Interactions between Nebulin-like Motifs and Thin Filament Regulatory Proteins*

Nebulin (600–900 kDa) and nebulette (107–109 kDa) are two homologous thin filament-associated proteins in skeletal and cardiac muscles, respectively. Both proteins are capped with a unique region at the amino terminus as well as a serine-rich linker domain and SH3 domains at the COOH terminus. Their significant size difference is attributed to the length of the central region wherein both proteins are primarily composed of ∼35 amino acid repeats termed nebulin-like repeats or motifs. These motifs are marked by a conserved SXXXY sequence and high affinity binding to F-actin. To further characterize the effects that nebulin-like proteins may have on the striated muscle thin filament, we have cloned, expressed, and purified a five-motif chicken nebulette fragment and tested its interaction with the thin filament regulatory proteins. Both tropomyosin and troponin T individually bound the nebulette fragment, although the affinity of this interaction was significantly increased when tropomyosin-troponin T was tested as a binary complex. The addition of troponin I to the tropomyosin-troponin T complex decreased the binding to the nebulette fragment, indicating an involvement of the conserved T2 region of troponin T in this interaction. F-actin cosedimentation demonstrated that the nebulette fragment was able to significantly increase the affinity of the tropomyosin-troponin assembly for F-actin. The relationships provide a means for nebulin-like motifs to participate in the allosteric regulation of striated muscle contraction.

Nebulin (600 -900 kDa) and nebulette (107-109 kDa) are two homologous thin filament-associated proteins in skeletal and cardiac muscles, respectively (1,2). cDNA sequences of nebulin and nebulette have shown that both proteins are primarily composed of ϳ35 amino acid repeats termed nebulin-like repeats or motifs, each marked by a conserved SXXXY sequence (2,3). Both nebulin and nebulette are capped by a unique region at the amino terminus in addition to serine-rich linker and SH3 domains at the carboxyl terminus (2)(3)(4)(5)(6). Immunoelectron microscopy has demonstrated that a single nebulin molecule extends from the Z-line (COOH terminus of the protein) to the pointed ends of the thin filament at its NH 2 terminus (7).
Nebulette is similarly anchored to the Z-line through its COOH terminus, but because of its relatively smaller size, it does not extend as far as the pointed ends of the actin filament. Experiments using in vitro systems have formed a consensus that nebulin-like motifs from both nebulin and nebulette associate with F-actin (4,6,8). Additional experiments have extended this observation to demonstrate that nebulin-like motifs may interact directly with tropomyosin (Tm) 1 and troponin (Tn) (4) along the striated muscle thin filament.
At the Z-line, the nebulin/nebulette interactions are probably mediated by contributions from the linker domain as well as interactions of the SH3 domains with ␣-actinin (6,9). However, results from Ojima and coworkers (10) suggest that neither the serine-rich linker domain nor the SH3 domain are obligatory elements for the incorporation of nebulin and presumably nebulette into the Z-line of striated muscle sarcomeres. Although both proteins are anchored at the Z-line in vivo, the primary sequence regions responsible for this interaction remain unclear. Recent experiments have also shown an association between the NH 2 terminus of nebulin and tropomodulin, a protein known to cap actin-Tm filaments (11). Therefore, emerging data demonstrate that the NH 2 -and COOH-terminal domains of nebulin and nebulette form unique interactions at both the pointed end of the thin filament and the Z-line of sarcomeres, complementing the established actin binding properties of these proteins.
Although the majority of experiments to date have focused on the binary interactions between nebulin/nebulette and other sarcomeric proteins, there is preliminary evidence to suggest that nebulin-like motifs may serve a role in the regulation of striated muscle contraction. These investigations have concentrated on the role that nebulin may play in regulating interactions within the framework of the isolated acto-S1 myosin system. Consistent with this hypothesis, it has been demonstrated that nebulin fragments of varying lengths are able to inhibit the acto-S1 myosin ATPase, an inhibition that can be reversed by calmodulin (12). Subsequent experiments by this group have suggested that nebulin fragments affect the acto-S1 myosin complex by optimizing the alignment of actomyosin interactions (13). To further clarify the interactions that nebulin-like repeats may have along the thin filament, we have characterized the binding of a chicken nebulette fragment to Tm and the Tn subunits. Our data suggest that Tm and troponin T (TnT) both contain interaction sites for the nebulin-like motifs. These interactions were strengthened when Tm and TnT were present as a binary complex and weakened by the addition of troponin I (TnI). Furthermore, the nebulette fragment was able to significantly increase the affinity of Tm⅐Tn for F-actin, demonstrating that nebulin-like motifs contribute to the assembly and allosteric property of the striated muscle thin filament.

