Nuclear genome organization in fungi: from gene folding to Rabl chromosomes

Abstract Comparative genomics has recently provided unprecedented insights into the biology and evolution of the fungal lineage. In the postgenomics era, a major research interest focuses now on detailing the functions of fungal genomes, i.e. how genomic information manifests into complex phenotypes. Emerging evidence across diverse eukaryotes has revealed that the organization of DNA within the nucleus is critically important. Here, we discuss the current knowledge on the fungal genome organization, from the association of chromosomes within the nucleus to topological structures at individual genes and the genetic factors required for this hierarchical organization. Chromosome conformation capture followed by high-throughput sequencing (Hi-C) has elucidated how fungal genomes are globally organized in Rabl configuration, in which centromere or telomere bundles are associated with opposite faces of the nuclear envelope. Further, fungal genomes are regionally organized into topologically associated domain-like (TAD-like) chromatin structures. We discuss how chromatin organization impacts the proper function of DNA-templated processes across the fungal genome. Nevertheless, this view is limited to a few fungal taxa given the paucity of fungal Hi-C experiments. We advocate for exploring genome organization across diverse fungal lineages to ensure the future understanding of the impact of nuclear organization on fungal genome function.


Introduction
In eukary otes, the DN A within the nucleus is or ganized as c hr omatin, a dynamic complex formed by DNA in association with proteins (Campos and Reinberg 2009 ). Approximately 146 bp of DNA is wr a pped ar ound an octamer of histone proteins, two copies of histones H2A, H2B, H3, and H4, that together form the nucleosome core particle (Luger et al. 1997 ), which represents the smallest organizational unit of chromatin (Campos and Reinberg 2009 ). DNA and histone proteins are subjected to a plethora of chemical modifications such as, but not limited to, methylation or acetylation of lysine residues in histone proteins, and specific combinations of these modifications-the histone codehave been associated with the formation and maintenance of, as well as dynamic transitions between, different local chromatin organization (Strahl andAllis 2000 , Jenuwein andAllis 2001 ). Depending on its local compaction state, c hr omatin can be br oadl y subdivided into two types where the DNA is more (euchromatin) or less (heter oc hr omatin) accessible to the transcriptional mac hinery, ther eby dir ectl y influencing gene expression (Bannister andKouzarides 2011 , Allshire andMadhani 2018 ). Euchromatin is considered to be transcriptionally active, while heterochromatin is typically transcriptionally silent (Jenuwein and Allis 2001 ). Het-er oc hr omatin can be further subdivided into constitutive heter oc hr omatin, i.e. lar gel y de void of genes, and facultativ e heter oc hr omatin that typically contains genes that are transcriptionall y r epr essed during specific de v elopmental or envir onmental conditions (Trojer and Reinberg 2007, Saksouk et al. 2015, Wang et al. 2016, Allshire and Madhani 2018.
Local c hr omatin can be folded in the thr ee-dimensional (3D) space of the nucleus to structur all y or ganize and pac ka ge DNA, while also enabling precise gene expression (Lieberman-Aiden et al. 2009, Sexton and Cavalli 2015, Bonev and Cavalli 2016, Hoencamp et al. 2021. For instance, higher-order c hr omatin structures allow the spatial association of genomic elements that are physicall y separ ated on the linear DNA strand or occur on differ ent c hr omosomes , and con v ersel y suc h structur es can physicall y segr egate nearby genomic sites thr ough specific folding barriers (Tolhuis et al. 2002, West and Fraser 2005, Lieberman-Aiden et al. 2009 ). Consequently, the disruption of DNA folding can result in ectopic interactions of genes and their regulatory regions (Sexton and Cavalli 2015, Bonev and Cavalli 2016, Krumm and Duan 2019, for instance, by hijacking enhancers as observed for proto-oncogenes (Northcott et al. 2014, Flavahan et al. 2016. Consequently, it is important to understand how the 3D genome organization in eukaryotes is established, how it affects genome function, and how this c hr omatin or ganization and the pr ocesses that establish it ar e conserv ed between eukaryotes. While m uc h about genome organization and its function is known in a few animal and plant model systems, there is a dearth of knowledge about these processes in most eukaryotes, including fungi. The fungal kingdom is estimated to contain millions of different species (Blackwell 2011 ). Fungi are important components of worldwide ecosystems as decomposers of organic material deriv ed fr om plants and animals, and man y hav e been exploited for decades in food production and biotechnology (Stajich et al. 2009 ). Many fungi are also symbionts that can engage in mutualistic to par asitic inter actions with other or ganisms, whic h can hav e significant economic and ecological impact (Fisher et al. 2012(Fisher et al. , 2020. For instance, in a gricultur e, plant pathogenic fungi can cause devastating epidemics in staple and commodity crops necessary for the survival of billions of humans (Fisher et al. 2020 ). Importantly, many fungi are also outstanding model systems for higher eukaryotes, including humans; essential eukaryotic functions are often conserved in fung i, yet fung i ar e typicall y mor e simplistic, geneticall y tr actable, and r esearc h is cost-efficient. In the last decade, r esearc h into c hr omatin and the or ganization of fungal genomes has lar gel y focused on the bakers' yeast Sacc harom yces cerevisiae and fission yeast Sc hizosacc harom yces pombe , giv en their ease of study and the plethora of available mutant strains and genetic tools. Due to the ecological and economic r ele v ance of fungi, it is important to further advance research into the function of fungal genomes with the goal of elucidating the regulation of gene expression and discovering how this regulatory control is influenced by nuclear genome organization. Here, we revie w the curr ent knowledge of fungal genome or ganization, organized hier arc hicall y fr om topological structur es at the le v el of individual genes through the association of c hr omosomes in the nucleus to form the Rabl c hr omosome conformation, in whic h centr omer es and telomer es cluster distinctl y at the nuclear periphery (Box 1). Recent work in fungal model systems and increasingly in more diverse fungi has started to elucidate pertinent subnuclear c hr omatin structur es that will be pivotal to our understanding of the function and conservation of genome organization and will provide the framework to detail how nuclear processes impact genome functions in fungi. To date, only a few highresolution fungal Hi-C datasets are a vailable , but an increasing number of datasets with lo w er resolution ( > 5 kb) from a wider diversity of fungi now enable researchers to draw general conclusions about hier arc hical 3D structur es observ ed that or ganize the fungal genome.

The composition of fungal genomes
Over the last few decades, advances in genome sequencing technologies hav e pr ovided a wealth of novel insights about the genome composition of fungi, which is a pr er equisite to study nuclear genome organization in detail. Yeasts are arguably the most well-studied fungal model organisms (Botstein andFink 2011 , Liti 2015 ), in part due to their r elativ el y small and well-c har acterized genomes . For example , the genome of the budding yeast S. cerevisiae is ∼12 Megabases (Mb) in size divided into 16 c hr omosomes, while the fission yeast S. pombe has a 13 Mb genome divided over onl y thr ee c hr omosomes (Goffeau et al. 1996, Wood et al. 2002, Engel et al. 2014. Relative to the more simplistic yeasts, the genomes of filamentous fungi are typically larger and more complex, but v astl y smaller than those of higher metazoans whose genomes often comprise billions of base pairs. Specifically, the genomes of filamentous fungi are typically 30-50 Mb in size, divided into variable c hr omosome numbers (Mohanta and Bae 2015 ), with se v er al species' genomes being m uc h lar ger (Kir an et al. 2016, Porto et al. 2019. Se v er al examples with nearly complete genome assemblies of more well-studied fungal organisms include the sa pr ophyte and model organism Neurospora crassa [41 Mb divided into se v en Linka ge Gr oups (LG) or c hr omosomes], the soil-borne plant pathogen Verticillium dahliae (36 Mb divided into eight c hr omosomes), the human pathogens Aspergillus fumigatus (28 Mb across eight c hr omosomes), and Cryptococcus neoformans (19 Mb across fiv e c hr omosomes), the wheat head blight fungus Fusarium graminearum (36 Mb across four chromosomes), and the grass endophyte Epichloë festucae (35 Mb across seven chromosomes) (Galagan et al. 2003, Loftus et al. 2005, Nierman et al. 2005, Faino et al. 2015, King et al. 2015, Winter et al. 2018, Bo wy er et al. 2022. Fungal genomes are w ell-kno wn to be highly dynamic with discr ete r egions enric hed with pol ymor phisms and c hr omosomal r earr angements (Raffaele and Kamoun 2012, Dong et al. 2015, Möller and Stukenbr oc k 2017, Torr es et al. 2020 ). These variable r egions ar e often embedded in differ ent c hr omosomes or e v en comprise complete c hr omosomes . For instance , r egions in pr oximity to telomeres can be highly variable in budding yeast, Aspergillus , Neurospora , or Magnaporthe species where these regions display frequent presence/absence of variation as well as chromosomal r earr angements (Farman and Kim 2005, Farman 2007, McDona gh et al. 2008, Br own et al. 2010, Chang and Ehrlich 2010, Starnes et al. 2012, Jamieson et al. 2013, Yue et al. 2017. Other species like V. dahliae contain similar ada ptiv e genomic regions (AGRs) that are embedded within different chromosomes and contain in planta expressed genes (de Jonge et al. 2012, Faino et al. 2016. Entire chromosomes are variable within strains of Fusarium oxysporum or Zymoseptoria tritici , as they contain genes important for the biology of these fungi (van Dam et al. 2017, Möller et al. 2019. It is becoming incr easingl y evident that dynamic regions in fungal genomes are often composed of highl y r e petiti v e r elicts of tr ansposable elements Stukenbr oc k 2017 , Fr antzeskakis et al. 2019 ). Consequently, these regions are known to have a high number of adenine and thymine (AT) base pairs (bp), which lo w er the overall guanine and cytosine (GC; can include G:C base pairs or GpC/CpG dinucleotides in a single strand) content of fungal genomes from an a ppr oximately 50% GC bp composition. These AT-ric h r egions, termed AT-isoc hor es to specificall y delineate small genomic regions depleted of GC bp (Testa et al. 2016 ), can be interspersed in the core c hr omosomes or localized on accessory c hr omosomes. For example, one-third of the genome of the fungus Leptosphaeria maculans , a Dothideomycete pathogen of the canola plant crops for producing r a peseed oil, contains blocks of dense AT-rich sequences deriv ed fr om tr ansposable elements that often house virulence genes (Fudal et al. 2007, Rouxel et al. 2011. Similarly, dense ATand r epeat-ric h r egions ar e distributed thr oughout the E. festucae genome, comprising r oughl y 25% of its genome (Testa et al. 2016, Winter et al. 2018, and ∼16% of the N. crassa genome is interspersed with AT-rich, yet gene poor isoc hor es ( Fig. 1 ) (Selker et al. 2003, Lewis et al. 2009, Testa et al. 2016 ). Man y AT-ric h sequences in fungal genomes are derived from the action of repeat-induced point-mutation (RIP) (Selker and Stevens 1987, Selker 1990, Fr eita g et al. 2002, where the GC nucleotides in duplicated or repeated sequences ar e heavil y m utated to contain numer ous AT tr ansition m utations, ther eby inactiv ating the underl ying sequence and increasing the local AT-content. RIP has been shown to be active in a plethora of fungal species (Cambareri et al. 1991, Clutterbuck 2011, Hane et al. 2015, Gladyshev 2017, and specifically in Neurospora , RIP'd transposable element relicts comprise essentially most of the AT-rich sequences across the genome (Lewis et al. 2009 ) (Fig. 1 ).

