Role of the Myosin Assembly Protein UNC-45 as a Molecular Chaperone for Myosin

See allHide authors and affiliations

Science  25 Jan 2002:
Vol. 295, Issue 5555, pp. 669-671
DOI: 10.1126/science.1066648


The organization of myosin into motile cellular structures requires precise temporal and spatial regulation. Proteins containing a UCS (UNC-45/CRO1/She4p) domain are necessary for the incorporation of myosin into the contractile ring during cytokinesis and into thick filaments during muscle development. We report that the carboxyl-terminal regions of UNC-45 bound and exerted chaperone activity on the myosin head. The amino-terminal tetratricopeptide repeat domain of UNC-45 bound the molecular chaperone Hsp90. Thus, UNC-45 functions both as a molecular chaperone and as an Hsp90 co-chaperone for myosin, which can explain previous findings of altered assembly and decreased accumulation of myosin in UNC-45 mutants ofCaenorhabditis elegans.

The motor protein myosin assembles into molecular machines essential for processes such as cell division, cell motility, and muscle contraction through a multistep pathway requiring additional proteins (1). UCS proteins (Caenorhabditis elegans UNC-45, Podospora anserinaCRO1, Saccharomyces cerevisiae She4p, andSchizosaccharomyces pombe Rng3p) are involved in myosin function and contain homologous COOH-terminal domains (2–6). Unc-45 and RNG3 are essential genes whose loss-of-function alleles implicate their gene products in myosin assembly in vivo; substitutions of conserved residues within or near their UCS domains cause defective assemblies of thick filaments during muscle development and of the contractile ring during cell division (Fig. 1A). UNC-45 and Rng3p interact functionally and specifically in vivo with muscle and cytoskeletal myosins, respectively (2,6, 7). UNC-45 also contains an NH2-terminal domain composed of three tetratricopeptide repeat (TPR) motifs and a newly discovered central region. This three-domain configuration is maintained in all UNC-45 animal homologs identified, including those of Drosophila,Xenopus, zebrafish, mouse, and human (8).

Figure 1

UNC-45 structure and function relationships. (A) Unc-45 and S. pombe RNG3 mutations show that UCS proteins are essential in myosin maturation and assembly in vivo. The three-domain structure is conserved in all animal homologs of UNC-45, from Drosophila to human (8). UNC-45 has an NH2-terminal TPR domain that closely resembles that of Hsp90-binding proteins. The COOH-terminal UCS region defines the UCS family of proteins and is also found in unicellular eukaryotes. Temperature-sensitive UNC-45 mutants (yellow arrowheads) show decreased accumulation and altered assembly of myosin and its filaments in C. elegans (2). Lethalunc-45 alleles (yellow crosses) are nonsense mutations in the central region that prevent myosin assembly (2,7). Temperature-sensitive mutations in the RNG3UCS domain (green arrowheads) disrupt actomyosin contractile ring assembly (6). (B) The TPR domain of UNC-45 interacted with Hsp90 but not Hsp70. Western blot analysis of proteins pulled down with full-length UNC-45 (UNC-45) or TPR(–) by antibodies to FLAG, which bound their COOH-terminal FLAG epitope tags from Sf9 cell lysates expressing each protein. The membranes were probed with antibodies to Hsp70 and Hsp90. (C) SDS-PAGE (Coomassie blue staining) of UNC-45 and TPR(–) (both at 6 μM) pull-downs of Hsp90 (5 μM) at 4°C. TPR(–) and Hsp90 can be resolved in this gel (Ref. lane). (D) Surface plasmon resonance analysis of TPR(o) protein (50 μM) binding to immobilizedC. elegans Hsp90 with various competing peptides. DVEEM and EEVD (14) peptides are control peptides; 70-C12 (GGAGGPTIEEVD) and 90-C12 (AEEDASRMEEVD) peptides represent the COOH-terminal amino acids of Hsp70 and Hsp90, respectively, and contain the sequences (GPTIEEVD for Hsp70 and MEEVD for Hsp90) that are necessary and sufficient for binding their corresponding TPR domains (13).

