| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
,1
* Center for Human and Clinical Genetics, Leiden University Medical Center, Leiden, The Netherlands;
Institute of Human Genetics, International Centre for Life, Newcastle-upon-Tyne, UK;
Department of Parasitology, Leiden University Medical Center, Leiden, The Netherlands; and
INSERM EMI-104, DRDC CEA-Grenoble, France
2Correspondence: Leiden Univesity Medical Center, Department of Human Genetics, Hlbinusdreef 2, 2333 ZA Leiden, The Netherlands. E-mail: maarel{at}lumc.nl
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: protein interaction • dysferlinopathies • membrane repair • LGMD2B • MM
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3)
. Regardless of phenotype, dysferlinopathy patients typically have absent or severely reduced levels of dysferlin in skeletal muscle.
There is a deficiency of muscle membrane repair in dysferlin-deficient mouse models, leading to the proposal that this defines a new pathogenic mechanism in muscular dystrophy (MD). However, the precise biochemical function of dysferlin remains unknown. Dysferlin, and its homologues otoferlin and myoferlin, show homology to the Caenorhabditis elegans spermatogenesis factor fer-1 that mediates spermatid vesicles/plasma membrane fusion (4)
. Dysferlin is widely expressed, and in skeletal muscle, it is primarily found at the plasma membrane (5)
and in cytoplasmic vesicles (6)
. Myoferlin is also expressed at the plasma membrane, and in addition, it is found at the nuclear envelope. Dysferlin and myoferlin contain six C2 domains, which are typically involved in the calcium-dependent binding to phospholipids and the most amino-terminal C2 domain, C2A, shows calcium-dependent phospholipid binding in vitro (7)
. Various proteins have been suggested to interact with dysferlin, including annexins A1 and A2, caveolin 3, calpain 3 (CAPN3), affixin (ß-parvin), and the dihydropyridine receptor (DHPR) (8
9
10
11
12
13)
.
Although dysferlin is not considered as an integral component of the dystrophin glycoprotein complex (DGC), patients with mutations in the DGC often show some reduced or altered dysferlin expression (6)
. Dysferlin-null mice develop a slowly progressive muscular dystrophy with vesicular accumulations and sarcolemmal disruptions similar to those observed in human dysferlinopathy, and the vesicular accumulations are particularly prominent during skeletal muscle regeneration. Isolated myofibers from dysferlin null mice show impaired resealing following high-intensity laser irradiation, implicating dysferlin in some aspect of calcium-dependent membrane repair (14)
.
By means of phage display, we recently reported the successful selection of Llama-derived heavy-chain antibody (Ab) (HCAb) fragments specific for dysferlin (11)
. We demonstrated these HCAb fragments to be functional in several immunological techniques, including immunoprecipitation. Using immunoprecipitation followed by mass spectrometry, we demonstrate here for the first time that AHNAK, a protein implied in cell membrane differentiation, repair, and signal transduction (15
16
17)
, is in complex with dysferlin in skeletal muscle. GST pull-down assays and coimmunofluorescence microscopy provide further support for a direct interaction between dysferlin and AHNAK, as does the secondary reduction in skeletal muscle labeling for AHNAK in muscle from dysferlinopathy patients. We also found that dysferlin and AHNAK relocalize in a rat model of muscle regeneration, providing further evidence for a functional relationship between both proteins.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
C2C12 cell cultures
C2C12 mouse cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM) with phenol red (Gibco BRL, Carlsbad, CA) supplemented with 10% fetal calf serum (FCS) (Gibco BRL), sodium pyruvate (final concentration 1 mM, Gibco BRL), and penicillin-streptomycin (100 IU/100 µg/ml, Gibco BRL). The cells were cultured in an incubator with 10% CO2 at 37°C.
Cells were induced to differentiate and fuse at 30–50% confluency by switching to serum-deprived medium (4% horse serum in DMEM). After differentiation for 7 days, cells were harvested, resuspended in lysis buffer 1 (50 mM Tris, pH 7.4, 150 mM NaCl, 0.15% CHAPS (3-[3-cholamidopropyl]dimethylammonio)-1-propanesulfonic acid)) 1x protease inhibitor cocktail (Roche Molecular Biochemicals, Basel, Switzerland), and incubated on ice for 10 min. Cellular debris was removed by brief centrifugation at 8,000 g for 5 min at 4°C, and the protein concentration of the supernatant was assayed with by SDS-PAGE.
