Aspergillus SUMOylation mutants exhibit chromosome segregation defects including chromatin bridges

Abstract

Functions of protein SUMOylation remain incompletely understood in different cell types. Via forward genetics, here we identified ubaB^Q247*, a loss-of-function mutation in a SUMO activation enzyme UbaB in the filamentous fungus Aspergillus nidulans. The ubaB^Q247*, ΔubaB, and ΔsumO mutants all produce abnormal chromatin bridges, indicating the importance of SUMOylation in the completion of chromosome segregation. The bridges are enclosed by nuclear membrane containing peripheral nuclear pore complex proteins that normally get dispersed during mitosis, and the bridges are also surrounded by cytoplasmic microtubules typical of interphase cells. Time-lapse sequences further indicate that most bridges persist through interphase prior to the next mitosis, and anaphase chromosome segregation can produce new bridges that persist into the next interphase. When the first mitosis happens at a higher temperature of 42°C, SUMOylation deficiency produces not only chromatin bridges but also many abnormally shaped single nuclei that fail to divide. UbaB-GFP localizes to interphase nuclei just like the previously studied SumO-GFP, but the nuclear signals disappear during mitosis when the nuclear pores are partially open, and the signals reappear after mitosis. The nuclear localization is consistent with many SUMO targets being nuclear proteins. Finally, although the budding yeast SUMOylation machinery interacts with LIS1, a protein critical for dynein activation, loss of SUMOylation does not cause any obvious defect in dynein-mediated transport of nuclei and early endosomes, indicating that SUMOylation is unnecessary for dynein activation in A. nidulans.

Introduction

Posttranslational modification of proteins by small ubiquitin-like modifier (SUMO) has been linked to many cellular processes, but the functions of SUMOylation in different cell types still need to be better understood (Vertegaal 2022). SUMOylation has recently been implicated in regulating cytoskeletal proteins or proteins involved in intracellular motility (Leisner et al. 2008; Meednu et al. 2008; Alonso et al. 2015; Greenlee et al. 2018). Importantly, the budding yeast (Saccharomyces cerevisiae) homolog of LIS1, which is a key dynein activator (Markus et al. 2020), has been shown to physically interact with the SUMOylation machinery (Alonso et al. 2012). However, LIS1 or proteins in the dynein pathway were not among the previously identified SUMO targets in the filamentous fungus Aspergillus nidulans (Harting et al. 2013; Horio et al. 2019). Despite this difference, it would be useful to test directly whether SUMOylation affects dynein function in different organisms and cell types.

SUMOylation has been known to affect mitosis in various ways (Biggins et al. 2001; Dasso 2008; Dawlaty et al. 2008; Zhang et al. 2008; Eifler and Vertegaal 2015; Sridharan and Azuma 2016; Eifler et al. 2018; Johansson et al. 2021; Kroonen and Vertegaal 2021; Ptak et al. 2021; Yatskevich et al. 2021; Liu et al. 2023; Saik et al. 2023). During a normal mitosis, replicated sister chromatids segregate and move to the opposite poles of the mitotic spindle during anaphase, which is followed by spindle disassembly and cytokinesis that divides the daughter cells. Chromatin bridges, a type of anaphase bridges, are caused by a failure in timely resolving of the linkage between the duplicated sister chromatids (Finardi et al. 2020). Such a failure can be caused by many factors, including, but not limited to, dicentric chromosomes caused by abnormal fusions between telomeres or broken chromosome ends and defects in resolving DNA catenanes caused by a topoisomerase II dysfunction or other problems (Dawlaty et al. 2008; Maciejowski et al. 2015; Nielsen et al. 2015; Finardi et al. 2020; Umbreit et al. 2020; Flynn et al. 2021). Most relevant to our current study, SUMOylation deficiency has been linked to the formation of anaphase chromatin bridges (Kroonen and Vertegaal 2021); for example, a defect in topoisomerase II SUMOylation causes chromatin bridges (Dawlaty et al. 2008). Chromatin bridges have been linked to chromothripsis (Maciejowski et al. 2015; Umbreit et al. 2020), which causes dramatic rearrangement of 1 or a few chromosomes (Leibowitz et al. 2015; Ly and Cleveland 2017). Chromatin bridges induced by some antimitotic drugs may also activate the cGAS-mediated type 1 interferon signaling (Flynn et al. 2021), which is a response to cytoplasmic DNA (Li and Chen 2018). Whether the bridges affect cytokinesis, especially abscission, has been studied in S. cerevisiae and mammalian cells (Shi and King 2005; Steigemann et al. 2009; Carlton et al. 2012; Amaral et al. 2016; Nähse et al. 2017; Finardi et al. 2020; Petsalaki and Zachos 2021), where a NoCut pathway or abscission checkpoint is implicated in abscission delay in response to chromatin bridges (Norden et al. 2006; Steigemann et al. 2009; Carlton et al. 2012; Amaral et al. 2016). However, this does not prevent the breakage of some chromatin bridges, which significantly enhances the genome instability observed in cancer cells (Maciejowski et al. 2015; Umbreit et al. 2020).

Many filamentous fungi contain hyphal syncytia with multiple well-spaced nuclei that are distributed by cytoplasmic dynein and other mechanisms (Xiang 2018; Mela et al. 2020). The presence of multiple nuclei in the same compartment should allow errors of chromosome segregation to be tolerated since the nuclei can complement each other for functions. The filamentous fungus A. nidulans is an established model organism for studying the regulation of cell cycle progression, among other interesting topics (Morris 1975; Morris and Enos 1992; Harris 2001; Osmani and Mirabito 2004; Osmani et al. 2006; De Souza et al. 2009; Nayak et al. 2010; De Souza et al. 2011; Edgerton-Morgan and Oakley 2012; Edgerton et al. 2015; Liu et al. 2015; Dörter and Momany 2016; Paolillo et al. 2018; Suresh and Osmani 2019; Etxebeste and Espeso 2020). A. nidulans cells undergo a “semiopen” mitosis with partial disassembly of the nuclear pore complex during mitosis (De Souza et al. 2004), a mode that differs from the more closed and more open modes of mitosis in yeast and mammalian cells (De Souza and Osmani 2009; Dey and Baum 2021). Unlike another filamentous fungus Ashbya gossypii where each nucleus undergoes mitosis autonomously within a syncytium (Gladfelter et al. 2006; Roberts and Gladfelter 2015), nuclei within a syncytium of A. nidulans undergo mitosis almost synchronously (Rosenberger and Kessel 1967; Suelmann et al. 1997). Syncytia are separated by a cross wall-like structure called the septum, whose formation is triggered by the septation initiation network (SIN), but it does not necessarily happen after each nuclear division, which differs from septation/cytokinesis in yeasts (Krapp et al. 2004; Wolfe and Gould 2005; Kim et al. 2006; Kim et al. 2009). More specifically, the first few rounds of mitosis in A. nidulans germ tubes are not coupled to septation, although septation may occur after each mitosis at later stages of hyphal growth, and this generates multinucleated hyphal compartments (Fiddy and Trinci 1976; Harris et al. 1994). In filamentous fungi, each septum contains a septal pore, which allows some intercompartmental exchanges during interphase, and the septal pore in A. nidulans closes during mitosis using a NimA kinase-dependent mechanism (Shen et al. 2014). This may help prevent the mitotic signals from spreading to adjacent hyphal compartments, which differs from the Woronin body-mediated septal pore sealing in response to hyphal injury (Jedd and Chua 2000; Tenney et al. 2000; Shen et al. 2014; Steinberg et al. 2017; Mamun et al. 2023; Songster et al. 2023).