Conserved heter oc hr oma tic region fea tures in fungal genomes
AT-ric h r e petiti v e isoc hor es dir ect the formation of the silent constitutiv e heter oc hr omatin (Miao et al. 2000, Lewis et al. 2009 ). As initially elucidated in Neurospora , AT-rich isochores, by an as of y et unkno wn mec hanism, r ecruit the histone methyltr ansfer ase complex DCDC ( D IM-5/-7/-9 C UL4 D DB1 dim-8 C omplex) (Miao et al. 2000, Lewis et al. 2009, 2010a, Rountree and Selker 2010, Courtney et al. 2020. DCDC catalyzes the tri-methylation of lysine 9 on histone 3 (H3K9me3) of nucleosomes within AT isoc hor es ( Fig. 1 ), with KMT1 DIM-5 (Lysine [ K ] M ethyl T r ansfer ase-1 / D efectiv e I n M ethylation-5 ) being the SET domain-containing subunit with histone methyltr ansfer ase catal ytic activity (Le wis et al. 2010a , b , Fr eita g 2017 ). H3K9me3 is bound by HP1 ( H eter oc hr omatin P rotein-1 ) to directly recruit the DNA methyltransferase DIM-2 for methylation on the fifth intracyclical atom of cytosine bases (5 m C) (Nielsen et al. 2002, Fr eita g et al. 2004, Honda and Selker 2008. Consequently , AT -rich isochores are readily observable across fungal c hr omosome sequences as they a ppear as peaks of enric hment of H3K9me3 (Fig. 1 ) and cytosine methylation interspersed among the gene-ric h euc hr omatin (Le wis et al. 2009 ). One important subclass of AT-rich isochores in fungal genomes are centr omer es, whic h ar e critical for c hr omosomal function as they facilitate homologous c hr omosome pairing during mitosis/meiosis (Stewart and Dawson 2004, Wells et al. 2006, Krassovsky et al. 2012, Kurdzo et al. 2017, Yadav et al. 2018a , Pr e viato de Almeida et al. 2019 ). Chromosomes of the budding yeast S. cerevisiae contain a "point" centr omer e that comprises ∼200 bp in a single nucleosome, while most other fungi have "regional" centromeres spanning thousands of base pairs (Malik and Henikoff 2009, Smith et al. 2012, Lefrançois et al. 2013, Yadav et al. 2018a, b , Seidl et al. 2020 ). Similar to AT-rich isochores, centr omer es ar e densel y enric hed for H3K9me3 and 5 m C (Le wis et al. , Smith et al. 2012, Seidl et al. 2020 ), yet ar e differ entiated fr om other AT-ric h tr ansposon relicts by the deposition of the centr omer e-specific histone variant CenH3 (Meluh et al. 1998, Malik and Henikoff 2009, Smith et al. 2012, Yadav et al. 2018a, b , Seidl et al. 2020. Facultativ e heter oc hr omatin can dynamicall y tr ansition between being densely compacted, thereby silencing the expression of the underlying DNA, to looser compaction resulting in c hr omatin being mor e open for activ ating tr anscription (Tr ojer and Reinberg 2007 ). Facultative heterochromatin in fungi is typically delineated by either the di-or tri-methylation of lysine 27 on histone H3 (H3K27me2/3) ( Fig. 1 ) catalyzed by PRC2 ( P ol ycomb R epr essiv e C omplex-2 ; the SET domain protein KMT6 SET-7 is the subunit that specifically methylates histones) (Fr eita g 2017 , Wiles andSelker 2017 ). This facultative heterochromatin mark is densely enriched over predominantly gene-rich regions of the genome (Fig. 1 ) to r epr ess gene expr ession (Jamieson et al. 2013, Basenko et al. 2015, Dumesic et al. 2015. In N. crassa , H3K27me2/3 is enriched over subtelomeric chromosomal regions that are positionally dependent on the presence of telomer e r epeats, although c hr omosome-internal, position independent peaks of H3K27me2/3 ar e r eadil y observ ed (Jamieson et al. 2013, 2016, Basenko et al. 2015, Klocko et al. 2016. In other fungi, lar ge str etc hes of H3K27me3 enric hment can be observ ed along cor e c hr omosome arms (Connoll y et al. 2013, Sc hotanus et al. 2015, Carlier et al. 2021, So y er et al. 2021, Kramer et al. 2022. For example, the fusion of multiple subtelomeric domains, which form the four chromosomes in F. graminearum , result in large chromosome-internal domains of H3K27me3 (Connolly et al. 2013 ), while H3K27me3 represses genes in the dynamic AGRs regions in V. dahliae , Kramer et al. 2022 ) and across accessory chromosomes in Z. tritici or F. oxysporum (Schotanus et al. 2015, Fokkens et al. 2018. Interestingly, the deposition of the H3K27me2/3 mark varies with the protein subunits associated with PRC2, as loss of PRC2-associated proteins NPF ( Neurospora p55 ortholog) or PAS ( P RC2 A ccessory S ubunit) abolishes subtelomeric H3K27me2/3 in N. crassa , while the conserv ed pr otein EPR-1 is known to "read" the H3K27me2/3 mark for gene r epr ession; in support, deletion of the EPR-1 homolog in F. graminearum , BP1, phenocopies loss of KMT6 and BP1 dir ectl y binds DNA to stabilize nucleosomes for enhancing transcriptional r epr ession (Jamieson et al. 2013, Mc-Naught et al. 2020, Tang et al. 2021. Genomic regions that display properties of both constitutive and facultative heter oc hr omatin ar e typicall y not observ ed  ), yet some species-specific exceptions occur. For instance, the genome of Z. tritici contains large domains enriched for both H3K27me2/3 and H3K9me3, while only a few larger interspersed heter oc hr omatic r egions in V. dahliae or onl y the telomer es of N. crassa ar e enric hed for both marks (Schotanus et al. 2015 ). In addition, a ne wl y emer ging yet understudied facultativ e heter oc hr omatic mark r equir ed for r epr essing gene expr ession in N. crassa and Fusarium fujikuroi is di-or tri-methylation of lysine 36 on histone H3 (H3K36me2/3), which is catalyzed by the SET-domain containing histone methyltr ansfer ase ASH1 (Bicocca et al. 2018, Janevska et al. 2018. In N. crassa , ASH1-specific H3K36me2 is enriched over lowly expressed genes that are also enriched with H3K27me2/3, and loss of ASH1 causes alter ed enric hment of H3K27me2/3 (Bicocca et al. 2018 ). Similarly , in F . fujikuroi , ASH1 methylates H3K36 to establish facultativ e heter oc hr omatin at subtelomer es; ASH1 deletion incr eases subtelomeric H3K27me3 enrichment, yet genes at the subtelomeres become unstable (Janevska et al. 2018 ). These data argue that ASH1-specific H3K36me2/3 is epistatic to the action of the PRC2 complex and highlighting the need for additional study on the dynamics of facultative heterochromatin marks in multiple fungal species.