EXPERIMENTAL PROCEDURES
Reverse Transcription (RT)-PCR Cloning of Mouse Nebulin and Chicken Nebulette cDNA Fragments-To clone mouse nebulin (MSN) and chicken nebulette (CN5) cDNAs, a partial sequence of nebulin cDNA from BALB/c strain mice (14) and a partial cDNA sequence from chicken nebulette (2) were used to design oligonucleotide primers suitable for RT-coupled PCR. Total RNA was prepared from 129 Sv/J mouse gastrocnemius muscle or white Leghorn chicken heart by the TRIzol reagent (Invitrogen) according to the manufacturer's instructions. Using 5 g of total RNA as template and 20 pmol of a reverse primer (mouse: Neb-R, 5Ј-CGGCATACTATGCCTTATACTGATTATCAC-3Ј, and chicken: CCN-R2, 5Ј-CCTTTGAATTCTCAAAATCTTAATGGTA-3Ј (mismatches are in bold to introduce restriction enzyme recognition sites and translation stop codons)), 20 units of Moloney murine leukemia virus reverse transcriptase was used for an RT reaction that was incubated overnight at 37°C. Subsequently, a PCR was set up using one-tenth of the RT products and 20 pmol each of two pairs of synthetic oligonucleotides, Neb-F/Neb-R and CCN-F/CCN-R2 (Neb-F, 5Ј-TCTC-CAGAGTAAACCAGATCCATATGAGTG-3Ј, and CCN-F, 5Ј-AGCACT-CATATGAACATC-3Ј (mismatches are in bold to introduce restriction enzyme recognition sites and translation initiation codons)). The PCR was done with a program of 35 cycles at 94°C for 1 min, 48°C for 1 min, and 72°C for 3 min. Aliquots of the PCR products were analyzed by agarose gel electrophoresis to confirm amplified cDNA fragments at the expected sizes of 470 bp (MSN) and 540 bp (CN5). The amplified mouse nebulin cDNA was precipitated with ethanol and cloned into the SmaI site of pBluescript II SK vector (Stratagene). The amplified chicken nebulette cDNA was precipitated, digested by NdeI and EcoRI endonucleases, and resolved by 0.7% agarose gel electrophoresis with Tris acetate-EDTA buffer (0.04 M Tris acetate, pH 8.0, 1 mM EDTA). The DNA band of interest was excised from the agarose gel and recovered using the Prep-A-Gene DNA recovery kit (Bio-Rad). The recovered cDNA was cloned into NdeI-and EcoRI-digested pAED4 expression vector (15). The insert sequences of the two constructs as well as the translation reading frame of the pAED4-CN5 construct were confirmed by dideoxy chain termination sequencing (15).
Isolation of a -DASHII Phage Containing Mouse Nebulin Genomic DNA-To isolate nebulin genomic DNA, a -DASHII 129 SvJ mouse genomic DNA phage library was screened by radioactive probe hybridization (16). The MSN cDNA was cut from the pBluescript II SK plasmid by BamHI and EcoRI restriction enzyme digestion, resolved by agarose gel electrophoresis, and recovered as described above. The purified DNA fragment was labeled with [␣-32 P]dCTP by random priming synthesis for use as a radioactive probe in screening nylon filter replicas of the phage library plated on agar plates. Based on the average insert size of the mouse genomic DNA library (12-15 kb pair), an aliquot of phage harboring inserts totaling 1.5ϫ the mouse genome (ϳ4.5 ϫ 10 9 bp) was plated out on twelve 150-mm agar plates. The nylon replicas were prehybridized in 100 ml of 30% (v/v) formamide, 5ϫ SSPE (0.9 M NaCl, 0.05 M NaH 2 PO 4 , 5 mM EDTA, pH 7.7), 5ϫ Denhardt's solution, and 100 g/ml denatured salmon sperm DNA at 50°C for 6 h after which time the heat-denatured radioactively labeled probe was added and incubated at 50°C for 16 h. Following the hybridization period, the filters were washed successively at room temperature with 1ϫ SSPE, 0.5ϫ SSPE, and 0.1ϫ SSPE followed by washing with 0.1 ϫ SSPE at 50°C. The filters were exposed to x-ray film overnight to reveal hybridization signals. A small agar plug was recovered from the plate corresponding to the area of each putative positive signal. The hybridization screening was then used in plaque purification to isolate the positive phage clones.
The positive phage DNA was prepared in large quantity using the Qiagen Lambda Midi Kit according to the manufacturer's instructions (Qiagen). An ϳ15-kb mouse genomic DNA insert was excised from the -phage arms by EcoRI digestion, and the EcoRI fragments were subcloned into the EcoRI site of pBluescript SK II vector. One subclone containing a 3.2-kb mouse genomic DNA fragment was sequenced by preparing a set of serial deletions using an exonuclease III-based nested deletion kit according to the manufacturer's instructions (Promega).