In this current study, we obtained an A. nidulans mutant exhibiting chromatin bridges and identified the causal mutation in a SUMO-activating enzyme UbaB. Our data suggest that SUMOylation in A. nidulans is not needed for dynein activation but important for resolving problems that generate chromatin bridges during mitosis. Our observations on the chromatin bridges suggest that majority of the bridges persist through the interphase prior to the next mitosis, and the bridges are enclosed by nuclear membrane with nuclear pore complexes. In one time-lapse sequence, we also found an event of septal pore sealing on a bridge.

Materials and methods

A. nidulans strains and media

A. nidulans strains used in this study are listed in Supplementary Table 1. The strains have been submitted to the Fungal Genetics Stock Center (FGSC), and each strain has a FGSC number. The The FungiDB (fungidb.org) designations for several Aspergillus nidulans (AN) genes mentioned in the strain list are as follows: AN7753 (nkuA) (Nayak et al. 2006), AN2450 (ubaB) (Harting et al. 2013), AN1191 (sumO) (Wong et al. 2008), AN2765 (hhoA or Histone-H1) (Ramón et al. 2000), AN0316 (tubA) (Doshi et al. 1991), AN2431 (An-Nup49) (Osmani et al. 2006), AN5627 (An-Nup96 or sonB) (De Souza et al. 2004), and AN0118 (nudA) (Xiang et al. 1994). The presence of the ΔnkuA::argB allele was determined by PCRs using the primers ktF (5′-CGTCGTACAGGTACCAGGACTTTC-3′) and ktR (5′-CTGCAATTATTGCATGCGTTCATC-3′), which gives a ∼3-kb product. The presence of the wild-type nkuA gene was determined by PCR using Ku70F (5′-GACGACTATCGCGAAGATGAC-3′) and Ku70R (5′-TCTCACAGTTGCCGCAGAA-3′), which gives a 0.7-kb product.

Detailed medium compositions can be found in a recent publication (Qiu, Zhang, et al. 2023). Solid rich medium was made of either YAG (0.5% yeast extract and 2% glucose with 2% agar) or YAG + UU (YAG plus 0.12% uridine and 0.11% uracil). Solid minimal medium containing 1% glucose was used for selecting progeny from a cross. For most live cells imaging experiments, cells were cultured in liquid minimal medium containing 1% glycerol for overnight at 32°C or containing 1% glucose for overnight at room temperature. For experiments involving the nimT23 mutation or for strains cultured at 42°C, we used liquid minimal medium containing 0.1% glucose.

Live cell imaging and analyses

Microscopic images used in Figs. 2, 4, 5, and 8 and Supplementary Figs. 3 and 4 were generated using a Nikon Ti2-E inverted microscope with Ti2-LAPP motorized TIRF module and a CFI apochromat TIRF 100 × 1.49 numerical aperture (N.A.) objective lens (oil). The microscope was controlled by NIS-Elements software using 488- and 561-nm lines of LUN-F laser engine and ORCA-Fusion BT cameras (Hamamatsu). For all images, cells were grown in the Lab-Tek Chambered #1.0 borosilicate coverglass system (Nalge Nunc International, Rochester, NY). Images were taken at room temperature. All the images were taken with a 0.1-s exposure time (binning: 2 × 2). For line scans presented in Fig. 8, we draw a line starting from the hyphal tip in the middle of the hypha and get the average intensity value for the width of 2 µm, which is normally the hyphal width. For Fig. 3 and Supplementary Fig. 5 and all 5 videos, live-cell imaging was performed as previously described (Bieger et al. 2021). Specifically, it was performed on a Nikon Ti-E Eclipse inverted epifluorescence microscope, equipped with a Perfect Focus System (Nikon) and a motorized Piezo stage, using a 100 × 1.49 N.A. oil immersion Apo TIRF Nikon objective. GFP and DsRed were excited using an AURA II triggered illuminator with 485- and 560-nm LEDs, respectively, and detected using a Zyla 4.2 sCMOS camera (Andor, Oxford Instruments). All hardware was controlled by NIS-Elements Advanced Research (version 4.60). Two-color 3D time-lapse data sets were deconvolved, with spherical aberration correction and background subtraction, using the “automatic” 3D deconvolution option in NIS-Elements Advanced. Micrographs represent maximum intensity projections of 2-color Z-series acquired 0.2 µm intervals spanning the entire depth of the hyphae. Supplementary Video 1 was generated in Imaris (9.5.1; Bitplane), while Supplementary Videos 2, 3, 4, and 5 were generated in Fiji (Schindelin et al. 2012). Images used in Figs. 1, 6, and 7 and Supplementary Fig. 6 were captured using an Olympus IX73 inverted fluorescence microscope linked to ORCA-FLASH4.0LT + sCMOS CAMERA and controlled by cellSens Dimension Version 3 software (Olympus Inc.). An UPlanSApo 100× objective lens (oil) with a 1.40 N.A. was used. An ET-EGFP/mCherry Dual Multiband Filter Set w/BX3 cube (purchased from Olympus Inc.) was used.

Fig. 2.

Images of the ΔubaB mutant showing that most of the chromatin bridges connect interphase nuclei. a) Images of Histone-H1-GFP (Histone-H1) and NLS-DsRed (NLS) in the wild-type control and the ΔubaB mutant. The 2 wild-type panels were from a time-lapse, and thus, the cell cycle stage of some nuclei is labeled by the letter(s) such as G1, M, or G2. Note that the NLS-DsRed signals are diffused in the cytoplasm during mitosis (M), and these signals have been enhanced in the top left panels to show the position of the septum (arrowhead). Two ΔubaB panels on the left show that nuclei connected by chromatin bridges can enter interphase as evidenced by the nuclear NLS-DsRed signals. Note that signals of these images have been enhanced to show the bridges more clearly. Unenhanced images (labeled using a smaller font size) of Histone-H1 and NLS are presented on the bottom of the enhanced images to allow better visualization of chromatin and nuclear signals of the connected nuclei. One ΔubaB panel on the right shows a mitotic chromatin bridge, and an unenhanced Histone-H1 image is presented on the bottom of the enhanced image. Bar: 10 µm. b) A quantitative analysis on the percent of interphase and mitotic chromatin bridges in the ΔubaB mutant (unpaired t-test, 2-tailed, Prism 9). The data were from 3 experiments. More than 15 bridges were counted in each experiment, and a total of 66 bridges were counted. Scatter plots with mean and SD values were generated by Prism 9. c) Images of GFP-TubA (microtubule) and NLS-DsRed (NLS) in the wild-type control and the ΔubaB mutant. The 2 wild-type panels were from a time-lapse, and the cell cycle progressed from M with mitotic spindles to G1 with cytoplasmic microtubules. The ΔubaB panel on the right shows an NLS-DsRed-labeled bridge surrounded by cytoplasmic microtubules typical of interphase cells. More examples of interphase bridges are shown in Supplementary Figs. 3 and 4.

