Interference of Amino-Terminal Desmin Fragments With Desmin Filament Formation
Harald Bär,1* Sarika Sharma, 1,2 Helga Kleiner,1 Norbert Mücke,3 Hanswalter Zentgraf,4 Hugo A. Katus,1 Ueli Aebi,5 and Harald Herrmann2
1 Department of Cardiology, University of Heidelberg, Heidelberg, Germany
2 Department of Molecular Genetics, German Cancer Research Center (DKFZ), Heidelberg, Germany
3 Division of Biophysics of Macromolecules, German Cancer Research Center (DKFZ), Heidelberg, Germany
4 Research Group Electron Microscopy, Applied Tumor Virology, German Cancer Research Center (DKFZ), Heidelberg, Germany
5 Maurice E. Müller Institute for Structural Biology, Biozentrum, University of Basel, Basel, Switzerland
Short polypeptides from intermediate filament (IF) proteins containing one of the two IF-consensus motifs interfere severely with filament assembly in vitro. We now have systematically investigated a series of larger fragments of the muscle-specific IF protein desmin representing entire functional domains such as coil1 or coil 2. “Half molecules” comprising the amino-terminal portion of desmin, such as DesAC240 and the “tagged” derivative Des(ESA)AC244, assembled into large, roundish aggregates already at low ionic strength, DesAC250 formed extended, rel- atively uniform filaments, whereas DesAC265 and DesAC300 were soluble under these conditions. Surprisingly, all mutant desmin fragments assembled very rapidly into long thick filaments or spacious aggregates when the ionic strength was raised to standard assembly conditions. In contrast, when these desmin mutants were assembled in the presence of wild-type (WT) desmin, their assembly properties were completely changed: The elongation of the two shorter desmin fragments was completely inhibited by WT desmin, whereas DesAC250, DesAC265 and DesAC300 coassembled with desmin into filaments, but these mixed filaments were distinctly disturbed and exhibited a very different phenotype for each mutant. After transfection into fibroblasts and cardiomyocytes, the truncated mutant Des (ESA)AC244 localized largely to the cytoplasm, as revealed by a tag-specific monoclonal antibody, and also partially colocalized there with the collapsed endog- enous vimentin and desmin systems indicating its interference with IF-organizing processes. In contrast, in cells without an authentic cytoplasmic IF system such as line SW13, Des(ESA)AC242 entered the nucleus and was deposited in small dot- like structures in chromatin-free spaces without any noticeable effect on nuclear morphology. Cell Motil. Cytoskeleton 66: 986-999, 2009. @ 2009 Wiley-Liss, Inc.
Key words: analytical ultracentrifugation; assembly; desmin; intermediate filaments; monoclonal antibody; premature stop codon
Contract grant sponsor: German Research Foundation (DFG); Contract grant numbers: BA 2186/2-1, BA 2186/3-1, HE 1853; Contract grant sponsor: Swiss National Science Foundation; Contract grant number: 3100A0-100448; Contract grant sponsors: Swiss Foundation for Research on Muscle Diseases, M.E. Mueller Foundation.
*Correspondence to: Harald Bär, M.D., Department of Cardiology, Medizinische Universitätsklinik Heidelberg, Im Neuenheimer Feld 410, 69120 Heidelberg, Germany. E-mail: h.baer@dkfz.de
Received 22 February 2009; Accepted 19 May 2009
Published online 15 June 2009 in Wiley InterScience (www. interscience.wiley.com). DOI: 10.1002/cm.20396
INTRODUCTION
In concert with microtubules and microfilaments, intermediate filaments (IFs) form an elaborate three- dimensional scaffold in metazoan cells [Goldman et al., 1979]. Different from the microtubule and microfilament constituents, IF proteins are expressed in complex cell- type specific patterns during embryogenesis and in the adult [Lazarides, 1982]. One of their major functions is probably to serve as a versatile absorber of mechanical stress [Fuchs and Cleveland, 1998; Herrmann et al., 2007]. In muscle tissue, desmin IFs form an extrasarco- meric cytoskeleton that interconnects neighboring myofi- brils and links the contractile apparatus to nuclei, mitochondria and cell-cell attachment sites such as the intercalated discs of cardiac muscle. Desminopathy, a subgroup of myofibrillar myopathy, is a rare devastating human disease affecting both skeletal and cardiac mus- cle. It is characterized by the formation of huge intracel- lular aggregates of desmin and associated proteins leading to myofibrillar misalignment [Bär et al., 2004; Goldfarb et al., 2004; Paulin et al., 2004]. A direct dem- onstration of the impact that desmin point mutations may exert on the sarcomeric organization in muscle cells has early on been experimentally achieved by the infec- tion of neonatal rat cardiac myocytes with adenoviral constructs coding for point-mutated desmins [Haubold et al., 2003].
Most of the mutations reported up to now result in an exchange of single amino acids in highly conserved parts of desmin’s «-helical, coiled-coil forming “rod” domain. In addition, several missense mutations located in the amino-terminal “head” and carboxy-terminal “tail” domains of the desmin molecule as well as some deletion mutations have been discovered to cause des- minopathy [Munoz-Marmol et al., 1998; Goldfarb et al., 2004; Arbustini et al., 2006; Bär et al., 2007]. In a series of investigations, we have determined the impact of vari- ous desmin mutations on the filament assembly proper- ties of the respective mutant protein both in vitro and in transfected cells [see Bär et al., 2007 and references therein]. A major outcome of these studies is that the effect of a given mutation on desmin’s filament and net- work forming capacity is not predictable by first princi- ples. Although mutations found in patients suffering from desminopathy were previously thought to abolish filament formation in general, our studies revealed that the majority of the corresponding proteins are assembly competent both in vitro and in transfected tissue culture cells. Hence, the respective pathomechanism may be based on changes of the mechanical properties of the fil- ament network or its interaction with other functional cellular entities such as mitochondria and the sarcomers, or both [Capetanaki et al., 2007; Herrmann et al., 2007].
From a structural point of view it was interesting to learn that a heterozygous single adenine insertion muta- tion in the very end of exon 3 of the desmin gene was reported to cause human muscle disease. This mutation was predicted to cause a frame shift and a premature ter- mination signal located four codons downstream of the insertion side [Schröder et al., 2003]. Hence, in the heter- ozygous patient this mutation could potentially result in the synthesis of a truncated desmin molecule with a predicted molecular weight of 27 kDa, i.e. Des(ESA) AC244. Although premature nonsense codons often trig- ger nonsense-mediated mRNA decay (NMD), thus aborting protein synthesis [Amrani et al., 2006], not all premature termination codons necessarily meet the same fate. Those escaping NMD might give rise to truncated proteins, which can influence cellular functions. Such an incident has been described for the neuronal IF protein NF-L in Japanese quails, where a premature stop codon leads to the production of an assembly-deficient frag- ment of the low molecular weight neurofilament triplet protein NF-L. In the homozygous state, its expression leads to a complete absence of neurofilaments in axons of both the central and the peripheral nervous system. Affected quails exhibit a characteristic generalized mus- cle tremor, leading to the naming of this mutant strain as “quiver” [Ohara et al., 1993].