Protein Expression and Purification-Escherichia coli BL21(DE3)-pLysS cells were transformed with the T7 polymerase-based pAED4-CN5 chicken nebulette expression construct and allowed to grow until the colonies were just visible. Four fresh colonies were randomly picked to inoculate 4 liters of NZ broth (10 g/liter casein hydrolysate, and 5 g/liter NaCl) containing 100 g/ml ampicillin and 25 g/ml chloramphenicol. The cultures were incubated overnight at 37°C with shaking until A 600 ϭ 0.8 at which time protein expression was induced by the addition of isopropyl-1-thio-␤-D-galactopyranoside to 0.4 mM. Following 3 h of growth, the bacteria were harvested by centrifugation and lysed in 20 mM imidazole, pH 7.0, and 1 mM EDTA by three passes through a French press at 1000 p.s.i. Successive ammonium sulfate cuts from 0 to 30% saturation and from 30 to 50% saturation were taken from the cleared bacterial lysate. The pH of the solution was maintained at 7 throughout the ammonium sulfate precipitation procedure. The 30 -50% ammonium sulfate precipitation pellet was resuspended in 30 ml of 20 mM imidazole, pH 7.0, 1 mM EDTA, and 6 mM ␤-mercaptoethanol and dialyzed at 4°C against 4 liters of 20 mM imidazole, pH 7.0, 1 mM EDTA, and 6 mM ␤-mercaptoethanol for two changes. To the dialyzed fraction, urea powder was added to 6 M, the pH was adjusted to 7, and the solution was clarified by centrifugation. Urea was used to minimize protein-protein interactions and maximize separation during ion-exchange chromatography. The supernatant was loaded onto a 100-ml (2.5 ϫ 20 cm) CM52 column equilibrated with 6 M urea, 20 mM imidazole, pH 7.0, 1 mM EDTA, and 15 mM ␤-mercaptoethanol. A linear KCl gradient of 0 -400 mM in the equilibration buffer was used to elute the CN5 protein from the column. The fractions from the CM52 column were analyzed by 12% SDS-polyacrylamide gel electrophoresis with an acrylamide to bisacrylamide ratio of 29 to 1, and fractions containing the protein of interest were collected, dialyzed against 4 liters of 0.1% formic acid, a volatile agent, for three changes, and lyophilized. Amino acid analysis of the purified CN5 protein verified the authenticity of cDNA expression and the effectiveness of protein purification (Molecular Biology Core Laboratory, Case Western Reserve University, Cleveland, OH).
The purification protocols for TnC, TnI, TnT, and ␣/␤-Tm from chicken skeletal muscle have been described previously, and the purity of these preparations have been demonstrated (17,18). Concentrations for all protein preparations were determined by the Nanoorange protein quantitation kit (Molecular Probes, Eugene, OR) with the exception of TnC, which was determined by absorption, assuming an extinction coefficient of 0.132 at 259 nm for a 1-mg/ml solution.
Generation of a Specific Antiserum and Monoclonal Antibody against CN5-Using the purified chicken nebulette CN5 protein, a female BALB/c mouse was immunized intraperitoneally to generate a specific antiserum. For the primary immunization, 200 g of the CN5 antigen was dissolved in phosphate-buffered saline and mixed into a water-inoil emulsion using an equal volume of Freund's complete adjuvant. Subsequent boosts using 100 g of the purified CN5 protein were given at 3-week intervals using Freund's incomplete adjuvant. To collect the anti-CN5 antiserum, the mouse was sacrificed for terminal bleeding and the blood was recovered and allowed to clot at 4°C overnight. The serum was separated by centrifugation (2000 ϫ g for 5 min), an equal volume of glycerol was added, and the antiserum was stored at Ϫ20°C.
As described previously (19), a short term immunization procedure was applied to develop a specific anti-CN5 monoclonal antibody (mAb). A 7-week-old female BALB/c mouse was injected intraperitoneally and intramuscularly with a total of 50 g of the purified CN5 antigen in 100 l of phosphate-buffered saline mixed with an equal volume of Freund's complete adjuvant. Ten days later, the mouse was intraperitoneally boosted twice with 200 g each of the CN5 antigen in 100 l of phosphate-buffered saline without adjuvant on two consecutive days. Two days following the last boost, spleen cells were harvested from the immunized mouse to fuse with SP2/0-Ag14 mouse myeloma cells using 50% polyethyleneglycol 4000 containing 7.5% dimethyl sulfoxide as described previously (20). Hybridomas were selected by HAT (0.1 mM hypoxanthine, 0.4 M aminopterin, and 16 M thymidine) media containing 20% fetal bovine serum and screened by indirect enzyme-linked immunosorbent assay against coated CN5 using horseradish peroxidase-labeled goat anti-mouse total immunoglobulin (Sigma) as the second antibody. The positive hybridoma culture supernatants were verified by Western blotting against the CN5 protein using alkaline phosphatase-labeled goat anti-mouse total immunoglobulin second antibody (Sigma) as described previously (17). The anti-CN5 antibodysecreting hybridomas were subcloned three times by limiting the dilution using young BALB/c mouse spleen cells as a feeder to establish stable cell lines. A high affinity anti-CN5 mAb (Ne-110) was characterized. Immunoglobulin subclass typing by sandwich enzyme-linked immunosorbent assay using a reagent kit from BD Biosciences determined that this mAb is mouse IgG1. The hybridoma cells were then cultured to produce high titer supernatant and introduced into 2,6,10,14-tetramethyl pentadecane (Pristane, Sigma) primed peritoneal cavity of BALB/c mice to produce mAb-enriched ascites fluids. The hybridoma culture supernatant was used in this study.