Fig. 4.

Chromatin bridges are surrounded by nuclear envelope with nuclear pore complexes. a) Images of Nup49-GFP (Nup49) and Histone-H1-mCherry (Histone-H1) in a wild-type control and the ΔsumO mutant. We have observed at least 20 bridges in the ΔsumO mutant, as evidenced from the Histone-H1 signals, and all contain Nup49-GFP (Nup49) signals on the bridges. b) Images of Nup96-GFP (Nup96), NLS-Ds-Red (NLS), and Histone-H1-mCherry (Histone-H1) in a wild-type control and the ΔsumO mutant. Note that the NLS and Histone-H1 signals on the right panel of the ΔsumO images have been enhanced to show the bridge. All bridges showed Nup96 signals (n > 20). Bar: 10 µm.

Fig. 5.

Nuclear UbaB-GFP disperses from nuclei during mitosis and is rapidly reimported upon completion of mitosis. At 0 min, the germling contains 1 nucleus, and the UbaB-GFP signals are localized to the nucleus that is also labeled by NLS-DsRed. At the 5-min time point, UbaB-GFP nuclear signals disappeared during mitosis. UbaB-GFP signals reappeared inside the 2 daughter nuclei at 10 min, after the nuclear pore complex fully assembled as evidenced by the NLS-DsRed signals. Bar: 5 µm.

Fig. 8.

Loss of SUMOylation does not cause early endosomes to accumulate abnormally at the hyphal tip. a) Distribution of mCherry-Rab5A-labeled early endosomes in the wild type, the ubaB^Q247* mutant and the ΔsumO mutant. The hyphal tip is indicated by a arrowhead. As a control exhibiting an abnormal hyphal-tip accumulation of mCherry-Rab5A signals, we used a recently studied kinesin-1 mutant, kinA^K895* (Qiu, Zhang, and Xiang 2023). The ubaB^Q247* and ΔsumO mutants show a normal distribution of early endosomes, as compared to the wild type and the kinA^K895* mutant. Bar: 10 µm. b) Line scans of mCherry-Rab5A (early endosomes) fluorescence intensity in the wild type, the ΔsumO mutant and the kinA^K895* mutant. All values are relative to the mean value for wild type at position 0, which is set as 1. XY graphs with mean (solid lines) and standard error of the mean (SEM, shading) were generated by Prism 9. The intensity of mCherry-Rab5A near the hyphal tip (between 0.455 and 3.185 µm from hyphal tip with 0.065 µm as intervals) was significantly higher in the kinA^K895* mutant than that in both the wild-type and the ΔsumO mutant (P < 0.0001, 2-way ANOVA with Tukey's multiple comparisons test, n = 31 hyphae for wild type, n = 31 hyphae for the ΔsumO mutant, and n = 29 hyphae for the kinA^K895* mutant). The overall intensity of mCherry-Rab5A appeared higher in the ΔsumO mutant than in the wild type along the hyphae, but the differences were not statistically significant (P > 0.05 for all the points from 0 to 10.595 µm from the hyphal tip).

Fig. 3.

Time-lapse images of the ubaB^Q247* mutant showing bridged interphase nuclei entering mitosis and septal pore sealing on a bridge (Supplementary Video 1). a) Bridged interphase nuclei can enter mitosis. The mutant images are parts of the time-lapse sequence (Supplementary Video 1) taken at 10-min intervals. Wild-type control panels are shown on the left. To help better visualizing chromatin, Histone-H1 images are shown on the right side of the merged images of Histone-H1 and NLS. The ubaB^Q247* mutant panels are shown on the right. Note that entry into mitosis is indicated by the disappearance of NLS-DsRed signals from the nuclei, and exit from mitosis is indicated by the reappearance of NLS-DsRed signals in the nuclei. Bar: 5 µm. b) Septum formation and septal pore closure on a bridge in the ubaB^Q247* mutant. Wild-type control panels are shown on the left. A closed septum is indicated by a arrowhead. Note that cytoplasmic NLS-DsRed signals only on 1 side of the hypha helped reveal the presence of a closed septum during mitosis (Shen et al. 2014). In the mutant panel, the appearance of NLS-DsRed signals only on the left compartment at the 20-min time point indicates septal pore closure during mitosis. The 2 nuclei on the right were connected by a bridge at time points 0 and 10 min. Note that the septum might have formed before this time point during interphase, but the cytoplasmic NLS-DsRed signals during mitosis helped reveal the presence of a closed septum (Shen et al. 2014). Please see Supplementary Fig. 5 for unmerged images of Histone-H1 and NLS. Bar: 5 µm.

Fig. 1.

Colony and chromatin bridge phenotypes of the ubaB^Q247*, ΔubaB, and ΔsumO mutants. a) The nuclear phenotype of the ubaB^Q247* mutant visualized by using the Histone-H1-GFP fusion. Hyphal shape is indicated by dotted lines. Bar: 10 µm. b) A diagram showing position of the Q247* mutation in UbaB (arrow). Two domains were drawn according to the SMART domain analysis (http://smart.embl-heidelberg.de/) and the NCBI annotation of UbaB (GenBank: CBF86880.1) (https://www.ncbi.nlm.nih.gov/protein/259487865): “UBA_e1_thiolCys” (from aa182 to aa370) contains the ubiquitin-activating enzyme active site. “UAE_UbL” (from aa441 to aa516) contains the ubiquitin/SUMO-activating enzyme ubiquitin-like domain needed for transferring SUMO to its conjugating enzyme E2. c) Colony phenotypes of wild-type, ubaB^Q247*, ΔubaB, and ΔsumO strains. Strains were grown at 37°C for 2 days on YAG rich medium. d) A quantitative analysis of colony diameter of wild-type, ubaB^Q247*, ΔubaB, and ΔsumO strains (ordinary 1-way ANOVA with Tukey's multiple comparisons test). Scatter plots with mean and SD values were generated by Prism 9. e) Chromatin bridges in the ubaB^Q247*, ΔubaB, and ΔsumO strains containing Histone-H1-GFP. Hyphal shape is indicated by dotted lines. Bar: 10 µm. f) A quantitative analysis on the percent of nuclei linked by chromatin bridges in wild-type, ubaB^Q247*, ΔubaB, and ΔsumO strains (ordinary 1-way ANOVA with Tukey's multiple comparisons test). Scatter plots with mean and SD values were generated by Prism 9. Note that this analysis was done using strains grown overnight at 32°C in liquid minimal medium containing 1% glycerol. Data were from 3 experiments. For each experiment, at least 125 total nuclei were counted for each strain. A total of 404, 506, 634, and 515 nuclei were counted for wild-type, ubaB^Q247*, ΔubaB, and ΔsumO samples.