In the meantime a correction for the above desmin disease mutation was published indicating that it did not represent a premature stop but instead the loss of a single amino acid [Schröder et al., 2007]. However, the drastic behavior of Des(ESA)AC244, this “first half” desmin, both in vitro and in transfected cells made us curious how amino-terminal fragments behave, in particular with respect to their association with wild-type (WT) desmin complexes and their ability to integrate into desmin IFs. Therefore, we followed up the molecular properties of desmin amino-terminal fragments that span the end of coil 1B, the linker 12, or even reach into coil 2 (see Fig. 1). In particular, we also analyzed the desmin amino- terminal fragment DesAC265 and the carboxy-terminal fragment AN264Des, which are generated by caspase cleavage in linker L12. Our data indicate that such large IF protein fragments have a strong potential to interfere with filament formation and network integrity both in vitro and in vivo.
MATERIALS AND METHODS Cloning and Mutagenesis
The mutation Des(ESA)AC244 was introduced by site-directed mutagenesis (Quickchange®, Stratagene, Germany) into the full-length clone of the human desmin WT cDNA. For protein expression, desmin WT and Des(ESA)AC244 cDNA were subcloned into the
A
WT
1A
1B
2A
2B
N
C
C
La
N
A
240
250
265
300
A C300
N
0
N
300
265
250
240
B
1B
L12
2A
L2
2B
efgabcdefgabcdefgabcdefgabc
cdefgabcdefgabcdefg
defgabcdefgabcdefgabcdefgabcde
Hs ERRIESLNEEIAFLKKVHEEEIRELQAQLQEQQVOVEHDMSKPDLTAALRDIRAQYETIAAKNISEAEEWYKSRVSDLTOAANKNNDALRQAKQEMHEYR
Mm ERRIESLNEEIAFLEKVHEEEIRELOAQLQECOVOVEHDMSKPDLTAALRDIRAQYETIAAKNISEAEEWYKSKVSDLTOAANKNNDALRQAKOEMMEYR
Ss
ERRIEALOEEIAFLKKVHEEELRELOAQLOECHVOVENDLSKPDLTAALRDIRAQYECISAKNVQEAEEWYKSKVSDLNONATKNNDALROAKOEVNEYR
240
250
265
300
prokaryotic expression vector pDS5 as described previ- ously [Herrmann et al., 1999; Bär et al., 2005a].
The constructs coding for the truncated desmin variants DesAC240, AN30DesAC240, DesAC250, DesAC265 and DesAC300 were generated by polymerase chain reaction via primers which introduced NdeI and HindIII sites at the 5’ and 3’ ends, respectively, as well as a C-terminal stop codon at the respective location. All sequences were verified by cDNA sequencing. These truncated desmin variants were then introduced NdeI-Hin- dIII into the pET 24a(+) vector (Novagen, USA) and pro- tein expression was induced by addition of 1 mM IPTG for 5-6 h (Isopropyl-thio-2-D-galactopyranoside; Carl Roth, Germany). For transfection studies, only the full- length clones of WT desmin and Des(ESA)AC244 were inserted into the unique EcoRI site of the eukaryotic expression vector p163/7, which contains the human major histocompatibility complex promoter to drive protein expression [Niehrs et al., 1992].
Protein Chemical Methods and Viscometry
The Escherichia coli strain TG1 (Amersham, Germany) was transformed with WT and mutant desmin plasmids, respectively. Recombinant desmin proteins were purified from inclusion bodies as described [Herr- mann et al., 1999; Bär et al., 2005a]. For in vitro recon- stitution of purified recombinant protein, the protein was dialyzed at a concentration of 0.5-1.0 mg/mL overnight
into a buffer containing 5 mM Tris-HCl (pH 8.4), 1 mM ethylenediaminetetraacetic acid, 0.1 mM ethylene glycol tetraacetic acid (EGTA) and 1 mM dithiothreitol (DTT) (“Tris-buffer”) using regenerated cellulose dialysis tub- ing (for WT-desmin: Spectra/Por®, MWCO 50,000, Carl Roth, Germany; for mutant or truncated desmin var- iants and their mixtures with: Servapor®, MWCO 12,000-14,000; Serva, Germany). Assembly was initi- ated by addition of equal volumes of “assembly buffer” (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) in 25 mM Tris-HCl (pH 7.5), 50 mM NaCl, at 37℃ as described [Bär et al., 2005a]. Assembly stud- ies and negative staining experiments were performed as described [Herrmann et al., 1999]. For mixing experi- ments, varying amounts of mutant or truncated desmin variants and WT protein were combined as indicated in 9.5 M urea prior to dialysis into “Tris-buffer” in order to allow heterodimer formation.
Analytical Ultracentrifugation
Analytical ultracentrifugation experiments with WT desmin and Des(ESA)AC244 were carried out in “Tris-buffer”, using an Optima XLA Beckman analyti- cal ultracentrifuge. Data analysis was performed as described [Mücke et al., 2004]. The large complexes
formed by Des(ESA)AC244 were recorded at 10,000 rpm. The analysis of the complexes formed by WT des- min was performed by centrifugation at 40,000 rpm. In order to analyze whether the two proteins are able to form heteropolymeric complexes, we combined WT des- min and Des(ESA)AC244 in varying ratios (5:4; 5:3 and 5:2) in urea prior to dialysis into “Tris-buffer”.
In parallel, samples of Des(ESA)AC244 were taken after stepwise dialysis from 8 M urea into “Tris-buffer” and subjected to centrifugation in an airfuge (Beckman Coulter, Germany) for 30 min at 10 psi and room tem- perature (RT). Samples of the total protein and the super- natant fraction were analyzed by gel electrophoresis.
Cosedimentation Assay
WT, mutant or truncated desmin variants as well as equimolar mixtures (performed in 9.5 M urea in order to allow heterodimer formation) of WT and the various desmin variants were dialyzed into “Tris-buffer” and subjected to centrifugation prior to addition of assembly buffer in order to analyze whether the two proteins inter- act significantly in order to form heteropolymeric complexes under low salt/high pH conditions. Samples of total protein, supernatant and pellet fractions were analyzed by gel electrophoresis.
