DnaJB6 is a RanGTP-regulated protein involved in dynein-dependent microtubule organization during mitosis

Bipolar spindle organization is essential for the faithful segregation of chromosomes during cell division. This organization relies on the collective activities of motor proteins. The minus-end directed dynein motor complex generates spindle inward forces and plays a major role in spindle pole focusing. The dynactin complex regulates many dynein functions increasing its processivity and force production. Here we show that DnaJB6 is a novel RanGTP regulated protein. It interacts with dynactin p150Glued in a RanGTP-dependent manner specifically in M-phase and promotes spindle pole focusing and dynein force generation. Our data suggest a novel mechanism by which RanGTP regulates dynein activity during M-phase. Summary statement We describe DnaJB6 as a novel RanGTP-regulated protein important for spindle assembly. Our data suggest that RanGTP regulates dynein-dependent inward spindle force generation and pole focusing through DnaJB6

. This process 64 involves NuMA that recruits the dynein--dynactin complex to the microtubule RanGTP regulated proteins have been identified and shown to play essential 77 roles in spindle assembly (Cavazza and Vernos, 2015). Recently we used a 78 mass spectrometry based proteomic approach to identify the full proteome 79 associated with RanGTP induced microtubules in Xenopus egg extracts and 80 validated a few selected proteins with no reported function in mitosis (Rosas--81 Salvans et al., 2018). One of them DnaJB6 (also known as MRJ, HSJ2 or MSJ1) 82 had also been identified in mitotic spindle proteomes (Bonner et al., 2011;83 Rao et al., 2016;Sauer et al., 2005) although its function in mitosis had not 84 been addressed. DnaJB6 is an HSP40 family protein with two alternatively 85 spliced isoforms (in human). These two isoforms share the first 239 amino 86 acids but differ at their C--terminus. The short isoform (DnaJB6--S, 241aa) is 87 cytoplasmic whereas the long isoform (DnaJB6--L, 326aa) that has a nuclear 88 localization signal at its C--terminus (Mitra et al., 2008)  breast cancer (Meng et al., 2016). Interestingly, DnaJB6 is highly expressed in 96 human testis, ovary, liver and placenta (Seki et al., 1999) and its expression is 97 increased in mitosis in HeLa cells (Dey et al., 2009;Seki et al., 1999). We 98 previously showed that DnaJB6 localizes to the spindle poles and its silencing 99

DnaJB6 is essential for microtubule organization during mitosis 143
To address the role of DnaJB6 in spindle assembly, we first examined more 144 carefully the spindle morphology defects associated with DnaJB6 silencing. 145 The major phenotypes were associated with microtubule organization 146 defects such as multipolar spindles (Fig 2B) or the presence of ectopic 147 microtubule clusters (Fig S1B). Indeed DnaJB6 silenced cells showed a 148 significant increase of multipolar spindles compared to control cells (30.5% 149 versus 5.6%) (Fig 2B). This phenotype was not associated to the 150 amplification of the number of centrosomes in the silenced cells as monitored 151 with anti--centrin antibodies. Indeed, 88% of the multipolar spindle 152 assembled in DnaJB6 silenced cells had only two centrin positive poles 153 whereas 78% of the few multipolar spindles assembled in control cells had 154 centrin signal at all the poles (Fig 2C,D). These results suggested that DnaJB6 155 is required for microtubule organization during mitosis. To gain further 156 support for this idea we monitored microtubule organization in control and 157 DnaJB6 silenced cells undergoing regrowth after nocodazole washout. Cells 158 were fixed at different time points after nocodazole washout and processed 159 for immunofluorescence ( Fig 3A). The efficiency of bipolar spindle 160 organization was measured by quantifying the number of microtubule asters 161 at the different time points (Fig S1C). In most control cells, the number of 162 microtubule asters was initially high due to the activity of the centrosomal 163 and chromosomal pathways (Meunier and Vernos, 2011) and reduced over 164 time to finally only two microtubule asters defining the poles of the bipolar 165 spindle. In contrast, the percentage of cells with only two microtubule asters 166 was significantly reduced in DnaJB6 silenced cells at all the time points 167 examined ( Fig 3B) and the difference with the control increased with time. At 168 60 minutes after nocodazole washout 57.7% of the control cells had two 169 asters organized into bipolar spindle, whereas this percentage was reduced 170 to 24.8% in DnaJB6 silenced cells (Fig 3C). 171 The strong defects in microtubule aster clustering and the presence of 172 centrin--negative extra spindle poles suggested that DnaJB6 could also be 173 required for spindle pole focusing. We therefore decided to look more 174 carefully at the morphology of the poles of bipolar spindles in the DnaJB6 7 silenced cells (Fig  3D). Indeed we found that 29% of the poles of the bipolar 176 spindles were not tightly focused in these cells whereas this was only 177 observed in 14% of the bipolar spindles assembled in control cells (Fig 3D). 178 Altogether, these results strongly suggest that DnaJB6 plays a role in 179 microtubule organization in mitosis, more specifically in microtubule aster 180 clustering and spindle pole focusing. 181 182

DnaJB6 has a direct role in microtubule organization in M--phase 183
Since DnaJB6 is a co--chaperon, its absence could potentially lead to an 184 accumulation of protein folding defects that could be related to the 185 phenotypes observed during mitosis. To determine whether the phenotypes 186 we observed in DnaJB6 silenced cells could be due to indirect effects on 187 protein stability we used the Xenopus egg extract system. Interestingly, p150Glued also accumulated to the spindle poles in DnaJB6 236 silenced cells (Fig 5F and S3A). Consistently with the results in egg extract, no 237 differences on the general levels of p150glued were detected by western blot 238 analysis of control and DnaJB6 silenced cell lysates (Fig 5E). Since it has been 239 recently shown that p150Glued is recruited to the microtubule minus--ends 240 by NuMA in mitosis (Hueschen et al., 2017), we then checked if NuMA localization was altered in DnaJB6--silenced cells. Immunofluorescence 242 analysis showed that NuMA was also significantly enriched at the spindle 243 poles in DnaJB6 silenced cells (Fig 5G and S3B). 244 Altogether, our data indicates that DnaJB6 interacts with p150Glued in a 245 RanGTP--dependent manner during mitosis. Moreover, this interaction may 246 be important for p150Glued and its targeting partner NuMa localization to 247 the spindle poles. 248 249

DnaJB6 increases the stability of dynactin specifically in mitosis 250
Since DnaJB6 is a co--chaperon, we reasoned that its interaction with 251 the regulation of inward force production during spindle assembly ( Fig  S4A,  294 B). Consistently we found a small but significant increase of the spindle 295 length in DnaJB6 silenced HeLa cells (10.4μm) compared to control cells 296 (9.9μm) ( Fig S4C). 297 Altogether, our results suggest that DnaJB6 is required for dynein--dependent 298 spindle inward forces generation during mitosis. 299 300

301
Here we showed that the nuclear protein DnaJB6 co--pulls down dynactin 302 purchased from Nasco and used at an age between 1 and 3 years. CSF 462 arrested egg extracts were prepared and used to performed cycled spindle 463 assembly assays as previously described (Desai et al., 1999). For anti--DnaJB6 immunostaining cells were pre--extracted for 6 seconds in 499 BRB80--1X, 0,5% triton and 1mM DSP. Cells were then fixed in PFA 4% for 7 500 minutes and then processed for IF as described.
Spindles assembled in egg extracts were centrifuged onto coverslips as 502 previously described, fixed in --20ºC methanol for 10 minutes and processed 503 for IF following the above protocol.