Fig. 6.

SUMOylation deficiency causes a severe defect in the first nuclear division at a higher temperature of 42°C. a) Nuclei labeled with Histone-H1-GFP in wild type, the ΔubaB single mutant, the nimT23 single mutant, and the nimT23, ΔubaB double mutant. Images were taken after a 6-h incubation at the restrictive temperature of 42°C (42°C) or after shifting the cells to 25°C for 0.5–1 h (42–25°C). Bar: 5 µm. b) A quantitative analysis on the percent of germ tubes (42–25°C) containing a chromatin bridge (ordinary 1-way ANOVA with Tukey's multiple comparisons test). Scatter plots with mean and SD values were generated by Prism 9. c) A quantitative analysis on the percent of germ tubes (42–25°C) containing an abnormally shaped nucleus (ordinary 1-way ANOVA with Tukey's multiple comparisons test). Scatter plots with mean and SD values were generated by Prism 9. For b andc, data were from 3 experiments. For each experiment, at least 40 germ tubes were counted for the ΔubaB single mutant, the nimT23 single mutant, and the nimT23, ΔubaB double mutant, and at least 24 germ tubes were counted for the wild-type strain. The total germ tubes counted were 109 for wild type, 229 for ΔubaB, 142 for nimT23, and 234 for ΔubaB, nimT23. d) Nuclei labeled with Nup49-GFP (top) and Histone-H1-mCherry (bottom) in wild type and the ΔsumO mutant grown at 42°C for 7 h.

Fig. 7.

Loss of SUMOylation does not cause an obvious defect in dynein-mediated nuclear migration. a) Histone H1-GFP-labeled nuclei in wild type, the ΔsumO mutant and the nudA1 (dynein heavy chain) mutant. Strains were grown overnight at 32°C in liquid minimal medium containing 1% glycerol. Small arrow indicates the spore swelling. Note that there is a chromatin bridge in the ΔsumO mutant. Bar: 10 µm. b) A quantitative analysis on the percent of germ tubes containing different numbers of nuclei in the spore swelling. Column bar graphs with mean and SD values were generated from 3 experiments. For each experiment, at least 40 germ tubes were counted from the wild-type and ΔsumO strains, and at least 20 were counted from the nudA1 mutant. Total germ tubes counted are 385 for the wild type, 486 for the ΔsumO mutant and 100 for the nudA1 mutant. P-values were generated from 2-way ANOVA with Tukey's multiple comparisons test.

Mutagenesis and identification of the mutation that causes the chromatin bridge phenotype

UV mutagenesis was performed as previously described (Qiu, Zhang, and Xiang 2023). For identifying the causal mutation, we first crossed the identified mutant to a wild-type strain. We then picked 5 mutant progeny from the cross with the same colony phenotype and picked 5 wild-type progeny (note that the chromatin bridge phenotype of the mutant is linked to the colony phenotype of poor conidiation). The spores of the 5 progeny in each group were mixed. Both the mutant and wild-type spores were cultured to make genomic DNA using the QIAGEN DNeasy Plant Mini Kit. The mutant and wild-type genomic DNAs were then sent to Otogenetics Corp for whole-genome sequencing. By comparing the whole-genome sequencing data from the 2 groups, we identified 1248 sequence variants in the mutant group but not in the wild type. We further annotated these variants in MySQL database using the A. nidulans gene features table (Supplementary excel file 1) downloaded from an Aspergillus genome database site that is no longer available and has been replaced by FungiDB (https://fungidb.org/fungidb/app). We identified 140 sequence variants positioned in the coding regions of 48 genes. We then manually checked these 48 genes using the IGV software and identified the single mutation in AN2450, which encodes UbaB (Supplementary Fig. 1).

Construction of the ΔubaB mutant

For constructing the ΔubaB mutant, we first made the ΔUbaB construct with the selective marker pyrG from Aspergillus fumigatus, AfpyrG, in the middle of the linear construct (Supplementary Fig. 2). Specifically, we used APYRGF (5′-TGCTCTTCACCCTCTTCGCG-3′) and APYRGR (5′-CTGTCTGAGAGGAGGCACTGA-3′) as primers and the pFNO3 plasmid (Yang et al. 2004) as template to amplify a 1.9-kb AfpyrG fragment. We used UBABNN1 (5′-CCGCTTGCTATAAGATTTACGATC-3′) and UBABNC (5′-CGCGAAGAGGGAGAAGAGCAAGTGCGATCAATAGTCAATAAAGC-3′) as primers and wild-type genomic DNA as template to amplify a 1.1-kb fragment upstream of the UbaB coding sequence. We used UBABCN (5′-ATCAGTGCCTCCTCTCAGACAGTAACGCACGCATATGCGCA-3′) and UBABCC1 (5′-CAACCAGTCGACATTCTCG-3′) as primers and wild-type genomic DNA as template to amplify a 1.1-kb fragment downstream of the UbaB coding sequence. We then used 2 primers: UBABNN (5′-CGCAAGCTCTATTTGTGCGAC-3′) and UBABCC (5′- GCAGCAGGCATTTCAACATCC-3′) to perform a fusion PCR to fuse the 3 fragments and obtained a 3.9-kb fragment, which we transformed into the A. nidulans strain XX357 containing ΔnkuA. Several transformants were obtained that show a “small-colony” phenotype. Homologous integration of the deletion construct was confirmed by PCR using the following 2 pairs of primers: UBABNN1 and AFpyrG3 (5′-GTTGCCAGGTGAGGGTATTT-3′); UBABCC1 and AfpyrG5 (5′-AGCAAAGTGGACTGATAGC-3′) (Supplementary Fig. 2).

Construction of a strain containing the ubaB-GFP allele at the ubaB locus

The construction was done by using standard procedures used in A. nidulans (Yang et al. 2004; Nayak et al. 2006; Szewczyk et al. 2006). For constructing the UbaB-GFP fusion, we used the following 6 oligos to amplify genomic DNA and the GFP-AfpyrG fusion (Yang et al. 2004): UBABGNC (5′-GGCTCCAGCGCCTGCACCAGCTCCGTCATCATCGATCAGAATAGCGC-3′), UBABGNN (5′-CACTGCAGTCCGAAGAAAGC-3′), UBABCN, UBABCC1, GAGAF (5′-GGAGCTGGTGCAGGCGCTG-3′), and pyrG3 (5′-CTGTCTGAGAGGAGGCACTGAT-3′). Specifically, UBABGNC and UBABGNN were used to amplify the 1-kb fragment in the codon region, and UBABCN and UBABCC1 were used to amplify the 1.1-kb fragment in the 3′ untranslated region, using wild-type genomic DNA as template, and GAGAF and PyrG3 were used to amplify the 2.7-kb GFP-AfpyrG fragment using the pFNO3 plasmid DNA as template (Yang et al. 2004). We then used 2 oligos, UBABGNN1 (5′-GGTTCTGGTGTTCGACAAGG-3′) and UBABCC, for a fusion PCR of the 5 fragments to generate the UbaB-GFP-AfpyrG fragment that we used to transform into a wild-type strain containing the ΔnkuA allele (Nayak et al. 2006). The transformants were then screened by microscopically observing the GFP signals in the cells, and the homologous integration of the fusion DNA into the endogenous UbaB locus was confirmed by PCR. Specially, a 1.1-kb product was obtained with the 2 following oligos: AfpyrG5 and UBABCC1.