Electron Microscopy
Protein samples were fixed with 0.1% glutaralde- hyde solution, absorbed to glow-discharged carbon- coated copper grids for 15 s and negatively stained by a 20 s treatment with 2% uranyl acetate (both solutions from Sigma-Aldrich, Germany). The specimen were recorded with a Zeiss electron microscope (Model 900; Carl Zeiss, Oberkochen, Germany) as described [Herrmann et al., 1996].
Immunological Procedures
The insertion of a single nucleotide was predicted to lead to a frame shift mutation and a premature stop signal after three codons. The mutant protein thus har- bors three amino acids at its carboxy-terminal end that are not found in WT desmin. A corresponding polypep- tide consisting of five WT and the three novel amino acids (CIAFLKESA) was synthesized and coupled to Keyhole limpet hemocyanin (KLH) via the amino-termi- nal cysteine that was added to the desmin sequence. The modified carrier protein was dialyzed against phosphate- buffered saline (PBS), pH 7.2 (10 mM Na2HPO4, 2 mM KH2PO4, 2.7 mM KCl, 150 mM NaCl; PBS) and then used to immunize male BALB/c mice following standard procedures. Initially, 50 µg of the polypeptide dissolved in Freund’s complete adjuvant (Sigma-Aldrich) was injected. This was followed by booster injections using antigen dis- solved in PBS. Monoclonal antibodies were raised essen-
tially according to the method of Köhler and Milstein [Köhler and Milstein, 1975]. Screening of the hybridomas for antibody production was performed by means of ELISA and Western blot using standard protocols.
Recombinant WT and mutant protein were sub- jected to SDS-PAGE (1 µg protein/lane). After electro- phoretic transfer to polyvinylidene difluoride membranes (Millipore, USA), nonspecific binding sites were blocked with 10% milk powder (Roth, Germany) in TBST (Tris- buffered saline Tween-20; 10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.05% Tween 20). The membranes were incubated with a monoclonal rabbit anti-desmin antibody (1:5000; Epitomics, purchased from Biomol, Germany), a polyclonal rabbit anti-desmin serum (1:5000; Progen, Germany) or with the mutant-specific anti-CIAFLKESA monoclonal mouse antibody DK26, prepared as described above (1:20) for 60 min at RT. After washing thrice with TBST (5 min), a horseradish peroxidase-con- jugated secondary antibody was applied (30 min at RT; Dianova, Germany). Antibody binding was detected af- ter additional washes in TBST using the ECL system from PerkinElmer Life Sciences, USA, according to the manufacturer’s protocol.
Three different cell lines - human adrenocortical carcinoma (SW13) cells completely devoid of cytoplas- mic IFs [Hedberg and Chen, 1986], murine fibroblast- derived 3T3 cells [Franke et al., 1978] as well as atrial cardiomycyte-derived HL-1 cells [Claycomb et al., 1998] - were employed for transient transfection studies. Cells were grown under standard conditions on glass coverslips and transiently transfected with either WT or mutant desmin constructs in vector p163/7 using Fugene 6® reagent according to the manufacturer’s protocol (Roche, Germany). 48 h after transfection, cells were fixed and permeabilized in ice-cold methanol and ace- tone for 5 min and 30 s, respectively. The cells were then blocked in 10% goat serum in PBS for 30 min at RT, incubated with either a polyclonal rabbit anti-desmin serum (1:100; Progen, Germany), a monoclonal rabbit anti-desmin antibody (1:100; Epitomics, USA), a mouse monoclonal anti-vimentin antibody, Vim3B4 (undiluted; Progen, Germany), or mouse monoclonal anti-CIAFL- KESA (undiluted; DK26) for 60 min at RT. After thoroughly rinsing in PBS, an Alexa 488 labeled goat- anti-rabbit antibody and Alexa 568 labeled goat-anti- mouse antibody (Invitrogen, Germany) along with the DNA stain DAPI (4,6-diamidino-2-phenylindole; Roche, Germany) were applied for 30 min. After washing three times in PBS the specimen were rinsed briefly in distilled water and dehydrated with 100% ethanol (1 min). Finally, the air-dried specimen were mounted in Fluoro- mount G (Southern Biotechnologies Associate, Birming- ham, USA), and viewed by confocal laser scanning fluorescence microscopy (DMIRE 2, Leica, Germany).
A
B
C
100 nm
50 nm
RESULTS Synthesis of Amino-Terminal Fragments of Desmin
Cytoplasmic IF proteins such as desmin form coiled- coil dimers that under conditions of low ionic strength and high pH associate to antiparallel, half staggered tetramers (Fig. 1, see also Strelkov et al. [2002]). A recently reported truncation mutation deviates after amino acid 239, in our sequence after amino acid 240 from the WT desmin sequence and exhibits the nondesmin sequence Glu-Ser- Ala (241-243) followed by a stop signal at position 244 [Schröder et al., 2003]. Because of the potential importance of the truncated part with respect to tetramer formation, we synthesized a series of the authentic desmin amino-termi- nal fragments DesAC240, DesAC250, DesAC265 and DesAC300 (Fig. 1). According to our present structural understanding of the IF dimer structure, the interaction between two dimers in a tetramer relies on arginine resi- dues residing in the “head” domain which interact mainly with the end of coil 1B, with coil 2A as well as the first part of coil 2B [Herrmann and Aebi, 2004]. In addition, we synthesized constructs for a partially head-deleted, amino- terminal desmin fragment, i.e. AN30DesAC240, and a des- min variant containing only approximately the second half of WT desmin, i.e. AN264Des. DesAC265 and AN264Des correspond to the two desmin fragments expected to arise after proteolytic cleavage of WT mouse desmin at the cas- pase cleavage site Asp264 that corresponds to Asp265 in man [Chen et al., 2003]. The recombinant expression yielded high amounts of pure protein that were all readily solubilized in 8 M urea.
Reconstitution of the Truncated Desmin, Des(ESA)AC244
After purification of Des(ESA)AC244 from bacte- rial inclusion bodies, we found that the mutant protein
precipitated to a significant amount as soon as the con- centration of urea was lowered to 2 M (data not shown) This is in stark contrast to the behavior of WT desmin. Here, the step-wise dialysis from 8 M urea to “Tris- buffer” is a renaturation schedule that yields soluble tetrameric coiled-coil complexes [Herrmann et al., 1996, 1999; Bär et al., 2006b]. The two major structures obtained in “Tris-buffer” were elongated “peanut”’- shaped aggregates with an approximate length of 120- 150 nm and a diameter of 20-30 nm and roundish aggre- gates with a diameter of ~30 nm (Figs. 2A-2C).
