Moss Kinesin-14 KCBP accelerates chromatid motility in anaphase

KCBP is a microtubule (MT) minus-end-directed kinesin widely conserved in plants. It was shown in Arabidopsis that KCBP controls trichome cell shape by orchestrating MT and actin cytoskeletons using its tail and motor domains. In contrast, the KCBP knockout (KO) line in the moss Physcomitrella patens showed a defect in nuclear and organelle positioning in apical stem cells. Moss KCBP is postulated to transport the nucleus and chloroplast via direct binding to their membranes, since it binds to and transports liposomes composed of phospholipids in vitro. However, domains required for cargo transport in vivo have not been mapped. Here, we performed a structure-function analysis of moss KCBP. We found that the FERM domain in the tail region, which is known to bind to lipids as well as other proteins, is essential for both nuclear and chloroplast positioning, whereas the proximal MyTH4 domain plays a supporting role in chloroplast transport. After anaphase but prior to nuclear envelope re-formation, KCBP accumulates on the chromosomes, in particular at the centromeric region in a FERM-dependent manner. In the KCBP knockout line, poleward chromosome motility in anaphase was reduced and lagging chromosomes occasionally appeared. These results suggest that KCBP binds to non-membranous naked chromosomes via an unidentified protein(s) for their transport. Finally, the liverwort orthologue of KCBP rescued the chromosome/chloroplast mis-positioning of the moss KCBP KO line, suggesting that the cargo transport function is conserved at least in bryophytes.


29
Microtubule (MT) and actin cytoskeletons function as tracks for various intracellular cargos 30 during transportation, such as large organelles in the cytoplasm. KCBP is a MT 31 minus-end-directed kinesin-14 family protein uniquely evolved in the plant lineage (Hamada, 32 2007;Reddy et al., 1996). In the moss Physcomitrella patens, KCBP has been shown to be 33 necessary for proper positioning of the nucleus and chloroplasts (Yamada et al., 2017). Several 34 data led us to propose a model in which KCBP directly binds to nuclear and chloroplast 35 membranes, walks along the MT towards minus ends, and transports those organelles to their 36 proper positions. First, intracellular motility of these organelles is MT-dependent, and not 37 actin-dependent (Miki et al., 2015). Second, the motility of the nucleus and chloroplasts is skewed 38 in the absence of KCBP (Yamada et al., 2017). Overall MT polarity in the observed knockout 39 (KO) cell supports the notion that MT minus-end-directed motility is specifically suppressed 40 without KCBP. Third, recombinant KCBP binds to and transports liposomes composed of acidic 41 phospholipids towards MT minus ends in vitro (Jonsson et al., 2015;Yamada et al., 2017). Fourth, minus-end-directed transporter of the nucleus during interphase (Yamada and Goshima, 2018). 1 The 'nuclear' enrichment of KCBP after chromosome segregation was concluded based solely on 2 its co-visualisation with histone-RFP, which is a chromosomal marker; therefore, it remains 3 unknown if the NE has been re-formed at the time of KCBP localisation and function. 4 Furthermore, nuclear migration is mediated by the NE-embedded LINC (Linker of Nucleoskeleton 5 and Cytoskeleton) or LINC-like complex in other animal, plant and yeast systems; LINC binds to 6 the tail region of motors (dynein, kinesin, myosin) or directly to actin filaments, thereby linking 7 the nucleus to cytoskeleton (Chang et al., 2015;Gundersen and Worman, 2013;Tamura et al., 8 2013). Second, it remains to be tested whether KCBP drives nuclear and chloroplast translocation 9 independently. There remains a possibility that mis-localised chloroplasts can affect nuclear  Finally, it has not been verified that walking ability is necessary for KCBP function in organelle 19 transport. An alternative possibility is that KCBP acts as a MT tether, while other motor(s) or MT 20 pushing/pulling force exerted at the cortex drives nuclear and chloroplast translocation. Indeed,

21
MT pushing/pulling is the mechanism of nuclear motility in fission and budding yeasts (Adames 22 and Cooper, 2000;Tran et al., 2001).