Conserved c hr omosomal fea tures in euc hr oma tic regions in fungal genomes
Euc hr omatic genes have the potential to be transcriptionally activ e, but pr esumabl y contr ol of gene expr ession can be driven by transcription factor recruitment, deposition of specific histone post-tr anslational modifications, or e v en the 3D genome or ganization. Many fungi share similar epigenetic marks as higher eukaryotes over actively transcribed genes, including the extensive acetylation of histone proteins and the di-or tri-methylation of lysine 4 on histone H3 (H3K4me2/3) (Pokholok et al. 2005, Lewis et al. 2009, Anderson et al. 2010, Xiong et al. 2010, Bicocca et al. 2018, Zhu et al. 2019, Courtney et al. 2020. As in metazoans, H3K4me3 is primarily enriched at promoter regions ( Fig. 1 ), while H3K4me2 is pr edominantl y deposited acr oss gene bodies (Liu et al. 2005, Pokholok et al. 2005, Cemel et al. 2017, Zhu et al. 2019 ). H3K4me2/3 is catal yzed by the conserv ed COMPASS complex, which contains the SET-domain protein KMT2 SET-1 as its catalytic subunit, but the composition of associated subunits can vary (Miller et al. 2001, Roguev et al. 2001, Freitag 2017. All four histone proteins within euchromatic nucleosomes are also substantially acetylated , with multiple lysine residues being subject for acetylation, including lysine 27 on histone H3 (H3K27ac; Fig. 1    Figur e 2. T he protein landscape of chromatin organization in fungi. Phylogenetic profile of 41 well-studied proteins known to impact chromatin organization encoded in 88 eukaryotes . T hese proteins are divided into the core histone proteins, CTCF, condensin subunits (type I and II), cohesin subunits , nuclear en v elope-associated pr oteins, and shelterin-associated pr oteins (see Table S2, Supporting Information, for specific gene information). Gene orthologs encoding most of the proteins involved in genome organization are conserved in fungi, with the exceptions of CTCF proteins associated with loop formation, condensin II for c hr omosome territories, Lamins at the nuclear periphery, and multiple components of the shelterin complex for telomere protection/organization. The heatmap color code reflects the fraction of species of a specific taxonomic lineage in which orthologs were found, and gray boxes indicate the absences of that specific protein. Known phylogenetic relationships between the different fungal lineages and other taxonomic groups (right) are shown on the left. Sc hematic r epr esentations that summarize the composition of the analyzed protein complexes are shown on the bottom; colored subunits depict presence in fungi, while gray colored subunits depict absence. Methods for ortholog searc hes ar e detailed in Supplemental File 1.

H 2 A H 2 B H 3 H 4 C T C F S M C 2 S M C 4 C A P H C A P G C A P D 2 C A P H 2 C A P G 2 C A P D 3 S M C 1 S M C 3 S C C 1 R E C 8 S C C 3 P D S 5 N IP B L M A U 2 W
transcription of genes necessary for plant infection . In general, acetylation appears dense over gene bodies and intergenic spacing, but little acetylation is found in AT-rich isoc hor es , as deacetylase complexes , such as the HCHC in N. crassa , act upon heter oc hr omatic r egions to r emov e acetyl gr oups to induce c hr omatin condensation (Honda et al. 2012(Honda et al. , 2016. The distribution of these "activating" histone modifications can influence the formation of nucleosomes in euc hr omatin by altering the association of histone proteins with DN A, thereb y changing the accessibility of the underlying DNA (Bannister and Kouzarides 2011 , Smolle and Workman 2013 ). Euc hr omatin acces-sibility has been assessed by A T AC-seq (Assay for Transposase Accessible Chromatin-sequencing). For example, Neurospora displays highl y accessible c hr omatin r egions (ACRs) that ar e often found upstream of genes: most are small, but a subset of these intergenic ACRs ar e lar ge ( > 2000 bp), pr esumabl y for m ultiple tr anscription factors to control of expression of ACR-proximal genes; these open c hr omatin r egions ar e enric hed with acetylation of H3K27 to possibly "open" the chromatin (Ferraro et al. 2021 ). Euc hr omatin is c har acterized by a mor e open nucleosome conformation that would also allow the underlying DNA to be more accessible. Specificall y, the pr omoter DNA of expr essed genes is typ-icall y de void of histone pr oteins, forming a nucleosome fr ee r egion that allows transcription factor binding to cis -regulatory elements for the recruitment of the RNA Polymerase II machinery for transcription initiation (Bai and Morozov 2010, Radman-Livaja and Rando 2010, Struhl and Segal 2013. Nucleosome-free regions (NFR) at promoters with more accessible chromatin have been widel y observ ed in fungi, including S. cerevisiae (Oberbec kmann  et (Jenull et al. 2020 ), Aspergillus niger , and Ustilaginoidea virens (Chen et al. 2021 ). Histone variants that alter nucleosome composition are also known to impact chromatin accessibility across eukaryotic genomes (Talbert and Henikoff 2010, Henikoff and Smith 2015. For example, the placement of the histone v ariant H2A.Z, whic h r eplaces H2A in the histone octamer, is enriched at the first nucleosome immediately downstream of the transcription start site of activ el y tr anscribed genes, possibly functioning to impact transcription by altering the c hr omatin structure at fungal promoters (Dong et al. 2018, Chen and Ponts 2020, Martire and Banaszynski 2020. Additionally, modifying the position of nucleosomes in the open reading frames of genes or within intergenic regions can impact the accessibility or composition of c hr omatin. Nucleosome positions ar e alter ed by c hr omatin remodelers, and in fungi, well-studied chromatin remodelers include the SWI-SNF complex and the AAA-ATP member DIM-1/CATP, the latter of which impacts the enrichment of H3K9me3 and 5 m C in intergenic regions in euchromatin (Cha et al. 2013, Kloc k o et al. 2019, Kamei et al. 2021, Wiles et al. 2022 ).

T he spa tial organiza tion of fungal c hr oma tin
During interphase, the nucleosomes package DNA into 10 nanometer (nm) c hr omatin fibers, a beads on a string conformation that in turn may a ggr egate further into 30 nm c hr omatin fibers (Bernardi 2015, Maeshima et al. 2016, Hansen et al. 2018b, Wako et al. 2020. Pr esumabl y, these c hr omatin fibers would be spatiall y or ganized in the nucleus to facilitate proper genome function, including the initiation of transcription to enable timely gene expression (Cook 1999, Misteli 2007. To facilitate this nuclear organization, the chromatin fibers in eukaryotic genomes form a series of subnuclear structures of hierarc hicall y incr easing sizes-in fungi, these r ange fr om c hr omatin loops to the Rabl c hr omosome conformation (Box 1)-that both allow critical genome features to a ggr egate into nuclear euc hr omatic/heter oc hr omatic compartments and pr e v ent the formation of knots (or other nonviable DNA strand folding). This hierarc hical or ganization of fungal c hr omatin folding allows euc hr omatic genes to have the potential to dynamically associate in spatially close proximity, thereby forming nonrandom interactions or higher-order 3D structures critical for gene expression. Howe v er, one can speculate that the large-scale folding of c hr omatin fibers is also necessary to compact the fungal genome in the nucleus to ensure the proper functioning of the DNA templated processes necessary for viability in an identical manner to the role of genome organization in metazoan genome function (Misteli 2007, Cavalli and Misteli 2013, Sexton and Cavalli 2015. It should be noted, ho w e v er, that most genome organization experiments in fungi are derived from Ascomycetes, and consequently additional work in Basidiomycetes as well as in earlier div er g ing fung i will be needed to determine which aspects of the hier arc hical genome organization detailed below are conserved across fungi.