Recombinant protein production
pcDNAI/Amp eukaryotic expression vectors containing N-AHNAK (residues 2–252), M-AHNAK (residues 821–1330), C-AHNAK (residues 4646–5643) were a kind gift from Dr. Takashi Hashimoto (Keio University School of Medicine, Tokyo, Japan). DNA fragments encoding the protein fragments N-AHNAK, M-AHNAK, and C-AHNAK, respectively, were obtained by BamHI/XhoI restriction from pcDNAI/Amp and ligated in the prokaryotic expression vector pET28a-GST (modified from pET28a (Novagen, Madison, WI) with an additional GST tag). Subsequently, pET28a-GST-C-AHNAK was split in two parts by BamHI/SacI digestion (designated C1-AHNAK, from residue 4646–5145) and by SacI/XhoI digestion (designated C2-AHNAK, from residue 5146–5643) and ligated in BamHI/SacI-digested, or SacI/XhoI-digested pET28c-GST, respectively.
To generate the carboxy terminus of AHNAK2, C2-AHNAK2 (aa 5146–5637), a 1487 bp fragment was polymerase chain reaction (PCR) amplified from cDNA clone (clone ID: DKFZp686J02145, RZPD, Berlin, Germany) with forward primer (5'-GGGTCGACCCTCTCCCTTTTCAGA-3') and reverse primer (5'-GCGGCCGCTCAGCCTTCATT-3') and cloned into TOPO blunt vector (Invitrogen, Paisley, UK). Subsequently, the fragment was digested with SalI/NotI and ligated in the SalI/NotI-digested prokaryotic expression vector pGEX 4T-3 (Amersham Biosciences, Uppsala, Sweden). Unfused GST protein was used as control.
To generate DYSF C2A (aa 1–130), a 392-bp fragment was PCR amplified from a plasmid that contains entire human dysferlin coding sequence with forward primer (5'- ATC GGG ATC CAT GCT GAG GGT CTT CAT C-3') and reverse primer (5'- ATC GCT CGA GCA CAG CTC CAG GCA GCG G-3') and cloned into the BamHI/XhoI-digested pET28a. Site-directed mutagenesis was performed to generate dysferlin C2A (V67D) by use of the QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA). DYSF2 (aa 2–1080) was generated by the plasmid that contains entire human dysferlin coding sequence with EcoRI/XhoI and ligated in the EcoRI/XhoI-digested prokaryotic expression vector pET28a. DYSF1 (aa 2–245) and DYSF3 (aa 1666–1788) were generated, as described previously (11)
. C2D (aa 1152–1285) and C2Q (aa 1314–1476) were generated in vitro using a TNT Coupled Rabbit Reticulocyte Lysate System as directed by the manufacturer (Promega, Madison, WI, USA).
Recombinant myoferlin proteins MYOF1 (aa 2–245), MYOF2 (aa 1–130) and MYOF C2A (aa 1–85) were produced by PCR amplification of 740-bp, 407-bp, 272-bp fragments from cDNA clone (clone ID: DKFZp686C16167Q2, RZPD), respectively. MYOF1 was generated with forward primer (5'-GGG AAT TCC TGC GAG TGA TTG T-3') and reverse primer (5'-GGG AAG CTT AAA CAA CTC ATC-3'); MYOF2 with forward primer (5'-GGGAATTCATGCTGCGAGTGA-3') and reverse primer (5'-GGGAAGCTTTGGAGCAGAAG-3'); and MYOFC2A with forward primer (5'-GGGAATTCATGCTGCGAGTGA-3') and reverse primer (5'-GGGAAGCTTAGTCGCCGTG-3'), respectively, and cloned into the TOPO 2.1 vector (Invitrogen). Subsequently, the fragments were digested with EcoRI/SalI, EcoRI/HindIII, and EcoRI/HindIII, respectively, and ligated in the prokaryotic expression vector pET28a (Novagen, Madison, WI, USA). All constructs were sequence-verified (LGTC, Leiden, The Netherlands).
The prokaryotic expression constructs were used for recombinant protein production in BL21 (DE3)-RIL E. coli (Stratagene, La Jolla, CA, USA) cells according to the manufacturer’s instructions. Briefly, bacterial cells were grown to log phase and recombinant protein production was initiated by the addition of IPTG (Fermentas, St. Leon-Rot, Germany) to a final concentration of 1 mM. After 2 h induction at 30°C, the cells were collected by centrifugation and lysed in lysis buffer 2 (50 mM Tris, pH 7.4, 1 mM EDTA, 1.5 mg/ml lysozyme, 0.15 M NaCl, 1% Triton X-100), and sonicated for 2 x 10 s with 15-s intervals on ice. For the pull-down assay, after centrifugation at 13,000 rpm for 30 min at 4°C, the supernatant containing soluble GST fusion was recovered. Fusion proteins were immobilized to Glutathione-Sepharose 4B (Amersham Pharmacia, Piscataway, NJ, USA), incubated at room temperature for 30 min and then centrifuged at 500 g for 5 min. The supernatant was removed and the glutathione-sepharose was washed 3 times with binding buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 0.2% Triton X-100). The GST/GST-fusion protein bound beads were ready for pull-down assay.
