Expansion microsopy reveals Plasmodium falciparum blood-stage parasites undergo anaphase with a chromatin bridge in the absence of mini-chromosome maintenance complex binding protein

The malaria parasite Plasmodium falciparum undergoes closed mitosis, which occurs within an intact nuclear envelope, and differs significantly from its human host. Mitosis is underpinned by the dynamics of microtubules and the nuclear envelope. To date, our ability to study P. falciparum mitosis by microscopy has been hindered by the small size of P. falciparum nuclei. Ultrastructure expansion microscopy (U-ExM) has recently been developed for P. falciparum, allowing visualization of mitosis at the individual nucleus level. Using U-ExM, three intranuclear microtubule structures are observed: hemispindles, mitotic spindles and interpolar spindles. A previous study demonstrated that the mini-chromosome maintenance complex binding-protein (MCMBP) depletion caused abnormal nuclear morphology and microtubule defects. To investigate the role of microtubules following MCMBP depletion and study the nuclear envelope in these parasites, we developed the first nuclear stain enabled by U-ExM in P. falciparum. MCMBP deficient parasites show aberrant hemispindles and mitotic spindles. Moreover, anaphase chromatin bridges, and individual nuclei containing multiple microtubule structures were observed following MCMBP knockdown. Collectively, this study refines our understanding of MCMBP-deficient parasites and highlights the utility of U-ExM coupled with a nuclear envelope stain for studying mitosis in P. falciparum.

microtubule defects. To investigate the role of microtubules following MCMBP depletion and 23 study the nuclear envelope in these parasites, we developed the first nuclear stain enabled by 24 U-ExM in P. falciparum. MCMBP deficient parasites show aberrant hemispindles and mitotic 25 spindles. Moreover, anaphase chromatin bridges, and individual nuclei containing multiple 26 microtubule structures were observed following MCMBP knockdown. Collectively, this study 27 refines our understanding of MCMBP-deficient parasites and highlights the utility of U-ExM 28 coupled with a nuclear envelope stain for studying mitosis in P. falciparum.

INTRODUCTION 30
Malaria is estimated cause over 400,000 deaths annually, these deaths are 31 predominantly in young children and are caused by the unicellular protozoan pathogen 32 Plasmodium falciparum [1]. Resistance against frontline antimalarials has emerged in many 33 parts of the globe and is spreading [2][3][4][5][6]. Moreover, there is no highly effective vaccine against 34 malaria, highlighting the need to develop new therapeutic interventions for ongoing and future 35 control of this disease. One therapeutic strategy is drug inhibition of DNA/RNA replication, 36 and/or cell division, a method that is commonly used for control of bacterial [7] and viral 37 diseases [8], along with many types of cancer [9]. The distinctive division method of P. 38 falciparum compared to its human host makes this an attractive strategy for P. falciparum drug 39 design, yet no current antimalarials directly target DNA replication or cell division [10]; 40 highlighting the need for further investigation into this pathway. 41 Plasmodium parasites undergo cell division by a process known as schizogony, whereby 42 a singly nucleated parasite undergoes repeated rounds of DNA replication and mitosis, within 43 a shared cytoplasm, followed by a single cytokinetic event that results in the formation of 16-44 32 daughter parasites [11][12][13]. Throughout division, Plasmodium undergoes closed mitosis, 45 where the nuclear envelope remains intact, as opposed to the open mitosis of its human host 46 [14]. Nuclear division during schizogony is orchestrated by a set of intranuclear microtubule 47 side up on the parafilm in the humidity chamber. 5 µL of both TEMED and APS were added 174 per 90 µL of monomer solution, which was briefly vortexed, and 35 µL pipetted onto the 175 parafilm before the coverslip was placed cell-side down onto the monomer solution. Gels were 176 then incubated at 37 °C for 1 hour before being transferred into the wells of a 6-well plate filled 177 with denaturation buffer for 15 minutes at room temperature (200 mM sodium dodecyl sulfate 178 (SDS), 200 mM NaCl, 50 mM Tris, pH 9). Gels were then separated from coverslips and 179 transferred into Eppendorf tubes containing denaturation buffer and denatured at 95 °C for 90 180 minutes. Denatured gels were transferred into 10 cm Petri dishes filled with 25 mL MilliQ 181 water and placed on platform shaker for 30 minutes, with the water replaced twice, each for a 182 further 30 minutes. After the first expansion in water, expanded gels were shrunk by adding 25 183 mL two PBS washes each for 15 minutes. Shrunken gels were placed into the wells of a 6-well 184 plate filled with blocking buffer (3% BSA-PBS) and blocked for 1 hour at room temperature 185 on a platform shaker. After blocking, primary antibodies were prepared in 1 mL of blocking 186 buffer and gels were incubated with primary antibody overnight at room temperature on a 187 platform shaker. 188 Gels were washed three times in 0.5 % v/v PBS-Tween 20 (PBS-T), each for 10 189 minutes, before being incubated with 1 mL of secondary antibodies, NHS ester and/or nuclear 190 stain diluted in PBS for 2.5 hours at room temperature on a platform shaker. Following 191 secondary incubation, gels were washed three times in PBS-T. Stained gels were then 192 transferred back to 10 cm Petri dishes and underwent a second round of expansion with three 193 30 minutes washes in 25 mL MilliQ water. 194 The diameter of fully expanded gels was measured using a tape measure and the 195 expansion factor determined by dividing the expanded gel size (in mm) by the initial coverslip 196 size (12mm). Gel diameter and expansion factor for all gels prepared in this study can be found 197 For gels stained with BODIPY TR Ceramide (BODIPY TRc), sections of the expanded 199 gel were cut and placed into the wells of a 6-well plate containing 1 mL 2 µM BODIPY TRc 200 in MilliQ and incubated on a platform shaker overnight. 201 To prepare gels for imaging, small sections were cut from the larger gel and gently dried before 202 being placed into 35 mm #1.5 coverslip bottomed imaging dishes (Cellvis;Fisher Cat. No. 203 NC0409658) that had been pre-coated with poly-D-lysine. 204