TPR motifs are protein-protein interaction modules of 34 amino acids, often found in tandem repeats of 3 to 16 in a diverse set of proteins (9). The UNC-45 TPR domain resembles that of Hop (Hsp70/Hsp90–organizing protein) and of protein phosphatase 5 (2, 10, 11), which bind conserved COOH-terminal sites in the molecular chaperones Hsp70 and/or Hsp90 (12, 13). Full-length UNC-45 and a TPR-deleted construct [TPR(–)] (11) were used to pull down endogenous Hsp70 and Hsp90 from Sf9 insect cell lysates. In our study, only full-length UNC-45 complexed with Hsp90 (Fig. 1B), indicating that this interaction required the TPR domain (11). Both constructs pulled down Hsp70, suggesting a possible Hsp70 binding site outside the TPR domain or chaperone-client interactions between these proteins. In the presence of only purified proteins, full-length UNC-45, but not TPR(–), was able to pull down recombinant C. elegans Hsp90 (Hsp90) (Fig. 1C) (11), indicating a direct interaction between the UNC-45 TPR domain and Hsp90. To determine whether the TPR domain preferentially interacts with Hsp90 or Hsp70, the binding of the recombinant UNC-45 TPR domain (TPR) (11) to immobilized Hsp90 (11) was competed byC. elegans Hsp70 or Hsp90 12-oligomer COOH-terminal peptides (Fig. 1D) (11). The structure of the MEEVD (14) Hsp90 COOH-terminal pentapeptide is conserved between C. elegans and humans, and its specific complex with the Hop TPR 2A domain has been determined crystallographically (13). The Hsp90 12-oligomer peptide (90-C12) was as effective as MEEVD at competing with Hsp90 for TPR binding (Fig. 1D), but the Hsp70 12-oligomer peptide (70-C12) showed a slight competitive effect only at high concentrations. The UNC-45 TPR domain therefore appears to bind Hsp90 preferentially.

Although genetic and immunolocalization experiments suggest that nematode body wall myosin interacts with UNC-45 (2,7, 15, 16), there has been no biochemical evidence demonstrating their association. Preliminary experiments showed that recombinant full-length UNC-45 and Hsp90 formed complexes with myosin in C. elegans lysates. Direct binding among these proteins was tested with purified components. Binary mixtures containing immobilized myosin and either Hsp90 or UNC-45, and a ternary mixture containing all three proteins, were incubated at 4°C, and the resulting complexes were analyzed by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 2A) (11). No interactions between myosin and UNC-45 or Hsp90 were detected. We next evaluated the effect of a 30°C incubation on these mixtures, because this temperature is required in vitro for Hsp90 and its associated co-chaperones to form complexes with other well-studied substrates (17) and would enhance hydrophobic interactions. At 30°C, approximately equimolar complexes were detected between myosin and both Hsp90 and UNC-45 in the binary mixtures and among the three proteins in the ternary mixture (Fig. 2B). There was no detectable change in binding of either UNC-45 or Hsp90 to myosin in the ternary mixture as compared to the binary mixtures (Fig. 2B). The ability of Hsp90 to bind myosin directly is similar to its ability to interact with the glucocorticoid receptor in the absence of the co-chaperone Hop (18). We next performed pull-down experiments to determine which region of UNC-45 interacted with myosin. We found that TPR(–), like full-length UNC-45, bound myosin. Thus, the myosin-binding site of UNC-45 lies within the central region/UCS domain fragment (Fig. 2C).

Figure 2

Myosin interacted with Hsp90 and the COOH-terminal region of UNC-45 in a temperature-dependent manner. All panels are Coomassie blue–stained SDS-PAGE analyses. (A) Myosin pull-downs of purified full-length UNC-45 and/or Hsp90 (both at 5.5 μM) at 4°C. There is no detectable interaction of UNC-45 and/or Hsp90. (B) Identical reactions to those in (A), except that they were incubated at 30°C for 30 min. Myosin bound both UNC-45 and Hsp90. (C) TPR(–) was substituted for full-length UNC-45 under reaction conditions described in (B), with similar results.