Antibodies
The antibodies used in this study were as follows. The monoclonal anti-dysferlin Ab NCL-Hamlet (Novocastra, Newcastle, UK) was used in a dilution of 1:300 for Western blot analysis and at a dilution of 1:50 or 1:100 for immunofluorescent microscopy. Monoclonal antibodies against ß-dystroglycan, dystrophin, ß- and 
-sarcoglycan, and spectrin were as described previously (18)
. Monoclonal antibody (mAb) against caveolin-3 (Becton Dickinson, Franklin Lakes, NJ) was used at a dilution of 1:1,000. Mouse monoclonal CD31 Ab (Abcam, Cambridge, UK) was diluted 1:500 for immunostaining. Mouse mAb against DHPR (Abcam, Cambridge, UK) was diluted 1:200 for immunostaining. Secondary antibodies goat anti-mousealexa488 (Molecular Probes, Eugene, OR), goat anti-rabbitalexa594 (Molecular Probes), sheep anti-mouse (Amersham), swine anti-rabbit (DAKO, Glostrup, Denmark) and rabbit anti-mouseHRP (DakoCytomation, Glostrup, Denmark) were diluted 1:250, 1:1,000, 1:500, 1:200 and 1:1,000, respectively. Affinity purified KIS/AHNAK polyclonal antibodies were used in appropriate dilutions as described previously (19)
. Mouse anti-T7HRP (Novagen, Madison, WI) was diluted 1:15,000 for Western blot analysis.
Coimmunoprecipitation and Western blot analysis
Immunoprecipitation of dysferlin was performed basically as described by Matsuda (13)
. Normal mouse or human skeletal muscle was homogenized on dry ice in the ratio of 1:10 in a lysis buffer 1 for 1 h at 4°C. 500 µl of protein homogenate was incubated first with 20 µg of either F4 or H7 HCAb fragment overnight (o/n) at 4°C and next precipitated with 25 µl of protein A Sepharose CL-4B for 1 h at 4°C. Immune complexes were eluted by boiling in 50 µl of 2 x SDS-PAGE sample buffer for 5 min and 20 µl was loaded on a 7% SDS-PAGE gels (Bio-Rad, Hercules, CA). After separation, proteins were transferred to PVDF membrane (Roche, Basel, Switzerland) and blocked with 4% skimmed milk (ELK, Campina, Woerden, The Netherlands) in PBS supplemented with 0.02% Tween 20 for 1 h at room temperature. Then, blots were incubated with NCL-hamlet (1:300) for 1 h at RT and rabbit anti-mouseHRP (1:2,000) for 1 h at RT. Enhanced chemiluminescence (ECL) plus (Amersham Pharmacia Biotech) was used for visualization.
Mass spectrometric analysis and protein identification
Immunoprecipitated proteins were separated on SDS-PAGE. The protein bands were excised from sypro-stained gel and trypsin digested, as described previously (20)
. Briefly, minced gel pieces were first washed with H2O and acetonitril, reduced with dithiothreitol at 56°C for 45 min, and then alkylated by iodoacetamide in the dark for 30 min. The gel was incubated in 30 µl of a 5 ng/µl modified trypsin solution in 50 mM ammonium bicarbonate, pH 8.6 and incubated at 37°C o/n. The digests were acidified with aqueous TFA to a final concentration of 0.1% and the peptides were extracted with one change of 50 mM ammonium bicarbonate. The sample was desalted and concentrated with a 10 µl ZipTip C18 (Millipore, Bedford, MA), following the instructions provided by the manufacturer. Peptides were eluted with 1.5 µl 
-cyano 4-hydroxycinnamic acid matrix (0.33 mg/ml in Aceton/Ethanol (1:3) onto the MALDI target plate. Mass spectrometry (MS) and tandem mass spectrometry (MS/MS) data were acquired with the MALDI-TOF-TOF mass spectrometer (Ultraflex TOF/TOF mass spectrometer, Bruker Daltonics, Bremen, Germany). Proteins were identified by peptide mass fingerprint (PMF) and MS/MS peptide sequencing with the searching program MASCOT (http:www.matrixscience.com). NCBInr was the database used with the following search parameters: a mass tolerance of 0.1 Da and one miscleavage were allowed and peptide mass changes due to carbamidomethylation of cysteine and oxidation of methionine were taken into account.