Statistical analysis

Statistical analyses were done using Prism 9. For the quantitative analysis on colony diameter of wild-type, ubaB^Q247*, ΔubaB, and ΔsumO strains (Fig. 1d), we used ordinary 1-way ANOVA with Tukey's multiple comparisons test. For the quantitative analysis on the percent of nuclei linked by chromatin bridges in wild-type, ubaB^Q247*, ΔubaB, and ΔsumO strains (Fig. 1f), we used ordinary 1-way ANOVA with Tukey's multiple comparisons test. For the quantitative analysis on the percent of interphase and mitotic chromatin bridges in the ΔubaB mutant (Fig. 2b), we used unpaired t-test (2-tailed). For quantitative analyses on the percent of germ tubes containing a chromatin bridge or abnormally shaped nuclei in wild-type, ΔubaB, nimT23, and ΔubaB, nimT23 strains (Fig. 6b and c), we used ordinary 1-way ANOVA with Tukey's multiple comparisons test. For the quantitative analysis on the percent of germ tubes containing different numbers of nuclei in the spore head in wild-type, ΔsumO, and nudA1 strains (Fig. 7b), we used 2-way ANOVA with Tukey's multiple comparisons test. For analyzing the line scans of mCherry-Rab5A (early endosomes) fluorescence intensity in the wild type, the ΔsumO mutant, and the kinA^K895* mutant (Fig. 8b), we used 2-way ANOVA with Tukey's multiple comparisons test.

Results

A loss-of-function mutation in a SUMO-activating enzyme UbaB produces abnormal chromatin bridges

From a UV mutagenesis, we obtained a mutant exhibiting an apparently increased number of binuclei pairs, some of which are connected by chromatin bridges (Fig. 1a), as judged by GFP-labeled Histone H1 (Xiong and Oakley 2009). Binuclei pairs can also be observed in wild-type cells as they represent chromosomes separated after anaphase. However, chromatin bridges are rarely observed in a wild-type control strain, and persistent chromatin bridges are consistent with a defect in segregating the duplicated chromosomes during anaphase (Finardi et al. 2020). To identify the causal mutation, we took an approach combining A. nidulans genetics, whole-genome sequencing, and bioinformatic analysis (see Materials and methods for more details). Using this approach, we identified the causal mutation in AN2450 (using the fungiDB.org gene designation), which encodes UbaB, the key component of the sole SUMO-activating enzyme, a heterodimer of AosA-UbaB in A. nidulans (Johnson et al. 1997; Harting et al. 2013; Horio et al. 2019). UbaB has 610 amino acids, and we identified a C to T mutation that changed the codon CAA (glutamine or Q) at aa247 to the stop codon TAA (Supplementary Fig. 1), thereby removing a big portion of UbaB (Fig. 1b). We named the mutant ubaB^Q247*.

To further address the cellular function of UbaB in A. nidulans, we constructed a ΔubaB mutant in which the coding region of the ubaB gene is deleted (Supplementary Fig. 2). We analyzed the phenotypes of this ΔubaB mutant together with a previously constructed ΔsumO mutant in which the only SUMO-encoding gene (AN1191) in A. nidulans is deleted (Wong et al. 2008). The ubaB^Q247*, ΔubaB, and ΔsumO mutants produced similar colonies that are mildly smaller than the wild-type colony (Fig. 1c and d), and they also appeared to have reduced conidiation (asexual spore production) as shown previously (note that another ΔubaB mutant was constructed previously) (Wong et al. 2008; Horio et al. 2019). This differs from the situation in budding yeast where SUMOylation is essential for colony growth (Dohmen et al. 1995; Johnson et al. 1997; Biggins et al. 2001). Importantly, the ΔubaB and ΔsumO mutants also exhibit chromatin bridges similar to those exhibited in the ubaB^Q247* mutant (Fig. 1e). In all 3 mutants grown in minimal glycerol liquid medium at 32°C (a condition that we use routinely for culturing cells overnight before live imaging), about 10% of the total nuclei are connected by visible chromatin bridges, a percentage that is significantly higher than that in wild type but not significantly different among the 3 mutants themselves (Fig. 1f). These data indicate that ubaB^Q247* is a loss-of-function mutation and that UbaB-mediated SUMOylation is important for the completion of chromosome segregation in A. nidulans.

Chromatin bridges in the SUMOylation-deficient cells can persist into interphase

To examine chromatin bridges during cell cycle progression, we observed both the ubaB^Q247* mutant and the ΔubaB mutant containing Histone-H1-GFP (Xiong and Oakley 2009) and the DsRed-labeled nuclear localization signals (NLS), which labels interphase nuclei (Suelmann et al. 1997; De Souza et al. 2004). Previous work has shown that the NLS-DsRed fusion leaks out of the nuclei to become cytoplasmic due to the partial disassembly of the nuclear pore complex during mitosis (De Souza et al. 2004). Importantly, the position of a sealed septum at mitosis can be inferred by the boundary of the NLS-DsRed signals (Shen et al. 2014). In our wild-type control strain, we were able to see the diffused NLS-DsRed signals during mitosis and the mitotic septal pore sealing just as described previously (Figs. 2a and 3b) (Shen et al. 2014). In the ΔubaB mutant, the vast majority of Histonh-H1-GFP-labeled bridges were surrounded by nuclear NLS-DsRed signals (Fig. 2a and b; Supplementary Fig. 3), indicative of these bridges being at interphase. Figure 2a also shows 1 anaphase bridge during mitosis, which was rarely observed, possibly because anaphase only lasts for a short time, and the bridges can persist into the next interphase, which is a relatively much longer period.

Although the nuclear NLS-DsRed signals are consistent with the nuclei being in interphase, one concern was that the bridges might have changed the nuclear envelope in a way that affects the NLS leakage dynamics. To address this concern, we observed ΔubaB cells containing GFP-tubA that labels microtubules (Horio and Oakley 2005; Xiong and Oakley 2009). We found that nearly 100% of hyphae containing the NLS-DsRed-labeled bridges (n = 27) exhibit cytoplasmic microtubules typical of interphase cells (Fig. 2c; Supplementary Fig. 4), further suggesting that these bridges are in interphase cells. This result also indicates that bridge persistence is not causally linked to a hyperstability of spindle microtubules in these mutants.

