Coxsackievirus B3 (CVB3) is known to cause a variety of human diseases including acute and chronic myocarditis as well as dilated cardiomyopathy (DCM). However, the mechanisms by which CVB3 causes diseases are not well understood.
Studies identifying protein–protein interactions during CVB3 infection are useful in delineating the pathogenesis of acute or chronic myocarditis. Screening a human heart cDNA library revealed a yet unknown interaction partner of the proapoptotic protein Siva. We demonstrate that Siva specifically interacts with the heart and skeletal muscle protein telethonin. The expression of Siva is increased in heart tissue of CVB3-infected mice and the proteins colocalize in cardiomyocytes.
telethonin might be involved in CVB3-mediated cell damage and in the resulting cardiac dysfunction due to the interaction with Siva. We suggest a molecular mechanism through which coxsackieviral infection contributes to the pathogenesis of chronic myocarditis and in particular of acquired forms of DCM.
Coxsackievirus B3 (CVB3), an enterovirus of the picornavirus family, has been implicated in a variety of serious diseases including acute and chronic myocarditis, dilated cardiomyopathy (DCM), type I insulin-dependent diabetes mellitus and pancreatitis.1–3 CVB3 is the serotype most frequently associated with myocarditis,4 and its pathogenesis has been studied extensively in mice. A majority of symptomatic patients recover well from acute myocarditis. However, 10–20% of patients will develop chronic disease, and a subset progresses over time to DCM.5–7 During DCM one or both ventricles dilate and decompensate, leading to progressive heart failure and significantly increased mortality.8 The mechanism by which CVB3 causes acute or chronic myocarditis, in particular leading to DCM, is not well characterized.4,9
Some forms of DCM involve the cytoskeleton. Mutations in genes, encoding sarcomeric proteins such as actin and titin are the cause of different forms of cardiomyopathy. Dystrophin was found to be responsible for Duchenne and Becker muscular dystrophy as well as for X-linked DCM.10–12 CVB3 induces DCM by protease 2A-mediated cleavage and inactivation of Dystrophin as well as dissociation of Dystrophin-associated glycoproteins from the sarcolemma essential for myocyte and cardiomyocyte integrity, due to the central role of cytoskeletal proteins in cardiac function.13
Infections of CVB3 are often accompanied by programmed cell death, and apoptotic processes were found in myocardial tissue of patients with DCM.14 Apoptotic cells are detectable in inflammatory lesions and outside of inflamed areas in the myocardial tissue of CVB3-infected mice.15 Previously, we identified the interaction between the CVB3 capsid protein VP2 and the proapoptotic host protein Siva,16 which is known to be involved in CD27- or GITR-induced apoptosis by providing the missing death domain or contact sites for downstream located proteins of different cell death pathways.17,18 Furthermore, transcription of the mouse equivalent of Siva (mSiva) is induced in pancreas and heart tissue after CVB3H3 infection.16 Efficient binding of VP2 to amino acids 118–136 of Siva is important for the activity of the host protein, yielding severe destruction and apoptosis in tissue. A comparison between CVB3H3 and the mutant CVB3H310A1, in which asparagine at position 165 in VP2 is changed into aspartate,19 shows differences in the apoptotic potential in vivo and in vitro as verified by viral spread and virus progeny production, Siva mRNA expression or DNA (deoxyribonucleic acid) fragmentation. Therefore, the results indicate a mechanism by which apoptosis contributes to coxsackievirus-dependent pathogenesis.20,21
To study more thoroughly the pathological processes caused by CVB3, we performed a two-hybrid screening of a heart cDNA (complementary DNA) library using Siva as bait and identified the cytoskeletal protein telethonin as a new binding partner. The interaction was confirmed in vivo and in vitro. We found that the binding site of telethonin to Siva is located in the N-terminal region, whereas the C-terminus of Siva is necessary and sufficient for interaction. We demonstrate colocalization of mSiva and mouse telethonin (mTelethonin) in infected cardiomyocytes. The interaction could help to explain cell damage and cardiac dysfunction of CVB3-infected animals. These findings indicate a molecular mechanism by which CVB3 contributes to the pathogenesis of myocardial diseases, and demonstrate a potential role of cytoskeletal protein alterations in acquired DCM.