Western Blot Identification of Nebulette-To monitor nebulette expression, Western blots were done using either the anti-CN5 antiserum or CN5 mAb Ne-110 on fresh SDS extracts of cardiac muscle from various species. The SDS-PAGE samples were prepared in the presence of 1 mM phenylmethanesulfonyl fluoride and 20 M leupeptin (Sigma) to ensure protein integrity. To better resolve large proteins (Ͼ100 kDa), 6% SDS-PAGE with an acrylamide:bisacrylamide ratio of 29:1 was used. However, in this system, the resolving gel does not retain proteins Ͻ60 kDa in molecular mass. To ensure that no proteolyic products Ͻ60 kDa in molecular mass were present, 14% (180:1) SDS-PAGE was used to present a complete SDS extract profile. After resolving total cardiac muscle extracts by SDS-PAGE, a three-buffer system was used to efficiently transfer the resolved protein bands onto nitrocellulose membranes (21). Following the transfer, the nitrocellulose replicas were blocked by incubation at room temperature for 3 h with Tris-buffered saline (150 mM NaCl, 2.5 mM KCl, and 25 mM Tris-HCl, pH 7.5) containing 1% bovine serum albumin. The blocked membrane was then incubated at 4°C overnight with either the anti-CN5 antiserum or Ne-110 mAb diluted in Tris-buffered saline containing 0.1% bovine serum albumin. The Western blots were developed via alkaline phosphatase-labeled second antibody as described previously (17).
Nebulette Affinity Chromatography-The relative binding strengths of ␣/␤-Tm and Tn subunits to the chicken nebulette fragment CN5 were determined using affinity chromatography. The purified nebulette fragment was coupled to CNBr-activated Sepharose 4B gel (Amersham Biosciences) according to the manufacturer's instructions using 9 mg of CN5 protein/0.5 ml of gel for coupling. A 0.4-ml column was packed and equilibrated with 100 mM NaCl, 15 mM PIPES, pH 7.0, 3 mM MgCl 2 , 1 mM K 2 EGTA, and 0.1 mM tris(carboxyethyl) phosphine. Proteins to be analyzed on the column were dissolved in 3 ml of the column buffer and dialyzed overnight at 4°C against 2 liters of the column buffer to ensure homogenous buffer conditions and complex formation. For binding experiments that test the interaction of ␣/␤-Tm or TnT with the CN5 affinity column, 0.4 nmol of protein was typically used. For experiments involving complexes of Tm with Tn subunits, 0.4 nmol of Tm dimer was used and the Tn subunits were mixed in a 1:1 molar ratio. After loading the dialyzed protein mixtures, the column was washed with 8 bed volumes of the equilibration buffer and eluted by a step gradient of 120 -680 mM NaCl in the equilibration buffer. All of the steps were carried out at room temperature. The fractions were examined by 12% (29:1) SDS-PAGE and visualized by silver staining to determine the protein peaks (22). A control experiment was done by mixing 0.4 nmol of TnT and bovine serum albumin (New England Biolabs) and loading onto the CN5 affinity column. Only the TnT bound specifically to the column since the bovine serum albumin was eluted in the flow-through and wash fractions only, indicating that nonspecific interactions in this setting were minimal (data not shown).
F-actin Cosedimentation-To determine the binding affinity of the CN5 nebulette fragment to F-actin and the interactions between thin filament regulatory proteins and CN5 in the presence of F-actin, a series of cosedimentation experiments were performed. Purified rabbit skeletal muscle F-actin (0.13 nmol) was incubated with serial concentrations of CN5, Tm, TnT, TnI, or whole troponin in either a 30-or 50-l volume for 45 min at 4°C. The cosedimentation buffer contained 10 mM potassium phosphate buffer, pH 7.4, 65 mM KCl, 2.5 mM MgCl 2 , 0.1 mM K 2 EGTA, and 1 mM dithiothreitol. In experiments in which CN5 was present during Tm or Tn subunit titrations, a constant concentration of 1.52 M CN5 was used. Following incubation, F-actin was sedimented by centrifugation at 100,000 ϫ g (45,000 rpm) for 30 min at 4°C in a Beckman Optima TL Ultracentrifuge using a TLA100.1 rotor. The supernatant and pellet were separated, and 15 l of 1ϫ SDS-PAGE sample buffer was used to solubilize the pellet. The supernatant and pellet were examined by either 14% (180:1) or 12% (29:1) SDS-PAGE. Resolved gels were silver-stained as described before and analyzed by gel densitometry. The titration curve was constructed from results of three experiments, and the data are presented as the mean Ϯ S.D. To estimate the binding affinity, the total [CN5] required for 50% maximum binding was obtained by averaging the data calculated from the fits for each individual experiment using single exponential fits.