To analyze the behaviors of the bridges during the cell cycle, we took time-lapse sequences of the mutants. One time-lapse series of the ubaB^Q247* mutant (with 10-min intervals) shows that interphase nuclei connected by a chromatin bridge can reenter mitosis, as judged by the disappearance of nuclear NLS-DsRed signals (Fig. 3a; Supplementary Video 1). However, from this series, we could not tell whether the bridges in mitotic cells were generated after anaphase chromosome segregation or represent the ones that persisted during mitosis (especially because nuclei can migrate in and out of frames). To better understand the behaviors of bridges during the cell cycle, we analyzed more time-lapse sequences with shorter intervals (5 or 2.5 min), starting with interphase bridges. In the majority of the cases (12 out of 17), interphase nuclei connected by bridges were observed to enter mitosis (Supplementary Videos 2 and 3). In 2 out of 17 cases, bridges seemed to be resolved (or broken) prior to mitotic entry (Supplementary Video 4). These results indicate that most bridges persist till mitotic entry. The bridges seemed quite diverse in shape and behavior, even when they were in the same hypha. Supplementary Video 2 shows an example of nuclei connected by bridges entering mitosis. After mitotic entry (as judged by the disappearance of nuclear NLS-DsRed signals), the bridge connecting the pair of nuclei on the right seemed to be resolved, while another pair of nuclei (on the left) connected by a much shorter bridge coalesced into one mass. Each pair then produced 4 daughter nuclei after anaphase chromosome segregation, consistent with the G2 identity of nuclei connecting the bridges prior to mitosis. The time-lapse sequences also show that interphase bridges are quite dynamic, as the connected nuclei can often be seen to join together and then separate again (Supplementary Videos 2 and 3). In addition, Supplementary Videos 4 and 5 show that new mitotic bridges formed during anaphase chromosome segregation can persist into the next interphase.

In one time-lapse sequence, we also detected a septum on a chromatin bridge during mitosis, as judged by the appearance of cytoplasmic NLS-DsRed signals only at one side of the septum (Fig. 3b; Supplementary Fig. 5 and Video 1) (Shen et al. 2014). That the NLS-DsRed signals are only present at one side is a strong indication that the septum pore is sealed (Shen et al. 2014).

Finally, we sought to determine if nuclear pore complexes are present on the chromatin bridges enclosed by nuclear envelope. To this end, we observed two previously studied nucleoporins (Nups), Nup96, which is associated with the nuclear pore complexes during both interphase and mitosis, and Nup49, a peripheral Nup that is dispersed during mitosis (Osmani et al. 2006). We found that both Nups label the chromatin bridges in almost 100% cases (n > 20 for both Nups) (Fig. 4).

The SUMO-activating enzyme UbaB localizes to nuclei during interphase

To observe UbaB localization, we constructed a strain containing a ubaB-GFP fusion gene integrated at the ubaB locus. UbaB-GFP localized to the nucleus in interphase cells (Fig. 5). However, the nuclear UbaB-GFP signals were not observed during mitosis, likely because of dispersal of the UbaB-GFP from nuclei when the nuclear pores partially disassemble (De Souza et al. 2004), and they only reappeared after the formation of daughter nuclei when the NLS-DsRed signals also reappeared (Fig. 5). This localization pattern is similar to the previously studied SUMO-GFP (Wong et al. 2008) and is consistent with many SUMO targets in A. nidulans being nuclear proteins including topoisomerase II (Horio et al. 2019). Because a defect in SUMOylation of topoisomerase II has been implicated in chromatin bridges in mammalian cells (Dawlaty et al. 2008), the chromatin bridge phenotype of our ubaB mutants could be explained by compromised topoisomerase II activity that leads to defect in resolving DNA catenanes. That UbaB-GFP appears quickly in the daughter nuclei upon postmitotic assembly of the nuclear pore complexes (De Souza et al. 2004) suggests that the proteins are imported robustly into the nuclei from the cytoplasm after mitosis.

SUMOylation deficiency causes a severe defect in the first nuclear division at a higher temperature of 42°C

To address the effect of SUMOylation deficiency on the first mitosis, we used the temperature-sensitive nimT23 (cdc25) mutant that is blocked in late G2 at the restrictive temperature of 42°C (O'Connell et al. 1992) and compared the nimT23 single mutant with a ΔubaB, nimT23 double mutant after shifting cells to the permissive temperature of 25°C. After a 6-h incubation at 42°C, germ tubes of both mutants showed a single G2 nucleus (labeled by Histone-H1-GFP) (Fig. 6a), while most wild-type germ tubes contained 2 nuclei (65%, n = 40) and some contained 4 nuclei (35%, n = 40). Within 0.5–1 h after the nimT23 mutants were shifted from 42°C to room temperature, 2 nuclei appeared in most germ tubes regardless of the ΔubaB allele, although ∼10% of the ΔubaB, nimT23 germ tubes had a chromatin bridge connecting the 2 nuclei (Fig. 6a and b). As a control, we also used the ΔubaB single mutant. Astonishingly, >50% (n = 72) of the ΔubaB germ tubes at 42°C contained an abnormally shaped single nucleus that looked much longer than a normal nucleus, and this abnormal pattern was still present after the cells were shifted to room temperature for 0.5–1 h. The abnormal single nucleus was also seen in some ΔubaB, nimT23 germ tubes, but the number was significantly lower than that in the ΔubaB single mutant (Fig. 6c). The abnormally shaped single nucleus was also observed in 60% (n = 35) of the ΔsumO germ tubes, and this number was reduced to only 17% (n = 61) in the ΔsumO, nimT23 mutant in a similar temperature-shift experiment. These abnormally shaped nuclei are mostly at interphase because ∼100% of them (n > 30) were decorated by Nup49-GFP (Fig. 6d), which is known to dissociate from the nuclear pores during mitosis (Osmani et al. 2006).

After a longer-time incubation (>8.5 h at 42°C), chromatin bridges (rather than the single abnormal nucleus) became a dominant feature in ΔubaB cells, as >50% (n = 59) of the ΔubaB germ tubes contained bridges (Supplementary Fig. 6). Nevertheless, colony size of the mutants was only mildly affected at 42°C (Supplementary Fig. 6), providing a strong support to the notion that multinucleated fungi can tolerate chromosome segregation defects during hyphal growth. Similar to what has been previously described for the ΔsumO mutant (Wong et al. 2008), the mutant colony obviously lacks pigmented asexual spores (conidia). Thus, compared to hyphal growth, the process of asexual spore formation seems less tolerant to chromosome segregation defects. This is likely due to the following: while functionally defective nuclei can compensate for each other in a multinucleated compartment during hyphal growth (Mela and Glass 2023), asexual spore production in A. nidulans consists a series of budding and nuclear division events similar to yeast budding, with daughter nuclei being delivered to the uninuclear conidia via nuclear division and positioning (Fischer and Timberlake 1995; Adams et al. 1998; Peñalva et al. 2012).