Moreover, analytical ultracentrifugation was car- ried out with the truncated protein in order to determine the physical state of the aggregates formed by the mutated protein. Depending on the protein concentration we obtained sedimentation coefficients of ~310 S (0.2 g/ L, open squares), ~400 S (0.34 g/L, filled squares) and ~490 S (0.48 g/L, open circles), respectively (Fig. 3A). Overall, s-values ranged from 200 to 500 S at lower and from 400 to 700 S at higher concentrations. This peak broadening and shift of the sedimentation curves for Des(ESA)AC244 with increasing protein concentration indicated that a heterogeneous population of higher- order complexes had formed and that the mutant aggre- gates interacted in a concentration-dependent fashion.
In order to investigate whether truncated and WT desmin protein are capable of interacting and form stable complexes, we denatured both proteins in 9.5 M urea and added increasing amounts of Des(ESA)AC244 to the WT (Figs. 3B and 3B’: ratio of WT:truncated protein: 5:2, open squares; 5:3, closed squares; and 5:4, open circles). Thereafter, we reconstituted these protein mixtures by dialysis into “Tris-buffer”. The size of complexes formed was then studied by sedimentation velocity ultra- centrifugation (Figs. 3B and 3B’). Notably, despite the presence of Des(ESA)AC244, WT desmin sedimented as a relatively homogenous species with peak s-values of
0.004
A
B
0.002
B’
0.04
g(s*)
0.002
0.001
0.02
0.00
0.000
0.000
200
400
600
800
2
4
6
8
10
200
400
600
800
s*(20,w) [Svedbergs]
s*(20,w) [Svedbergs]
s*(20,w) [Svedbergs]
~5.2 S (Fig. 3B). Moreover, at any of the three concentra- tions of the added mutant desmin, the concentration of WT desmin-as determined by the area under the curve - did not change, indicating that WT desmin did not form complexes with Des(ESA)AC244 in any significant amount during reconstitution. The asymmetric peak pro- file of the sedimentation velocity plot with a shift to higher s-values for WT desmin was similar to that meas- ured for desmin alone [Bär et al., 2006b]. In addition, Des(ESA)AC244 sedimented in very similar complexes both in the presence and absence of WT desmin (Figs. 3A and 3B’). The analysis of these mixtures after centrifuga- tion (30 min at RT, 10 psi in the Airfuge) by gel electro- phoresis revealed that Des(ESA)AC244 was completely recovered in the pellet after reconstitution of both proteins in “Tris-buffer”, whereas WT desmin was found exclu- sively in the supernatant fraction (data not shown). These results confirmed once again that heteropolymeric com- plexes did not form in significant amounts, therefore sug- gesting that the two proteins do not interact in a stable manner under low salt conditions.
In Vitro Assembly Studies of the Truncated Desmin, Des(ESA)AC244
Under standard assembly conditions, WT desmin assembled into smooth, extended filaments upon addi- tion of “assembly buffer” starting with unit-length filaments (ULFs; data not shown). At this time point, however, Des(ESA)AC244 already exhibited very long fibrous structures (Fig. 4A). These “thick filaments” were obviously formed by annealing of both the round and the extended precursors obtained after dialysis into
low salt/high pH buffer as they completely disappeared after addition of salt (compare Fig. 2A with Fig. 4A). The increase in length of the mutant “filaments” over time was difficult to follow since they rapidly started to form net-like fiber systems. However, we observed that by 60 min of assembly many of the strands were thinner than at the earlier time points whereas other parts of the “network” had significantly increased in diameter (Fig. 4B). In order to investigate the influence of the mu- tant protein on WT desmin assembly, we combined equi- molar amounts of Des(ESA)AC244 with WT desmin in 9.5 M urea prior to dialysis into “Tris-buffer”. After ini- tiation of assembly, the equimolar mixture the truncated protein with WT desmin exhibited only “peanut”- shaped aggregates with a length of ~120-150 nm and a diameter of ~20-30 nm in contrast to the extensively elongated fibers observed for Des(ESA)AC244 when assembled on its own (compare Figs. 4A and 4C). Hence, after establishing assembly conditions, the mu- tant protein aggregates are capable to interact with WT desmin. Thereby, the assembly of WT desmin is termi- nated during the elongation phase of IFs; vice versa, WT desmin completely blocked the elongation of the trunca- tion mutant aggregates. Instead of forming extended “filaments”, these peanut-shaped aggregates were often found to either terminate single short WT filaments or to be embedded in otherwise “normal” WT filaments (Fig. 4D). Hence, we speculate that upon addition of assembly buffer, WT desmin ULFs or their assembly precursors can associate with active binding sites on the aggregates and thereby block their assembly. In the equi- molar mixture these ULFs are present in vast numerical
A
B
C
D
100 nm
excess. Therefore, they are able to escape the deleterious effects of the mutant aggregates and can assemble into IFs to some extend.
In order to follow desmin assembly quantitatively and to monitor potential influences of the mutant on des- min IF network formation, we performed viscometric analyses as optimized previously for desmin [Bär et al., 2005a,b]. Roughly 20 to 30 min after initiation of assem- bly, WT desmin reaches a final relative viscosity of 0.1. The viscosity profiles obtained with mixtures of WT and mutant desmin clearly indicated that as little as ~10% of the mutant protein is sufficient to significantly obstruct the formation of long IFs and filament networks. In this case the relative viscosity was roughly lowered by 30% (data not shown). An equimolar mixture or the mutant alone showed the same low relative viscosity of 0.02, indicating rapid formation of protein aggregations.
In Vitro Assembly Studies of Desmin Fragments
In order to further understand the lack of interaction of Des(ESA)AC244 with WT desmin during renaturation and its unexpected assembly characteristics, we recombi- nantly produced the authentic desmin part represented by the mutant, DesAC240, and progressively longer trun-
cated desmin variants, namely, DesAC250, DesAC265, DesAC300, (Fig. 1). In addition, we investigated an amino-terminally truncated version of the DesAC240 pro- tein missing the first 30 amino acids of the head domain that are assumed to be active in complex interaction above the tetramer level, i.e. AN30DesAC240 (see [Herr- mann and Aebi, 2004]). Finally, we synthesized the “second half” of desmin, AN264Des, which comple- ments the “first half”, DesAC265.
Assembly studies of these truncated desmin variants revealed that DesAC240, very much like Des (ESA)AC244, exhibited roundish aggregates already after dialysis into Tris-buffer (Fig. 5, upper row, 0 s; for com- parison with Des(ESA)AC244 see Fig. 2A). However, unlike Des(ESA)AC244, this truncated desmin does not form extended thick filamentous strands upon initiation of assembly (see Fig. 4), but the individual aggregates fuse instantaneously into large proteinaceous masses (Fig. 5, 10 s and 10 min). Quite differently, DesAC250 already forms filamentous structures in “Tris-buffer” that fuse into large masses within minutes after addition of assem- bly buffer (Fig. 5, second row). The protein consisting of the first 264 amino-terminal amino acids, DesAC265, is partially soluble in “Tris-buffer” (Fig. 5, third row),
0 sec
10 sec
10 min
AC240
AC250
AC265
AC300
200 nm
but again leads to rapid formation of large aggregates after addition of salt. In contrast, DesAC300 is soluble under these ionic conditions as it is found exclu- sively in the supernatant in a sedimentation assay (see below). However, it rapidly aggregates into large ball-shaped structures after initiation of assembly. The C-terminal desmin fragment AN264Des remains com- pletely soluble after dialysis into “Tris-buffer” and does not form larger aggregates after addition of salt (data not shown).
