wlament formation of the heteropolymers of mutant and wild type protein is obstructed. Alternatively, the mutant protein may segregate from wild type protein and interfere, in a “gain of function” mode, with cellular homeostasis. IFs might exhibit either compromised biophysical or binding properties, ultimately leading to enhanced degradation by the ubiquitin-proteasome pathway. Moreover, misfolded oligomers may block the proteasome, thus eventually leading to mitochondrial dysfunction (Liu et al., 2006). Segregation of misfolded proteins into pathognomonic aggregates is currently thought to be a protective mechanism, but their presence between individual sarcomers might lead to impairment of the mechanical properties of myocytes.

Acknowledgments

References

We gratefully acknowledge the technical assistance of Tatjana Wedig. Harald Bär and Harald Herrmann acknowledge grants from the German Research Foundation (DFG; BA 2186/2-2 to H.B. and HE 1853 to H.H.). Ueli Aebi is funded by a research grant and a NCCR program grant on “Nanoscale Science” by the Swiss National Science Foundation, the Swiss Society for Research on Muscular Diseases, the M.E. Müller Foundation of Switzerland, and the Canton Basel-Stadt.

Ahmad, F., Seidman, J.G., Seidman, C.E., 2005. The genetic basis for cardiac remodeling. Annu. Rev. Genom Hum. Genet. 6, 185–216. Arbustini, E., Pasotti, M., Pilotto, A., Pellegrini, C., Grasso, M., Previtali, S., Repetto, A., Bellini, O., Azan, G., ScaYno, M., Campana, C., Piccolo, G., Vigano, M., Tavazzi, L., 2006. Desmin accumulation restrictive cardiomyopathy and atrioventricular block associated with desmin gene defects. Eur. J. Heart Fail 8, 477–483. Bär, H., Strelkov, S.V., Sjöberg, G., Aebi, U., Herrmann, H., 2004. The biology of desmin Wlaments: how do mutations aVect their structure, assembly, and organisation? J. Struct. Biol. 148, 137–152.

