A conserved CaM- and radial spoke associated complex mediates regulation of flagellar dynein activity

Abstract

For virtually all cilia and eukaryotic flagella, the second messengers calcium and cyclic adenosine monophosphate are implicated in modulating dynein- driven microtubule sliding to regulate beating. Calmodulin (CaM) localizes to the axoneme and is a key calcium sensor involved in regulating motility. Using immunoprecipitation and mass spectrometry, we identify members of a CaM-containing complex that are involved in regulating dynein activity. This complex includes flagellar-associated protein 91 (FAP91), which shares considerable sequence similarity to AAT-1, a protein originally identified in testis as an A-kinase anchor protein (AKAP)- binding protein. FAP91 directly interacts with radial spoke protein 3 (an AKAP), which is located at the base of the spoke. In a microtubule sliding assay, the addition of antibodies generated against FAP91 to mutant axonemes with reduced dynein activity restores dynein activity to wild-type levels. These combined results indicate that the CaM- and spoke-associated complex mediates regulatory signals between the radial spokes and dynein arms.

Figure 1.

Silver-stained gels of immunoprecipitation experiments using anti-CaM antibodies and extracts isolated from wild-type, radial spokeless (pf14), and central pairless (pf18) axonemes as well as axonemes lacking both the outer dynein arms and inner arm isoform I1 (pf28pf30). The precipitates were resolved on either a 7% (top) or 12% (bottom) acrylamide gel. Central apparatus proteins are indicated by dots; spoke proteins are indicated by triangles. CaM-IP2, -IP3, and -IP4 are indicated by asterisks. The darkly stained bands are the heavy (HC) and light (LC) chains from the antibodies. CaM-IP2, -IP3, and -IP4 are precipitated from all of the mutant extracts tested, indicating that these polypeptides are not components of the radial spokes, central apparatus, outer dynein arms, or inner dynein arm I1.

Figure 2.

CaM-IP2, -IP3, and -IP4 form a single complex that is associated with the radial spokes. (A) Western blots of axonemes (8 μg per lane) probed with antibodies generated against CaM-IP2 and -IP3. Both polypeptides are present in all axonemes in approximately equal amounts. (B) Silver- stained gel (top) and corresponding Western blots (middle and bottom) of immunoprecipitation experiments using anti–CaM-IP2 antibodies and wild-type, radial spokeless (pf14), and central pairless (pf18) axonemal extracts. CaM-IP2 antibodies precipitate the CaM-IP2, -IP3, and -IP4 complex (asterisks) as well as RSPs (triangles). Blots were probed with antibodies generated against CaM-IP2, -IP3, RSP2, or RSP3. These results indicate that CaM-IP2, -IP3, and -IP4 form a single complex that is associated with the radial spokes.

Figure 3.

CaM-IP2 is associated with RSP3. Western blots of extracts isolated from wild-type (A), spoke headless (B; pf17), and spokeless (C; pf14) axonemes fractionated on sucrose gradients. Blots were probed with antibodies generated against CaM-IP2, -IP3, RSP2, RSP3, or CaM. In extracts isolated from both wild-type or pf17 axonemes, CaM-IP2 and -IP3 cosediment with spoke stalk components. In extracts isolated from both pf17 and pf14 mutant axonemes, the sedimentation value for the CaM-IP2 complex is considerably lower than that in wild type. (D) Gel overlay of expressed RSP3. Bacterial extracts expressing fragments of CaM-IP2 (37 kD), -IP3 (32 kD), -IP4 (60 kD), and RSP3-GST (110 kD) were resolved on polyacrylamide gels and either transferred to nitrocellulose (left) or stained with Coomassie blue (right). The membrane (left) was incubated with expressed RSP3 and probed with antibodies generated against RSP3. These results indicate that RSP3 binds to CaM-IP2.

Figure 4.