All plasmids used were generated by standard PCR cloning methods. The cDNA molecules were amplified by Pwo polymerase (Roche Diagnostics, Mannheim, Germany) and ligated into the target vectors (Table 1). The correct insertion was confirmed by DNA sequencing.
|Name||Vector||Brief description||aa Sequence|
|(A) Two-hybrid experiments|
|(B) Binding assays|
|Name||Vector||Brief description||aa Sequence|
|(A) Two-hybrid experiments|
|(B) Binding assays|
Yeast two-hybrid experiments
The yeast strain KFY3 (MATa, ura3–52, his3–200, ade2–101, lys2–801, trp1–901, leu2–3, -112, gal4Δ, gal80Δ, cyh2r, LYS2::GAL1UAS-HIS3TATA-HIS3, URA3::GAL1UAS-GAL1TATA-lacZ, ADE2::ADE2, leu2::PGAL1-yEGFP3-loxP-kanMX-loxP)22 was used. For directed two-hybrid experiments, cDNA molecules were cloned into the bait vector pAS2-1, or the prey vector pACT2 (BD Biosciences Clontech, Heidelberg, Germany), respectively, to yield fusions with the DNA-binding domain (BD) or activation domain (AD) of the yeast Gal4 transcription factor. Yeast cells were transformed as described,23 plated onto selective agar dishes and incubated at 30°C for 3 days.
A human heart cDNA library (BD Biosciences Clontech) was screened for proteins that interact with Siva. Agar plates with cotransformants were probed for LacZ activity by a filter lift assay according to Breeden and Nasmyth.24 Plasmid DNA from positive clones was rescued from yeast cells, transformed into Escherichia coli (E. coli) strain HB10125 and isolated with standard procedures. In a second two-hybrid assay, three single colonies of KFY3 transformants containing the bait plasmid and one of the obtained library plasmids were analysed. Positive clones were selected and the inserts of library plasmids sequenced. These sequences were compared by BLAST search (http://www.ncbi.nlm.nih.gov/BLAST/).
Verification of interactions and identification of the human Siva and telethonin BDs were investigated with full length and truncated fusion proteins. Plasmids containing the Gal4-DNA-BD or -AD fusion proteins were transformed into KFY3 cells. Three single colonies of KFY3 transformants were picked and grown in liquid minimal medium at 30°C overnight. The activity of β-galactosidase was determined using a quantitative liquid assay according to Breeden and Nasmyth.26
Protein expression and binding assays
Telethonin was cloned into pET-32b (Novagen, Merck, Darmstadt, Germany). Standard protocols were used for His6 fusion protein expression and purification with a cobalt-based immobilized metal affinity chromatography resin (BD Biosciences Clontech).27 All subfragments of Siva were cloned into pGEX-4T-3. Fusion proteins were purified from E. coli lysates by using Glutathione Sepharose® 4B (GE Healthcare, Munich, Germany). Eluate purity was assessed by SDS–PAGE (sodium dodecyl sulphate polyacrylamide gel electrophoresis).28
Interaction between human Siva and telethonin was verified by in vitro pull-down assays. Briefly, 30 µg of fusion proteins, bound to their solid matrix (cobalt-based resin or Glutathione Sepharose, respectively), were incubated with 45 µg of eluted fusion protein of the interaction partner in a total volume of 150 µL binding buffer (50 mM Tris–HCl pH 7.5; 150 mM NaCl; 0.1% NP40; 1 mM EDTA). Subsequent steps were performed as previously described.22
Slot blot assays were performed while 1.5 µg of glutathione S-transferase (GST) fusion proteins, GST and BSA (bovine serum albumin) were blotted on nitrocellulose membranes. The membranes were blocked with 2.5% BSA in PBS (phosphate buffered saline) at 4°C overnight and incubated at room temperature for 2 h in presence or absence of human recombinant telethonin at a concentration of 15 µg. After washing with PBS and 0.05% Tween, the bound telethonin was visualized using anti-telethonin antibody (BD Biosciences Clontech).