Pull-down assays
For GST-binding assay, normal human skeletal muscle and mouse C2C12 cells were homogenized on dry ice and lysed in the ratio of 1:10 of lysis buffer 1 plus 1x protease inhibitor cocktail (Roche Molecular Biochemicals, Basel, Switzerland). Fusion proteins produced in E. coli were lysed in buffer 2. For calcium-dependent binding assay, fusion proteins were lysed in buffer 2 supplemented with either 1 mM Ca2+ or 2 mM EDTA. Lysates were centrifuged for 15 min at 4°C, and the supernatants were precleared with glutathione-sepharose 4B (Amersham) for 1 h at 4°C. Five-hundred-microliter aliquots of precleared supernatant were incubated with 10 µg of purified GST fusion protein affinity beads in binding buffer supplemented 2% skimmed milk for muscle lysates and 1% BSA for fusion protein lysates for o/n at 4°C or with unfused GST protein as a control. For calcium-dependent binding assays, binding buffer was supplemented with 1% BSA and either 1 mM Ca2+ or 2 mM EDTA. After binding, the beads were washed 3 times in binding buffer without milk/BSA and 1 time in 50 mM Tris-HCl pH 7.4, then eluted by boiling in 50 µl of 2xSDS-PAGE sample buffer for 5 min. Ten microliters were loaded on 7% SDS-PAGE gels to detect endogenous dysferlin or on 10% SDS-PAGE gels to detect recombinant dysferlin protein fragments. After separation, proteins were transferred to PVDF membranes and dried o/n. Then, blots were incubated with NCL-hamlet (1: 300) for 2 h at RT and rabbit anti-mouseHRP (1: 2,000) for 1 h at RT for detection of endogenous dysferlin; anti-T7HRP Ab for the detection of recombinant dysferlin protein fragments. ECL plus was used for visualization.
Immunohistochemistry
For immunohistochemical examinations, transverse, or longitudinal muscle cryosections of 5 µm thickness were fixed in 3.7% formaldehyde containing 0.1% triton X for 30 min, following by preincubation with PBS containing 4% skimmed milk at room temperature for 2 h. The sections were next incubated with primary Ab fragments overnight at 4°C, and subsequently with affinity-purified secondary antibodies at room temperature for 40 min, following by incubation of fluorescein-labeled tertiary Ab for 40 min at RT. Background staining was performed by omitting the primary Ab from the first step. The sections were washed with PBS, dehydrated with 70, 90, 100% ethanol and mounted in a DAPI (50 ng/µl)/ Vectashield mounting medium (Burlingame, CA). Final preparations were analyzed with a Leica DMRA2 fluorescence microscope, and images were obtained using Leica CW4000 digital system.
Regenerating rat muscle
Female Wistar rats (90–100 g) were obtained from an accredited breeder and maintained in accordance with local procedures for animal handling. Notexin (200 µl of a 10 µg/ml solution in 0.9% w/v NaCl) was injected subcutaneously (s.c.) into the dorsolateral aspect of one hind limb, so as to bathe the underlying soleus muscle in toxin. Four days after toxin injection, the soleus muscle was removed and frozen in isopentane; sections were cut for routine immunohistochemistry. The contralateral soleus muscle acted as a control. Thawed sections were treated in 0.1% Triton X-100 in 1x PBS for 30 min at room temperature and endogenous peroxidase blocked using DAKO peroxidase block. Sections were incubated in DAKO protein block for 5 min and overnight in primary Ab diluted in DAKO background reducing Ab diluent. After washing HRP-labeled secondary Ab previously preabsorbed in normal rat serum was added and incubated for 1 h at room temperature. After washing TSA amplification reagent was added as the manufacturer’s instructions (Perkin-Elmer). Images were collected and analyzed on a Zeiss LSM-510 confocal microscope.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, a most interesting novel component of the immunoprecipitation (IP)-complex was AHNAK, which was chosen for further characterization.
To confirm that AHNAK was coimmunoprecipitated with dysferlin, independent IPs were analyzed by Western blot analysis with affinity purified polyclonal antibodies raised against AHNAK (19)
. As shown in Fig. 1
B, multiple protein bands corresponding to different AHNAK protein isoforms were detected in the IP fraction with HCAb fragment F4, but not in the IP fraction of unrelated HCAb fragments or in the IP fraction without primary HCAb fragments. This specific immunoreactivity indicates an interaction between dysferlin and AHNAK.
|
Interaction of GST AHNAK fusion proteins with endogenous dysferlin
Two homologues of AHNAK exist in the human and mouse genome: AHNAK localized at human chromosome 11 and AHNAK2 at human chromosome 14 (21)
. Both proteins are composed of a large number of highly conserved repeat segments but differ in their C-terminal domain. The polyclonal AHNAK antisera react to both proteins and cannot discriminate them.