RESULTS
Exon Organization of the Nebulin Gene-A mouse nebulin cDNA fragment was cloned by RT-PCR according to a published cDNA sequence (14). This cDNA was used as a probe to isolate an ϳ15-kb mouse nebulin genomic DNA fragment from a phage library, and an ϳ3.2-kb EcoRI fragment of the genomic DNA was subcloned and sequenced. The nucleotide sequence has been submitted to the GenBank Data Base (GenBank TM accession number AF440239). Two consecutive exons of the nebulin gene were found in the genomic DNA fragment encoding 69 amino acids and 35 amino acids, which correspond to two and one of the ϳ35 amino acid-repeating motifs, respectively (Fig. 1). Both exons had splice boundaries within the conserved pentapeptide, specifically at SX/XXY, indicating this sequence to be at the boundary between the ϳ35 amino acid motifs. Partial sequencing of other EcoRI subclones of the ϳ15-kb genomic DNA showed similar exon boundaries (data not shown). Because exon organization reflects the evolutionary and potentially functional units of multi-domain proteins, the genomic structure of the mouse nebulin gene indicates that the single ϳ35 amino acid motif (Fig. 1, exon B) represents the fundamental functional unit of nebulin-like proteins.
Cloning, Expression, and Purification of the CN5 Nebulette Fragment-Using oligonucleotide primers designed according to a published partial cDNA sequence (2), a chicken nebulette cDNA encoding a polypeptide beginning and ending with the conserved SXXXY pentapeptide was amplified. This frame coincides with the exon boundaries of the nebulin gene (Fig. 1). The chicken nebulette cDNA encodes a 168-amino acid protein spanning five nebulin-like motifs with a theoretical M r of 19398 and an isoelectric point of 9.73 and is shown to be aligned to the area of highest homology within the repeating domain of human nebulette (Fig. 2) (5, 6). Similar to the human nebulin fragments as reported previously (8), CN5 expressed in E. coli at high levels. The CN5 protein was readily purified by standard biochemical procedures. Purified CN5 migrated as a single band in SDS-PAGE (Fig. 3) and was found to be relatively soluble under non-denaturing conditions (solubility Ͼ0.5 mM). This was in contrast to the MSN fragment, which was confined to the inclusion bodies upon large scale expression in E. coli (data not shown), similar to previously published data on mammalian nebulin fragments (6,8,24). A mouse polyclonal antiserum was generated using the purified CN5 protein as the immunogen. As shown in Fig. 3, the polyclonal antiserum was specifically identified the CN5 protein in Western blot. This finding was also recapitulated by the anti-CN5 mAb Ne-110 (data not shown).
Expression of Nebulette in Cardiac Muscle and during Development-As shown in Fig. 4, the anti-CN5 mAb strongly identified nebulette proteins from chicken and bovine heart extracts. The monoclonal antibody did not strongly recognize nebulette in mouse, sheep, or rabbit cardiac muscle extracts, implying epitope diversity of nebulette proteins across species. Initially, 14% (180:1) SDS-PAGE was used to resolve the isoforms to ensure that no proteolytic fragments were present. To further increase the resolution of isoforms, 6% (29:1) SDS-PAGE was used. Although the chicken nebulette isoforms were identified as two very closely migrating bands under 14% (180:1) SDS-PAGE, 6% (29:1) SDS-PAGE was better able to resolve the isoform diversity. Using the chicken as a developmental model (total extracts from embryonic days 14 to 20), neonatal and adult chicken hearts were analyzed for nebulette expression (Fig. 5). Multiple bands were identified by the anti-CN5 mAb, suggesting nebulette isoform diversity (5). A higher M r isoform appeared to be predominant throughout heart development but was down-regulated by 6 weeks post-hatch. This developmental pattern of high to low M r isoform expression is reflective of the regulation of TnT expression with development (23), suggesting that this type of isoform switch is fundamentally conserved in the myofilament proteins of striated muscle. Two closely migrating isoforms, a predominant lower M r isoform and a less expressed higher M r isoform, were identified in extracts from bovine heart chambers, suggesting that isoform diversity was not confined to the avian heart. It was also noted that the ratio of the isoforms varied between chambers with the lower M r isoform being predominant in the left ventricle. The primary sequence differences between the nebulette isoforms are of interest but remain to be determined.
TnT Interaction with CN5 Is Strengthened by Tm and Weakened by TnI-Using affinity chromatography and F-actin cosedimentation, the interactions between CN5 and the thin filament proteins F-actin, TnT, TnI, and Tm were tested. Consistent with data previously demonstrated for other nebulin fragments, CN5 specifically bound to F-actin in cosedimentation experiments with half-maximal binding occurring at 2.09 Ϯ 0.6 M CN5 (Fig. 6). The maximum stoichiometry of CN5 binding to undecorated F-actin was ϳ0.4 mol CN5/mol actin or 2.8 mol of the 5-unit CN5 protein/7 actin monomers that defines the striated muscle thin filament regulatory unit. This finding suggests that in the absence of regulatory proteins on the actin filament, the CN5 protein binds at approximately twice the stoichiometry that would be expected if each of the five nebulin motifs bound one actin monomer (1.4 mol CN5/7 mol actin). Assuming that the ϳ35-amino acid sequence motif represents an actin-binding unit, this 2:1 binding ratio to naked actin filaments may be a reflection of the multiple binding sites available for nebulin on each actin monomer as well as additional nonphysiological binding modes that nebulin fragments may undertake under these conditions (25). It should also be noted that the interaction between two or three of the nebulin-like motifs with adjacent actin monomers may be sufficient to sustain a stable association between CN5 and F-actin in vitro, allowing an overestimation of the stoichiometry with the unregulated actin filament. This is supported by the observation that a 3-4-unit nebulin fragment can cross-link adjacent F-actin strands (25). Although the binding between CN5 and F-actin was analyzed simply as a one-step reaction, it remains possible that the binding of nebulin and nebulette fragments containing multiple actin binding motifs to F-actin may be a corporative multiple-step process.