Loss of SUMOylation does not apparently affect dynein function in A. nidulans

In filamentous fungi, dynein and its activator LIS1 are critical for nuclear distribution and the transport of early endosomes and hitchhiking cargoes such as peroxisomes (Plamann et al. 1994; Xiang et al. 1994; Xiang et al. 1995; Lenz et al. 2006; Abenza et al. 2009; Zekert and Fischer 2009; Egan et al. 2012; Guimaraes et al. 2015; Salogiannis et al. 2016; Qiu et al. 2019; Müntjes et al. 2021; Salogiannis et al. 2021; Christensen and Reck-Peterson 2022). As shown earlier (Fig. 6a), after shifting the ΔubaB, nimT23 double mutant to room temperature, nuclei were able to migrate into the germ tube, unlike those in the nudA1 (dynein heavy chain), nimT23 double mutant (Fig. 6a). To more carefully address whether loss of SUMOylation affects nuclear distribution, we quantified dynein-mediated nuclear distribution using a ΔsumO mutant and a wild-type strain grown overnight at 32°C. A nudA1 (dynein heavy chain) single mutant was used as a control. Although the restrictive temperature for the nudA1 mutation is 42°C, a nuclear distribution phenotype of the mutant is still obvious at a lower temperature. In A. nidulans, when conidia (asexual spores) germinate, they swell and a germ tube extends. The spore is visible as a swelling in the germling. After overnight growth at 32°C, although some nuclei in the nudA1 mutant distributed along the germ tube, multiple nuclei were clustered in the spore swelling (Fig. 7a and b). In contrast, the ΔsumO mutant contains mostly 1 nucleus in the spore swelling while other nuclei distributed along the germ tube just like in the wild-type strain (Fig. 7a and b). Thus, loss of SUMOylation does not apparently affect dynein-mediated nuclear migration into the germ tube.

As a defect in dynein or LIS1 function causes early endosomes to accumulate abnormally near the hyphal tip, early endosome distribution is an excellent readout of dynein or LIS1 function (Lenz et al. 2006). Compared to dynein-mediated nuclear distribution, early endosome distribution in A. nidulans appears to be even more sensitive to dynein abnormalities (Tan et al. 2014). Similar to mutations in dynein or LIS1, kinesin-1 mutations also affect dynein-mediated early endosome transport as kinesin-1 delivers dynein to the microtubule plus end to receive the early endosome cargo (Zhang et al. 2003; Lenz et al. 2006; Abenza et al. 2009; Zekert and Fischer 2009; Zhang et al. 2010; Egan et al. 2012; Qiu, Zhang, and Xiang 2023). As a control for comparing with the SUMOylation-deficient mutants, we used a recently studied kinesin-1 mutant, kinA^K895*, which exhibits only a mild nuclear distribution defect and grows better than the nudA1 mutant but shows an obvious hyphal-tip accumulation of early endosomes at 32°C (Qiu, Zhang, and Xiang 2023) (Fig. 8a and b). In both the ubaB^Q247* and ΔsumO mutants, we have not detected any defect in the distribution of early endosomes labeled by mCherry-Rab5A (formerly called RabA) (Pinar and Peñalva 2021) (Fig. 8a). Specifically, loss of SUMOylation does not cause any obvious hyphal-tip accumulation of early endosomes (Fig. 8a and b). Thus, SUMOylation does not play an important role in dynein-mediated early endosome transport in A. nidulans.

Discussion

In this study, we isolated a UV mutagenesis-generated A. nidulans mutant exhibiting chromatin bridges, and we identified the causal mutation (ubaB^Q247*) in the ubaB gene encoding a key subunit of the SUMO-activating enzyme. The ubaB^Q247*, ΔubaB, and ΔsumO mutants all exhibit abnormal chromatin bridges (Fig. 1), demonstrating the importance of SUMOylation in the completion of anaphase chromosome segregation in a multinucleated filamentous fungus.

Anaphase DNA bridges are known to be caused by various factors including telomere fusion and a failure in resolving DNA catenanes in a timely manner (Dawlaty et al. 2008; Maciejowski et al. 2015; Nielsen et al. 2015; Finardi et al. 2020; Umbreit et al. 2020). SUMOylation or SUMO-interaction has been known to affect telomeres as well as enzymes required for resolving DNA catenanes such as topoisomerase (Tanaka et al. 1999; Dawlaty et al. 2008; Ryu et al. 2010; Miyagawa et al. 2014; Yalçin et al. 2017; Pandey et al. 2020; Zhang et al. 2020; Zhang et al. 2021; Hertz et al. 2023). However, there is no obvious SUMO target involved in telomere integrity in A. nidulans, while both Topo I (AN0253) and Topo II (AN5406) are SUMO targets (Horio et al. 2019), making it possible that the bridges are caused by a failure in a timely decatenation of replicated DNAs. Indeed, a deficiency in RanBP2-mediated SUMOylation of Topo II alpha gives rise to anaphase bridges in mammalian cells (Dawlaty et al. 2008). Possibly, SUMOylation of Topo II alpha allows it to accumulate at inner centromeres to decatenate sister centromeres before anaphase (Dawlaty et al. 2008), although recent studies have implicated more factors affected by Topo II alpha SUMOylation at the C-terminal domain (Clarke and Azuma 2017). Other SUMO targets in A. nidulans include 2 subunits of the condensin complex, subunit 3 (AN8056) and SMC2 (AN5899) (Horio et al. 2019), which is relevant because a condensin defect is involved in chromatin bridge formation (Hartl et al. 2008), and SUMOylation has been implicated in regulating condensin (Stephens et al. 2015). However, SUMOylation of condensin subunits promotes condensin turnover (Psakhye and Branzei 2021). Thus, condensin stability may be increased in SUMOylation-deficient cells, which may affect chromosome decondensation but seems unlikely to cause anaphase bridge formation. Future work will be needed to identify the exact SUMO target(s) whose lack of SUMOylation results in the chromatin bridges.

While chromatin bridges are known to be formed during anaphase chromosome segregation (Finardi et al. 2020), the vast majority of the SUMOylation deficiency-caused bridges in A. nidulans appear to persist into interphase. This was suggested by the nuclear localization of the NLS-DsRed signals and presence of cytoplasmic microtubules typical of interphase cells (Fig. 2). By taking time-lapse sequences, we found that a majority of the bridges persist until the next mitotic entry (Fig. 3; Supplementary Videos 1–3). Anaphase chromosome segregation can produce new bridges, which can persist into the next interphase (Supplementary Videos 4 and 5). To examine whether the bridges are enclosed by nuclear envelope with nuclear pore complexes, we observed 2 previously studied Nups, Nup49 and Nup96 (Osmani et al. 2006). Almost all the chromatin bridges were decorated by both Nup49 and Nup96 (Fig. 4). Since Nup49 is a peripheral Nup that is dispersed during mitosis (Osmani et al. 2006), this observation further indicates that the bridges are in interphase cells. At the end of A. nidulans mitosis, daughter nuclei are generated after nuclear membrane restriction and fission, but 2 restrictions (instead of 1) occur first to generate 3 nuclear membrane-enclosed compartments, with the old nucleolus residing in the middle compartment (Ukil et al. 2009). This is followed by the disassembly of the old nucleolus into the cytoplasm and the subsequent reassembly of the new nucleolus in the G1 nucleus (Ukil et al. 2009). Although Nup96 signals can be seen to enclose the middle compartment with the old nucleolus at anaphase (Ukil et al. 2009), we have never caught a bridge with a third compartment in the middle. Instead, in many nuclei connected by bridges, a nucleolus-like small region devoid of Histone-H1 signals can be seen, which is also consistent with these nuclei being at interphase.