Chromatin loops/globules
Chromatin loops (often called "globules") are among the most pr e v alent 3D structur es in metazoan genomes (Kadauke and Blobel 2009, Lieberman-Aiden et al. 2009, Heger et al. 2012. Here, c hr omatin loops occur when two loci physically separated on the linear c hr omosome associate, and the intervening chromatin condenses . In metazoans , these ∼300 kilobases (kb) c hr omatin loops may functionally influence gene expression by facilitating interactions between distant r egulatory sequences, suc h as enhancers and silencers, although the possibility exists that c hr omatin loops only form to structurally organize metazoan genomes (Cavalli and Misteli 2013, Rao et al. 2014, Dekker and Hear d 2015. Tw o critical components in the formation of c hr omatin loops in humans are cohesin and condensin I, which are architectural DNAbinding protein complexes important for DNA replication and c hr omosome folding, as well as for condensing c hr omosomes for meiosis and mitosis (Green et al. 2012, Hirano 2012, Haarhuis et al. 2017, Rao et al. 2017, Davidson and Peters 2021, Hoencamp et al. 2021, Jeppsson et al. 2022 ). In the cohesin complex, the S tructural M aintenance of C hromosomes (SMC)-1 , SMC-3, Scc1, and Scc3 proteins form the core machinery for chromatin looping (Dorsett 2007, Peric-Hupkes and van Steensel 2008, Skibbens 2019. In addition, five to six accessory proteins are responsible for dynamically loading or unloading the core cohesin complex onto chromatin and activating the cohesin ATPase activity (Haarhuis et al. 2017, Davidson and Peters 2021, Yoshida et al. 2022 ). The loading and activation of the SMC complex onto chromatin in budding yeast is facilitated by the yeast homologs Scc2 and Scc4, as signaled by the acetyltr ansfer ase Eco1, while cohesin dissociation fr om c hr omatin is mediated by WAPL and PDS5 (Ciosk et al. 2000, Rolef Ben-Shahar et al. 2008, Chan et al. 2012, Murayama and Uhlmann 2014, Çamdere et al. 2015, Haarhuis et al. 2017, Petela et al. 2018 ). Together, SMC proteins comprise a ring-shape structure in the cohesin complex thr ough whic h c hr omatin is activ el y extruded, thereby forming chromatin loops (Fudenberg et al. 2016, Ganji et al. 2018, Bauer et al. 2021. Chromatin extrusion continues until a boundary element is encounter ed, whic h both signals for cessation of extrusion mechanism and the anchoring of the r esulting c hr omatin loop (Ganji et al. 2018 ). Cohesin remains associated with c hr omatin thr ough inter phase, potentiall y to stabilize loops, until mitosis dependent Scc1 cleav a ge occurs (Uhlmann et al. 1999, Nasmyth 2001, Murayama and Uhlmann 2014. Borders of c hr omatin loops ar e typicall y demarcated in metazoan Hi-C datasets by the visualization of a "focus" or point of enriched contacts off diagonal (Rao et al. 2014 ). These strong interactions at loop bases are often centered over binding sites for the CTCF (CCCTC binding factor) (Fudenberg et al. 2016, Br ac kley et al. 2018, Banigan and Mirny 2020, Davidson and Peters 2021. CTCFmediated loops are known to insulate nearby active and repressiv e c hr omatin r egions by bloc king further c hr omatin extrusion (de Wit et al. 2015, Sanborn et al. 2015. Her e, two CTCF pr oteins bind to conv er gentl y oriented asymmetric 14 bp sequences and dimerize to form a loop of c hr omatin in the genomes of higher eukaryotes (Ong and Corces 2014, Rao et al. 2014, de Wit et al. 2015, Nor a et al. 2017. Consequentl y, depletion of CTCF disrupts loop formation in many cases (Zuin et al. 2014, Nora et al. 2017, Xu et al. 2021. The cohesin core complex and most accessory proteins are highl y conserv ed thr oughout eukaryotes (Fig. 2 ), so it is r easonable to expect that c hr omatin loop formation occurs by a loop extrusion mechanism in fungi as well (Fig. 3 A). In fungi, the role of cohesin has been only examined in the fission yeast S. pombe , where Figur e 3. T he formation of regional structures in several fungal organisms. In filamentous fungi, TAD-like structures generally form through hier arc hical clustering of local c hr omatin structur es. (A) A gener alized sc hematic of the folding of c hr omatin fr om the "beads-on-a-string" model of the 10-nm fiber in which DNA is bound to histone proteins to form nucleosome, to the formation of loops/globules with the action of cohesin, to the folding of those globules into TAD-like/regional globule clusters in the fungal genome . T he folding of TAD-like structures inferred from Hi-C contact data in (B) N. crassa (Rodriguez et al. 2022 ) and (C) V. dahliae (Torres et al. 2023 ). (D) The compartmentalization of heter oc hr omatic (str onger interactions) and euchromatic regions can be observed in the genome of Rhizophagus irregularis (Yildirir et al. 2022 ). Methods for Hi-C data processing and image generation are detailed in Supplemental File 1. in a temper atur e-sensitiv e cohesin m utant str ain the ∼40 kbsized globules are no longer visible across the genome (Mizuguchi et al. 2014, Tanizawa et al. 2017. Chromatin loops/globules have been also observed in filamentous fungi. In N. crassa slightly larger ( ∼60-80 kb) globules ar e r eadil y visible in Hi-C datasets (Rodriguez et al. 2022 ). Giv en the conserv ation of cohesin subunits in these filamentous fungi (Fig. 2 ), cohesin should also act across these genomes. Ho w e v er, it is possible that cohesin and its activity are not fully conserved throughout all fungal lineages. In Microsporidia, almost all cohesin complex components are absent (Fig. 2 ), suggesting that these spore-forming unicellular fungal par asites e volv ed distinct mec hanisms to topologicall y or ganize their genome. Further, WAPL and Eco1 seem to be absent in fungi belonging to the Zoopagomycetes (Fig. 2 ), suggesting novel mechanisms for signaling loop dissociation might exist in these obligate soil nematode par asites. Inter estingl y, and in contr ast to cohesin subunits, the gene encoding CTCF is restricted to higher metazoan species, whic h typicall y hav e lar ger genomes. One can speculate that the CTCF bound to the base of a c hr omatin loop anchors, and possibly stabilizes, the larger-sized globules/loops in metazoans (Li et al. 2020, Pugac he v a et al. 2020. Other boundary elements a part fr om CTCF hav e been found in plants (Dong et al. 2017, Xie et al. 2019, Drosophila melanogaster (Ramírez et al. 2018 ), Caenorhabditis elegans (Anderson et al. 2019 ), and humans (Anania et al. 2022, Valton et al. 2022. Ho w e v er, the specific boundary elements restricting chromatin loop size by repressing cohesin loop extrusion in fungi have not been fully characterized (Mizuguchi et al. 2014, Schalbetter et al. 2019, Rodriguez et al. 2022. Fungi may employ diverse mechanisms to negatively regulate loop formation, including the enrichment of conv er gent genes that delineate loop "boundaries" in S. pombe (Mizuguchi et al. 2014 ), or the incor por ation of AT-ric h, r e petiti v e heter oc hr omatic isoc hor es in the genome, as observed in N. crassa (Rodriguez et al. 2022 ) and E. festucae (Winter et al. 2018 ). While future experiments are needed to assess the r equir ement of cohesin for forming c hr omatin loops in filamentous fungi (Schalbetter et al. 2019 ), the hypothesis that constitutiv e heter oc hr omatic r egions possibl y act as loop anchors would render CTCF or similar boundaries unnecessary in some fungi.

Topologicall y associa ted domains
Topologically associated domains (TADs) are 3D chromosomal structures that regionally organize the genome by subdividing c hr omatin compartments (Dixon et al. 2012 , Acemel andLupiáñez 2023 ) suggesting that a single TAD may contain se v er al c hr omatin loops or become visible when multiple loops are averaged over a cell population Heard 2015 , Hansen et al. 2018a ) (Fig. 3 ). TADs in metazoans ar e typicall y megabase-sized genomic r egions in whic h c hr omatin is mor e a pt to contact yet is insulated fr om c hr omatin outside of the TAD (Lieberman-Aiden et al. 2009, Dixon et al. 2012 ). In metazoan Hi-C datasets, TADs appear as triangles of increased contact probability immediately off-diagonal, with few distant contacts with chromatin beyond the TAD border (Fig. 3 ). Inter estingl y, TAD borders in metazoans are often enriched for housek ee ping genes , tRNA genes , and r etr otr ansposons (Dixon et al. 2012 ). Ho w e v er, it is curr entl y unclear if TADs ar e primaril y structur al, i.e. TAD formation or ganizes the genome in the nucleus, or are essential for genome function (Beagan and Phillips-Cr emins 2020 ), possibl y r equir ed to r egulate gene expr ession by increasing the contact probability of distant enhancers/silencers with cognate promoters (Flavahan et al. 2016, Dixon et al. 2018. While altered TAD borders may cause misregulation of gene ex-pression in some human cancers (Flavahan et al. 2016, Taberlay et al. 2016, Valton and Dekker 2016, Dixon et al. 2018, Akdemir et al. 2020, m ultiple datasets r eport onl y minimal c hanges in gene expression when individual TADs are altered (Rao et al. 2017, Ghavi-Helm et al. 2019. Additionall y, TADs c hange little thr ough interphase of the cell-cycle, but TADs ar e consistentl y lost upon entry into mitosis when c hr omosomes ar e locall y condensed, and ar e r eformed befor e entry to G1, suggesting TADs ar e primaril y structural in nature (Naumova et al. 2013, Abramo et al. 2019.