To test the direct interactions between CN5 and TnT or Tm, a CN5 affinity chromatography column was used. Purified ␣/␤-Tm dimers only moderately interacted with the nebulette fragment as evidenced by the flow-through and subsequent rapid elution from the CN5 affinity column between 120 and 160 mM NaCl (Fig. 7A). In contrast, TnT alone interacted strongly with the nebulette fragment and was eluted with a primary peak at 280 -360 mM NaCl (Fig. 7B). Combining TnT and Tm in a 1:1 molar ratio resulted in both proteins binding to the column and a noticeable shift to higher affinity as compared with Tm or TnT alone (Fig. 7C). The main peak of the Tm⅐TnT complex was eluted at 480 -560 mM NaCl. Early elution of some ␣/␤-Tm alone was observed in Fig. 7C, corresponding to the elution of uncomplexed ␣/␤-Tm as seen in Fig. 7A. However, the overlapping elution profiles suggest that ␣/␤-Tm bound TnT as expected, forming a binary complex that interacted with the CN5 fragment, and remained resistant to higher salt concentrations. Troponin I significantly reduced the binding affinity between CN5 and the Tm⅐TnT complex, resulting in peak elution of Tm⅐TnT⅐TnI between 280 and 400 mM NaCl (Fig. 7D) well below the elution range for Tm⅐TnT. The complete overlap of the Tm, TnT, and TnI peaks under this condition demonstrates that TnI stabilized the interaction between Tm and TnT as expected to form a Tm⅐TnT⅐TnI complex and subsequently minimized the early elution of Tm seen in Fig.  7C. The flow-through of some Tm⅐TnT⅐TnI complexes from the CN5 affinity column was observed in Fig. 7D, indicative of TnI affecting the Tm-TnT-CN5 interaction. When TnC was included to form a Tm⅐Tn complex, no binding to the CN5 column was detected (data not shown).
Cosedimentation experiments verified the affinity chroma- tography results within the framework of the actin filament. In the presence of serial concentrations of the thin filament regulatory proteins and a fixed concentration of CN5, two likely classes of interactions were reflected by the cosedimentation experiments. Because the thin filament regulatory proteins bind CN5 with various affinities, the free proteins during the cosedimentation experiments may compete with F-actin for CN5 binding. In addition, the F-actin-bound Tm and/or Tn subunits may affect the interaction between CN5 and F-actin, either through steric blocking, an increase in the number of available CN5 binding sites beyond those on F-actin alone, or conformational modulation of the CN5-F-actin interaction. The results in Fig. 8 show that the binding of CN5 to F-actin decreases with increasing concentrations of Tm and/or Tn subunits. However, CN5 binding remained largely unchanged once the physiological stoichiometric decoration of F-actin (7:1) by the regulatory protein(s) was reached. This finding suggests that the free Tm and/or Tn subunits were not able to compete with F-actin and eliminate CN5 binding. As demonstrated in Fig. 8, the decoration of F-actin with Tm reduced concurrent nebulette binding, possibly because of Tm constraining CN5 to binding at its physiological sites on F-actin while limiting nonphysiological interactions between CN5 and the actin filament (25). In contrast, TnT was more efficient in reducing the binding of nebulette to F-actin, suggesting that the strong interaction between nebulette and TnT weakened nebulette-actin binding. However, the Tm⅐TnT binary complex resulted in only a moderate effect on nebulette binding to F-actin (Fig. 8B) and provided an approximate stoichiometry of 0.7 CN5/7 actin regulatory units, or half of the theoretical stoichiometry expected if each of the five nebulin motifs bound an actin monomer (1.4 CN5/regulatory unit). This lower stoichiometry may be explained by the hypothesis that in contrast to the head-tailregistered binding of Tm to F-actin, the unorganized binding of nebulin fragments would not fully decorate the actin filament. These experiments demonstrated that the Tm⅐TnT complex and nebulette were able to decorate F-actin simultaneously. The higher amounts of CN5 binding to Tm⅐TnT-decorated F-FIG. 7. Binding of Tm, TnT, or TnI to a CN5 affinity column. The interaction of Tm and troponin subunits, alone or in combination, with the CN5 fragment was tested by CN5 affinity chromatography. The column profiles were analyzed by SDS-PAGE. A, ␣/␤-Tm alone showed weak binding to the immobilized CN5 fragment, demonstrated by Tm in the flow-through (F/T) as well as an early elution peak at 120 -160 mM NaCl. B, fast TnT (fTnT) alone bound strongly to the CN5 column, because it was absent in the flow-through and showed a primary peak eluting at 280 -360 mM NaCl. C, the Tm⅐TnT complex bound more effectively to the CN5 column, resulting in a main Tm⅐TnT peak eluting at 480 -560 mM NaCl significantly higher than that observed for Tm or TnT alone. D, addition of fast TnI (fTnI) weakened the binding of Tm⅐TnT with the immobilized CN5 protein as evidenced by the elution of Tm⅐TnT⅐TnI from the affinity column between 280 and 400 mM NaCl. All SDS-PAGE was 12% gel with an acrylamide:bisacrylamide ratio of 29:1 and was silver-stained to visualize the resolved proteins.