To investigate whether AHNAK can bind dysferlin and which domains are responsible for specific binding to dysferlin, a series of recombinant GST-AHNAK fusion proteins (Fig. 2
A) representing the amino-terminal domain of AHNAK (N-AHNAK: aa 2–252), a central repeat unit (M-AHNAK: aa 821–1330), and two carboxyterminal domains (C1-AHNAK: aa 4646–5145 and C2-AHNAK: aa 5146–5643) were applied in a GST pull-down assay with human muscle (Fig. 2B
) or mouse C2C12 protein extracts (Fig. 2C
). Western blot analysis of the pull-down fractions using the dysferlin-specific mAb NCL-Hamlet showed that only the most carboxyterminal end of AHNAK, C2-AHNAK, was able to bind to dysferlin. No binding was observed for equivalent amounts of the GST fusion proteins N-AHNAK, M-AHNAK, C1-AHNAK, or for the control unfused GST protein.
|
In C2C12 cells, we identified two protein bands reactive to NCL-Hamlet, a prominent high MW isoform and a less prominent low MW isoform of dysferlin. Although the low MW isoform is most abundant in the precleared protein fraction, in our pull-down assay from C2C12 cells, only the high MW isoform of dysferlin was pulled down by GST-C2-AHNAK but not the low MW isoform (Fig. 2C
).
Interaction of GST-C2-AHNAK with T7-tagged dysferlin fusion proteins
To determine whether both AHNAK proteins can bind dysferlin, whether this interaction is direct or indirect, and to further define the interaction domain of dysferlin, we produced HIS- and T7-tagged fusion proteins, representing different domains of dysferlin, including C2A (aa 2–130), C2D (aa 1152–1285) and C2Q (aa 1314–1476), as well as the dysferlin protein fragments DYSF1 (aa 2–245), DYSF2 (aa 1666–1788), and DYSF3 (aa 2–1080). After incubation with the C-terminal GST-AHNAK fusion construct GST-C2-AHNAK or control GST protein preimmobilized to glutathione Sepharose 4B beads, the GST pull-down fractions were analyzed on Western blot analysis probed with anti T7HRP antibodies. Three recombinant protein fragments all incorporating the C2A domain of dysferlin (C2A, DYSF1, and DYSF3) were specifically pulled down by GST-C2-AHNAK (Fig. 3
A–C). No binding was observed for equivalent amounts of the fusion proteins lacking the C2A domain of dysferlin (DYSF2, C2D, C2Q; data not shown) and the control GST protein. Furthermore, a C-terminal GST-fusion construct of AHNAK2 (GST-C2-AHNAK2: aa 5146–5637), was also able to pull down dysferlin fragments incorporating the C2A domain of dysferlin. This interaction of dysferlin with AHNAK2 seemed even stronger than with AHNAK (compare lanes 6 and 5 in Fig. 3C
). Therefore, this observation indicates that the carboxyterminal domains of both AHNAK proteins specifically and directly associate with C2A-dysferlin. A scheme in Fig. 3G
represents the direct interaction of AHNAK proteins with dysferlin.
|
Interaction of GST-C2-AHNAK with T7-tagged myoferlin fusion proteins
To elucidate whether both AHNAK proteins can also interact with myoferlin, a similar pull-down strategy was performed by replacing the T7-tagged amino-terminal constructs of dysferlin with those of myoferlin. By analogy to dysferlin, there is a direct interaction of the N-terminal myoferlin with the C–terminal domains of both AHNAK proteins (Fig. 3D-F
). In general, the interaction between myoferlin and AHNAK2 was weaker than for AHNAK and as shown in Fig. 3F
, only GST C2-AHNAK effectively pulled down T7-tagged C2A-myoferlin. A scheme in Fig. 3H
represents the direct interaction of AHNAK proteins with myoferlin.