TAD formation in higher metazoans appears to require the action of condensin complexes. Specifically, the SMC complexes condensin type I and II are known to shape individual chromosomes by forming higher order looping structures (Hagstrom andMeyer 2003 , Hirano 2012 ). In both complexes, the yeast homologs SMC2 and SMC4 form a highl y conserv ed cor e ring structur e (Sc hleiffer et al. 2003 ), but individual condensin types can be distinguished by other subunits: condensin I contains C APH, C APG, and C APD2, while condensin II has C APH2, C APG2, and C APD3 (Ono et al. 2003, Schleiffer et al. 2003. Condensin I has the greatest impact during the mitotic phase of the cell cycle, where it later all y compacts sister c hr omatids by forming smaller supercoiled c hr omatin loops (Kim ur a and Hirano 1997, Hagstrom and Meyer 2003, Green et al. 2012, Golfier et al. 2020, Kong et al. 2020. Condensin II affects high order c hr omosome or ganization during all cell cycle sta ges, as it is found in the nucleus during S phase; during mitosis, condensin II ma y in volv e the formation of lar ge c hr omatin loops, in conjunction with condensin I activity, for establishing a "nested loop arc hitectur e" critical for sister c hr omatid/homologous c hr omosome pairing (Yu and Koshland 2005, Ono et al. 2013, Kong et al. 2020. Emer ging e vidence fr om m ultiple species suggests that filamentous fungi have TAD-like structures analogous to metazoan TADs (Mizuguchi et al. 2014, Eser et al. 2017, Tsochatzidou et al. 2017, Winter et al. 2018, Schalbetter et al. 2019, Rodriguez et al. 2022. Fungal TAD-like structures are several hundred kilobases in size and the euc hr omatin internal to these TAD-like structures is mor e a pt to contact (Dixon et al. 2012, Rodriguez et al. 2022. In N. crassa , TAD-like structures were originall y termed r egional globule clusters (RGCs) named for the aggregation of several ∼40 kb euchromatic globules into larger, compact structures ∼250 kb in size that could be inter pr eted as a lar ge c hr omatin a ggr egates analogous to metazoan TADs (Fig. 3 B). RGCs display extensive yet random internal euc hr omatic contacts that are not restricted from outside chromatin, as strong inter-RGC contacts r eadil y occur, ar guing pr oteins insulating internal euc hr omatin ar e not encoded in N. crassa (Rodriguez et al. 2022 ). RGCs are flanked by constitutive heterochromatic regions to delineate RGC borders . T he clustering of heter oc hr omatin r egions, possibl y thr ough liquid-liquid phase separ ation (LLPS) condensates (Larson et al. 2017 ), may act as RGC anchor (Fig. 3 B) in an analogous manner to CTCF at c hr omatin loops (Rodriguez et al. 2022 ); consequently, cohesin would act specifically to form the smaller globules internal to and comprising this TAD-like structure. Similar patterns of TAD-like structures have been observed in S. cerevisiae , where globule structures are delimited by transcriptionall y activ e genes that are often in a conv er gent orientation (Tsochatzidou et al. 2017, Schalbetter et al. 2019. Additionall y, TAD-like structur es can be seen in E. festucae , where RIP'd ATric h heter oc hr omatic r egions str ongl y inter act to form lar ge structures to compact chromatin (Winter et al. 2018 ), and in the fungal pathogen Puccinia striiformis , which may form uninsulated RGCs acr oss eac h c hr omosome arm (Xia et al. 2022 ). Another example of TAD-like structures is found in V. dahliae and related Verticillium species where Hi-C datasets display TAD-like structures with increased internal chromatin contact probabilities and few intr ac hr omosomal contacts beyond the TAD-like structure boundaries (Fig. 3 C) (Torres et al. 2023 ). In contrast to the situation in N. crassa , (differ entiall y) expr essed genes occur at, or in pr oximity to, TAD-like boundaries in V. dahliae and E. festucae , suggesting that TAD-like structures in these fungi are necessary for proper gene expression. Additional evidence from yeasts suggests that TAD-like structur es ar e critical for other genome functions apart from gene expression, such as repressing recombination or promoting genome evolution (Mizuguchi et al. 2014, Tsochatzidou et al. 2017, Gu et al. 2022. For example , TADs ma y be essential for fungal c hr omosome r eplication during S-phase of the cell cycle, as ∼200-kb TAD-like structures across the S. cerevisiae genome separate clusters of early or late timed origins of r eplication acr oss its 16 c hr omosomes (Eser et al. 2017 ).

Organization of interspersed constituti v e heter ochr omatic regions
AT-ric h isoc hor es comprising constitutiv e heter oc hr omatin can be found embedded throughout fungal chromosomes as well as at centr omer es and telomer es (Le wis et al. 2009, Winter et al. 2018, Seidl et al. 2020. Neurospor a cr assa has a ppr oximatel y 300 ATand H3K9me3-enriched isochores interspersed throughout the genome, whic h r ange in size fr om < 1 to ∼400 kb  ). These regions readily associate in the nucleus, as exceptionall y str ong contacts between silent c hr omatin regions both within a chromosome and across chromosomes are fr equentl y observ ed in fungal Hi-C datasets (Fig. 4 ), ar guing that the clustering of constitutiv e heter oc hr omatic r egions in the fungal nucleus may be particularly important for the proper chromosome conformation in fungi. Ho w e v er, it is r eadil y a ppar ent in fungal Hi-C datasets that any interspersed constitutive heter oc hr omatic r egion has the potential to inter act, as uniforml y strong contacts between all H3K9me3-marked silent regions are r eadil y observ ed within Hi-C data deriv ed fr om a population of fungal nuclei , Winter et al. 2018, Rodriguez et al. 2022. Mor eov er, the c hr omatin inside e v ery constitutiv e heter oc hr omatic r egion str ongl y and consistentl y inter acts acr oss the entire length of that silent region, with the strongest interactions occurring on the le v el of indi vidual n ucleosomes, suggesting that silent c hr omatin forms dense globule-like structur es consisting of a stochastic nucleosome aggregation , Winter et al. 2018, Rodriguez et al. 2022. At the highest resolutions, dense globules are visible at the boundaries between heter oc hr omatic and euc hr omatic r egions, impl ying the formation of 3D c hr omatin structures to prevent heterochromatin spread (Rodriguez et al. 2022 ). Ho w e v er, the loss of the known constitutive heterochromatin machinery has little impact on the folding of individual heter oc hr omatic r egions. In Neurospora , deletion of the gene encoding the KMT1 DIM-5 histone methyltr ansfer ase or its cognate binding partner HP1 reduces the dense internal compaction of heter oc hr omatic r egions and leads to r educed contacts between the euc hr omatin bordering these silent regions  ). This suggests that folding of constitutive heterochromatic regions is dependent on proper deposition of different chromatin modifications and that the primary role of H3K9me3 and HP1 is to compact individual silent r egions, ther eby r estricting contacts e v en between distant heter oc hr omatic r egions , Zenk et al. 2021. Notably , AT -rich DNA forms few contacts with the surr ounding euc hr omatin despite activ e and silent c hr omatin being in close proximity on the linear chromosome , Rodriguez et al. 2022, suggesting the heter oc hr omatin-internal nucleosomes are isolated from active c hr omatin in fungal nuclei. Similarl y, H3K9me3-enric hed AT-ric h sequences in V. dahliae and AT-ric h isoc hor es in E. festucae appear to be insulated from euchromatic contacts (Winter et al. 2018, Seidl et al. 2020 Despite the ov er all segr egation of heter oc hr omatic and euc hr omatic DNA, recent work has shown that contacts that form between active and silent c hr omatin may r egulate fungal gene expression (Rodriguez et al. 2022 ). Specifically, in N. crassa , small "bands" of str ong inter actions between H3K9me3-marked constitutiv e heter oc hr omatic r egions and select genes in mor e GC-ric h genomic r egions, whic h ar e possibl y marked with a unique combination of histone post-translational modifications, are readily observed at the highest resolution Hi-C datasets (Fig. 1 ) (Rodriguez et al. 2022 ). Man y of these genes display dr astic c hanges in gene expression when constitutive heterochromatin is compromised (e.g. in a dim-5 / kmt1 mutant strain) (Rodriguez et al. 2022 ). The possibility of constitutive heterochromatin regulating gene expression has been observed previously (Yang et al. 2014 ) and may not be limited to Neurospora as se v er al hundr ed genes that significantl y c hange expr ession hav e been observ ed in V. dahliae upon loss of the lysine 9-specific histone methyltransferase Dim-5 (Kramer et al. 2021 ).
To segr egate heter oc hr omatic genomic loci fr om those that ar e activ el y tr anscribed, and possibl y to pr e v ent aberr ant RNA synthesis of silent c hr omatin, heter oc hr omatin clusters at the nuclear membrane in virtually all eukaryotic nuclei (Gonzalez-Sandoval and Gasser 2016, Solovei et al. 2016, Falk et al. 2019. In mammals, most heter oc hr omatin associates with lamin filaments and additional anchor proteins to form LADs ( L amina-A ssociated D omains) at the nuclear envelope (Guelen et al. 2008 , Briand andCollas 2020 ). Mammalian proteins required for LAD formation include the lamin B receptor, Emerin, and L amina-A ssociated-P olypeptide 2 -b (LAP-2), all of which contain a LEM ( L AP-2, E merin, M AN1) domain capable of binding c hr omatin at the nuclear envelope (Lin et al. 2000, Wagner and Krohne 2007, Buchwalter et al. 2019. Howe v er, fungi do not encode these proteins (Fig. 2 ), nor has the LEM domain been observed in any fungal pr oteins (Wa gner and Kr ohne 2007 , Kor en y and Field 2016 ). Ho w e v er, fungi encode two general classes of integral nuclear membr ane pr oteins that facilitate heter oc hr omatin inter actions with the nuclear membrane: MSC and SUN proteins (Koreny and Field 2016 ). Members of the MSC family of proteins include the integr al membr ane pr oteins Src1/Heh1, Heh2, Man1, and Lem2 (Br ac hner et al. 2005, King et al. 2006, Wagner and Krohne 2007, Grund et al. 2008, Mekhail and Moazed 2010, Taddei and Gasser 2012. All MSC proteins contain N-terminal LEM-like and MSC ( M an1-S rc1p C -terminal) domains in their primary structures. In S. pombe , Lem2 facilitates the anchoring of heterochromatic regions to the nuclear en velope , where the chromatin silencing machinery, including histone deacetylase complexes targeted to the telomer es, ar e r ecruited for c hr omatin r epr ession (Sugiyama et al. 2007, Barrales et al. 2016, Hirano et al. 2020. Further, individual heter oc hr omatic r egions r equir e specific pr oteins for nuclear envelope association: centromeres require Csi1 and telomeres use Dsh1 and Bqt3 (or Bqt4) (Barrales et al. 2016, Harr et al. 2016, Ebrahimi et al. 2018. The SUN (Sad-1, Unc-84) protein family similarly mediates heterochromatin interactions with the nuclear membrane (Tzur et al. 2006 ). Specifically, the S. cerevisiae SUN protein Mps3 associates with Sir silencing proteins, including the Sir4-Sir3 complex that binds deacetylated histone H4 in silent c hr omatin (Bupp et al. 2007   crassa genome is c har acterized by interc hr omosomal centr omeric contacts, as w ell as inter c hr omosomal contacts between telomer es in a Rabl c hr omosome conformation. Individual c hr omosomes hav e str ong telomeric inter actions, while centr omeric c hr omatin str ongl y self-inter acts yet is isolated from other genomic foci. Interestingly, the strongest long-range interactions occur between H3K9me3-enriched constitutive heterochromatic r egions. (A) Corr ected in situ Hi-C data of combined DpnII (euc hr omatin-specific) and MseI (heter oc hr omatin-specific) Hi-C data (Rodriguez et al. 2022 ) at 20 kb resolution, showing the interactions across the entire genome (all seven chromosomes indicated on the top and right of the Hi-C contact ma p), one c hr omosome (Linka ge Gr oup III), and a zoomed ima ge of the right arm of LG III. Blue arr o wheads indicate inter c hr omosomal centr omer e inter actions, pur ple arr owheads indicate inter-or intr ac hr omosomal telomeric inter actions, and white arr owheads indicate str ong intr ac hr omosomal heter oc hr omatic inter actions. ChIP-seq tr ac k of H3K9me3 enric hment shows r egions of constitutiv e heter oc hr omatin. (B) Aggr egate c hr omosome plots, at 5 kb resolution, of DpnII (euc hr omatin-specific, left) or MseI (heter oc hr omatin-specific, right) in situ Hi-C data (Rodriguez et al. 2022 ). Two a ggr egate c hr omosome sc hematics ar e below the Hi-C contact ma p, with centr omer es denoted by r ed boxes and telomer es sho wn b y gr een ov als. Blue arrowheads indicate interc hr omosomal centr omer e inter actions and pur ple arr owheads indicate intr a-or interc hr omosomal telomeric inter actions. Methods for Hi-C data processing and image generation are detailed in Supplemental File 1.
