Woronin bodies move dynamically and bidirectionally by hitchhiking on early endosomes in Aspergillus nidulans

The proper functioning of organelles depends on their intracellular localization, mediated by motor protein-dependent transport on cytoskeletal tracks. Rather than directly associating with a motor protein, peroxisomes move by hitchhiking on motile early endosomes in the filamentous fungus Aspergillus nidulans. However, the cellular function of peroxisome hitchhiking is unclear. Peroxisome hitchhiking requires the protein PxdA, which is conserved within the fungal subphylum Pezizomycotina, but absent from other fungal clades. Woronin bodies are specialized peroxisomes that are also unique to the Pezizomycotina. In these fungi, multinucleate hyphal segments are separated by incomplete cell walls called septa that possess a central pore enabling cytoplasmic exchange. Upon damage to a hyphal segment, Woronin bodies plug septal pores to prevent widespread leakage. Here, we tested if peroxisome hitchhiking is important for Woronin body motility, distribution, and function in A. nidulans. We show that Woronin body proteins are present within all motile peroxisomes and hitchhike on PxdA-labeled early endosomes during bidirectional, long-distance movements. Loss of peroxisome hitchhiking by knocking out pxdA significantly affected Woronin body distribution and motility in the cytoplasm, but Woronin body hitchhiking is ultimately dispensable for septal localization and plugging.


Introduction
Within a eukaryotic cell, organelles are precisely positioned to regulate their cellular function. The intracellular trafficking of these cargos is primarily achieved by molecular motors that bind different cargos and directly transport them along cytoskeletal filaments like microtubules or actin 1,2 . Microtubules are polar structures that serve as the primary track for transporting cargo long distances in eukaryotes. They possess a fast-growing plus end typically oriented towards the cell periphery and a minus end originating at a microtubule organizing center. Kinesin motors transport cargo predominantly towards the plus end of microtubules (anterograde) 3 . Conversely, cytoplasmic dynein motors ("dynein" here) transport cargo toward microtubule minus ends (retrograde) 1 .
The canonical view of intracellular transport is that a cargo attaches directly to a motor protein for movement, typically via a protein adaptor 1,4 . However, a novel form of intracellular transport termed "hitchhiking" was recently discovered, whereby a primary cargo directly interacts with a motor and a secondary hitchhiking cargo is transported indirectly via association with the primary cargo (review 5 ). Hitchhiking-like phenomena have been observed in diverse cell types and organisms such as filamentous fungi 6,7 , animal neurons 8,9 , and plant cells 10,11 .
In Aspergillus nidulans, peroxisomes move long distances along microtubules by hitchhiking on motile early endosomes transported by kinesin-3 or dynein 7 . We previously identified a putative molecular tether (peroxisome distribution mutant A; PxdA) and a phosphatase (DenA-interacting phosphatase A; DipA) that are essential for peroxisome hitchhiking on early endosomes 7,12 . Both PxdA and DipA are found on early endosomes and loss of either compromises long distance motility of peroxisomes, resulting in an accumulation of peroxisomes in the hyphal tip without affecting early endosome motility or distribution 7,12 . Peroxisomes are the only organelle demonstrated to hitchhike on early endosomes in A. nidulans 12 , but the cellular basis for why peroxisomes hitchhike is unclear.
Using comparative genomics, we found that PxdA is conserved within the Pezizomycotina subphylum of ascomycete filamentous fungi, suggesting that a function for peroxisome hitchhiking might also be conserved within this clade. We therefore investigated if hitchhiking was required for septal plugging, a peroxisome function unique to the Pezizomycotina.
Septal plugging is a critical process within Pezizomycotina fungi as they possess continuous hyphal compartments separated only by a septal wall containing a pore 13 . This interconnected hyphal network or "mycelium" has the advantage of quickly exchanging components throughout the organism 14,15 . However, if one portion of the organism is injured or lysed, significant cytoplasmic leakage can occur. In many species within the Pezizomycotina, peroxisome derivatives known as Woronin bodies plug septal pores upon mycelial injury, allowing the rest of the mycelium to remain intact 16 . Woronin bodies can also function to dynamically open and close pores to control hyphal heterogeneity 14 . Woronin body biogenesis is thought to initiate at the growing hyphal tip, based on a spatial bias in mRNA expression of the essential Woronin body component Hex-1 in Neurospora crassa 17 . After biogenesis at the hyphal tip, Woronin bodies are thought to be transported to nascent septal sites, where they bud from peroxisomes and remain tethered at the septa 18,19 . However, the mechanism by which Woronin bodies are transported to septa is unknown.
In this study, we investigated whether peroxisome hitchhiking is important for Woronin body motility, positioning, and function in A. nidulans. We find that Woronin body proteins associated with peroxisomes move dynamically and bidirectionally, and that this motility is dependent upon PxdA-mediated hitchhiking on early endosomes. We show that Woronin bodies separate from peroxisomes once localized to septa, but this fission process does not require PxdA. Finally, we show that the distribution of Woronin bodies in hyphal tips requires hitchhiking, but this is ultimately dispensable for the positioning of Woronin bodies at septa and their function in septal closure after hyphal injury.