These data obtained by electron microscopy, indi- cating different solubility properties for the individual mutants, were further specified by centrifugation studies. The proteins were reconstituted from urea into “Tris- buffer” either on their own (Fig. 6, panels designated
Note the formation of large roundish aggregates for DesAC240 and fila- mentous assemblies for DesAC250 already at low salt/high pH conditions. In contrast, DesAC265 only forms some small filamentous assemblies whereas DesAC300 remains largely soluble. Upon addition of salt, all proteins associated into large proteinaceous masses.
“mut”) or after mixing with equimolar amounts of WT desmin in 9.5 M urea (Fig. 6, designated “mix”) and then subjected to centrifugation in “Tris-buffer”. WT desmin is soluble under these conditions (left panel, S). In contrast, DesAC240 and DesAC250 are completely insoluble (panels DesAC240, mut and DesAC250, mut: closed arrowheads). Mutant DesAC265 distributes equally to the soluble and the insoluble fraction (panel DesAC265, mut: S and P, closed arrowheads), whereas DesAC300 is completely soluble again (panel DesAC300, mut: S, closed arrowhead). When the mutant proteins were mixed with WT desmin before centri- fugation, only DesAC250 and DesAC265 altered their sedimentation behavior, i.e. were now partly recovered in the soluble fraction (panels DesAC250, mix and
WT
AC240
AC250
AC265
AC300
| TSP | Mut Mix | Mut Mix | Mut Mix | Mut Mix |
|---|---|---|---|---|
| TSP T SF P | TSPTSP | T S P T S P | TS PTSP | |
Fig. 6. Sedimentation analyses of wild-type desmin, the truncated des- min variants DesAC240, DesAC250, DesAC265, DesAC300, AN30DesAC240, AN264Des (indicated by “Mut”) and of their equimo- lar mixtures with wild-type desmin (indicated by “Mix”) after re-consti- tution into “Tris-buffer”. The total protein (T), supernatant (S) and pellet fractions (P) are shown after separation by gel electrophoresis. Notice that wild-type desmin is largely soluble in Tris-buffer. DesAC240
and DesAC250 are found entirely in the pellet fraction (P, closed arrow- heads). A protein only 15 amino acids longer, i.e. DesAC265, is found both in the S and the P fraction. DesAC300 is completely soluble (closed arrowhead). After combining the respective fragments with wild-type desmin (“Mix”), the cosedimentation assay indicates that fragments must be at least as long as DesAC250 in order to be able to heteropoly- merise with wild-type protein (open arrowheads).
DesAC265, mix: S; compare closed and open arrow- heads). These data strongly indicate that the mutants DesAC250 and DesAC265 formed soluble complexes with WT desmin. For DesAC300 this was very likely also the case, but this could not be revealed with this assay, as the mutant protein was already soluble. How- ever, at least we could reveal that no insoluble com- plexes were generated as a result of the co-reconstitution (panel DesAC300, compare “mut” and “mix”).
In order to inspect if these associations had an effect on the assembly properties of both mutant and WT desmin, we performed an analysis of their assembly products by electron microscopy. Like Des(ESA)AC244, DesAC240 was found both to inhibit desmin WT fila- ment elongation and to be prevented from associating into larger scale structures by the presence of WT des- min (Fig. 7, panel WT+AC240). Hence, also after 60 min of assembly, desmin IFs appeared rather normal, except for a higher background of open filaments indi- cating some interference of the mutant protein with assembly. With DesAC250 the situation changed: fila- ments appeared indeed “normal” after 10 min of assem- bly but at 60 min they associated massively laterally into band-like fibers that partially started to unravel, too (panel WT+DesAC250). DesAC265 exhibited a com- pletely different phenotype, as filaments did not associ- ate longitudinally after they had reached an average length of about 600 nm (panel WT+DesAC265). In stark contrast, DesAC300 interacted intensively with WT des- min early on so that already after 10 s long, very open structures were observed that, however, at no time point organized into bona fide IFs but instead associated into open arrays of laterally associated fibers (panel WT+DesAC300). In summary, the mutant proteins DesAC250, DesAC265 and DesAC300 heavily interfere with WT desmin assembly when reconstituted together
in equimolar amounts. Vice versa, WT desmin drasti- cally influences the association behavior of these trun- cated proteins as they aggregate less massively but integrate into filamentous structures instead (compare e.g. DesAC250 and DesAC300 at 60 min in Figs. 5 and 7).
In order to understand the contribution of the first 30 amino acids of the “head”-domain of desmin to the aggregation of the truncated proteins DesAC240, we investigated another mutant protein, AN30DesAC240, lacking these amino-terminal amino acids. In “Tris- buffer”, this protein associated into small, fiber-like pro- tein structures already very different from the large, pea-like aggregates seen with DesAC240 (see Fig. 2). However, even after addition of assembly buffer, these structures did not assemble any further (data not shown) - quite in contrast to DesAC240. Hence, the first 30 amino acids are certainly engaged in the super-aggrega- tion of the first-half fragments of the desmin molecule (DesAC240, DesAC250) under low ionic conditions. As soon as part of the linker domain L12 (DesAC265) or even more so coil 2A and linker L2 (DesAC300) are present, this effect of the first 30 amino acids is drastically reduced (DesAC265) or even neutralized (DesAC300). Correspondingly, when reconstituted and assembled together with WT desmin, AN30DesAC240 did not integrate into desmin IFs but partially collapsed onto normal looking desmin IFs (data not shown).
Expression of Des(ESA)AC244 in Cultured Cells
We have generated a monoclonal antibody that specifically recognizes the mutant protein in order to distinguish its localization in transfected cells from that of endogenous WT desmin. For this purpose we immu- nized mice with the terminal eight amino acids of Des (ESA)AC244, i.e. “IAFLKESA”, coupled via a N-termi- nal cysteine to KLH. From the large number of
10 sec
10 min
60 min
WT+ AC240
WT+ AC250
WT+ AC265
2
WT+ AC300
antibody-producing cell clones that reacted in immuno- blotting with the “first half”-mutant, only one was entirely negative with WT desmin (Fig. 8, lanes 5 and 6). For control staining reactions, a commercially avail- able monoclonal rabbit anti-desmin antibody, raised against the carboxy-terminus of desmin and thus recog- nizing only WT desmin (Fig. 8, lanes 3 and 4), and a pol- yclonal anti-desmin serum recognizing both proteins were used (data not shown).