vations highlight a direct nuclear membrane-heterochromatin contact being important for genome organization. The MSC and SUN pr oteins ar e also widel y conserv ed acr oss filamentous fungi (Fig. 2 ), impl ying similar mec hanisms might also be emplo y ed in these species to tether heter oc hr omatic r egions to the nuclear membrane.

Chroma tin compartmentaliza tion in fungi
Euc hr omatin and heter oc hr omatin in metazoans typicall y segregates into two distinct nuclear compartments: the euc hr omatic (active) "A" and heterochromatic (silent) "B" compartments. This compartmentalization of c hr omatin is r eadil y observ ed as a "c hec kerboard" pattern in Hi-C contact maps (Fig. 3 D). The functional inter pr etation of this pattern is that genomic regions that have a similar transcriptional activity (e.g. heterochromatic re-gions that are silent) are spatially interacting within the nucleus (Lieberman-Aiden et al. 2009, Dixon et al. 2012, Rao et al. 2014, Dong et al. 2017, Rowley et al. 2017, Nichols and Corces 2021. Further studies have demonstrated that chromatin in each compartment physically associates: in the A-compartment euchromatin interacts in the central nucleus region, while in the Bcompartment heter oc hr omatin associates at the nuclear periphery (Lieberman-Aiden et al. 2009, Rao et al. 2014, Buchwalter et al. 2019, Beagan and Phillips-Cremins 2020. Mechanistically, this segregation may occur due to the aggregation of heterochromatic r egions, possibl y thr ough LLPS, at the nuclear membr ane causing euc hr omatin to associate in the nucleus center (Larson et al. 2017, Falk et al. 2019, or due to the forces emerging from the activity of DNA-templated proteins in euchromatin forcing the segregation of silent chromatin to the nuclear periphery (Mahajan et al. 2022 ). In contrast to the extensive compartmental-ization seen in metazoan Hi-C datasets, few fungi have evidence of prominent compartmentalization. To date, only the arbuscular mycorrhizal fungus Rhizophagus irregularis , a member of the Glomeromycetes clade , displa ys clear A/B compartments (Fig. 3 D) (Xia et al. 2022, Yildirir et al. 2022. In contrast, other fungi display minimal c hr omatin compartmentalization, possibl y r eflecting the presence of smaller heter oc hr omatic r egions integr ated among larger euchromatin domains (Xia et al. 2022 ). Ho w ever, segregation of fungal chromatin in a manner analogous to A/B compartments, wher e heter oc hr omatic r egions a ggr egate yet ar e separ ated fr om euc hr omatic TAD-like structur es, has been observ ed in the Hi-C datasets of multiple fungal species, including S. cerevisiae (Duan et al. 2010), S. pombe (Mizuguchi et al. 2014, N. crassa , Rodriguez et al. 2022 ), E. festucae (Winter et al. 2018 ), A. bisporus (Hoencamp et al. 2021 ), and V. dahliae (Seidl et al. 2020 ). Pr esumabl y, inter actions between heter oc hr omatic r egions, e v en when the small AT-rich isochores across fungal genomes are formed into heterochromatin, could be crucial for phase separation into the active "A" and silent "B" compartments (Falk et al. 2019 ). Chromatin compartmentalization is also supported by historical electron microscopy data, which shows clusters of densely stained heterochromatin, which are often at the nuclear periphery but can be in the nucleus center, interspersed with lightl y stained euc hr omatin (Shatkin and Tatum 1959 ), thus arguing the c hr omatin composition or ganizes the nuclear genome in fungi.

Organization of fungal chromosomes into a Rabl conformation
The compartmentalization of the heter oc hr omatic centr omer es and telomeres of fungal chromosomes onto the nuclear membrane would facilitate the formation of Rabl c hr omosome conformations within the fungal nucleus (Box 1) (Duan et al. 2010, Schalbetter et al. 2019, Hoencamp et al. 2021. Rabl c hr omosome conformation is typicall y c har acterized by the clustering of centr omer es on one side of the nuclear envelope and c hr omosomal arms extending outw ar ds to w ar ds the opposing nuclear periphery on which the (sub)telomeres associate (Fig. 5 A; e.g. Jin et al. 2000 ). Microscopic (e.g. Guacci et al. 1994, Lar oc he et al. 1998, Jin et al. 2000, Goto et al. 2001, Gasser 2002, Schober et al. 2008 ) as well as Hi-C experiments (Duan et al. 2010, Schalbetter et al. 2019, Hoencamp et al. 2021 ) also confirmed that S. cerevisiae organizes its 16 chromosomes in Rabl conformation during interphase, with a centr omer e cluster in proximity to the spindle pole body while the 32 (sub)telomeres associate nonr andoml y in four to eight foci at the nuclear membr ane opposite the centr omer e bundle (Bystric ky et al. 2004, Duan et al. 2010, Therizols et al. 2010, Schalbetter et al. 2019. Similarl y, S. pombe or ganizes its thr ee c hr omosomes into a Rabl structure (Mizuguchi et al. 2014 ). In filamentous fungi, the Rabl conformation was initially observed in N. crassa Hi-C experiments , Rodriguez et al. 2022. These data wer e corr obor ated by fluor escent micr oscopy of N. crassa nuclei in which a single centr omer e focus and three to four telomeric foci associate with the nuclear membrane (Fig. 5 , Bo x 1). Ad ditional Hi-C data from a plethora of filamentous fungi show that the centr omer es contact independent of-and distinct fr om-telomer e clustering and thus confirmed the existence of Rabl c hr omosomes, including in ascomycetes [ Penicillium oxalicum  ), E. festucae (Winter et al. 2018 (Liang et al. 2022 ), Puccinia graminis (Sperschneider et al. 2021, Henningsen et al. 2022, Austropuccinia psidii (Edw ar ds et al. 2022 ), and Agaricus bisporus (Hoencamp et al. 2021 )] (Table S1, Supporting Information). Inter estingl y, R. irregularis does not show clear centr omer e bundling for organizing its chromosomes, but it does seem to exhibit telomere bundling (Yildirir et al. 2022 ). T hus , these observations collectiv el y suggest that the vast majority of fungi exhibit at least some of the features associated with Rabl-like chromosomal conformation.