Human embryonic kidney 293 (HEK293) cells were transfected with plasmid pVITRO2–mcs (Invivogen, San Diego, CA, USA) containing Siva and telethonin sequences using Effectene Transfection Reagent (Qiagen, Hilden, Germany). Transfection efficiency of HEK293 cells was determined with EGFP by fluorescence microscopy after 48 h. The cells were treated for immunofluorescence detection as described later.
Adult (8–10 weeks of age) male inbred BALB/c (H-2dd) mice were inoculated by intraperitoneal injection of 0.2 mL saline containing the stated amount of virus (infectious dose 0.5 LD50 of CVB3H3, virus titer: 4.7 × 104pfu/100 mg tissue).19 All procedures conform to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85–23, revised 1996). For each group (infected or non-infected), four mice were treated. For indirect immunofluorescence, frozen 10 µm thick sections were prepared from heart tissue of non-infected and of CVB3-infected mice 7 days p.i. These experiments have been repeated three times. The sections were fixed with 3.8% paraformaldehyde in PBS for 12 min and permeabilized with 0.5% saponine in PBS for 11 min at room temperature. After washing in 0.1% saponine/PBS, sections were blocked with 1% FCS in 0.1% saponine/PBS at 4°C overnight. The sections were washed and then incubated with the primary anti-Siva antibody from goat (1/25 dilution) and the anti-telethonin antibody from rabbit (1/50 dilution, Santa Cruz, Heidelberg, Germany). Detection of the primary antibody was performed using indocarbocyanine-conjugated donkey anti-goat and fluorescein isothiocyanate-conjugated donkey anti-rabbit immunoglobulin (Jackson Immunoresearch, West Grove, PA, USA). An LSM 510 laser scanning microscope (Carl Zeiss Microscopy, Jena, Germany) was used for confocal microscopy at ×63 magnification.
Interaction between human Siva and telethonin
Screening for novel interaction partners of Siva
To identify interaction partners involved in CVB3 pathogenesis and apoptosis induced by Siva, a human heart cDNA two-hybrid library was screened using Siva as bait. A total of 210 000 independent yeast colonies were analysed. After retransformation of isolated cDNA clones into yeast cells, the cytoskeletal protein telethonin was identified as a potential binding partner. Results of LacZ filter assays with corresponding clones are shown in Figure 1A.
Verification of interaction and determination of Siva and telethonin binding domains
DNA sequencing revealed that the encoded fusion protein of the original two-hybrid clone lacked the first four residues of telethonin. To reduce the probability of a false positive interaction, the entire telethonin sequence was fused with Gal4-AD and probed with full-length Siva using a yeast two-hybrid liquid assay (Table 1A). Both fusion proteins interacted, as determined by the β-galactosidase activity of the plasmid containing cells (Figure 1B and C). The data revealed that the first four amino acids of telethonin were not necessary for Siva binding. To identify the binding regions of Siva and telethonin, two-hybrid liquid assays with full-length proteins and truncated fragments were performed. As shown in Figure 1B, only full-length Siva interacted with telethonin, whereas fragments thereof, i.e. Siva-1 and Siva-2 did not. We concluded that the C-terminus of Siva is necessary and sufficient for telethonin binding. Furthermore, the binding region of telethonin was restricted to the N-terminus (Figure 1C). The cells containing the plasmids telethonin-1 or telethonin-2 showed less β-galactosidase activity compared with cells expressing the full-length protein suggesting that these telethonin fragments lack amino acids required for Siva binding.