23
In this study, we addressed these questions through a structure-function analysis of KCBP. 24 We identified the FERM-C domain of moss KCBP as critical for cargo transport, which is a 25 dispensable domain of Arabidopsis KCBP for MT/actin organisation. In addition, our results 26 suggest that the naked chromatid, rather than the membranous nucleus, binds to KCBP as a cargo 27 in anaphase/telophase, ensuring rapid and robust chromosome segregation to daughter cells.

31
FERM-C domain is required for chromosome and chloroplast translocation 32 KCBP contains three recognisable domains that are known to bind to other molecules: 33 MyTH4, FERM, and motor domains. To identify the domain(s) responsible for nuclear and 34 chloroplast transport, we made six different truncation constructs of KCBPb, in which one or 35 multiple domains were deleted and Cerulean fluorescent protein was attached (Fig. 1A). We also 36 constructed a control full-length Cerulean-KCBPb and a 'rigor' mutant. The rigor mutant 37 possesses a mutation in a residue critical for its motor activity; hence, the motor binds to but does 38 not walk along MTs. The constructs were transformed into a moss line that expresses GFP-tubulin 39 and histoneH2B-mRFP and has had all of its KCBP paralogues deleted, and stable transgenic lines 40 were selected. The expression of each truncated protein was confirmed by immunoblotting ( Fig.   41 1B).

42
For each truncation, we selected two to four clonal lines and performed long-term time-lapse 43 imaging of the chromosome (histone-RFP) and chloroplast (autofluorescence) with a low 44 magnification lens ( Fig. 2A; lines used are #757-764 as listed in Table S1). Applying the fragments that lack FERM-C, which constitutes the lipid-and protein-binding interface, failed to 3 rescue either phenotype, suggesting that this domain is responsible for the organelle attachment of 4 the motor (Fig. 2D, E).  To correlate the function and localisation of KCBP, we observed chromosomal dynamics as 19 well as mutant KCBP localisation with high-resolution confocal microscopy ( Fig. 3). Since our 20 microscope setting could not prevent GFP signal leakage into the Cerulean channel, background 21 GFP-tubulin signal was always detectable when Cerulean-KCBPb fragments were imaged ( Fig. 3;

22
'Control' shows a cell that does not express Cerulean). Nevertheless, we were able to observe the 23 chromosomal enrichment of full-length Cerulean-KCBPb during telophase ( Fig. 3 'Full-length').

24
The motor-deleted fragment also showed signals on stalled chromosomes, indicating that the   Our previous study showed that, in vitro, KCBP binds to phospholipids in a manner 27 dependent on its tail region (Yamada et al., 2017). This led to the model that KCBP might 28 transport membranous organelle, such as nuclei and chloroplasts, through direct binding to their 29 membranes. Although this is still a viable model, our current study urges its reconsideration, at 30 least for the nucleus. We observed that KCBP accumulates on migrating chromosomes before the 31 completion of NE reformation, as indicated by the absence of CRWNa and SUN1. The 32 accumulation of KCBP, particularly near the centromeres, is also not readily explained if KCBP 33 directly binds to NE. The result essentially rejects the hypothesis that LINC acts as the nuclear 34 adaptor for the KCBP motor, unlike in other species. Furthermore, the rate of poleward motility of 35 chromatids reduced and lagging chromatids sometimes appeared during anaphase in the absence 36 of KCBP.

37
Thus, KCBP likely binds to chromosomal protein(s) at anaphase/telophase and transports 38 chromosomes that are not yet enveloped by lipid membranes (Fig. 7). The importance of the 39 FERM-C domain is not inconsistent with this model, since FERM-C can bind to proteins as well 40 as lipids. In a well-studied case, the crystal structure of the MyTH4-FERM domain attached to a 41 cargo protein was determined for myosins (Hirano et al., 2011;Wei et al., 2011;Wu et al., 2011). 42 Intriguingly, the cargo peptides mainly bind to FERM-N and FERM-C of myosin VII and myoxin 43 X, respectively. In the former, the MyTH4 domain also associates with a part of the cargo protein,  re-formation cannot be excluded entirely.