H. Bär et al. / Journal of Structural Biology 158 (2007) 107–115 Bär, H., Mücke, N., Kostareva, A., Sjöberg, G., Aebi, U., Herrmann, H., 2005a. Severe muscle disease-causing desmin mutations interfere with in vitro Wlament assembly at distinct stages. Proc. Natl. Acad. Sci. USA 102, 15099–15104. Bär, H., Kostareva, A., Sjöberg, G., Sejersen, T., Katus, H.A., Herrmann, H., 2006a. Forced expression of desmin and desmin mutants in cultured cells: impact of myopathic missense mutations in the central coiled-coil domain on network formation. Exp. Cell Res. 312, 1554–1565. Bär, H., Mücke, N., Ringler, P., Müller, S.A., Kreplak, L., Katus, H.A., Aebi, U., Herrmann, H., 2006b. Impact of disease mutations on the desmin Wlament assembly process. J. Mol. Biol. 360, 1031–1042. Bär, H., Fischer, D., Goudeau, B., Kley, R.A., Clemen, C.S., Vicart, P., Herrmann, H., Vorgerd, M., Schröder, R., 2005b. Pathogenic eVects of a novel heterozygous R350P desmin mutation on the assembly of desmin intermediate Wlaments in vivo and in vitro. Hum. Mol. Genet. 14, 1251–1260. Fidzianska, A., Kotowicz, J., Sadowska, M., Goudeau, B., Walczak, E., Vicart, P., Hausmanowa-Petrusewicz, I., 2005. A novel desmin R355P mutation causes cardiac and skeletal myopathy. Neuromuscul. Disord. 15, 525–531. Franke, W.W., Schmid, E., Osborn, M., Weber, K., 1978. The intermediatesized Wlaments in rat kangaroo PtK2 cells. II. Structure and composition of isolated Wlaments. Cytobiologie 17, 392–411. Goldfarb, L.G., Vicart, P., Goebel, H.H., Dalakas, M.C., 2004. Desmin myopathy. Brain 127, 723–734. Goldfarb, L.G., Park, K.Y., Cervenakova, L., Gorokhova, S., Lee, H.S., Vasconcelos, O., Nagle, J.W., Semino-Mora, C., Sivakumar, K., Dalakas, M.C., 1998. Missense mutations in desmin associated with familial cardiac and skeletal myopathy. Nat. Genet. 19, 402–403. Gotzmann, J., Foisner, R., 2006. A-type lamin complexes and regenerative potential: a step towards understanding laminopathic diseases? Histochem. Cell Biol. 125, 33–41. Hedberg, K.K., Chen, L.B., 1986. Absence of intermediate Wlaments in a human adrenal cortex carcinoma-derived cell line. Exp. Cell Res. 163, 509–517. Herrmann, H., Aebi, U., 1999. Intermediate Wlament assembly: temperature sensitivity and polymorphism. Cell Mol. Life Sci. 55, 1416–1431. Herrmann, H., Aebi, U., 2004. Intermediate Wlaments: molecular structure, assembly mechanism, and integration into functionally distinct intracellular scaVolds. Annu. Rev. Biochem. 73, 749–789. Herrmann, H., Hofmann, I., Franke, W.W., 1992. IdentiWcation of a nonapeptide motif in the vimentin head domain involved in intermediate Wlament assembly. J. Mol. Biol. 223, 637–650. Herrmann, H., Kreplak, L., Aebi, U., 2004. Isolation, characterization, and in vitro assembly of intermediate Wlaments. Method Cell Biol. 78, 3–24. Herrmann, H., Haner, M., Brettel, M., Ku, N.O., Aebi, U., 1999. Characterization of distinct early assembly units of diVerent intermediate Wlament proteins. J. Mol. Biol. 286, 1403–1420. Herrmann, H., Hesse, M., Reichenzeller, M., Aebi, U., Magin, T.M., 2003. Functional complexity of intermediate Wlament cytoskeletons: from structure to assembly to gene ablation. Int. Rev. Cytol. 223, 83–175. Herrmann, H., Haner, M., Brettel, M., Müller, S.A., Goldie, K.N., Fedtke, B., Lustig, A., Franke, W.W., Aebi, U., 1996. Structure and assembly properties of the intermediate Wlament protein vimentin: the role of its head, rod and tail domains. J. Mol. Biol. 264, 933–953. Hofmann, I., Herrmann, H., Franke, W.W., 1991. Assembly and structure of calcium-induced thick vimentin Wlaments. Eur. J. Cell Biol. 56, 328–341. Kaminska, A., Strelkov, S.V., Goudeau, B., Olive, M., Dagvadorj, A., Fidzianska, A., Simon-Casteras, M., Shatunov, A., Dalakas, M.C., Ferrer, I., Kwiecinski, H., Vicart, P., Goldfarb, L.G., 2004. Small deletions disturb desmin architecture leading to breakdown of muscle cells and development of skeletal or cardioskeletal myopathy. Hum. Genet. 114, 306–313. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685.