Peroxisome hitchhiking proteins and Woronin body specialization are conserved within the Pezizomycotina fungi
Peroxisome hitchhiking has been directly observed in two species of fungi, A. nidulans and Ustilago maydis 6,7 . However, a cellular function for peroxisome hitchhiking has not been established. Two early endosome-associated proteins, a putative molecular tether (PxdA) and a phosphatase (DipA), have been shown to be essential for peroxisome hitchhiking in A. nidulans (Fig. 1A-B) 7,12 .
To search for a potential cellular role for peroxisome hitchhiking, we used comparative genomics to identify which fungal clades likely exhibit PxdAdependent peroxisome hitchhiking. We used BLASTp to search for DipA and PxdA protein homologs in 33 fungal species representing distinct clades across the fungal tree, with the underlying assumption that fungal species encoding both PxdA and DipA in their genomes would exhibit peroxisome hitchhiking. Putative PxdA homologs were only identified in the Pezizomycotina clade and were present in all 15 Pezizomycotina species we analyzed ( Fig. 1C; Table S1). Peroxisome hitchhiking has also been observed outside of the Pezizomycotina, in the basidiomycete U. maydis 6 . We did not identify a homolog of PxdA in U. maydis, suggesting that peroxisome hitchhiking may function by a distinct mechanism in this species. However, it is possible that U. maydis possesses a PxdA homolog that is sequence divergent and therefore not identified by our methods.
Putative DipA homologs were identified in all queried Pezizomycotina species as well as in other fungal clades (Table S1). DipA functions in cellular processes beyond peroxisome hitchhiking, including septal positioning 20 , . CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for this this version posted January 21, 2023. ; https://doi.org/10.1101/2023.01.20.524968 doi: bioRxiv preprint likely explaining why DipA is found more broadly in the fungal kingdom.
As putative homologs for both PxdA and DipA were found only in the Pezizomycotina, we hypothesized that a cellular function for peroxisome hitchhiking may also be specific to these fungi. The plugging of septal pores by Woronin bodies, which are derived from peroxisomes, also occurs exclusively within the Pezizomycotina ascomycetes 21 . Additionally, previous work has suggested that Woronin bodies might require active transport to reach nascent septal assembly sites prior to budding from peroxisomes 19 . We therefore sought to test if peroxisome hitchhiking affected Woronin body localization and function in septal pore plugging in A. nidulans.

Characterization of Woronin body biogenesis and distribution in Aspergillus nidulans
To determine if Woronin body function requires hitchhiking, we first needed to reliably distinguish peroxisomes and Woronin bodies from one another in live cells. We achieved this by quantifying the colocalization and co-occurrence of fluorescently tagged proteins that were either an essential Woronin body component or a peroxisome marker. We chose SspA and HexA as representative Woronin body proteins for A. nidulans based on their involvement in Woronin body formation and function, which has been most extensively examined in N. crassa, another member of the Pezizomycotina ( Fig.  1C; Table S1).
In N. crassa, the proteins WSC (Woronin sorting complex; SspA in A. nidulans) and Hex-1 (HexA in A. nidulans) are essential for Woronin body formation and function, respectively. Previous work has shown that peroxisomes begin developing Woronin bodies with the import of Hex-1 via its C-terminal peroxisome targeting signal (PTS1) 16 . After import, Hex-1 self-assembles into a dense crystalline core, the main structural component of the Woronin body 16,22 . The Hex-1 complex then polarizes to one side of the peroxisome due to the presence of the WSC protein in the peroxisome membrane 23 . This polarization is followed by organelle fission and separation of the Woronin body from the peroxisome 17,24 .
. CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in These colocalization experiments in live cells established that SspA could serve as a sufficient marker for later stages of Woronin body maturation. We therefore used this as a tool to visually distinguish how Woronin body (SspA-mGFP5) and peroxisome (mCherry-PTS1) colocalization varies at different intracellular positions in wild type hyphae ( Fig. 2A). We began by scoring the fraction of colocalized Woronin body and peroxisome puncta in the cytoplasm (Fig. 2B) versus at septa (Fig. 2C-D). Cytoplasmic Woronin bodies were frequently associated with a peroxisome in wild type hyphae (Fig. 2E) and rarely found without an attached peroxisome (Fig. 2B). In comparison, at septa, about half of Woronin bodies had no associated peroxisome signal ( Fig. 2D-E). For those septal Woronin bodies with corresponding peroxisome signal, the Woronin body signal was always closer to the septa ( Fig. 2C; white arrow). This preferred orientation is likely due to Woronin body tethering at septal sites by HexA 27 .
To better measure the relative amounts of peroxisome versus Woronin body proteins in the cytoplasm and at septa, we quantified the fluorescent intensity of SspA and HexA relative to PTS1. The ratio of SspA to HexA was consistent at both septa and in the cytoplasm (Fig. S1G), suggesting Woronin body composition does not vary much between both locations. In contrast, septal puncta exhibited a significantly higher intensity ratio of SspA to PTS1 than cytoplasmic puncta (Fig. 2F). We observed a similar pattern for the ratio of HexA to PTS1 (Fig. S1H). Together, this data supports the idea that most Woronin bodies have separated from peroxisomes by the time they associate with septa.
Finally, we wanted to estimate how many cytoplasmic peroxisomes were actively developing Woronin bodies. We observed that 67% of all cytoplasmic peroxisomes had an associated Woronin body (Fig. 2G). Overall, these data show that not all peroxisomes in the cytoplasm are actively developing Woronin bodies in A. nidulans. However, for peroxisomes that have begun developing Woronin bodies, the two organelles remain attached in the cytoplasm, and only separate shortly before or immediately following recruitment to septa.
Previous work has established that in most Pezizomycotina including A. nidulans, Woronin bodies are tethered at septa 18 . However, a small number of species including N. crassa use a divergent mechanism of Woronin body tethering, whereby Woronin bodies are delocalized from septa and instead are tethered to the actin cortex throughout a hyphal compartment 28 . Our observations in A. nidulans on the localization of Woronin body maturation and separation from peroxisomes after septal association is similar to what was previously reported for N. crassa 28 , despite these species exhibiting different Woronin body localization and tethering mechanisms.