Stains and antibodies 205
The following primary antibodies were used in this study: Mouse IgG1 anti-alpha 206

Image acquisition 228
All microscopy presented in this study was performed on a Zeiss LSM800 229 AxioObserver microscope that had an Airyscan detector. Additionally, all images were 230 acquired using a 63x Plan-Apochromat (NA 1.4) objective lens. All images presented in this 231 study were acquired as Z-stacks with an XY pixel size of 0.035 µm and a Z-step size of 0.15 232 µm. All images then underwent Airyscan processing using ZEN Blue (Version 3.1). 233

Image analysis 234
All image analysis performed in this study used ZEN Blue (Version 3.1). All 235 measurements of length were made using the "profile" function of ZEN Blue. 236 To measure hemispindle branch length, maximum intensity projections were made of 237 Airyscan-processed images. Hemispindle branches were first counted and then were measured 238 from the edge of the tubulin staining closest to the MTOC (visible on the NHS Ester channel) 239 to the edge of the tubulin staining furthest away from the MTOC. Nuclei that contained both 240 hemispindles and interpolar spindles were excluded from this analysis. In nuclei that contained 241 multiple MTOCs, it could not always be determined which MTOC each branch was coming 242 from, and so these nuclei were excluded from this analysis. 243 To measure mitotic spindle length, maximum intensity projections were made of 244 Airyscan-processed images. Mitotic spindle size was measured as the greatest distance between 245 the edge of the tubulin staining that was adjacent to each of the two MTOCs. 246

Statistical analyses 247
This study reports both the measured distances of mitotic and hemispindles, and then 248 estimated actual distances in unexpanded parasites. To estimate actual distances, the mean 249 expansion factor of all gels used in this study was determined (4.3x; Supplementary Figure 1c). 250 All actual distances were then divided by this mean expansion factor to get the actual estimated 251 distances reported in this study. 252 All graphs and statistical analyses in this study were performed and generated using 253 GraphPad PRISM 9. All values of statistical significance in this study were determined using 254 an unpaired, two-tailed T-test. 255

RESULTS 256
Ultrastructure expansion microscopy (U-ExM) significantly enhances visualization of 257 microtubule structures in P. falciparum 258 To validate the utility of U-ExM for visualizing microtubules in P. falciparum, we first 259 confirmed that we could visualize all previously identified microtubule structures 260 (hemispindle, mitotic spindle, interpolar spindle, subpellicular microtubules) in unexpanded 261 parasites ( Figure 1a). Additionally, we incorporated a general protein stain (N-262 hydroxysuccinimide (NHS) ester). In unexpanded parasites NHS ester staining did not produce 263 a staining pattern that obviously represented a particular organelle. Despite no obvious 264 demarcation of organelles, NHS ester staining appeared slightly denser in the chromatin-free region of the nucleus (Figure 1a), which has previously been shown to contain the microtubule 266 organizing center (MTOC) [16]. Additionally, in segmented schizonts NHS ester staining 267 appeared denser at the apical tip of merozoites, likely corresponding to the merozoite secretory 268 organelles rhoptries, micronemes, or dense granules (Figure 1a).

