Caldesmon and the regulation of cytoskeletal functions

Abstract

Caldesmon (CaD) is an extraordinary actin-binding protein, because in addition to actin, it also bindsmyosin, calmodulin and tropomyosin. As a component of the smoothmuscle and nonmuscle contractile apparatus CaD inhibits the actomyosin ATPase activity and its inhibitory action is modulated by both Ca2+ and phosphorylation. The multiplicity of binding partners and diverse biochemical properties suggest CaD is a potent and versatile regulatory protein both in contractility and cell motility. However, after decades ofinvestigation in numerous laboratories, hard evidence is still lacking to unequivocally identify its in vivo functions, although indirect evidence is mounting to support an important role in connection with the actin cytoskeleton. This chapter reviews the highlights of the past findings and summarizes the current views on this protein, with emphasis of its interaction with tropomyosin.

Figure 1

CaD has two isoforms resulting from alternative splicing. The domain structures of mammalian h-CaD (upper bar) and l-CaD (lower bar) indicate the common functional regions for myosin-binding (light green), CaM-binding (red) and actin-binding (blue and turquoise). All functional domains are shared between both isoforms, except that the central ‘spacer’ (yellow) is missing in l-CaD. Also shown are the two phosphorylation sites common for ERK and cdc2 kinase. A color version of this figure is available at www.Eurekah.com.

Figure 2

The compensatory isoform switchover depends on the tissue type. Western analysis by Odyssey imaging of tissue extracts from bladder and aorta of wild-type (^+/+) and CaD-deficient (^−/−) mice. Note that the up-regulation of nonmuscle CaD is more robust in bladder than in aorta. There appeared to be more severe proteolysis in bladder than in aorta. β-actin (green) was used as a reference. A color version of this figure is available at www.Eurekah.com.

Figure 3

CaD binds actin longitudinally with the N-terminus appearing on one side of the actin filament. A model illustrates the position of h-CaD on the two-stranded actin filament. CaD molecules on one strand are shown in solid lines and those on the opposite strand in broken lines. Note that the ends of all CaD molecules appear on the same side of an actin filament. Such an arrangement is only possible with h-CaD. Taken from Mabuchi et al.

Figure 4

A model with obliquely arranged contractile units allows clustered distribution of h-CaD along both actin and myosin filaments. Actin filaments (thin lines attached to the dense body) and myosin filaments (thick lines with a helical twist to represent the side-polar filaments) intersect each other at an angle. Both the intersecting angles and the twist of myosin filaments are exaggerated in this drawing. The myosin heads on opposite sides of the myosin filament interact with two bundles of actin filaments in reverse orientation, allowing muscle shortening. As suggested by this model, because of the nonparallel alignment, a given myosin filament is bound to interact with multiple actin filaments in the same bundle. Assuming that h-CaD (red dots) is only present at the junctions of these filaments, it naturally has a clustered distribution along either the actin or the myosin filament. Taken from Wang. A color version of this figure is available at www.Eurekah.com.

Figure 5

3-D reconstruction shows an unusual binding mode of CaD on the actin filament. Surface views (a–c) of thin filament reconstructions showing the position of the C-terminal CaD fragment, H32K and phospho-H32K on F-actin and transverse sections (d–f) through maps of 3D reconstructions. The extra density (open bold arrows in b and c) contributed by H32K is associated with subdomains 1 and 2 of actin. In b the density that spans like a staple from the back of subdomain 1 to subdomain 3 of the neighboring actin monomer of the genetic helix (red ellipse). This inter-strand connectivity is present in b (red arrow), but is absent in a and c (green arrows). Open bold arrows in e indicate regions of significant H32K density and the red arrow points to the inter-strand density. Taken from Foster et al. A color version of this figure is available at www.Eurekah.com.

Figure 6

CaD is able to crosslink neighboring actin subunits. Disulfide crosslinking occurs between a C-terminal CaD fragment that contains two cysteine residues (at positions 595 and 766; H32Kqc), or its variant that contains only one cysteine residue (at position 766; H32Kqc/ca) and actin, as shown by the SDS-PAGE with Coomassie staining in the absence of reducing agent. Note that ERK-induced phosphorylation resulted in less H32Kqc/ca•actin crosslinking and almost no H32Kqc•actin₂ species.

Figure 7

Actin filaments form bundles in the presence of CaD. CaD is known to bundle actin filaments owing to its multiple actin-binding sites. However, full-length h-CaD forms loose bundles (top panel), whereas the C-terminal fragment forms tight and straight bundles (bottom panel). Both samples were processed for rotary shadowing and viewed by electron microscopy. Scale bar: 1 μm.(Mabuchi and Wang, unpublished results).

Figure 8

A model depicting how CaD could change the configuration of myosin. Smooth muscle myosin heavy chains (orange) undergo a conformational change upon phosphorylation (red dots) at the light chains (yellow) from an elongated (6S) to a compact (10S) configuration. In this model binding of the N-terminal domain of CaD (blue) to the S-2 region of myosin is thought to interfere with the head-rod interaction and shift the myosin conformation to the 6S form even in the absence of phosphorylation. A color version of this figure is available at www.Eurekah.com.