Co-translational protein targeting facilitates centrosomal recruitment of PCNT during centrosome maturation

As microtubule-organizing centers of animal cells, centrosomes guide the formation of the bipolar spindle that segregates chromosomes during mitosis. At mitosis onset, centrosomes maximize microtubule-organizing activity by rapidly expanding the pericentriolar material (PCM). This process is in part driven by the large PCM protein pericentrin (PCNT), as its level increases at the PCM and helps recruit additional PCM components. However, the mechanism underlying the timely centrosomal enrichment of PCNT remains unclear. Here we show that PCNT is delivered co-translationally to centrosomes during early mitosis by cytoplasmic dynein, as evidenced by centrosomal enrichment of PCNT mRNA, its translation near the centrosome, and requirement of intact polysomes for PCNT mRNA localization. Additionally, the microtubule minus-end regulator, ASPM, is also targeted co-translationally to mitotic spindle poles. Together, these findings suggest that co-translational targeting of cytoplasmic proteins to specific subcellular destinations may be a generalized protein targeting mechanism.


193
To further test the dependency of centrosomal enrichment of PCNT mRNA on intact, actively 194 translating polysomes, we treated the cultured cells with either emetine, which stabilizes 195 polysomes by irreversibly binding the ribosomal 40S subunit and thus "freezing" translation 196 during elongation (Jimenez, Carrasco, & Vazquez, 1977), or harringtonine, which disrupts 197 polysomes by blocking the initiation step of translation while allowing downstream ribosomes to 10 run off from the mRNA (Huang, 1975). We found that PCNT mRNA localization patterns in 199 emetine-and harringtonine-treated cells resembled those observed in vehicle-(control) and supplement 2). Because the microtubule nucleation activity is often positively correlated with the 208 centrosome size, we speculated that centrosomal enrichment of pericentrin mRNA/polysomes 209 might be a microtubule-dependent process. We thus tested if the localization of pericentrin 210 mRNA would be perturbed when microtubules were depolymerized. We found that in both 211 zebrafish and cultured human cells, pcnt/PCNT mRNA was no longer enriched around the 212 centrosome upon microtubule depolymerization ( Figure 4C and D). In contrast, a cytochalasin B 213 treatment, which disrupts the actin cytoskeleton, had no effect on the centrosomal enrichment of 214 PCNT mRNA (Figure 4-figure supplement 3). These results suggest that microtubules, but not 215 actin filaments, serve as "tracks" on which pericentrin mRNA/polysomes are transported.

217
Given that cytoplasmic dynein is a common minus-end-directed, microtubule-based motor that 218 transports cargo toward the microtubule minus end (i.e., toward the centrosome), we next tested 219 whether centrosomal localization of PCNT mRNA is a dynein-dependent process. We treated

229
To determine the functional significance of translation of centrosomally localized PCNT mRNA 230 during early mitosis, we compared centrosomal PCNT levels shortly before and after mitotic 231 entry (i.e., late G2 vs. early M phase). We arrested cultured human cells from progression out of    Using this strategy, we found that approximately 2-fold more PCNT proteins were incorporated 248 into the centrosomes in early mitotic cells as compared to late G2 cells ( Figure 5A). Importantly, 249 the numbers of PCNT mRNA did not significantly differ between late G2 and early M phases, 250 even though there was an approximately 4-fold increase from G1 to late G2 phases ( Figure 5B).

251
Therefore, these results indicate that the increase in centrosomal PCNT protein levels when 252 cells progress from G2 to M phases (e.g., the 25-minute period after RO-3306 washout) is due 253 to upregulation of translation and not to altered mRNA abundance.

255
To independently assess the impact of translation during early mitosis on PCNT incorporation 256 into the centrosomes, we disrupted this process by pulsing the RO-3306 synchronized cells with 257 puromycin to inhibit translation for two minutes, followed by immediate fixation and anti-PCNT

274
To determine if the cell uses a similar co-translational targeting strategy to target other large 275 proteins to the centrosome, we examined the distribution of CEP192, CDK5RAP2/CEP215, and 276 ASPM mRNA in cultured human cells. We found that while CEP192 and CEP215 mRNA did not

290
Here we report that PCNT protein is delivered co-translationally to the centrosome during to speculate that co-translational targeting of PCNT (and its orthologous proteins) to the 300 centrosome is an evolutionarily conserved process. In addition to PCNT, the cell appears to use 301 a similar co-translational targeting strategy to deliver the large microtubule minus-end 302 regulator/spindle-pole focusing factor, ASPM, to mitotic spindle poles, as ASPM mRNA is 303 strongly enriched at mitotic spindle poles in a translation-dependent manner, concomitantly with 304 the ASPM protein level reaching its maximum at the same place. We suspect that co-

N a s c e n t P C N T ( P C N T N + / C -) F u l ll e n g t h P C N T ( P C N T N + / C + )
Multiple ribosomes on a large PCNT mRNA

Control + Puromycin
Merge PCNT-C-term PCNT-N-term PCNT mRNA   Representative confocal images are shown for each condition. A "fire" lookup table (LUT) was used to show PCNT signal intensities. The sum intensity of anti-PCNT signals from both centrosomes of each cell was measured and plotted. (B) Numbers of PCNT mRNA at different cell cycle stages of Centrin-GFP RPE-1 cells were determined by PCNT smFISH. S phase/early G2 cells were identified by EdU labeling for 30 minutes. (C) HeLa cells were treated with vehicle control or 300 µM puromycin for 2 minutes before anti-PCNT immunostaining. The sum intensity of anti-PCNT signals from both centrosomes of each prophase or prometaphase cell was measured and plotted. Data are represented as mean ± 95% CI. "n" indicates the total number of cells analyzed from at least two independent experiments. p ' t-test (two-tailed). Scale bar: 10 µm. Subsequently, this nascent polypeptide-dynein interaction allows the entire polysome, which is still actively translating PCNT mRNA, to be transported along the microtubule toward the centrosome. This co-translational targeting mechanism may maximize efficiency of PCNT production and delivery to the centrosome, prevent ectopic accumulation of PCNT outside of centrosomes, and/or facilitate integration of PCNT into the expanding PCM during early mitosis.