309
All microtubule structures were also observed following U-ExM but could be observed 310 in far greater detail with less confounding complexity from neighboring nuclei ( Figure 1b).
Notably, all the branches of a hemispindle could be readily differentiated including many small 312 branches that previously would have been below the limit of detection ( Figures 3 & 4). Notably, however, centrin 334 staining did not colocalize with the entirety of the MTOC, with centrin foci contained within a 335 small portion of the whole MTOC. By comparing with the nuclear stain, it could be seen that 336 centrin foci localized towards the cytoplasmic side of the MTOC structure, suggesting that P. 337 falciparum may compartmentalize subsets of proteins inside the MTOC. 338 In MCMBP deficient parasites, MTOC staining often appeared aberrant with misplaced 339 centrin foci (Figure 2e) (Supplementary Figures 3 & 4). However, these defects were not 340 consistent or easily quantifiable by regular microscopy measurement techniques. The nature of 341 these defects is unclear but suggest that MCMBP knockdown may alter the formation or 342 integrity of the MTOC. 343 In the presence of Shld1 (Figure 2a  Plasmodium undergoes mitosis without breakdown of the nuclear envelope, and in 378 doing so the nuclear envelope provides a critical barrier for the compartmentalization of the 379 nucleus from the cytoplasm. Therefore, nuclear envelope integrity and remodeling are critical 380 during Plasmodium mitosis. Despite the importance of nuclear envelope dynamics during 381 schizogony, and Plasmodium mitosis, there is currently no reliable marker of the P. falciparum 382 nuclear envelope for microscopic visualization. Previous studies have localized a few 383 nucleoporins (nups) to the nuclear envelope of either P. falciparum [16,39] or P. berghei [40], 384 but their distribution and number is dynamic across the lifecycle, limiting the robustness of 385 nups as nuclear envelope markers. Therefore, we wanted to identify a uniform, U-ExM 386 compatible stain for the P. falciparum nuclear envelope to allow us to study nuclear envelope 387 changes in the context of MCMBP deficient parasites. 388 BODIPY TR ceramide (BODIPY TRc) is a commonly used fluorescent lipid stain, 389 which has previously been used to stain live parasites from multiple different parasite lifecycle 390 stages, across P. falciparum and P. berghei, and imaged in both fixed and live-cell microscopy 391 [40][41][42][43][44]. Despite its extensive use, BODIPY TRc has not previously been reported to stain the 392 nuclear envelope of P. falciparum. We coupled BODIPY TRc with U-ExM, with BODIPY 393 TRc staining occurring post-expansion. Remarkably, we found that the P. falciparum nuclear 394 envelope is consistently and reliably labelled by BODIPY TRc (Figure 3a). In addition to 395 staining the nuclear envelope, BODIPY TRc enabled observation of the RBC membrane, 396 PVM, PPM, and endoplasmic reticulum as previously demonstrated when staining live 397 parasites [43][44][45][46][47] (Supplementary Figure 5). Together, we demonstrate that BODIPY TRc is the 398 first Plasmodium nuclear envelope stain enabled by U-ExM. 399 400 401  Figure 6); reminiscent of chromatin bridges that occur during a 436 defective anaphase of other organisms [48,49]. This contrasts with MCMBP HADD parasites 437 grown in the presence of Shld1, where DNA staining was not observed inside this bridge region 438 (Figure 1b,3a). We sometimes observed dividing nuclei with interpolar spindles where each 439 nucleus was of vastly different size, potentially indicating uneven DNA segregation in some nuclei following MCMBP knockdown (Supplementary Figure 6). Moreover, we also observed 441 interpolar spindles connecting MTOCs in nuclei that did not appear to be separating from each 442 other at all (Supplementary Figure 6)  In MCMBP HADD parasites grown in the absence of Shld1, we also frequently observed 456 a single nuclear envelope that contained multiple microtubule structures and multiple MTOCs 457 not connected by an interpolar spindle (Supplementary Figure 6). Nuclei were observed that 458 contained two hemispindles and two MTOCs, which likely represent nuclei where nuclear 459 fission either did not occur or had occurred aberrantly (Figure 4) (Supplementary Figure 6). 460 Additionally, nuclei that contained two mitotic spindles and four MTOCs were also observed 461 ( Figure 4). This suggests that in the aberrant nuclei that contain two MTOCs after mitosis, both 462 MTOCs can duplicate and form mitotic spindles in the same nucleus. Collectively, these 463 observations suggest that MCMBP deficient parasites undergo uneven DNA segregation, 464 leading to aneuploidy and nuclear fission defects, but that these defects don't inhibit further 465 rounds of mitosis as shown previously [25].

