Brush border myosin-I structure and ADP-dependent conformational changes revealed by cryoelectron microscopy and image analysis

Abstract

Brush border myosin-I (BBM-I) is a single-headed myosin found in the microvilli of intestinal epithelial cells, where it forms lateral bridges connecting the core bundle of actin filaments to the plasma membrane. Extending previous observations (Jontes, J.D., E.M. Wilson-Kubalek, and R.A. Milligan. 1995. Nature [Lond.]. 378:751-753), we have used cryoelectron microscopy and helical image analysis to generate three-dimensional (3D) maps of actin filaments decorated with BBM-I in both the presence and absence of 1 mM MgADP. In the improved 3D maps, we are able to see the entire light chain-binding domain, containing density for all three calmodulin light chains. This has enabled us to model a high resolution structure of BBM-I using the crystal structures of the chicken skeletal muscle myosin catalytic domain and essential light chain. Thus, we are able to directly measure the full magnitude of the ADP-dependent tail swing. The approximately 31 degrees swing corresponds to approximately 63 A at the end of the rigid light chain-binding domain. Comparison of the behavior of BBM-I with skeletal and smooth muscle subfragments-1 suggests that there are substantial differences in the structure and energetics of the biochemical transitions in the actomyosin ATPase cycle.

Figure 1

Images of actin filaments decorated with BBM-I. (a) A cryoelectron micrograph of actoBBM-I in the absence of nucleotide (rigor), which has been digitized and computationally straightened. There is very low contrast, due to an excess amount of protein in the background. (b) A cryoelectron micrograph of an actin filament decorated with BBM-I.ADP. As with the rigor images, actoBBM-I.ADP exhibits very low image contrast, due primarily to a high background. (c and d) Computed Fourier transforms of the straightened filaments shown in a and b, respectively.

Figure 2

The final layer line data for actin filaments decorated with BBM-I. The layer line data in a and b were used in the Fourier-Bessel synthesis of the rigor and MgADP maps shown in Figs. 3 and 6, respectively. The data were truncated to a uniform resolution of 30 Å. The solid lines represent the amplitudes and the dotted lines represent the phases for each layer line. The ordered pairs of numbers are the Bessel order (n) and the layer line number (l) for a 54:25 helical selection rule.

Figure 3

3D map of actoBBM-I in the absence of nucleotide, calculated from the layer line data shown in Fig. 2. Densities can be identified in the 3D map that are attributable to the catalytic domain, each of three calmodulin light chains, and a lipid binding domain, as indicated. C, catalytic; 1, 2, and 3, indicate the three calmodulin light chains; LB, lipid-binding domain.

Figure 6

3D map of actin filaments decorated with BBM-I.ADP. The map was calculated using the layer line data shown in Fig. 2 b. Like the rigor map, mass can be seen to extend to high radius. The gross features of both the rigor and ADP maps are essentially the same: a globular catalytic domain and a long, irregular light chain– binding domain. Although the lipid-binding domain is not visible in the ADP map, mass protrudes from the end of the light chain– binding domain, at the position where the lipid-binding domain is found in the rigor map. The large conformational change is quite striking (compare to Fig. 3), giving the impression that the filament has reversed polarity. C, catalytic domain; 1, 2, and 3 indicate the three light chains; LB, the lipid-binding domain.

Figure 4

3D model of BBM-I. This model was built by fitting the x-ray structures of the myosin catalytic domain and the ELC of skeletal muscle myosin into the EM density map (magenta wire cage) for BBM-I. At the COOH-terminal end of the molecule, an extra density is found which can be attributed to the basic, lipid-binding domain. In addition to the catalytic domain, the three light chains (LC1, LC2, and LC3) and the lipid-binding domain (LB) are indicated. The Cα backbones of the catalytic domain and the three light chains are displayed in alternating yellow and white for clarity, and the HC helix is shown in green.

Figure 5

Comparison of the BBM-I and skeletal muscle S1 light chain– binding domains. (a) Fit of the myosin catalytic domain (yellow), the BBM-I LCBD helix (green), and the skeletal muscle S1 light chain–binding domain helix (white). There is clearly a difference in the position of the LCBD between the two myosins, indicating a different exit point of the long α-helix from the catalytic domain. (b) A stereo pair of the backbones shown in a, rotated ∼90° about the filament axis. The EM density has been omitted for clarity. The S1 helix also appears displaced laterally relative to the BBM-I helix. (c) A view of the carbon backbones, looking down the filament axis, rotated 90° about the horizontal. Again, the EM density was omitted for clarity.

Figure 8

Comparison of the ADP and rigor 3D maps. (a) Direct comparison of the rigor and ADP maps clearly reveals the large swing of the BBM-I LCBD. In addition to the swing, there also appears to be a 20–30° rotation of the LCBD about its long axis. The dotted and solid red bars represent the orientation of the ADP (left) and rigor (right) LCBDs, respectively. (b) A superposition of the two x-ray fits for direct comparison of the two conformations. The common catalytic domain is in dark blue, the ADP LCBD is in cyan, and the rigor LCBD is in magenta. Shown in black are five monomers of the actin filament. This view highlights the magnitude of the BBM-I tail swing, as well as the fact that the tail moves as a rigid unit.

Figure 7

3D model of BBM-I in the presence of 1 mM MgADP. The EM density is shown in magenta, the myosin catalytic domain is in yellow, the CaM light chains are in cyan, and the HC helix is shown in green. The results of the rigor fitting (Fig. 4) were rotated as a rigid unit to obtain the fit to the ADP map. The light chain–binding domain was rotated by ∼31° with respect to the rigor orientation, in addition to a rotation of 20 to 30° about its long axis.

Figure 9

Structural interpretation of the events occurring during force production. This is a graphic representation of our interpretation of the force- generating cycle, as presented in the discussion. The transition from AM′′ADP.P_i to AM′.ADP.P_i is based on mechanical studies performed on skeletal muscle fibers. The representation of this as an additional rotation of the LCBD is purely hypothetical (emphasized by ?), as no structural rearrangements have yet been demonstrated. In this scheme, each force-producing step can be considered to be a two-step process: a force-generating isomerization followed by a ligand release step. This diagram also emphasizes that force is not directly coupled to products release, but occurs in separate steps. The steps in the cycle leading from rigor to the weakly bound states have been omitted, as indicated by the dotted arrow.