In vitro binding between Siva and telethonin
The interaction between Siva and telethonin in vitro was confirmed by pull-down experiments. Siva variants used in these studies are shown in Table 1B and Figure 2A. We found that His6–telethonin was able to bind Siva-DD-C, Siva-C1 and Siva-C2, but did not show any interaction with Siva-DD-ΔC and Siva-ΔDD (Figure 2B and C). GST and full-length GST–TopBP1 [GST–topoisomerase ii (beta)-binding protein 1] served as negative Controls. Slot blot experiments revealed interaction between telethonin and Siva-DD-C or Siva-C2 only (Figure 2D). In brief, the results show the C-terminus of Siva is necessary and sufficient for telethonin binding.
Siva and telethonin localization in transfected HEK293 cells
Immunofluorescence experiments demonstrated colocalization of Siva and telethonin in the cytoplasm of transfected HEK293 cells. A remarkably higher Siva expression was found in cells transfected with pVITRO2–mcs–Siva–telethonin compared with the Control (Figure 3A and B) suggesting a low basal expression of Siva in HEK293 cells per se. Therefore, a HEK293 lysate was tested for Siva by dot blot experiments. The HEK293 cells contained a small amount of Siva proteins (data not shown).
Localization of Siva and telethonin in cardiomyocytes of mice
The mouse proteins of Siva and telethonin are highly homologue to the human gene products.29,30 Two-hybrid and pull-down experiments verified the interaction between the mouse equivalents of both proteins (data not shown). To describe the protein interaction in a natural environment, frozen heart sections of non-infected and CVB3-infected mice were stained for mSiva and mTelethonin and the proteins were subsequently visualized by confocal microscopy. For secondary antibody examination several sections of CVB3-infected mice were not incubated with primary antibodies (Figure 4B). The two proteins colocalized in distinct areas of infected cardiomyocytes (Figure 4A). Non-infected cardiomyocytes also expressed both proteins but Siva was found in a considerable lower amount than in the infected cells and no colocalization could be observed (Figure 4C). Transcription of mSiva, CVB3-VP2, and β-actin in tissue of individually non-infected or CVB3-infected mice was analysed by reverse transcription polymerase chain reaction (RT–PCR) as previously described (data not shown).16
Coxsackievirus infections have been shown to cause acute or chronic myocarditis in humans. Mouse models were established to study the complexity of CVB3-caused pathogenesis, in which apoptosis plays a considerable role. Several apoptosis-regulating proteins were found to be involved in host cell fate regulation after viral infection. Previously, we demonstrated that the structural protein VP2 of CVB3H3 binds specifically to the proapoptotic protein Siva. The interaction is required for induction of caspase-catalysed apoptosis linked to severe tissue destruction and accelerated viral spread.16,20,21 Binding of Siva to the antiapoptotic members Bcl-XL and Bcl-2 of the Bcl-2-family results in a loss of protective function and in sensitization of cells to apoptosis.31 Proapoptotic activities of the BNip proteins were also observed during interactions with Bcl-XL and Bcl-2. Zhang et al.32 showed that murine Nip21 overexpression promotes CVB3-induced apoptosis via a caspase-dependent mitochondrial pathway.
To get more insights into cell damage processes and heart failure of CVB3-infected humans, we screened a human heart cDNA library for binding partners of Siva. The cytoskeletal protein telethonin was found to interact with Siva (Figure 1A). Telethonin is located inside the sarcomeres of striated muscles, as a part of the Z-disc.33 The Z-disc is the main anchoring point of the molecular machinery that underlies muscle contraction and plays a pivotal role in muscle structure and function,34 including sarcomeric assembly and organization. Defects in the components of the Z-disc can trigger pathological pathways. Telethonin is associated with the autosomal recessive disorder limb girdle muscular dystrophy 2G.35
The present study demonstrated an interaction between Siva and telethonin using yeast two-hybrid analysis, in vitro GST pull-down and slot-blot experiments as well as colocalization studies. The domain relevant for the interaction with Siva is located within the N-terminus of telethonin. Moreover, we showed that adjacent amino acids are also involved in Siva binding (Figure 1C). There is no difference in enzyme expression between cells which contain the two fragments of telethonin. Hence, the functional C-terminus is not necessary for interaction.