DISCUSSION 513
The development of U-ExM and application to P. falciparum parasites, have allowed 514 us to understand the functions of proteins and processes of P. falciparum to a level of detail 515 not previously possible. We applied U-ExM in the context of MCMBP deficient parasites to 516 significantly refine our understanding of the function of this protein during blood-stage 517 replication of P. falciparum. 518 The U-ExM protocol used in this study is largely similar previously published protocols 519 [16,22,36], with the notable modification of changing the protein crosslinking (FA/AA) 520 incubation step from 5 hours to overnight, which significantly shortened day 1 of the U-ExM 521 protocol. In this study, we harvested parasites at multiple timepoints throughout the lifecycle, 522 and all -Shld1 cultures were harvested 3 hours after their +Shld1 counterparts due to the 523 documented growth delay [25]. This shortening of day 1 of the U-ExM protocol made the 524 protocol far more practical, enabling the progressive study of different lifecycle stages in the 525 same U-ExM experiment, rather than having to harvest different lifecycle stages as 526 independent experiments. 527 In addition to highlighting new biology, our application of U-ExM allowed us to 528 identify some potential drawbacks of this technique. Notably absent in all images of U-ExM 529 parasites were the food vacuole and hemozoin crystal. Through NHS ester staining of 530 unexpanded parasites (Figure 1a), the likely location of the hemozoin crystal and food vacuole 531 could be inferred due to a distinct lack of staining. By contrast, there was no indication on U-532 ExM parasites of where the hemozoin crystal or food vacuole membrane would be located 533 based on NHS ester or BODIPY TRc staining. We hypothesize that the hemozoin crystal either 534 does not get anchored, or does not expand with the gel, potentially limiting the utility of U-535 ExM for studies of hemoglobin catabolism and hemozoin biomineralization.
In this study a wide range of nucleic acid stains were used on U-ExM parasites: DAPI, 537 DRAQ5, SYBR Green and SYTOX red. Notably, all these stains showed considerably more 538 photobleaching than we would observe in unexpanded parasites; particularly SYBR Green, 539 which began visibly photobleaching almost immediately. The reason for this is not clear, but 540 it should be noted that while in unexpanded MCMBP deficient parasites a clear nuclear staining 541 defect was observed [25], the same could not be readily observed in U-ExM parasites. At the 542 concentrations used in this study, we found SYTOX red to be the brightest and most 543 photostable of the nucleic acid stains used. Potentially related to the changes in chromatin was 544 our inability to localize MCMBP by U-ExM. MCMBP has previously been localized in 545 unexpanded parasites, showing nuclear and cytoplasmic foci [25]. Despite this, our attempts to 546 localize MCMBP by U-ExM showed no significant signal. It has been noted previously that 547 some antibodies appear to be incompatible with U-ExM [36], although the reasons for this are 548 unclear, but this does not appear to be the case as the anti-HA antibody we used to detect 549 MCMBP has previously been used successfully on U-ExM samples [16]. Given that the 550 canonical role of MCMBP is to bind DNA, and we observe significant differences in the 551 appearance of DNA in U-ExM parasites, it is possible that some DNA-binding proteins are not 552 retained after U-ExM. 553 BODIPY TRc stained parasites presented in this study were stained post-expansion. 554 We attempted to stain live cells with BODIPY TRc or include BODIPY TRc with the primary 555 or secondary antibody incubations but this uniformly resulted in extremely faint staining (data 556 not shown). Additionally, we tried to stain parasites with Nile red, which has previously been 557 shown to stain some organelles in unexpanded P. falciparum blood-stage parasites [45]; but this 558 was also unsuccessful (data now shown). Overall, this suggests that potentially large 559 differences exist in the fluorescent stains that are compatible with unexpanded P. falciparum 560 compared to with U-ExM. 561 merozoite, the PVM, RBC membrane and nuclear envelope, but did not reveal any structures 563 that could be characteristically identified as the apicoplast or mitochondrion as have been 564 identified by EM studies [12]. Additionally, both the surface and lumen the rhoptries of 565 merozoites stained very strongly with BODIPY TRc, which would support previous 566 observations that the rhoptries contain membranous whorls [37,51]. To date, these membranous 567 whorls have not been observed by light microscopy [52], highlighting the use of U-ExM 568 coupled with BODIPY TRc for studying merozoite physiology. 