(A) Chlamydomonas cell geometry. Flagella (f) extend from the centrioles (white), which are located apically and are attached to the nucleus (yellow) by centrin-containing fibers. The pyrenoid (p; blue), a starch-containing structure, is located basally and is embedded in a cup-like mass of chloroplast (green). The eyespot (e; red), the light-sensing organelle, is located laterally at a reproducible angle relative to the centrioles.
(B) DIC image of a wt Chlamydomonas cell. The pyrenoid (p), eyespot (e), and flagella (f) are indicated. All DIC images are sections through full 3D datasets.
(C) In asq1 cells, mother–daughter centriole pairs are randomly localized on the cell surface.
(D) In asq2 cells, centrioles are independently positioned on the cell surface and no longer found in pairs.
(E) Defining θ_centriole. A 3D vector reflecting the long axis of the cell is drawn from the center of mass (yellow circle) of the pyrenoid (blue) to the cellular center of mass (purple circle). θ_centriole is the angle between the vector defining the long axis of the cell and the vector from the cellular center of mass to each centriole (white). All angle measurements are made in three dimensions using 3D image datasets.
(F) The mean θ_centriole (black line) for wt cells is 20.5 ± 9.0° (n = 62),
(G) θ_centriole increases to 42.3 ± 21.3° for asq1 cells (n = 54). Angles are biased toward the top half of the cell, presumably because the lower half of the cell is occluded by the chloroplast, which is localized in the basal half of the cell body (see [J]) and closely apposed to the plasma membrane, thus reducing access of basal bodies to the cell surface.
(H) θ_centriole increases to 61.7 ± 32.3° for asq2 cells (n = 71).
(I) Defining θ_chloroplast. The cell center–pyrenoid axis is defined as described in (E). θ_chloroplast is defined as the angle between the vector defining the long axis of the cell and the vector from the cellular center of mass to each plastid genome (large green circle).
(J) The θ_chloroplast for wt cells is shown in green (mean = 112.1 ± 36.0°, n = 181). Each line represents the position of one plastid genome. The yellow-shaded area represents the area of the cell occupied by the pyrenoid. The non-180° edge of this shaded region indicates the mean position of the pyrenoid boundary (mean = 139.0 ± 14.4°, n = 90).

Images of fixed cells stained with DAPI and antibodies against centrin and acetylated tubulin (green) and Bld10p (red). DIC images are shown in the top panels, and fluorescence images of the same cells are shown below. All images are positioned so that the pyrenoid is located at the bottom.
(A) wt cells have two centrioles located together at the apical side of the cell. One pro-centriole can be seen in the foreground of this image (red, marked by antibodies recognizing Bld10p). The other pro-centriole is occluded from view by the centriolar pair.
(B and C) asq1 cells have two centrioles that are positioned together at random locations on the cell surface.
(D) In asq2 cells, centrioles can be found at locations independent of one another. In this cell, both centrioles appear to be mispositioned.
(E) In this asq2 cell, one centriole along with its pro-centriole (marked by Bld10p staining in red) is found at the correct apical location. Another mispositioned centriole is found on the left, shifted off the long axis.

(A) DIC image of a wt cell. wt cells have two flagella located at the apical side of the cells.
(B and C) DIC images of asq2 cells. asq2 cells have variable numbers of centrioles and therefore make variable numbers of flagella. Some centrioles are randomly localized, whereas some are in the correct apical position.
(D) Distribution of flagellar number in asq2 cells is reminiscent of the vfl (variable flagellar number) phenotype. asq2 cells (white) have a mean of 1.46 ± 1.1 flagella per cell (n = 1,274). This distribution is similar to that of vfl2 cells (green; mean = 1.33 ± 1.05, n = 593) and vfl3 cells (purple; mean = 1.12 ± 1.8, n = 466), but is in contrast to wt cells, which make two flagella (black; mean = 1.94 ± 0.34, n =1,005).
(E) vfl2 cells, previously identified as defective in centriole segregation, have a mean θ_centriole of 55.2 ± 28.8° (n = 64).
(F) vfl3 cells, defective in mother–daughter centriole cohesion, have a mean θ_centriole of 59.4 ± 35.2° (n = 90).

(A) Electron micrograph showing electron-dense connecting fibers (distal striated fiber, denoted by arrow) joining mother and daughter centrioles.
(B) Electron micrographs of asq2 cell showing that centriole connecting fibers are missing.
(C) Model for centriole positioning by mother centriole. In wt cells (left box), two centrioles are localized to the apical pole. These centrioles are connected by electron-dense connecting fibers (see [A], arrow). During duplication, each centriole will serve as a mother (white) to give rise to a daughter centriole (blue). New connections will form between each new mother–daughter pair. One mother–daughter centriole pair will be segregated to each cell following cell division. Each centriole will give rise to a flagellum, resulting in two cells with two centrioles and two flagella. In asq2 cells (right box), centrioles are no longer connected (see [B], arrow). As in wt cells, each centriole will serve as a mother (white) to give rise to a daughter centriole (blue). However, because mother and daughter centrioles are no longer connected, centrioles will not segregate properly following mitosis, resulting in cells with variable numbers of centrioles. Among the centrioles that are distributed between cells, there will be a mix of mother and daughter centrioles. If the mother centrioles contain the necessary mark (purple) that allows them to find their proper subcellular location, whereas daughter centrioles are naive and unable to track to the correct place in the cell, then cells will have a population of properly positioned mother centrioles and a population of randomly localized daughter centrioles.