Transient transfection of SW13 cells, which are devoid of an endogenous cytoplasmic IF system, with a plasmid encoding Des(ESA)AC244 revealed a punctate
distribution pattern for the truncated protein in the nucleoplasm upon staining with mutant-specific mono- clonal antibody DK26 (Fig. 9A). Notably, all transfected cells exhibited predominant nuclear staining with few granules occasionally found in the cytoplasm. It was moreover of interest that these accumulations appeared to emerge through progressive fusion of small-sized granules different from the situation encountered after ectopical expression of tailless keratins or vimentin engi- neered to harbor a nuclear localization signal. In both cases rather few and distinct fragments with diameters of up to 2 um were formed [Bader et al., 1991; Herrmann
1
2
3
4
5 6
WT
(ESA)AC244
et al., 1993]. Nevertheless, the size or shape of nuclei in transfected SW 13 cells was not influenced significantly (data not shown), but was as heterogeneous as nuclear shape in this cell line usually is (see [Sarria et al., 1994]).
In contrast, transfected WT desmin generated a fil- amentous network in this adrenocortical carcinoma cell line (data not shown, see [Bär et al., 2006a]). In 3T3 fibroblasts the mutant also formed nucleoplasmic aggre- gates, partially displacing the chromatin scaffold (Fig. 9B, arrow). In addition, smaller aggregates were distrib- uted throughout the cytoplasm. Most notably, the mutant protein caused a severe reorganization of the endogenous vimentin IF network in 3T3 cells, but only few aggre- gates stained simultaneously for both vimentin and the mutant protein. This effect indicated that Des(E- SA)AC244 had induced the collapse but largely segre- gated from the vimentin system (Fig. 9B). Nuclear integrity was not affected in the transfected cells as revealed by staining for the nuclear envelope-specific proteins lamin B receptor and lamin A (data not shown). On occasion, large cytoplasmic accumulations of Des(E- SA)AC244 directly neighboring the nucleus induced small indentations of the nuclear shape in this region, however, without impacting the integrity of the nuclear envelope.
A similar behavior was also manifested in trans- fected HL-1 atrial cardiomyocyte-derived cells, where the endogenous desmin cytoskeleton exhibited a severe reorganization. Here, the mutant protein, however, did not appear in the nucleoplasm but was concentrated in small perinuclear aggregates. The colocalization of both
WT and mutant protein in these aggregates is visualized by a double-immunolabeling with our mutant-specific antibody. In summary, despite the fact that Des(E- SA)AC244 interacts with WT desmin only to a limited extent during assembly in vitro, the transfected Des(E- SA)AC244 is indeed “noxious” to cellular IF systems of the vimentin- and desmin-type as it strongly interferes with the in vivo organization of these cytoskeletal elements. Furthermore, Des(ESA)AC244 can clearly enter the nucleus of SW13 cells, but it does not carry endogenous cytoplasmic IF proteins in any large amount into the nuclei of these cells.
DISCUSSION
Mutations in the desmin gene can cause severe myopathies [Goldfarb et al., 1998; Munoz-Marmol et al., 1998]. As the pathomechanism of such mutations is far from being understood, we recently investigated the fundamental molecular properties of desmin, i.e. its complex formation and the mechanism of filament assembly. We found that at least half of the known muta- tions actually allow filament formation, although with some alterations in filament architecture. Moreover, through these studies it became clear that the consequen- ces of a single amino acid change are not predictable. Even the occurrence of a proline residue in the a-helical rod domain, normally considered as a “helix breaker”, was found to be compatible with filament assembly both in vitro and in transfected cells [Bär et al., 2005a].
With this background information, we investigated the only insertion mutation of the desmin gene reported up to now which leads to a frame-shift and hence a pre- mature termination signal in the desmin mRNA due to insertion of a single nucleotide. We found this particu- larly interesting from a structural point of view as we had previously demonstrated that both the isolated rod domain as well as coil 1A of vimentin are able to “trap” WT vimentin into a non assembly-competent complex [Herrmann and Aebi, 1998; Mücke et al., 2004]. More- over, in order to determine the effect of expressing the truncated desmin in cultured cells, we generated mono- clonal antibodies that recognize mutant but not WT des- min based on three unique amino acids (Glu-Ser-Ala) in the carboxy-terminus of the mutant protein. Surprisingly, this newly generated monoclonal antibody did not stain the desmin-containing aggregates in patient muscle tis- sue [Schröder et al., 2003]. Therefore, a reevaluation of the patients desmin gene was initiated which revealed that the mutated gene had been sequenced incorrectly and that the authentic disease mutation represented a deletion mutation of three nucleotides and not an inser- tion of one nucleotide [Schröder et al., 2007]. Neverthe- less, despite the lack of relation to an existing disease
A
(ESA)AC245
DAPI
Merge
B
(ESA)AC245
Vim
Merge
C
(ESA)AC245
Des
Merge
mutant, the results obtained with this near “first half” of desmin were indeed striking and motivated us to analyze the role of the end of coil 1B and that of coil 2A when exposed in a truncation situation for the assembly of desmin.
Based on molecular size alone, a 240 amino acids long fragment is slightly larger than the numerical first half of desmin, which would contain 235 amino acids. However, it does not enclose the entire coil 1B and breaks up one heptad before the supposed end of coil 1B
vimentin cytoskeleton, competition with binding sites of vimentin leads to a breakdown of this IF network and to perinuclear aggregation. Like in SW13 cells, the intranuclear protein aggregates displace the hetero- chromatin (B, arrow). Blue, DAPI; green, Des(ESA)AC244; red, vimentin (Vim). Bar, 10 um. (C) In HL-1 cardiomyocytes, the mutant protein forms aggregates in the perinuclear region but is not found in the nucleoplasm. The endogenous desmin, specifically stained by a monoclonal anti-desmin antibody, is starting to collapse and colocalizes partially with the desmin variant in perinuclear aggregates. Blue, DAPI; red, wild-type desmin (Des); green, Des(ESA)AC244. Bar, 10 um.