49
In contrast to the nucleus/chromosomes, we could not gain much insight into the molecular 1 mechanism of chloroplast transport in this study. The major obstacle was that we could not 2 observe the localisation of KCBP on the chloroplasts, since the autofluorescence of this organelle 3 was overwhelming. However, the essentiality of FERM-C and motor activity and the importance 4 of MyTH4 suggest that KCBP recognises the proteins and/or lipid surface of chloroplasts and 5 transports them to the MT minus-ends, similar to anaphase/telophase chromosomes. Identifying 6 the interacting partner of KCBP at the chloroplast surface could be an interesting future research 7 avenue.

27
This is highly analogous to what we observed for KCBP in the present study (Fig. 7). More

35
Moss culture and transformation 36 Moss lines used in this study are listed in Table S1; all lines originated from the Physcomitrella patens 37 Gransden2004 strain. Methodologies of moss culture, transformation, and transgenic line selection have 38 been thoroughly described in previous studies (Yamada et al., 2016). Briefly, cells were cultured on BCD 39 agar medium for imaging. Transformation was performed by the standard PEG-mediated method and stable 40 lines were selected with antibiotics. The mCherry gene was inserted into the C-termini of CRWNa and 41 SUN1 via homologous recombination. Citrine and mCherry tag insertion were confirmed by PCR. 42 Integration of Cerulean-KCBP truncation constructs was confirmed by PCR followed by immunoblotting.

44
Plasmid construction 45 Plasmids and primers for plasmid construction used in this study are listed in Table S2. To generate 46 truncation/rescue plasmids, KCBP sequences were amplified from a cDNA library and ligated into the 47 pENTR/D-TOPO vector containing Cerulean sequences, followed by a Gateway LR reaction into a 48 pMN601 vector that contains the EF1α promoter, nourseothricin-resistance cassette, and PTA1 sequences 49 designated for homologous recombination-based integration (Miki et al., 2016). Arabidopsis and 50 Marchantia KCBP sequences were similarly cloned into a pTM153 vector that contains the EF1α promoter, 1 blasticidin-resistance cassette, and PTA1 sequences designated for homologous recombination-based 2 integration (Miki et al., 2016). For endogenous CRWNa or SUN1 tagging with mCherry, a plasmid was 3 constructed, where the mCherry gene and the blasticidin-resistance cassette were flanked by ~1 kb 4 sequences within the ORF and 3"UTR. CENP-A-mCherry driven by actin promoter was targeted to hb7 5 locus and selected by nourseothricin resistance. 6 7 Immunoblotting 8 Cell extracts were prepared by grinding protonema colonies (Yamada et al., 2016). Immunoblotting of 9 Cerulean-tagged proteins was performed with home-made anti-GFP antibody (rabbit "Nishi", final bleed, 1: 10 500). 11 12 In vivo microscopy 13 Methods for epifluorescence and spinning-disc confocal microscopy were previously described ( Data analysis 25 Quantification of the chromosome and chloroplast movement following cell division was performed for 26 apical cells, following the method described in Yamada et al. (2017). Images were acquired every 3 min for 27 10 h under dark condition and kymographs were generated. The distance from the metaphase plate to the 28 nuclear centre was measured after generating kymographs. To quantify chloroplast distribution, the 29 intensity of chloroplast autofluorescence was measured. Images were acquired every 3 min for 10 h under 30 dark condition. Chloroplast signal intensity was measured along cells' long axes by drawing a 31 one-pixel-wide line using ImageJ. Each cell was divided into four regions, and the mean signal intensity 32 within each region was divided by the average intensity of the whole cell.      0.17 µm/min (n = 11); ΔFERM-C add-back, 2.0 ± 0.12 µm/min (n = 9). **p < 0.01, ***p < 0.001, 17 ****p < 0.0001 (unpaired t-test, two-tailed). Horizontal bars, 5 µm; Vertical bar in D, 5 min.