115

Lazarides, E., 1982. Intermediate Wlaments: a chemically heterogeneous, developmentally regulated class of proteins. Annu. Rev. Biochem. 51, 219–250. Liu, J., Chen, Q., Huang, W., Horak, K.M., Zheng, H., Mestril, R., Wang, X., 2006. Impairment of the ubiquitin-proteasome system in desminopathy mouse hearts. FASEB J. 20, 362–364. Mattout, A., Dechat, T., Adam, S.A., Goldman, R.D., Gruenbaum, Y., 2006. Nuclear lamins, diseases and aging. Curr. Opin. Cell Biol. 18, 335–341. Morita, H., Seidman, J., Seidman, C.E., 2005. Genetic causes of human heart failure. J. Clin. Invest. 115, 518–526. Mücke, N., Wedig, T., Burer, A., Marekov, L.N., Steinert, P.M., Langowski, J., Aebi, U., Herrmann, H., 2004. Molecular and biophysical characterization of assembly-starter units of human vimentin. J. Mol. Biol. 340, 97–114. Munoz-Marmol, A.M., Strasser, G., Isamat, M., Coulombe, P.A., Yang, Y., Roca, X., Vela, E., Mate, J.L., Coll, J., Fernandez-Figueras, M.T., Navas-Palacios, J.J., Ariza, A., Fuchs, E., 1998. A dysfunctional desmin mutation in a patient with severe generalized myopathy. Proc. Natl. Acad. Sci. USA 95, 11312–11317. Muntoni, F., Bonne, G., Goldfarb, L.G., Mercuri, E., Piercy, R.J., Burke, M., Yaou, R.B., Richard, P., Recan, D., Shatunov, A., Sewry, C.A., Brown, S.C., 2006. Disease severity in dominant Emery Dreifuss is increased by mutations in both emerin and desmin proteins. Brain 129, 1260–1268. Niehrs, C., Huttner, W.B., Ruther, U., 1992. In vivo expression and stoichiometric sulfation of the artiWcial protein sulfophilin, a polymer of tyrosine sulfation sites. J. Biol. Chem. 267, 15938–15942. Omary, M.B., Coulombe, P.A., McLean, W.H., 2004. Intermediate Wlament proteins and their associated diseases. N. Engl. J. Med. 351, 2087–2100. Park, K.Y., Dalakas, M.C., Goebel, H.H., Ferrans, V.J., Semino-Mora, C., Litvak, S., Takeda, K., Goldfarb, L.G., 2000. Desmin splice variants causing cardiac and skeletal myopathy. J. Med. Genet. 37, 851–857. Parry, D.A., 2005. Microdissection of the sequence and structure of intermediate Wlament chains. Adv. Protein Chem. 70, 113–142. Patterson, C., 2006. Search and destroy: the role of protein quality control in maintaining cardiac function. J. Mol. Cell Cardiol. 40, 438–441. Paulin, D., Huet, A., Khanamyrian, L., Xue, Z., 2004. Desminopathies in muscle disease. J. Pathol. 204, 418–427. Pinset, C., Montarras, D., Chenevert, J., Minty, A., Barton, P., Laurent, C., Gros, F., 1988. Control of myogenesis in the mouse myogenic C2 cell line by medium composition and by insulin: characterization of permissive and inducible C2 myoblasts. DiVerentiation 38, 28–34. Pruszczyk, P., Kostera-Pruszczyk, A., Shatunov, A., Goudeau, B., Draminska, A., Takeda, K., Vicart, P., Strelkov, S.V., Goldfarb, L., Kaminska, A., 2006. Autosomal dominant restrictive cardiomyopathy with atrioventricular conduction block caused by a mutation located in a highly conserved segment of desmin coiled-coil domain. Int. J. Cardiol. (Epub ahead of print). Rogers, K.R., Herrmann, H., Franke, W.W., 1996. Characterization of disulWde crosslink formation of human vimentin at the dimer, tetramer, and intermediate Wlament levels. J. Struct. Biol. 117, 55–69. Schröder, R., Goudeau, B., Simon, M.C., Fischer, D., Eggermann, T., Clemen, C.S., Li, Z., Reimann, J., Xue, Z., Rudnik-Schöneborn, S., Zerres, K., van der Ven, P.F., Fürst, D.O., Kunz, W.S., Vicart, P., 2003. On noxious desmin: functional eVects of a novel heterozygous desmin insertion mutation on the extrasarcomeric desmin cytoskeleton and mitochondria. Hum. Mol. Genet. 12, 657–669. Seidman, J.G., Seidman, C., 2001. The genetic basis for cardiomyopathy: from mutation identiWcation to mechanistic paradigms. Cell 104, 557–567. Strelkov, S.V., Herrmann, H., Geisler, N., Wedig, T., Zimbelmann, R., Aebi, U., Burkhard, P., 2002. Conserved segments 1A and 2B of the intermediate Wlament dimer: their atomic structures and role in Wlament assembly. EMBO J. 21, 1255–1266. Worman, H.J., Courvalin, J.C., 2005. Nuclear envelope, nuclear lamina, and inherited disease. Int. Rev. Cytol. 246, 231–279.