Interaction of GST-C2-AHNAK with dysferlin C2A and its mutant in the presence or absence of Ca2+
To determine whether the direct interaction between both C2-AHNAK proteins and dysferlin C2A is Ca2+-dependent, we performed a similar pull-down assay in the presence (Fig. 4
A) or absence (Fig. 4A'
) of Ca2+. As shown in Fig. 4A
, we observed that both C2-AHNAK and C2-AHNAK2 maintained binding affinity for the C2A domain of dysferlin under both conditions. No binding was observed for equivalent amounts of unfused GST protein in both conditions. These results indicate a Ca2+-independent interaction between both AHNAK proteins and dysferlin.
|
To determine whether mutations in the C2A domain of dysferlin affect the binding with both AHNAK proteins, we introduced the V67D mutation into dysferlin C2A (V67D, GenBank accession number AF075575). This point mutation is associated with both mild and severe phenotypes (22)
. Previously, it was shown that the V67D mutation in dysferlin C2A leads to reduced calcium-sensitive phospholipid binding (7)
. We tested the interaction of GST C2-AHNAK proteins with dysferlin C2A-V67D in a similar pull-down assay in the presence (Fig. 4B
) or absence (Fig. 4B'
) of Ca2+. Figure 4B and 4B'
shows that dysferlin C2A-V67D was unable to bind to C2-AHNAK2 in the presence or absence of Ca2+ and the interaction between C2-AHNAK and dysferlin C2A-V67D was reduced in the absence of Ca2+.
Immunolocalization of AHNAK in normal human skeletal muscle sections
Previously, it was shown that AHNAK localizes to the plasma membrane of skeletal muscle (19)
. To further investigate the localization of AHNAK in skeletal muscle, we performed colocalization studies in transverse and longitudinal human muscle cryosections. In cross sections, double labeling with AHNAK and dysferlin showed that both proteins are primarily localized at the sarcolemma of skeletal muscle (Fig. 5
A). In longitudinal sections, AHNAK and dysferlin showed colocalization (Fig. 5C
). Double labeling of AHNAK and anti CD31 antibodies showed that AHNAK is also present in the intermediate blood vessels and capillaries (Fig. 5B
). Recently, it was shown that dysferlin coprecipitates with the DHPR by immunoprecipitation and partially colocalizes with DHPR in longitudinal sections of skeletal muscle. We observed that AHNAK and DHPR clearly colocalize in longitudinal sections of normal skeletal muscle by double immunofluorescent labeling (Fig. 5D
). Taken together, AHNAK colocalizes with dysferlin at the sarcolemma of muscle fibers in cross sections and both AHNAK and dysferlin colocalize with DHPR at T-tubules in longitudinal sections of skeletal muscle.
|
Immunostaining of dysferlin and AHNAK in patient muscle sections
Immunostaining analysis of frozen muscle sections of three unrelated LGMD2B patients (patients 1, 2, and 3 from Huang et al. (11)
) showed a comparable and similar reduction of AHNAK and dysferlin, with some patchy trace staining at the sarcolemma of these three dysferlinopathy patients (Fig. 6
A). Interestingly, although AHNAK is secondarily reduced at the sarcolemma in the absence of dysferlin, its presence in the blood vessels persists (Fig. 6B
), suggesting a muscle-specific reduction of AHNAK in these patients.
|
To confirm the secondary reduction of AHNAK, as observed by immunostaining, we performed quantitative Western blot analysis. Figure 6A'
shows a single section Western blot analysis of the three LGMD2B patients and one normal skeletal muscle as a control. Full-length AHNAK was only observed in the control but not in the muscle of the LGMD2B patients. Ponceau S staining of blots after transfer confirmed equivalent loading of proteins. This result is in agreement with the observation of secondary reduction of AHNAK in LGMD2B patients by immunostaining (Fig. 6A
).
Along with the analysis of dysferlinopathy patients, we also performed a random labeling of seven additional nondysferlin muscular dystrophies with dysferlin and AHNAK antibodies. These patients had genetically confirmed diagnoses of BMD, FSHD, DMD, -sarcoglycanopathy, ß-sarcoglycanopathy, -sarcoglycanopathy, and LGMD1C (Fig. 6C
). Both dysferlin and AHNAK showed normal expression at the sarcolemma in all cases except LGMD1C. In this case, both dysferlin and AHNAK showed a secondary overall reduction along with the primary loss of caveolin-3, while the residual AHNAK staining was also restricted to the blood vessels (Fig. 6C
).