The N-terminus of telethonin (aa 1–90) interacts specifically with the N-terminal immunoglobulin domains Z1 and Z2 of the giant muscle protein titin.36 Titin plays a role in elasticity and muscle assembly by providing a scaffold for other sarcomeric proteins. The interaction between titin and telethonin is required for the structural integrity associated with muscle function of sarcomeres.29,37,38 Another protein that associates with telethonin (aa 53–81) is the muscle LIM protein (MLP) which stabilizes the telethonin–titin complex. A human MLP mutation (W4R) results in a marked defect in telethonin interaction/localization. Knöll et al.39 propose that the complex contributes to the regulation of cardiac mechanical stress sensing and cardiac hypertrophy. Malfunction of this control machinery can lead to human DCM and associated heart failure. Numerous mutations in the genes of titin and telethonin are associated with DCM and hypertrophic cardiomyopathy.40,41
Furthermore, the characterization of the Siva BD revealed that the C-terminus (aa 151–175) is necessary and sufficient for the interaction (Figures 1B and 2B and C). The high number of cysteine residues and potential zinc-binding motifs at the C-terminus suggest a complex three-dimensional structure.22 Little is known about the signalling pathway leading to Siva-induced apoptosis and contradictory results exist about the concerned protein domains. Siva was found to be involved via death domain-like structures in the CD27-transduced apoptotic pathway. In contrast with previous reports, Py et al.42 show that the death domain-like structure is dispensable for induction of apoptosis in lymphoid cells; instead, the main determinants are located in both the N- and C-terminal regions. Siva is an ubiquitous protein that could participate in various apoptotic pathways in a cell-type-dependent manner.17
The transcription of mouse SIVA is increased in pancreas and heart tissue in the presence of infectious virus particles as shown by RT–PCR and TUNEL assays.16 Expression of Siva leads to programmed cell death17 and may contribute to cell damage in CVB3-infected cells and organs.20 In this report we compared the expression levels of Siva and telethonin in hearts of non-infected and CVB3-infected mice by immunofluorescence. Only infected hearts showed a high level of Siva expression in distinct areas (Figure 4A and C), whereas the expression level of telethonin could not be quantified in more detail in the current study. At the beginning only a few cells are infected with CVB3 but during the course of an infection the number of distinct areas will increase.43 The outcome of a CVB3 infection is determined by complex interactions among several variables, such as the virus genotype,44,45 or the age,3,46 and immune status47 of the host.9 Furthermore, we were able to detect colocalization of Siva and telethonin in transfected HEK293 cells as well as in distinct areas of hearts of CVB3-infected mice only (Figures 3A and 4A). Therefore, we assume that an increase in Siva expression during a CVB3 infection leads to Siva association with telethonin.
It is conceivable that Siva translocates in CVB3-infected cardiomyocytes to the Z-disc by an unknown mechanism and competes with the binding of MLP and titin to telethonin. Defects in the stabilization of the complex which is required for sarcomeric function can lead to human DCM and associated heart failure. Our results indicate a new molecular mechanism through which coxsackieviral infection could contribute to the pathogenesis of chronic myocarditis and in particular of acquired forms of DCM via apoptosis-regulating proteins. We propose that the development of an inhibitory constraint of the Siva–telethonin association will support the treatment of these deleterious diseases.
This work was supported by an internal project of the Leibniz Institute for Natural Product Research and Infection Biology – Hans Knöll Institute.
We thank Andrea Hartmann for performing confocal microscopy of mice heart sections. We acknowledge Uta Schmidt and Frank Hänel for providing GST-TopBP1 protein and for helpful advice.
Conflict of interest: none declared.