actin versus TnT-decorated F-actin suggest relative increase in the CN5-actin affinity. This effect was confirmed by assays using the Tm⅐TnT⅐TnI complex. The higher levels of CN5 binding to F-actin in the presence of Tm⅐TnT or Tm⅐TnT⅐TnI rather than in the presence of TnT alone exclude the possibility that steric blocking by TnT competes with the CN5-actin interaction. Because the binding of CN5 to TnT is stronger than that observed for Tm (Fig. 7, A and B), the difference between CN5 interaction with F-actin-Tm versus F-actin-TnT is not likely due to the number of binding sites on these decorated actin filaments since it would predict stronger binding to F-actin-TnT, which was not observed. Therefore, the results support the hypothesis that the Tm⅐Tn complex and nebulin have an allosteric interactive relationship dictating the interaction of nebulin-like motifs with the thin filament.
Additional cosedimentation experiments tested the effect of CN5 on the affinity of Tm⅐Tn assembly onto the actin filament. Tropomyosin-troponin concentration was titrated in the presence or absence of a fixed concentration (1.52 M) of CN5. The CN5 fragment decreased the concentration of Tm⅐Tn required to achieve half-maximal binding to F-actin (147.4 Ϯ 32.1 nM without CN5 versus 70.8 Ϯ 4.9 nM with CN5; p ϭ 0.05 by Student's t test) (Fig. 9), further suggesting that nebulin-like motifs participate in the striated muscle thin filament assembly.

DISCUSSION
The role of nebulin-like proteins in the sarcomere of striated muscles has been reported to be primarily a structural one. Evidence has been mounted by studies showing the interaction of nebulin or nebulette fragments with F-actin (6,8), protein constituents of the Z-line (6,9), and with tropomodulin at the pointed ends of the thin filaments (11). Correlations between nebulin/nebulette expression and Z-line thickness as well as nebulin protein size and actin filament length in skeletal muscle sarcomeres (5,26) have all contributed to the large body of evidence in support of a structural role for nebulin/nebulette.
Through genomic DNA cloning, we showed in this study that the exons of the nebulin gene are arranged in a manner reflective of the ϳ35-amino acid nebulin-like motifs that serve as the basic functional units. The partial genomic DNA sequence data suggest that the conserved SXXXY sequence straddles the junction between adjacent exons (Fig. 1) as also suggested by Millevoi et al. (5), providing a compelling case for the SXXXY sequence serving as the boundary between the repeating nebulin-like motifs. Although the use of intact proteins remains desirable, the insolubility of intact nebulin and nebulette in physiological buffers necessitates the use of protein fragments during functional characterizations (6,8,24). This afforded the rationale in designing a five-motif nebulette fragment defined by the conserved pentapeptide flanking each motif (Fig. 2).  52 M). In each panel, the precipitates from ultracentrifugation were resolved by SDS-PAGE and silver-stained. The molar ratio of F-actin to the titrated protein during the cosedimentation is indicated under each lane. Bound fractions were assessed by densitometry of silver-stained gels, and the binding of CN5 was plotted against total concentration of the thin filament regulatory proteins. Tm moderately reduced CN5 binding to F-actin in contrast to TnT, which more effectively decreased the binding of CN5 to F-actin (A). The addition of Tm⅐TnT binary complexes to the titration also demonstrated a saturable inhibitory effect. (B) addition of TnI to the Tm⅐TnT binary complex slightly increased the inhibitory effect on CN5 binding to F-actin.
Similar to published observations with other nebulin fragments (4,8,24), we found that the nebulette CN5 fragment bound F-actin with affinity in the low micromolar range (Fig.  6). Although unsurprising given the homology between nebulin and nebulette, this finding confirmed that F-actin binding is a conserved function of the nebulin-like motifs. For the CN5 protein, F-actin binding is relevant because this region of nebulette localizes to actin filaments when transfected into primary cultures of cardiomyocytes (6).