Immunostaining of dysferlin and AHNAK on regenerating mouse skeletal muscle sections
We finally examined the expression of dysferlin and AHNAK during muscle regeneration in a rat model system. Four days after notexin treatment, dysferlin shows extensive cytoplasmic staining, consistent with relocalization of this protein to the cytoplasm during regeneration. As a marker of membrane integrity, we used spectrin, which was also shown to be increased, confirming the regenerative process. Interestingly, although the staining for ß-dystroglycan and dystrophin is predominantly at the membrane in control and notexin-injected muscle, AHNAK is redistributed to the cytoplasm in parallel with dysferlin (Fig. 7
).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Dysferlin is ubiquitously expressed, with highest expression in skeletal muscle and heart (5)
. Dysferlin-null mice progressively develop pathological characteristics of muscular dystrophy. In dysferlin-null mice, there are subsarcolemmal accumulations of yet uncharacterized vesicles, and dysferlin-deficient muscle fibers are defective in Ca2+-dependent sarcolemma resealing (14)
. To gain further insights into the molecular mechanisms of dysferlin function, we have searched for proteins that interact with dysferlin in skeletal muscle. In this paper, we describe a novel interaction of dysferlin with AHNAK based on IP coupled to MALDI-MS analysis, further confirmed by coimmunoprecipitation and GST pull-down assays, and by immunofluorescence microscopy analysis. CAPN3, whose mutations cause LGMD2A, has been confirmed to interact with dysferlin in a similar way (11)
. Caveolin-3, affixin, and annexins, another three known dysferlin-interacting proteins, were undetectable because of their locations in the smears caused by the cross-reactivity of the secondary Ab against coimmunoprecipitated IgG molecules from the muscle homogenate.
In this paper, we describe a novel interaction of dysferlin with AHNAK based on IP coupled to MALDI-MS analysis, further confirmed by coimmunoprecipitation and GST pull-down assays, and by immunofluorescence microscopy analysis. The interaction sites between dysferlin and AHNAK have been defined as the carboxyterminal 500 aa of both AHNAK variants and the C2A domain of dysferlin and its homologue myoferlin. In C2C12 cells, we detected two isoforms of dysferlin. Recently, Salani et al. (23)
also presented evidence for two dysferlin isoforms, of which the low-MW variant is progressively replaced by the full-length dysferlin as myoblast fusion proceeds and completely disappears in adult skeletal muscle. Regardless of the precise differences between both isoforms we detect, AHNAK specifically interacts with the full-length dysferlin despite the much higher abundance of the low MW isoform in the precleared protein fraction. Although the low MW isoform is thought to preserve the carboxyterminal region, as it can be recognized by NCL-Hamlet, the binding with AHNAK is completely abolished. This may suggest that the low MW isoform of dysferlin may have either lost the C2A domain that is recognized by AHNAK or has undergone a conformational change that impairs the binding between both proteins.
In proteins containing a single C2 domain, this domain is often involved in calcium-dependent phospholipid binding. However, C2 domains can also be involved in protein-protein binding, in a Ca2+-dependent or independent fashion. Previously, it was shown that in dysferlin and myoferlin, the C2A domain showed Ca2+-dependent phospholipid binding, while the other C2 domains showed no affinity for phospholipids (7)
. It was speculated that these domains are involved in protein-protein interactions. Moreover, the V67D mutation in the C2A domain demonstrated reduced phospholipid binding. Our studies show a calcium-independent interaction between AHNAK and the C2A domains of dysferlin. Interestingly, the V67D mutation disrupts the binding between dysferlin and AHNAK2, while the affinity for AHNAK is reduced in a calcium-dependent manner.
AHNAK is a family of two proteins (AHNAK and AHNAK2) of exceptionally large size (
600–700 kDa) and characterized by common amino acid sequences and structural features (21)
. The AHNAK proteins have a tripartite structuring, including the amino-terminal 251 amino acid large head, a large central region of 4390 amino acids composed of 26 repeated elements, and the carboxyl-terminal tail of 1002 amino acids. Like dysferlin, high expression levels of AHNAK are observed in all muscular cells, including cardiomyocytes and skeletal muscle cells (19)
. The uniform distribution of AHNAK at the plasma membrane of skeletal muscle is also observed in normal human skin (24)
and mouse keratinocytes (25)
. However, structure prediction algorithms do not provide evidence for the presence of transmembrane domains in AHNAK to explain the plasma membrane localization. It was suggested that its membrane association is probably dependent on specific interactions with the annexin A2/S100A10 complex in human epithelial MCF-7 cells through the AHNAK C-terminus (26)
. Nevertheless, annexin A2/S100A10 complexes do not contain transmembrane domains either. Dysferlin has also been shown to associate with the annexin 2/S100A10 complex in a Ca2+- and membrane injury-dependent manner (12)
. We thus propose a model in which the sarcolemmal localization and stabilization of AHNAK can be controlled by its direct interaction with dysferlin and myoferlin, which are anchored to the sarcolemma through their single transmembrane domain (Fig. 8
).
|
The functional significance of the interaction between dysferlin and AHNAK comes from the observation that AHNAK is, like dysferlin, primarily localized at the sarcolemma and is secondarily reduced in muscle from patients with genetically confirmed dysferlinopathy, generally comparable to the loss of dysferlin. Moreover, in LGMD1C, dysferlin and AHNAK showed a secondary overall reduction to the primary loss of caveolin-3, for which a functional relationship with dysferlin has already been suggested (10)
. In contrast, in unrelated muscular dystrophies, dysferlin and AHNAK showed normal sarcolemma staining. This secondary reduction is muscle specific, as we do not observe a loss of AHNAK in the blood capillaries of LGMD2B and LGMD1C patients.