One of the most prominent features observed in Hi-C contact maps of species employing a Rabl c hr omosome conformation is the strong interchromosomal interactions indicative of centr omer e bundling (Fig. 4 ). Centr omer e bundles manifest as dark "spots" between c hr omosomes in whole genome contact ma ps, as observed in yeasts ( S. cerevisiae and S. pombe ), N. crassa (Fig. 4 ), V. dahliae , and E. festucae , among others (Mizuguchi et al. 2014, Belton et al. 2015, Tanizawa et al. 2017, Winter et al. 2018, Seidl et al. 2020, Rodriguez et al. 2022. The size of the interc hr omosomal centr omer e contacts is dependent on whether the centr omer es in that species are point (small intercentr omer e inter actions) or r egional (lar ge intercentr omer e inter actions) centr omer es (Belton et al. 2015, Taniza wa et al. 2017, Yada v et al. 2018a, b , Seidl et al. 2020, Rodriguez et al. 2022. Work in budding yeast has shown that centr omer es bundle at the spindle pole body during interphase, an inter action whic h may r equir e the centr omer es to associate with microtubules as the application of nocodazole, a toxin that disrupts micr otubule assembl y, leads to r educed centr omer e clustering (Jin et al. 2000, Goto et al. 2001. The yeast CBF3 complex, which deposits the centromere specific histone variant Cse4 in centromeric DNA, also contains the critical kinetochore protein Ndc10p, highlighting the connection between centr omer e clustering and microtubule binding (Lechner andCarbon 1991 , Yan et al. 2018 ). Sur prisingl y, fe w inter actions between the heter oc hr omatic centr omer es and other interspersed heter oc hr omatic r egions ar e observ ed in N. crassa , e v en when heter oc hr omatic featur es ar e in close pr oximity on the linear c hr omosome , Rodriguez et al. 2022. This argues that the centromeric bundle is refractory to interacting with other heter oc hr omatic r egions, impl ying that the centr omer e bundle forms a dense, compact structure that isolates centromeric DNA (Figs 4 and 5 ). Specificity for centr omer e bundling may be deriv ed fr om deposition of CenH3 into centr omeric nucleosomes, as CenH3 enrichment is only observed at the centromeres in N. crassa and other fungi (Smith et al. 2012, Seidl et al. 2020. Pr esumabl y, kinetoc hor e pr oteins that specifically associate with these centromeric histone variants might play a role in establishing and maintaining the centr omer e bundles at the nuclear membr ane acr oss filamentous fungi, in a manner similar to that of yeast (Pidoux and Allshire 2004, Westermann et al. 2007, Biggins 2013. Centr omer e clustering may also r equir e some gener al pr operty of heter oc hr omatin independent of H3K9me3 deposition, as yeasts lacking H3K9me3 and SETdomain histone methyltr ansfer ases still hav e extensiv e interc hr omosomal centr omer e bundling, suggesting centr omer e clusters could r el y on the silent c hr omatin in these r egions forming LLPS condensates. Increased contacts between telomeres that indicate clustering of c hr omosome ends ar e another pr e v alent featur e of most fungal Hi-C interaction maps (Fig. 4 ), and strong interactions originating at c hr omosome ends in Hi-C datasets can extend ∼200 kb internal to the c hr omosome into the subtelomeres (Fig. 4 ) Figur e 5. T he formation of the Rabl c hr omosome conformation in fungal nuclei. (A) A schematic representation of two chromosomes (colored brown and gray for distinction) in the Rabl chromosome conformation (Mizuguchi et al. 2015 ), where centromeres (red circles) cluster on one side of the nucleus and the telomeres (green circles) cluster on the opposite side; the nucleolus is a distinct structure apart from these interphase chromosomes. (B) Detailed schematic of how interspersed heterochromatic regions (black circles) facilitate chromatin associating with the nuclear membrane and the compartmentalization of active and silent chromatin. Each chromosome forms a weak territory, with more intrachromosomal contacts, but some interc hr omosomal inter actions can occur. (Mizuguchi et al. 2014, Tanizawa et al. 2017, Rodriguez et al. 2022 ). The r elativ e position and distance between telomere bundles in the nucleus ar e gov erned by the c hr omosomal arm length, the position of the centr omer e, and the nuclear v olume (Bystrick y et al. 2004, Therizols et al. 2010. Individual telomeres may also have a unique landscape of chromatin modifications. For instance, in Neurospora these c hr omosome ends ar e the onl y loci in the genome enriched for both H3K9me3 andH3K27me2/3 (Kloc k o et al. 2016 , Jamieson et al. 2018 ). Ho w e v er, the terminal H3K9me3enric hed telomer e r epeats minimall y inter act with nearby interspersed facultative or constitutive heterochromatic regions (Rodriguez et al. 2022 ). Importantly, no experimental data thus far has observed the colocalization of both centromeres and telomeres in the same nuclear region in wild type cells (Fig. 4 ), suggesting that mechanisms exist that independently organize these c hr omosomal structur es . One possibility is that fungi ma y coopt facultativ e heter oc hr omatin to r estrict (sub)telomeric contacts, thereby ensuring proper genome organization. In support, the set-7 ( kmt6 ) N. crassa strain devoid of H3K27me2/3 can hav e micr oscopic colocalization of centr omer es and telomer es when facultativ e heter oc hr omatin is compr omised . In Hi-C datasets, this loss of H3K27me2/3 reduces subtelomeric interactions and causes general genome organization disorder consistent with r educed inter actions between the subtelomeres and the nuclear membrane (Klocko et al. 2016 ). Howe v er, onl y subtelomeric H3K27me2/3 is critical for nuclear membr ane inter actions, as a npf ( Neurospora p55 homolog) strain that loses subtelomeric H3K27me2/3 exhibits genome disorder consistent with compr omised subtelomer e inter actions despite internal H3K27me2/3 enrichment being maintained (Klocko et al. 2016 ). T hus , current data seems to suggest that the unique c hr omatin landsca pe at the centr omer es and telomer es ensure the isolation of these chromosomal features in a Rabl conformation despite both exhibiting properties of constitutive heter oc hr omatin.
Loss of shelterin complex members eliminates telomere clustering and impacts the localization of telomeres to the nuclear periphery, as increased numbers of telomere foci are observed in loss of function shelterin mutants (Palladino et al. 1993, Chikashige and Hiraoka 2001, Kanoh and Ishika wa 2001. T he heterodimeric Yku complex (Yku70/80) may be r equir ed for telomer e clustering, as Yku70/80 anchors telomeres to the nuclear envelope (Laroche et al. 1998, Hediger et al. 2002, but also plays a role in shielding telomeric ends from shortening or fusion, and in telomere silencing (Boulton and Jackson 1998, Polotnianka et al. 1998, Ponn usam y et al. 2008. In filamentous fungi, recent work elucidated the role of an unexpected protein contributing to telomere tethering. In the Neur ospor a dim-3 str ain, whic h encodes a m utant allele of Importin ɑ /Karyopherin ɑ , some telomere foci no longer associate with the nuclear membrane, and a dim-3 strain exhibits a dr asticall y alter ed genome or ganization in a manner consistent with compromised telomere anchoring  ). This phenotype may be indir ectl y caused by an increase in nuclear volume, which under normal situations would physicall y constr ain telomer e anc horing: dim-3 nuclei consistentl y hav e a larger nuclear membrane diameter than the nuclei in wildtype strains . Consistently, previous work in metazoans suggest that Importin ɑ /Karyopherin ɑ is necessary to constrict nuclear volumes (Levy and Heald 2010 ). Further experiments aimed at elucidating the proteins necessary for telomer e-nuclear membr ane anc horing, based on the groundwork established in yeast systems, should pr ov e fruitful to uncover if filamentous fungi use similar mechanisms for telomere anchoring.
Quantitative computational models suggest that the specific tethering of heter oc hr omatic c hr omosomal r egions, including centr omer es and telomer es to distinct nuclear landmarks, is sufficient to explain higher order organization of fungal genomes (Tjong et al. 2012, Wong et al. 2012. For example, in budding yeast, the complex Rabl nuclear organization emerges in computational models in which chromosomes are allo w ed to form random contacts, yet chromosomes are constrained by the tethering of c hr omosomal featur es to the nuclear envelope, by the distances between telomeres, and by the colocalization of functionally related loci (Tjong et al. 2012, Wong et al. 2012. Ther efor e, these geometrical constraints alone are sufficient to explain highl y or ganized nuclear genome organization, including the Rabl-like c hr omosomal conformation, in S. cerevisiae (Tjong et al. 2012 ). Despite the genomes of filamen-tous fungi being larger and often containing more chromosomes, the bundling of centr omer es, telomer es, and interspersed heter oc hr omatic r egions at the nuclear periphery would be expected to drive the Rabl c hr omosome or ganization in a similar manner.
One corollary effect of a Rabl conformation in fungal nuclei is that c hr omosomes cannot form distinct territories, a property of higher metazoan genomes in which each chromosome occupies a defined space in the nucleus (Manuelidis 1985, Cr emer and Cr emer 2001, Tanabe et al. 2002, P ar ada et al. 2004. In humans, c hr omosomal territories ar e e vident by an enhanced intr ac hr omosome contact fr equency, and minimal interc hr omosomal inter actions (Lieberman-Aiden et al. 2009, Imakaev et al . 2012, Falk et al. 2019, Hoencamp et al. 2021. Eukaryotes encoding a complete condensin II complex form these territories, arguing that the presence of condensin II either str engthens c hr omosomal territories or suppr esses Rabl conformation formation (Hoencamp et al. 2021 ). Specifically, deletion of the condensin II subunit CAPH2 in human cells promotes the formation of a Rabl-like chromosome conformation by incr easing interc hr omosomal and tr ans-centr omeric contacts while lo w ering lengthwise compaction of c hr omosomes (Hoencamp et al. 2021, Yoshida et al. 2022. Ho w e v er, nuclear arc hitectur e is mor e v ariable when longer evolutionary time scales are considered (Hoencamp et al. 2021 ). Specifically, fungal genomes do not encode condensin II accessory subunits (Fig. 2 ), as pr e viousl y noted in S. cerevisiae , S. pombe , and N. crassa (Hudson et al. 2009, Hirano 2012, Hoencamp et al. 2021, Rodriguez et al. 2022, and consequently strong chromosome territories rarely form in fungal nuclei. T hus , the formation of a Rabl c hr omosome conformation (Fig. 5 ) in fungi is near ubiquitous. In this model, eac h c hr omosome exhibits extensiv e interc hr omosomal contacts (Fig. 4 ) that could be necessary for proper genome function, including the regulation of fungal gene expression. Ho w ever, several intriguing observations directly contrast the Rabl chromosome model being applicable to all fungi. First, in R. irregularis , no clear Rabl c hr omosomes can be observed in Hi-C experiments (Fig. 3 D) (Yildirir et al. 2022 ) yet condensin II orthologs are absent (Fig. 2 ), suggesting novel proteins facilitate chromatin compartmentalization into novel subnuclear structures. Further, orthologs of the condensin II accessory subunits C APH2, C APG2, and C APD3 are present in species of the Monoble pharidom ycetes clade (Fig. 2 ), suggesting that c hr omosome territories may exist in these taxa, but no Hi-C data curr entl y exists to test this hypothesis.