We observed that the binding of CN5 and ␣/␤-Tm to F-actin (Fig. 8A) was not mutually exclusive, although there was apparent competition consistent with both proteins binding to subdomain 1 on the actin monomer (25). However, it was evident that CN5 binding to undecorated F-actin was at approximately twice the stoichiometry when compared with Tm, Tm⅐TnT, or Tm⅐TnT⅐TnI-decorated F-actin (Fig. 8). This finding suggests that decoration of naked actin filaments with nebulinlike proteins is not fully indicative of their physiological position on F-actin and that the regulatory Tm⅐Tn filament on F-actin is able to further constrain nebulin binding sites. In addition, the results demonstrate that both nebulin and Tm have binding domains on F-actin that are not mutually exclusive, although the cosedimentation experiments did demonstrate some overlap (Fig. 8). This recapitulates an in vivo setting in which nebulin-like motifs and Tm dimers run parallel along the F-actin filament. Using affinity chromatography, we found a weak but reproducible association between the five-motif nebulette fragment and ␣/␤-Tm (Fig. 7A). Although this interaction was a relatively low affinity, it opened the possibility that head-to-tail-linked Tm dimers and repeating nebulin-like motifs running parallel along the actin filament may provide a higher total avidity of interaction. Because the allosteric regulation of striated muscle contraction is known to involve shifts of the position of Tm on F-actin (27, 28), nebulinlike motifs and Tm may come closer into contact or distance themselves depending on the dynamic position of Tm on the actin filament, thus allowing the nebulin-Tm association to contribute reversibly to thin filament dynamics. This model agrees with previously published data (25) demonstrating that one site of nebulin-actin interaction lies at subdomain 1 of the actin, a position believed to be occupied by tropomyosin during the blocked or resting state of the striated muscle thin filament.
In contrast to Tm, the interaction between the CN5 fragment and TnT was stronger with the peak of elution occurring at 280 -360 mM NaCl by affinity chromatography (Fig. 7B). Troponin T complexed to Tm showed even higher affinity binding to CN5 (peak at 480 -560 mM NaCl) (Fig. 7C). Three possible mechanisms may underlie this observation: (a) higher avidity from the sum of two independent CN5 binding sites on Tm and TnT; (b) Tm and TnT as a binary complex possibly presenting a unique interaction pocket with increased affinity for nebulinlike motifs; or (c) the binding of Tm to TnT resulting in a conformational change in TnT, consequently increasing the binding affinity of the protein for CN5. F-actin cosedimentation results further supported the formation of a Tm⅐TnT binding pocket for the nebulin-like motifs. Although TnT alone caused a large decrease in the binding of nebulette to F-actin (Fig. 8A), the addition of Tm moderated the effect and allowed the assembly of a thin filament accommodating both Tm⅐TnT and CN5 (Fig. 8B). Additional affinity chromatography experiments demonstrated that TnI decreased the affinity of Tm⅐TnT for the nebulette fragment (Fig. 7D), suggesting that the interaction pocket formed by the Tm⅐TnT complex is either directly competed by TnI or altered by a change in the conformation of TnT upon TnI binding. In either case, these results underscore the role of the conserved COOH-terminal T2 region of TnT in mediating the interaction with CN5. These findings were verified by cosedimentation experiments in which Tm⅐TnT⅐TnI was also found to modulate CN5 binding to F-actin, similar to the results using Tm⅐TnT (Fig. 8B). Therefore, the association of CN5 with an F-actin-Tm⅐TnT filament demonstrated both a Tm⅐TnT and an F-actin binding component with the former being challenged moderately in the presence of TnI. The interaction between CN5 and the thin filament was extended by additional cosedimentation experiments demonstrating that the nebulette fragment was able to increase the affinity of Tm⅐Tn for F-actin (Fig. 9). This increase suggests that the documented interactions, particularly between nebulin-like motifs and Tm⅐TnT, provided a tangible means of increasing the affinity of the thin filament assembly for F-actin. This observation also supports the hypothesis that nebulin-like motifs are able to interact with and contribute to the regulation of the Tm⅐Tn assembly on the actin filament.
The data collected indicate that during the allosteric control of striated muscle contraction, a reversible interaction may occur along the thin filament between nebulin and the Tm⅐Tn regulatory system, affecting its interaction with Factin. The binding between Tm⅐TnT and nebulin-like motifs may be the pivotal association underlying this observation and offers a role for regulation by TnC or an analogous Ca 2ϩ -binding protein such as calmodulin (12). Experiments by Root and Wang (13) detailing changes in actomyosin in the presence of nebulin-like motifs may be a preliminary account of how these interactions integrate within the actin filament. The data presented in Figs. 8 and 9 necessitate future experiments in testing the acto-S1 interaction in an intact regulated filament. We propose that allosteric changes during muscle contraction due to Ca 2ϩ binding to TnC and/or strong cross-bridge attachment may contribute to the cooperativity of striated muscle, possibly by allowing a reversible nebulin/ nebulette interaction with the Tm⅐Tn complex and in turn facilitating a more pronounced propagation of the transition from the blocked to the closed or from the closed to the open states of the thin filament (29). We hypothesize that the putative effect of the nebulin-like motifs on the coordinated regulation of the thin filament may contribute to the higher cooperativity of skeletal versus cardiac muscles (30 -32), a possibility that will be tested most effectively in the framework of an integrated muscle fiber.