In addition, we have provided further evidence for a functional cooperation between dysferlin and AHNAK during muscle regeneration. Dysferlin and AHNAK show a marked increase and cytoplasmic localization during regeneration, consistent with the direct interaction between them and mobilization of the AHNAK-dysferlin complex during repair and regeneration. Regulation of cellular AHNAK content in relationship with cell membrane remodeling and specialization has already been observed in epithelial and endothelial cells (26)
. In these cells, AHNAK stabilization is dependent on its recruitment to the plasma membrane through interaction with partner proteins, including the annexin2/S100A10 complex (26)
. Dysferlin has also been shown to associate with the annexin 2/S100A10 complex in a Ca2+- and membrane injury-dependent manner (12)
. Thus, we suggest that at the sarcolemmal membrane dysferlin, AHNAK and the annexin2/S100A10 may act in a single functional complex, including caveolin-3, affixin, and CAPN3, with a role in renewal of the complex structures of the sarcolemma.
In agreement with the proposed membrane repair function for AHNAK, it was demonstrated that AHNAK is also an integral component of a newly discovered Ca2+-regulated vesicle capable of rapid exocytosis, called enlargosome. Enlargosome exocytosis is induced by plasma membrane disruption and is thought to be involved in both plasma membrane differentiation and repair (16)
. In this line, an interaction of dysferlin and AHNAK in striated muscle is particularly intriguing, as dysferlin was already implied in patch fusion repair. Like the uncharacterized vesicular accumulation in SJL-dysf mice, AHNAK-positive enlargosomes are concentrated in the cytoplasmic rim below the plasmalemma (16)
. It is, therefore, imperative to further investigate the role of enlargosomes in muscle membrane maintenance and repair.
Another intriguing hypothesis for the function of the dysferlin complex is a putative involvement in signal transduction in addition to plasma membrane repair. Recently, it was demonstrated that dysferlin coprecipitates with the dihydropyridine receptor (DHPR) by immunoprecipitation. And double immunofluorescent staining revealed that dysferlin was observed to partially colocalize with DHPR in skeletal muscle fibers (8)
. The DHPR forms the L-type Ca2+ channel, acts as the voltage sensor and initiates the cascade of events leading to excitation-contraction (EC) coupling by the strict functional and structural relationship of DHPR and ryanodine receptor type 1 (RyR1) (27)
. In agreement with the interaction between dysferlin and DHPR, the association of AHNAK with the regulatory ß2 subunit of L-type Ca2+ channel indicated a potential role of AHNAK in the regulation of channel activity (17)
. Moreover, it has been demonstrated that AHNAK binds and activates phospholipase C-1 in the presence of arachidonic acid (28)
and interacts with the EF hand Ca2+ binding protein S100B (29)
. In this study, we also demonstrated that AHNAK colocalizes with DHPR in longitudinal sections of normal human skeletal muscle. Thus, it is intriguing to suggest that the dysferlin-AHNAK complex may have implications in signal transduction in skeletal muscle.
In summary, we demonstrate that AHNAK is a novel member of the dysferlin protein complex in skeletal muscle. In this paper, we report that AHNAK directly interacts with dysferlin and that both proteins colocalize in skeletal muscle by double immunofluorescence labeling. Moreover, we have provided evidence for a functional cooperation between dysferlin and AHNAK during muscle regeneration. Our studies suggest that the dysferlin complex, including a novel component, AHNAK, may act simultaneously in Ca2+-mediated muscle membrane homeostasis processes and signal transduction events.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication June 23, 2006. Accepted for publication October 25, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. J. Hernandez-Deviez, M. T. Howes, S. H. Laval, K. Bushby, J. F. Hancock, and R. G. Parton Caveolin Regulates Endocytosis of the Muscle Repair Protein, Dysferlin J. Biol. Chem., March 7, 2008; 283(10): 6476 - 6488. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. H. Lee, H. J. Lim, S. Yoon, J. K. Seong, D. S. Bae, S. G. Rhee, and Y. S. Bae Ahnak Protein Activates Protein Kinase C (PKC) through Dissociation of the PKC-Protein Phosphatase 2A Complex J. Biol. Chem., March 7, 2008; 283(10): 6312 - 6320. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