[Herrmann and Aebi, 2004]. The novel last three amino acids in Des(ESA)AN243, Glu-Ser-Ala, fit well into a coiled-coil topology. They replace Lys-Leu-His, which are in f, g and a position of the heptad. The first and most remarkable feature of this desmin fragment was its pronounced insolubility in urea, as it started to precipi- tate out of solution in 2 M urea. This behavior is indeed unusual compared to WT desmin, which is completely soluble even in the absence of urea as long as the ionic strength is low. We conclude that the head domain of
this truncation polypeptide which lacks its binding region on a neighboring desmin dimer can now interact with neighboring dimers well beyond the tetrameric complex as observed with WT IF proteins (see Fig. 1). Therefore we assume that the absence of coil 2A, linker L2 and the start of coil 2B allows the head domains to engage in further interactions with additional dimers. The size of the structures eventually formed depends strongly on the protein concentration of the desmin mu- tant as revealed by analytical ultracentrifugation. How and why these huge complexes are able to assemble so rapidly by longitudinal annealing after raising the ionic strength is not clear but predicts some kind of order within these peanut-shaped “assembly-precursors”. Most notably, its elongation reaction after salt addition is completely blocked in the presence of desmin WT tet- ramers. Very likely, the higher ionic condition renders these proteins reactive enough to bind to tetramers or precursors of ULFs. Interestingly, the “first half” aggre- gates are, at early assembly time points, not decorated by ULFs, which would be easily detected by electron mi- croscopy (see Fig. 4C). At later time points, however, these reactive aggregates get incorporated into the IF system such that they either act as centers for the origin of assembly or by blocking elongation.
In close proximity to the truncation reported here, a potential cleavage site of caspase 6 has been reported at Asp264 in the linker region L12 [Chen et al., 2003]. Cleavage of desmin at this site was expected to produce a dominant-negative inhibitor of IF formation, resulting in the breakdown of the desmin cytoskeleton and the propagation of apoptosis. Our transfection experiments suggest that the isolated “first half” desmin is indeed capable to negatively interfere with desmin and vimentin filament system organizing factors. We have demon- strated that the amino-terminal part of desmin is not able to integrate into the desmin cytoskeleton but that it forms aggregates, which associate with desmin filaments, thus preventing extended filamentous network formation. Hence, the truncated desmin is a true disruptor of endog- enous IF systems as shown here for the vimentin cyto- skeleton in fibroblasts and the desmin cytoskeleton in cardiomyocyte-derived cells. Both its cellular distribution and its effect on the IF system truly mirrors the outcome of transfection experiments with certain point mutated desmins, in particular those that are neither able to form filaments on their own in vitro nor get incorporated into an endogenous vimentin system [Bär et al., 2006a]. Since the observed reorganization of the vimentin system is very similar to that exhibited after destruction of the microtubule system with drugs such as colcemid, part of the effects that these mutant desmin aggregates exert may affect the binding of IFs to microtubules and micro- tubule-associated protein complexes compounds.
A very interesting transition indeed occurs as the truncated desmin is slightly increased in size. From DesAC250 on, i.e. with only 10 amino acids more, the mutant proteins associate into complexes with WT des- min and are integrated into growing filaments. We con- clude from this behavior that as soon as the complete coil 1 is present in the amino-terminal desmin fragments, they form complexes that can associate with desmin tet- ramers to soluble complexes well below ULF size, which we did not observe before raising the ionic strength. Moreover, as soon as the fragments include the linker L2 region and the first heptad of coil 2B, they form entirely soluble complexes. It has been reported previously that this short linker segment, which probably forms an «-he- lix, may have an important function in the dimerization of IF molecules, well before coiled-coil formation is complete [Hess et al., 2006]. In vimentin, the Tt - Tt stack- ing of aromatic residues in linker L2 (Trp290, Tyr291, and Phe295 - corresponding to Trp295, Tyr296 and Val300 in desmin) have been demonstrated to play an eminent role in this association. The molecular details for these interactions will be addressed in future studies employing techniques such as analytical ultracentrifuga- tion and electron paramagnetic resonance with corre- sponding spin-labeled fragments.
ACKNOWLEDGMENTS
HL-1 cells were kindly provided by Dr. W. C. Claycomb, Louisiana State University Health Science Center, New Orleans, LA, USA. We gratefully acknowl- edge the technical assistance of Tatjana Wedig and Michaela Hergt.
REFERENCES
Amrani N, Sachs MS, Jacobson A. 2006. Early nonsense: mRNA decay solves a translational problem. Nat Rev Mol Cell Biol 7(6):415-425.
Arbustini E, Pasotti M, Pilotto A, Pellegrini C, Grasso M, Previtali S, Repetto A, Bellini O, Azan G, Scaffino M, Campana C, Piccolo G, Viganò M, Tavazzi L. 2006. Desmin accumulation restric- tive cardiomyopathy and atrioventricular block associated with desmin gene defects. Eur J Heart Fail 8(5):477-483.
Bader BL, Magin TM, Freudenmann M, Stumpp S, Franke WW. 1991. Intermediate filaments formed de novo from tail-less cytokeratins in the cytoplasm and in the nucleus. J Cell Biol 115(5):1293-1307.
Bär H, Strelkov SV, Sjöberg G, Aebi U, Herrmann H. 2004. The biol- ogy of desmin filaments: How do mutations affect their structure, assembly, and organisation? J Struct Biol 148(2): 137-152.
Bär H, Fischer D, Goudeau B, Kley RA, Clemen CS, Vicart P, Herr- mann H, Vorgerd M, Schröder R. 2005a. Pathogenic effects of a novel heterozygous R350P desmin mutation on the assembly of desmin intermediate filaments in vivo and in vitro. Hum Mol Genet 14(10):1251-1260.
Bär H, Mücke N, Kostareva A, Sjöberg G, Aebi U, Herrmann H. 2005b. Severe muscle disease-causing desmin mutations inter- fere with in vitro filament assembly at distinct stages. Proc Natl Acad Sci USA 102(42):15099-15104.
Bär H, Kostareva A, Sjöberg G, Sejersen T, Katus HA, Herrmann H. 2006a. Forced expression of desmin and desmin mutants in cul- tured cells: Impact of myopathic missense mutations in the central coiled-coil domain on network formation. Exp Cell Res 312(9):1554-1565.
Bär H, Mücke N, Ringler P, Müller SA, Kreplak L, Katus HA, Aebi U, Herrmann H. 2006b. Impact of disease mutations on the desmin filament assembly process. J Mol Biol 360(5):1031- 1042.
Bär H, Goudeau B, Wälde S, Casteras-Simon M, Mücke N, Shatunov A, Goldberg YP, Clarke C, Holton JL, Eymard B, Katus HA, Fardeau M, Goldfarb L, Vicart P, Herrmann H. 2007. Conspic- uous involvement of desmin tail mutations in diverse cardiac and skeletal myopathies. Hum Mutat 28(4):374-386.
Capetanaki Y, Bloch RJ, Kouloumenta A, Mavroidis M, Psarras S. 2007. Muscle intermediate filaments and their links to mem- branes and membranous organelles. Exp Cell Res 313(10): 2063-2076.