Our finding that chromatin bridges can persist into interphase is consistent with observations made in mammalian cells (Steigemann et al. 2009; Maciejowski et al. 2015; Pampalona et al. 2016; Umbreit et al. 2020). In both S. cerevisiae and mammalian cells, an Aurora B kinase-mediated NoCut pathway or abscission checkpoint has been proposed to delay abscission and prevent premature cutting of DNAs in the chromatin bridges during the final step of cytokinesis (Norden et al. 2006; Steigemann et al. 2009; Carlton et al. 2012) (although Aurora B inhibition in the presence of bridges prevents the furrow regression rather than abscission, Steigemann et al. 2009). In filamentous fungi, a process similar to cytokinesis is septation (although septation only occurs after a few rounds of mitosis after spore germination, Fiddy and Trinci 1976; Harris et al. 1994). Interestingly, septation in a multinucleated filamentous fungus such as A. nidulans always leaves a hole (or septal pore) in the middle of a septum, which would be transiently sealed upon the next mitotic entry (Shen et al. 2014). This pore sealing depends on the NimA kinase required for mitotic entry, and it is most likely the reason why mitosis is only synchronous within the same hyphal compartment (Shen et al. 2014). In this study, we detected septal pore sealing on a bridge. This is a rare event caught by serendipity (especially because mitotic cells are very rare (<5%) in a cell population and most bridges are in interphase), but it argues against the presence of an effective checkpoint that would prevent bridge cutting in these mutants. In A. nidulans, loss of a core Nup also causes the formation of chromatin and nuclear membrane bridges (Chemudupati et al. 2019). Specifically, it causes the old nucleolus to be removed early, thereby changing the “double-restriction” pattern of nuclear division to a dumbbell-shaped nuclear envelope morphology of anaphase nuclei connected by a bridge (Chemudupati et al. 2019). Interestingly, these bridges are accompanied by hyperstable spindles, and the only A. nidulans Aurora kinase (De Souza et al. 2017) localizes to these bridges until the bridges are resolved (Chemudupati et al. 2019). In S. cerevisiae, hyperstability of mitotic spindles has been causally linked to Aurora B-mediated bridge monitoring implicated in “NoCut” (Amaral et al. 2016). Since septal pore sealing in A. nidulans happens upon mitotic entry prior to the formation of an elongated spindle, there may not be a similar monitoring mechanism at this stage. However, telophase nuclear fission does not happen in the presence of a bridge, similar to what has been observed previously (Chemudupati et al. 2019). As proposed previously, a karyokinesis checkpoint may prevent nuclear fission in the presence of a chromatin bridge or other mitotic defects that occurred after the spindle assembly checkpoint has been satisfied (Markossian et al. 2015; Chemudupati et al. 2019). Alternatively, the bridge may interfere with the normal “double-restriction” pattern of nuclear fission that happens during telophase in A. nidulans (Ukil et al. 2009).

In mammalian cells, chromatin bridges lead to chromothripsis, which is partly due to bridge-initiated micronuclei formation in cells of the second generation (Umbreit et al. 2020). In the SUMOylation-deficient A. nidulans cells, although we have observed structures resembling lagging chromosomes or chromatin fragments in the middle of the dividing nuclei during the anaphase subsequent to an interphase with bridges (Supplementary Videos 2–5), micronuclei were not routinely found in interphase hyphae (Supplementary Videos 2 and 3). This could be due to the difference between closed and open mitosis. In mammalian cells, a new nuclear envelope forms on both the primary nuclei and micronuclei after mitosis (Liu et al. 2018; Zhao et al. 2023). During mitosis in A. nidulans, the old nuclear envelope remains intact despite the partial disassembly of nuclear pore complexes (De Souza et al. 2004; Osmani et al. 2006), and micronuclei formation would depend on nuclear membrane fission between the primary- and micronuclei, which may not always happen successfully. Alternatively, our current method might not be sensitive enough to detect micronuclei in fungal hyphae, and future experiments will be needed to better address this issue.

SUMOylation deficiency significantly delays the first nuclear division, and this problem was highly noticeable at 42°C, which produces abnormally shaped single interphase nuclei in ∼50% of the germ tubes (Fig. 6). Compared to the bridged nuclei, these single nuclei may be due to a more severe defect in chromosome separation, for example, the chromosomes may be held together more tightly or the bridged daughter nuclei may coalesce more effectively. Time-lapse imaging at 42°C will be needed in future to examine the formation and dynamics of these nuclei. This kind of abnormal nuclei were also seen at 32°C in ΔubaB germ tubes, but the incidence is much lower (∼19%, n = 59), possibly due to a slower speed of growth (note that a ∼8.5-h incubation at 32°C in the same liquid medium produces germ tubes similar to those produced after ∼6 h at 42°C). It is also interesting to note that the nimT23 (cdc25) mutation blocking the cell cycle at G2 (O'Connell et al. 1992) reduces the incidence of either abnormal nuclei or chromatin bridges in SUMOylation-deficient cells (Fig. 6). Possibly, an extended G2 helps resolve problems made during DNA replication and better prepare the cells for anaphase chromosome segregation. Alternatively, as the nimT23-containing cells enter mitosis after they are shifted to room temperature, the lower temperature may produce a less severe defect in anaphase chromosome segregation. Future work will be needed to examine these possibilities. As the SUMO-pathway has been pursued as cancer drug targets (Kroonen and Vertegaal 2021), it would be worthwhile to study factors modulating the mitotic defects in SUMOylation-deficient cells.

Finally, we showed that SUMOylation is not critical for LIS1-mediated dynein activation in A. nidulans. Whether SUMOylation regulates LIS1 still needs to be tested in yeast where LIS1 interacts with proteins in the SUMOylation pathway (Alonso et al. 2012). Yeast dynein is almost exclusively required for nuclear migration/spindle orientation (Yeh et al. 1995; Winey and Bloom 2012). Prior to being offloaded with dynactin to the cortex to bind the cortical adapter Num1 (Omer et al. 2018), dynein is accumulated at the microtubule plus end in a LIS1-dependent but dynactin-independent fashion (Lee et al. 2003; Sheeman et al. 2003). This differs from A. nidulans where the plus-end dynein accumulation depends on dynactin but not LIS1 (Zhang et al. 2003).

Data availability

The authors affirm that all data necessary for confirming the conclusions of this article are represented fully within the article, its figures, and its supplemental data including Supplementary Table 1, 6 supplementary figures, and 5 supplementary videos. Supplementary File 1 contains detailed descriptions of all the Supplementary data. Strains used and created in this work (listed in Supplementary Table 1) are available upon request. In addition, the strains have also been submitted to the FGSC, and the FGSC numbers are shown in Supplementary Table 1.

Supplemental material available at GENETICS online.

Funding

This work was funded by the National Institute of General Medical Sciences (grant R35GM140792 to XX and grant 1R15GM132869 to MJE) and the Irving S. Johnson Fund of the University of Kansas Endowment (to BRO).