We investigated purified UNC-45 for chaperone activity in vitro, because several proteins that bind directly to Hsp90 show such activity, including p23, the large immunophilins, and the Hsp90 protein kinase–targeting subunit Cdc37 (19–21). Molecular chaperones possess two fundamental biochemical characteristics: (i) the ability to prevent aggregation of partially unfolded proteins and (ii) the ability to maintain partially unfolded proteins in a state competent for refolding (22, 23). UNC-45 demonstrated these characteristics when tested with the well-studied chaperone substrate citrate synthase (11, 23). The chaperone activity of UNC-45 mapped to the central region/UCS domain fragment, the same region responsible for binding myosin (11).

Certain myosins and myosin subfragments containing the myosin head are not functional when recombinantly expressed (24,25), unlike myosin rod fragments and associated light chains (26, 27), and might require interaction with chaperones. We studied the binding of UNC-45 to myosin heads using myosin subfragment 1 (S1) that had been enzymatically cleaved from intact myosin. Full-length UNC-45, immobilized via its FLAG tag, was used to pull down scallop muscle S1 with or without Hsp90 at 30°C (11). UNC-45 formed stoichiometric complexes with S1 and Hsp90 (Fig. 3A). Thus, UNC-45 directly binds the myosin head in addition to Hsp90, which is consistent with the formation of a ternary complex. S1, like citrate synthase, aggregated when incubated at 43°C, but the addition of increasing amounts of UNC-45 reduced S1 aggregation in a concentration-dependent manner (Fig. 3B) (11), whereas bovine serum albumin had no effect. Thus, the myosin head is a substrate for the UNC-45 chaperone activity. The biological specificity of this UNC-45 activity toward myosin suggested by the genetic and cell biological experiments may be determined by parts of the molecule that make additional contacts with myosin, as well as by the actual localization of UNC-45 in vivo.

Figure 3

UNC-45 prevented the thermal aggregation of S1 and formed stoichiometric complexes with Hsp90 and S1. (A) SDS-PAGE analysis of S1 (10 μM) and/or Hsp90 (1 μM) interacting with UNC-45 (1 μM) at 30°C for 30 min. (B) The aggregation of S1 (1.0 μM) at 43°C was measured by light scattering (320 nm) (11) with no additional protein (solid circles), 2.0 μM bovine serum albumin (open circles), 0.5 μM UNC-45 (squares), 1.0 μM UNC-45 (diamonds), or 2.0 μM UNC-45 (triangles).

We have shown that UNC-45 binds myosin through its COOH-terminal regions and binds Hsp90 through its TPR domain. Both interactions appeared to be stoichiometric, similar to those of the progesterone receptor with Hsp90 and interacting co-chaperones (28). The interaction of UNC-45 with myosin required above-ambient temperatures consistent with a chaperone:substrate relationship. The hypothesis that UNC-45 is a myosin chaperone is supported by the fact that UNC-45 interacted and showed molecular chaperone activity in vitro with the myosin head. Thus, in conjunction with previous genetic studies that implicated UNC-45 in myosin assembly in vivo, our results suggest that the UCS proteins may function as myosin-directed chaperones analogously to Cdc37, which targets Hsp90 to protein kinases and exhibits chaperone activity in vitro (21). Previous studies show that the UNC-54 myosin heavy chain in unc-45 mutant backgrounds acts as a poison for thick filament assembly (7), accumulates at 50% lower levels than in the wild type (2), localizes abnormally (2, 15), and produces structurally altered thick filaments (2). These results may be explained by defects in the chaperone or co-chaperone activity of UNC-45 that lead to altered folding and assembly of myosin.

  • * These authors contributed equally to this work.

  • Present address: Max Planck Institute for Biochemistry, D82152 Martinsried, Germany.

  • To whom correspondence should be addressed. E-mail: hepstein{at}


View Abstract

Stay Connected to Science

Navigate This Article