Chen F, Chang R, Trivedi M, Capetanaki Y, Cryns VL. 2003. Caspase proteolysis of desmin produces a dominant-negative inhibitor of intermediate filaments and promotes apoptosis. J Biol Chem 278(9):6848-6853.
Claycomb WC, Lanson NA, Jr, Stallworth BS, Egeland DB, Delcarpio JB, Bahinski A, Izzo NJ, Jr. 1998. HL-1 cells: A cardiac mus- cle cell line that contracts and retains phenotypic characteris- tics of the adult cardiomyocyte. Proc Natl Acad Sci USA 95(6):2979-2984.
Franke WW, Schmid E, Osborn M, Weber K. 1978. Different interme- diate-sized filaments distinguished by immunofluorescence microscopy. Proc Natl Acad Sci USA 75(10):5034-5038.
Fuchs E, Cleveland DW. 1998. A structural scaffolding of intermedi- ate filaments in health and disease. Science 279(5350):514- 519.
Goldfarb LG, Park KY, Cervenakova L, Gorokhova S, Lee HS, Vas- concelos O, Nagle JW, Semino-Mora C, Sivakumar K, Dalakas MC. 1998. Missense mutations in desmin associated with fami- lial cardiac and skeletal myopathy. Nat Genet 19(4):402-403.
Goldfarb LG, Vicart P, Goebel HH, Dalakas MC. 2004. Desmin my- opathy. Brain 127 (Part 4):723-734.
Goldman RD, Milsted A, Schloss JA, Starger J, Yerna MJ. 1979. Cytoplasmic fibers in mammalian cells: Cytoskeletal and con- tractile elements. Annu Rev Physiol 41:703-722.
Haubold K, Herrmann H, Langer SJ, Evans RM, Leinwand LA, Klymkowsky MW. 2003. Acute effects of desmin mutations on cytoskeletal and cellular integrity in cardiac myocytes. Cell Motil Cytoskeleton 54(2):105-121.
Hedberg KK, Chen LB. 1986. Absence of intermediate filaments in a human adrenal cortex carcinoma-derived cell line. Exp Cell Res 163(2):509-517.
Herrmann H, Aebi U. 1998. Structure, assembly, and dynamics of intermediate filaments. Subcell Biochem 31:319-362.
Herrmann H, Aebi U. 2004. Intermediate filaments: Molecular structure, assembly mechanism, and integration into function- ally distinct intracellular scaffolds. Annu Rev Biochem 73:749-789.
Herrmann H, Eckelt A, Brettel M, Grund C, Franke WW. 1993. Tem- perature-sensitive intermediate filament assembly. Alternative
structures of Xenopus laevis vimentin in vitro and in vivo. J Mol Biol 234(1):99-113.
Herrmann H, Haner M, Brettel M, Müller SA, Goldie KN, Fedtke B, Lustig A, Franke WW, Aebi U. 1996. Structure and assembly properties of the intermediate filament protein vimentin: The role of its head, rod and tail domains. J Mol Biol 264(5):933-953.
Herrmann H, Haner M, Brettel M, Ku NO, Aebi U. 1999. Characteri- zation of distinct early assembly units of different intermediate filament proteins. J Mol Biol 286(5):1403-1420.
Herrmann H, Bär H, Kreplak L, Strelkov SV, Aebi U. 2007. Interme- diate filaments: From cell architecture to nanomechanics. Nat Rev Mol Cell Biol 8(7):562-573.
Hess JF, Budamagunta MS, Shipman RL, FitzGerald PG, Voss JC. 2006. Characterization of the linker 2 region in human vimen- tin using site-directed spin labeling and electron paramagnetic resonance. Biochemistry 45(39):11737-11743.
Köhler G, Milstein C. 1975. Continuous cultures of fused cells secret- ing antibody of predefined specificity. Nature 256(5517):495- 497.
Lazarides E. 1982. Intermediate filaments: A chemically heterogene- ous, developmentally regulated class of proteins. Annu Rev Biochem 51:219-250.
Mücke N, Wedig T, Burer A, Marekov LN, Steinert PM, Langowski J, Aebi U, Herrmann H. 2004. Molecular and biophysical char- acterization of assembly-starter units of human vimentin. J Mol Biol 340(1):97-114.
Munoz-Marmol AM, Strasser G, Isamat M, Coulombe PA, Yang Y, Roca X, Vela E, Mate JL, Coll J, Fernandez-Figueras MT, Navas-Palacios JJ, Ariza A, Fuchs E. 1998. A dysfunctional desmin mutation in a patient with severe generalized myopa- thy. Proc Natl Acad Sci USA 95(19):11312-11317.
Niehrs C, Huttner WB, Ruther U. 1992. In vivo expression and stoichiometric sulfation of the artificial protein sulfophilin, a polymer of tyrosine sulfation sites. J Biol Chem 267(22): 15938-15942.
Ohara O, Gahara Y, Miyake T, Teraoka H, Kitamura T. 1993. Neuro- filament deficiency in quail caused by nonsense mutation in neurofilament-L gene. J Cell Biol 121(2):387-395.
Paulin D, Huet A, Khanamyrian L, Xue Z. 2004. Desminopathies in muscle disease. J Pathol 204(4):418-427.
Sarria AJ, Lieber JG, Nordeen SK, Evans RM. 1994. The presence or absence of a vimentin-type intermediate filament network affects the shape of the nucleus in human SW-13 cells. J Cell Sci 107 (Part 6):1593-1607.
Schröder R, Goudeau B, Simon MC, Fischer D, Eggermann T, Clemen CS, Li Z, Reimann J, Xue Z, Rudnik-Schöneborn S, Zerres K, van der Ven PF, Fürst DO, Kunz WS, Vicart P. 2003. On noxious desmin: Functional effects of a novel heterozygous desmin insertion mutation on the extrasarcomeric desmin cyto- skeleton and mitochondria. Hum Mol Genet 12(6):657-669.
Schröder R, Goudeau B, Casteras-Simon M, Fischer D, Eggermann T, Clemen CS, Li Z, Reimann J, Xue Z, Rudnik-Schöneborn S, Zerres K, van der Ven PF, Fürst DO, Kunz WS, Vicart P. 2007. Corrigendum: On noxious desmin: Functional effects of a novel heterozygous desmin insertion mutation on the extrasar- comeric desmin cytoskeleton and mitochondria. Hum Mol Genet 12(21):1-2.
Strelkov SV, Herrmann H, Geisler N, Wedig T, Zimbelmann R, Aebi U, Burkhard P. 2002. Conserved segments 1A and 2B of the intermediate filament dimer: Their atomic structures and role in filament assembly. Embo J 21(6):1255-1266.