Concluding remarks and future research
The spatial organization of the eukaryotic nuclear genome is closely linked to its biological functions (Lieberman-Aiden et al. 2009, Sexton and Cavalli 2015, Bonev and Cavalli 2016, Hoencamp et al. 2021. Her e, we addr essed the occurr ence, formation, and functional implications of the spatial organization on the nuclear genome in fungi. Yeasts have been important model systems to stud y n uclear genome organization in the last decades, but only r ecentl y data on the composition and organization of c hr omatin in more diverse fungi became a vailable . Based on these, we sought to summarize and discuss structures homologous to those found in model eukaryotes and examined the protein complexes that are implicit in establishing these structures. We argued that the folding of c hr omatin fibers in fungi is similarl y hier arc hical as in other eukaryotes, ranging from small-scale chromatin loops of a few kilobases to large-scale subdomains comprising hundreds of kilobases that segregate chromatin into A or B compartments . T he self-interacting domains similar to metazoan TADs have been observed in yeasts (Duan et al. 2010, Mizuguchi et al. 2014, Tsochatzidou et al. 2017, Schalbetter et al. 2019, and TAD-like structures that r egionall y or ganize the genome ar e pr e v alent in most studied filamentous fungi. Ho w e v er, the pr ecise natur e of the boundary or insulator regions that allow the loading of cohesin or restrict c hr omatin extrusion r emains to be examined in detail. Furthermore, based on experimental data from several diverse fungi, we argued that Rabl chromosome conformation is the hallmark of fungal genome organization, and the independent bundling of centr omer es and telomer es driv es the ov er all nuclear or ganization of the fungal genome. Implicit in this hier arc hical or ganization is that differ ent "le v els" ar e interconnected and that changes in local c hr omatin or ganization hav e significant impact on the global nuclear organization, e.g. in formation of heterochromatic structure at centromeres and telomeres and vice versa . In addition to the critical link between gene expression and chromatin folding, se v er al conserv ed DNA-templated pr ocesses may be dir ectl y tied to genome organization, including DNA replication and repair that can alter genome and nuclear or ganization. DNA r epair e v ents can contribute to fungal evolution, as the occurrence of a double stranded DNA break and its concomitant repair onto DNA str ands spatiall y close in 3D may be the first step in structur all y varying a species' genome (Zhang et al. 2012, Hanson et al. 2021, Huang and Cook 2022. T hus , DNA-templated processes in the nucleus are often influenced by the organization of the genome, yet improper functioning of these genomic functions can feedback and alter genome organization, highlighting the intricate inter connection betw een c hr omosome conformations and genome function. While we her e striv ed to paint a complete picture on the genome organization in fungi, it is important to note that detailed, high-resolution datasets on chromosomal conformation as well as c hr omatin composition and or ganization ar e onl y av ailable for v ery fe w model fungal species. Ev en though the hier arc hical or ganization of nuclear genome organization is largely conserved between the fungal species examined to date, we also highlighted intriguing differences between species, including the speciesspecific c har acteristics defining TAD-like structur es. Recent work comparing the position of TAD-like structures and conservation of TAD boundaries between fungal strains and species of the same genus suggests that there is limited TAD variation (Torres et al. 2023 ), whic h is r eminiscent of observ ations in other eukaryotes (Rao et al. 2014, Harmston et al. 2017, Rowley et al. 2017, Krefting et al. 2018, Fudenberg and Pollard 2019, Liao et al. 2021, McArthur and Ca pr a 2021. T hese data fa vor a hypothesis where TAD-like structur es ar e gener all y conserv ed in r elated fungi but may be less conserved when greater fungal diversity is considered. Nevertheless , in Drosophila species , r earr angements occur pr edominantl y at TAD boundaries and not in TAD bodies (Liao et al. 2021 , Wright andSchaeffer 2022 ), suggesting that TAD-like structures play important roles for the evolution of genome organization. Therefor e, compar ativ e studies that systematically probe genome organization throughout the fungal lineage will help deepen understanding of the establishment, conservation, and functional implications of genome or ganization. We, ther efor e, advocate that exploring genome organization in the context of the extensive fungal biodiversity will be essential to uncover in the future how nuclear organization impacts fungal genome function and evolution.

Box 1 : The history of deciphering genome organization
Historically, studies on genome organization date to the late 1800s when Carl Rabl made his seminal observations on c hr omosome organization in eukaryotic nuclei (Cremer andCremer 2006 , 2010 ). Using light microscopy, he observed that centr omer es ar e located on one side of the nucleus and the telomer es ar e found at the opposing side, an observ ation that was consistently maintained during the cell cycle. Further advances in the 1950s, made by examining cells with electron microscopy, elucidated the partitioning of the active and silent c hr omatin: the densely compacted heterochromatin predominantly localizes at the nuclear periphery while the more-open euchromatin is mostly found in nucleus center (Shatkin andTatum 1959 , Cremer andCremer 2006 ). The advent of fluorescent microscopy allo w ed individual genomic features to be examined (Renz 2013 ), either with dyes that gener all y stained the DNA (e.g. DAPI), fluor escent pr obes to examine individual loci [e.g. fluor escent in situ hybridization (FISH) experiments], fluor escentl y ta gged pr oteins to examine the nuclear location and dynamics of proteins in vivo in live cells, or fluor escentl y labeled antibodies that highlight the localization of individual proteins in the nucleus in fixed cells . T hese advances elucidated the alternating banding pattern of euc hr omatin and heter oc hr omatin in Drosophila salivary gland polytene chromosomes (Bridges 1938, Zhimulev et al. 2014, how some genes colocalize with RNA Pol ymer ase II in possible cotranscriptional hubs (Schoenfelder et al. 2010 ), and the Rabl c hr omosome conformation that organizes fungal genomes into weak c hr omosome territories; in a Rabl conformation, the centr omer es fr om eac h c hr omosome cluster into a single focus at the nuclear periphery, while at a different inner nuclear membrane location, the telomeres of each chromosome associate into se v er al foci (Funabiki et al. 1993, Guacci et al. 1994, Goto et al. 2001, Gasser 2002 ) (Box 1 figure). Ho w ever, these methods are limited by the number of fluorescent dyes with unique emission wavelengths and the resolution of microscopic images (Lichtman andConchello 2005 , Carlton 2008 ), thereby restricting the number of loci and/or proteins that could be examined at the same time and the resolution by which we can study the interactions of individual proteins and/or colocalization of genes.
Recentl y, these c hallenges hav e been addr essed by the intr oduction of c hr omosome conformation ca ptur e (3C)-based experiments, whic h r e volutionized the field of nuclear genome organization (Dekker et al. 2002 ). In 3C, c hr omatin is cr oss-linked with nonspecific cross-linking agents (e.g. formaldehyde), underlying DNA is digested with restriction enzymes, and interacting genomic loci are ligated into a single DNA molecule. Traditional 3C uses a PCR reaction with target-specific primers to show interacting genomic loci. Coupling 3C to high-throughput sequencing (Hi-C) allo ws resear chers to examine on a genomewide scale how genomic regions interact (Lieberman-Aiden et al. 2009 ). Typicall y, Hi-C r esults ar e displayed as a two dimensional (2D) heatmap, with the x-and y-axes plotting the genomic location on the c hr omosome(s), and the intensity of color showing the strength of interactions as measured by the contact probability (Fig. 4 ), although the 3D chromosome folding can also be modeled from Hi-C data , Oluwadare et al. 2019, thereby allowing r esearc hers to infer the folding of c hr omosomes in exquisite detail and derive functional and structur al mec hanisms for how this or ganization occurs.
Box 1 F igure. Fluorescent microscop y can assess specific c hr omosomal featur es in fungi. The clustering of centr omer es, which is independent of the colocalization of telomeres, and the association of these features with the nuclear membrane is instrumental to the formation of a Rabl c hr omosome conformation in fungal nuclei. Fluorescent micrograph of conidia from a N. crassa str ain expr essing CenH3::iRFP (r ed) to illuminate centr omer es, TZA1::GFP (pr e viousl y r eported as TRF1::GFP; gr een) to highlight telomeres, and ISH1::BFP (blue) demarcate the nuclear membrane . The image on the left is an enhanced image of a single nucleus from the conidia image on the right. Methods for ima ging ar e detailed in Supplemental File 1.