(A) If mother centrioles (white) are properly positioned, then the distance between flagellated centrioles in asq2uni1 mutant cells should be much smaller and less variable than that of single-mutant cells.
(B) The distance between flagellated centrioles in biflagellate asq2uni1 cells is much smaller and less variable (mean = 0.89 ± 0.36 μm, n = 85) than in biflagellate asq2 cells (mean = 1.48 ± 0.85 μm, n =88). This difference is highly significant (one-tailed t-test, p < 1.23 e⁻⁸)
(C) If mother centrioles (white) are properly localized, then the θ_centriole should be much smaller and less variable for flagellated centrioles in asq2uni1 cells than for asq2 cells.
(D) θ_centriole for flagellated centrioles in asq2uni1 cells is significantly (one-tailed t-test p < 2.02 e⁻¹⁰) smaller (green lines, mean θ_centriole = 32.4 ± 13.1°, n = 60) and less variable than in asq2 cells (grey lines, mean θ_centriole = 61.7 ± 32.4°, n = 71).
(E) Flagellated mother centrioles (m; white arrow) are properly localized in asq2uni1 cells, whereas unflagellated daughter centrioles (d; blue arrow) are not. Cells are labeled with anti-acetylated tubulin and centrin antibody (green), anti-Bld10p antibody specific for centrioles (red) and DAPI (blue). Misplacement of nonflagellated daughter centrioles in vfl2uni1 indicates that uni1 does not simply suppress the centriole positioning phenotype of vfl2.

(A) In wt cells, centrioles (red) are attached to the nucleus (blue) via centrin fibers (red). Both the centrioles (C) and nucleus (N) are properly localized near the apical part of the cell. Other plastid genomes are visible with DAPI staining (smaller blue dots).
(B) asq1 cell showing centrioles (red) and nucleus (blue) mislocalize together.
(C) When centrioles are uncoupled from the nucleus in vfl2uni1 cells, flagellated (green) mother centrioles (red) are properly localized to the apical side of the cell, whereas the nucleus (blue) can visit variable positions.
(D) Mean θ_centriole for mother centrioles in vfl2uni1 cells is 24.9 ± 14.7° (n =49, orange lines), which is significantly less than θ_centriole for vfl2 cells (grey lines, one-tailed t-test p < 2.71 e⁻¹¹), but not significantly different from wt.
(E) wt cells have a mean θ_nucleus of 15.5 ± 8.1° (n = 58). θ_nucleus was determined by measuring the angle between the vector defining the long axis of the cell and a vector from the nuclear center of mass to the pyrenoid center of mass.
(F) vfl2uni1 cells have a significantly higher (one-tailed t-test, p < 2.9 e⁻⁶) mean θ_nucleus of 25.0 ± 11.8° (n = 49) compared to wt.
(G) θ_nucleus and θ_centriole are correlated in wt cells, indicating that the position of the two organelles is coupled.
(H) When the nucleus is detached from the centrioles in vfl2uni1 cells, the nuclear position no longer correlates to centriolar position. Scatter plot visually shows loss of correlation between θ_centriole and θ_nucleus. Points are color coded into two groups of cells, those with a nucleus whose position is within the correct wt range (defined as θ_nucleus less than one standard deviation from wt mean, and plotted in orange) and those with a nucleus whose position is incorrect (defined as θ_nucleus more than one standard deviation from wt mean, and plotted in gray). The two groups of points classified in this manner span the same range of values for θ_centriole, further supporting a lack of correlation between nuclear and centriolar position when the nucleus is detached from the centriole. Inset: the mean θ_centriole (mean θ_centriole = 25.7 ± 11.3°, gray bar, n = 29) of cells with an improperly positioned nucleus (NI) is indistinguishable from the mean θ_centriole (mean θ_centriole = 23.7 ± 18.8°, orange bar, n = 20) of cells with a correctly positioned nucleus (NC). This shows that the mother centrioles can still attain the correct localization regardless of nuclear position.

(A) Centrioles contain nine triplet microtubule blades (yellow) arranged around a central cartwheel that sits on an amorphous disc structure (blue). At the most distal ends of centrioles in the region just proximal to the site of flagellar assembly, transition fibers are assembled (black ellipses) near the apical membrane. bld1 mutant cells have normal centrioles and transition fibers, but are defective in flagellar assembly due to a loss of intraflagellar transport.
(B) bld2 cells are defective in centriole assembly and lack the B- and C-tubule of the triplet microtubule blades. As a result, the distal portion of bld2 centrioles is missing.
(C) bld10 cells lack centriolar microtubules and have just the very proximal portion of the centriolar structure.
(D) bld1 centrioles (green represents centrin/acetylated tubulin labeling) localize to the apical membrane.
(E and F) bld2 and bld10 cells have mispositioned centrioles (green) that appear in the cell interior. They are still found closely apposed to the nucleus (blue).
(G) bld1 cells have normally positioned centrioles (mean θ_centriole = 19.8 ± 8.0°, n =52) despite their lack of flagella. This demonstrates that neither flagella themselves, nor the intraflagellar transport machinery, is required for centriole positioning.
(H) bld2 centrioles lack the distal region and are mispositioned (mean θ_centriole = 45.9 ± 26.9°, n = 44).
(I) bld10 centrioles are also mispositioned (mean θ_centriole = 40.2 ± 30.8°, n = 46).