Binding of CaM-IP2, -IP3, and -IP4 to CaM is calcium sensitive. (A) Silver-stained gel of CaM immunoprecipitation from radial spokeless axonemal extracts (pf14) followed by treatment with CaCl₂ (top). After precipitation (CaM-IP), the protein A beads were washed twice with 2 mM CaCl₂ (Ca²⁺1 and Ca²⁺2), and the resulting extract was loaded onto the gel. Proteins remaining (lane R) associated with the beads after the CaCl₂ wash were eluted with sample buffer. CaM-IP2, -IP3, and -IP4 (asterisks) are extracted from the beads. In contrast, the central pair proteins (dots) are not extracted with CaCl₂. Western blots using anti-CaM antibodies (bottom) reveal that CaCl₂ extracts very little CaM from the beads. These results indicate that binding of CaM-IP2, -IP3, and -IP4 to CaM is calcium sensitive. (B) Silver-stained gel of immunoprecipitation experiments performed in 1 mM CaCl₂ using anti–CaM-IP2 antibodies and wild-type, spokeless (pf14), and central pairless (pf18) axonemal extracts. The CaM-IP2 antibodies precipitate the CaM-IP2, -IP3, and -IP4 complex (asterisks) and spokes (triangles), indicating that they do not dissociate from each other or the spokes in high calcium buffer conditions.

Figure 5.

Association of the CaM-IP2 complex with CaM is disrupted in DRC mutants. (A) Silver-stained gel of immunoprecipitates using either anti-CaM or anti–CaM-IP2 antibodies and axonemal extracts isolated from wild-type (wt), spokeless (pf14), or DRC mutant (pf2 and pf3) axonemes. Compared with wild-type and pf14 precipitates, the CaM-IP2 complex is substantially reduced in precipitates of pf2 and pf3 extracts using the anti-CaM antibodies. However, immunoprecipitation using the anti–CaM-IP2 antibodies indicates that relatively equal amounts of CaM-IP2, -IP3, and -IP4 (asterisks) are precipitated in wild-type, pf2, and pf3 axonemes. (B) Western blot of isolated axonemes using anti-CaM antibodies. The pf2 and pf3 mutant axonemes have wild-type levels of CaM. These combined results indicate that association of the CaM-IP2 complex with CaM is disrupted in DRC mutants.

Figure 6.

CaM-IP2 modulates dynein-driven microtubule sliding. (A) Microtubule sliding velocities of wild-type and pf18 axonemes. The CaM-IP2 (IP2) antibodies have no effect on the sliding velocity of wild-type axonemes. In contrast, sliding velocities of pf18 axonemes increase upon addition of the CaM-IP2 antibody. This increase is not observed in the presence of C1a-32 (C1a; Wargo et al., 2005) or CaM-IP3 (IP3) antibodies. (B) Microtubule sliding velocities of pf18 axonemes incubated with varying concentrations of CaM-IP2 antibody. The increase in dynein activity is dose dependent, with a maximal increase in velocity at 0.2 μM CaM-IP2 antibody. (C) Microtubule sliding velocities of mutants lacking the outer dynein arms (pf28), outer dynein arms and the central apparatus (pf18pf28), and outer dynein arms and spokes (pf14pf28) in the presence (IP2; patterned bars) and absence (black bars) of the CaM-IP2 antibody. (D) Sliding velocities of a mutant lacking the central pair and I2 inner arm heavy chains (pf18ida4) increase upon the addition of anti–CaM-IP2 antibodies. Velocities of a mutant lacking the central apparatus and the I1 inner dynein arm (pf18ida1) are significantly higher than velocities of pf18 (P < 0.001 by t test). However, sliding velocities of pf18ida1 axonemes incubated with CaM-IP2 antibodies are significantly lower than those of pf18 incubated with CaM-IP2 antibodies (P < 0.001 by t test). Sliding velocities of a radial spokeless mutant (pf14) increase upon the addition of anti–CaM-IP2 antibodies (IP2). However, the addition of CaM-IP2 antibodies to axonemes that lack the radial spokes and inner dynein arm I1 (pf14pf30) have no effect on sliding velocity. All bars represent the mean of >70 measurements ± SEM (error bars) from three or more independent experiments. (E) Summary of microtubule sliding experiments. The CaM-IP2, -IP3, and -IP4 complex (purple) is localized to the base of the spoke (blue).

Figure 7.

Western blots of isolated axonemes (10 μg/lane) probed with antibodies generated against the I1 intermediate chain IC138. Intermediate chain IC138, left. As indicated, some axoneme samples were treated with either anti–CaM-IP2 or anti–C1a-32 antibodies (C1a; Wargo et al., 2005) in the presence or absence of calf intestinal alkaline phosphatase (CIP). In the absence of calf intestinal alkaline phosphatase treatment, IC138 appears hyperphosphorylated in both central pairless (pf18) and spokeless (pf14) axonemes compared with wild type. No observable difference in phosphorylation is detected after the addition of CaM-IP2 antibodies. The 120–150-kD molecular mass range of Coomassie blue–stained gels of the same samples are shown as a loading control (right).