Space-dependent formation of central pair microtubules and their interactions with radial spokes

Abstract

Cilia and flagella contain nine outer doublet microtubules and a pair of central microtubules. The central pair of microtubules (CP) is important for cilia/flagella beating, as clearly shown by primary ciliary dyskinesia resulting from the loss of the CP. The CP is thought to regulate axonemal dyneins through interaction with radial spokes (RSs). However, the nature of the CP-RS interaction is poorly understood. Here we examine the appearance of CPs in the axonemes of a Chlamydomonas mutant, bld12, which produces axonemes with 8 to 11 outer-doublets. Most of its 8-doublet axonemes lack CPs. However, in the double mutant of bld12 and pf14, a mutant lacking the RS, most 8-doublet axonemes contain the CP. Thus formation of the CP apparently depends on the internal space limited by the outer doublets and RSs. In 10- or 11-doublet axonemes, only 3-5 RSs are attached to the CP and the doublet arrangement is distorted most likely because the RSs attached to the CP pull the outer doublets toward the axonemal center. The CP orientation in the axonemes varies in double mutants formed between bld12 and mutants lacking particular CP projections. The mutant bld12 thus provides the first direct and visual information about the CP-RS interaction, as well as about the mechanism of CP formation.

Figure 1. Abnormal features of bld12 axonemes.

(A) Length distributions of wild type (gray) and bld12 (black) flagella. (B) Axonemes with 8, 9, 10, and 11 outer doublet microtubules. Bar, 100 nm. (C) Diameter and inter-doublet spacing in axonemes with various doublet numbers. Diameter increases with the doublet numbers, while the inter-doublet spacing is constant. The linear regression slope of the diameter is 22.14±1.69 nm (R² = 0.900). The difference between the inter-doublet spacing is not significant (ANOVA, P = 0.59).

Figure 2. CPs in RS-lacking axonemes.

Cross section images of bld12pf14 axonemes with 7 to 10 outer doublets are shown. The frequency of each pattern in the observed images is indicated in the parenthesis (n = 1,369, Table 1). All CPs have the normal polarity in the axoneme as judged by their CP projections. Bar, 100 nm.

Figure 3. Frequency of CP-RS association in different sectors of the CPs.

(A) The CP cross section was divided into 12 sectors by six lines including the one connecting the centers of the two central microtubules. The center of the consecutive spokeheads in contact with the CP was defined as the spokehead-interaction site. Major CP projections are designated . (B) A polar histogram representing the distribution of the spokehead-interaction sites on the CP in cross section images of 10- or 11-doublet axonemes of bld12 (n = 56). The length of each bar represents relative frequency to locate at the spokehead-interaction site. (C) Axoneme images that correspond to the two peaks in the histogram in (B). (D) Distribution of the CP region in contact with the wall of outer doublets in pf14 axonemes (n = 80). The histogram suggests that the C1 microtubule is located on the outer surface on the helical CP. (E) A cross section image that represents the peak distribution in the histogram in (D). The differences between the distributions of bld12 (B) and pf14 (D) are significant (χ² test, p<0.05).

Figure 4. Removal of CP projections manifests weak CP-spokehead association.

(A) Distributions of the spokehead-interaction sites on the CP in cross section images of 10- or 11-doublet axonemes of bld12pf6, bld12cpc1, bld12cpc1pf6, and bld12pf16 (n = 59, 49, 40, and 41), which lack specific CP projections. (B) Cross section images that represent the peaks in the histograms in (A). (C) Distributions of contact sites on the CP surface with the outer doublet wall in cross section images of pf14pf6, pf14cpc1, and pf14pf16 (n = 33, 69, and 46). The differences between the distributions of bld12pf6 and pf14pf6; between bld12cpc1 and pf14cpc1, and between bld12pf16 and pf14pf16, are significant (χ² test, p<0.05 for each of the three pairs). These distribution patterns are similar to the pattern in pf14.