569 We show that MCMBP knockdown results in the aberrant formation of all intranuclear 570 microtubule structures but not of subpellicular microtubules, which are formed when MCMBP 571 is no longer expressed [25]. Moreover, the combination of NHS ester and BODIPY TRc with 572 U-ExM allowed us to assess the MTOC and nuclear envelope at a level of detail reminiscent 573 of electron microscopy. This also confirmed that the in the blood-stage of P. falciparum the 574 MTOC spans the nuclear envelope, as previously reported [16]. The combination of BODIPY 575 TRc and NHS ester staining allowed us to show that aneuploidy in MCMBP deficient parasites 576 is due formation of anaphase chromatin bridges and/or a lack of nuclear fission ( Figure 5). We 577 hypothesize that these lead to the downstream phenotypes we see of wildly varied nuclear size 578 and zoid merozoites following cytokinesis. 579 The observation of anaphase chromatin bridges in MCMBP-deficient parasites is 580 supported by the canonical function in the MCM complex [26]. In other organisms, the presence 581 of MCMBP has been shown to promote dissociation between the MCM complex and 582 chromatin, allowing separation of sister chromatids [53][54][55]. Moreover, MCMBP of P. 583 falciparum has been shown to interact with the members of the condensin complex structural 584 maintenance of chromosomes (SMC) 2 and 4 [25]. SMC2 and SMC4 have canonical roles in 585 chromosome condensation [56,57], and have recently been shown to be involved in Plasmodium chromosome separation [58]. Inhibition of SMC2 [59], SMC4 [60] and MCM complex member 587 MCM7 [61] have all been shown to lead to the formation of anaphase chromatin bridges in 588 other organisms. Therefore, we suggest that the observation of anaphase chromatin bridges in 589 MCMBP-deficient P. falciparum may be caused by either an ability to properly separate sister 590 chromatids, or a defect in the detachment of microtubules from chromosomes. Currently, the 591 relationship between the formation of anaphase chromatin bridges and nuclear fission is 592 unclear. But given that inhibition of SMC2, SMC4, and MCM7 causes anaphase chromatin 593 bridges in organisms that undergo open mitosis, and therefore do not undergo nuclear fission, 594 defective nuclear fission is not a pre-requisite for the formation of anaphase chromatin bridges. 595 While we observe anaphase chromatin bridges and multiple microtubule structures in 596 a single nucleus following depletion of MCMBP, using MCMBP HADD parasites these events 597 do not occur in every round of mitosis. If this were the case, we would expect to see all DNA 598 staining contained within a single, giant, nuclear envelope, but we do not. Given that the 599 knockdown system used leads to imperfect and uneven depletion of MCMBP, it is possible 600 that the phenotypic heterogeneity we observe is a product of differing levels of MCMBP. 601 MCMBP is likely essential for growth in the blood-stage of Plasmodium [62,63], and we 602 hypothesize that the complete removal of MCMBP would lead to the formation of anaphase 603 chromatin bridges and inhibited nuclear fission in every round of mitosis. 604 Our observations of microtubules, following U-ExM, were largely concordant with 605 recent studies that also made measurements of branch and spindle lengths [16,22]. 606 Measurements for mitotic spindle length, hemispindle branch length and hemispindle branch 607 number all reported similar results [16,22]. Neither data set, however, controlled for the number 608 of nuclei per cell and so it is currently unclear whether any of these measurements change later 609 in the parasite lifecycle. One difference observed in our study however, was the presence of 610 hemispindle-like branches in nuclei connected by interpolar spindles (referred to as anaphase spindles in that study) [16]. Previous images have only observed the long interpolar branches 612 connecting the MTOCs, without smaller branches in each nucleus [16]. Critically, a previous 613 hypothesis suggested that the hemispindle formed as a remnant of the retraction of the 614 interpolar spindle [64]. Our observation that the two seem to co-exist would suggest that this is 615 not the case. Moreover, this suggests that whatever the function(s) of hemispindles are, likely 616 begin immediately following nuclear segregation and before nuclear fission. 617 Overall, this study provides insight into the poorly understood, yet therapeutically 618 attractive and biologically fascinating, process of mitosis in P. falciparum. Our findings 619 significantly further our understanding of the phenotype of parasites following knockdown of 620 MCMBP. Importantly, these insights were only possible because of the application of U-ExM 621 to P. falciparum. Moreover, we developed BODIPY TRc as the first U-ExM-compatible stain 622 to visualize the nuclear envelope and used this to develop our understanding of both MCMBP 623 deficient parasites, and parasite physiology more broadly.