Woronin bodies move bidirectionally using the peroxisome hitchhiking protein PxdA
Woronin body biogenesis is thought to occur primarily in the cytoplasm near the hyphal tip 17,18 . Woronin bodies are then localized to their tethering site at or near septal pores, where they function to seal septal pores following injury 19 . However, the mechanism by which they are transported to septa is unknown. We therefore wanted to characterize Woronin body movement in hyphae and determine if peroxisome hitchhiking was responsible for Woronin body movement and transport to septa.
We showed that Woronin bodies frequently colocalize with peroxisomes in the cytoplasm but are present alone at septa (Fig. 2E). This suggests that Woronin bodies could either (1) move in conjunction with hitchhiking peroxisomes and bud off once present at septa or (2) bud off peroxisomes in the cytoplasm and be transported to septa by a distinct mechanism. To test this, we first compared Woronin body and peroxisome movement in adult hyphae. In A. nidulans, only a small population of peroxisomes exhibit long-distance movements at a given time 7,29 . We found that Woronin bodies also exhibit infrequent but dynamic bidirectional movements, like peroxisomes (Movie S1). As expected 7 , 10% of total movements for both peroxisomes and Woronin body movements were processive over long-distances (>3 µm), indicative of microtubule-driven movement. During these long-distance movements, Woronin bodies moved at similar speeds ( Fig. 3A) and over similar distances to peroxisomes (Fig. 3B). Both peroxisomes and Woronin bodies also exhibited bidirectional movements with no directional bias (Fig. 3C).
Next, we sought to test how frequently Woronin bodies and peroxisomes colocalize during long-distance runs (>3 µm travelled in one direction). The large majority of moving Woronin bodies were colocalized with a peroxisome, and vice versa ( Fig. 3D-E). This finding was surprising as we previously observed that only two thirds of cytoplasmic peroxisomes had associated Woronin body signal (Fig. 2G). Furthermore, we found no obvious bias in whether the Woronin body was 'leading' or 'trailing' the peroxisome during motile runs, and instead observed the signals frequently overlapped during movement (Fig. 3F). This contrasts with our observations on stationary cytoplasmic peroxisomes and Woronin bodies, which frequently exhibited Woronin body signal that was polarized and offset to one side of the peroxisome signal (Fig. 2B). Together, this data suggests that Woronin body proteins move via association with a hitchhiking peroxisome, and not as a separate / distinct organelle.
We next tested if Woronin body movement required the essential peroxisome hitchhiking protein PxdA. Compared to wild type hyphae, in which peroxisomes and Woronin bodies are consistently distributed across hyphae (Fig. 3G), pxdAΔ hyphae had an accumulation of both peroxisomes and Woronin bodies at the hyphal tip (Fig. 3H) and a drastic reduction in the number of Woronin bodies and peroxisomes moving in either direction ( Fig. 3C; Movie S2). Furthermore, motile Woronin bodies colocalized with the putative hitchhiking tether PxdA ( Fig. 3I-J). These data demonstrate that Woronin bodies require PxdA for longdistance, bidirectional movements in hyphal tips.
While loss of PxdA abolished most long-distance movements, a small number of peroxisomes and Woronin bodies still moved processively in pxdAΔ hyphal tips (Fig. S1K-L). However, these Woronin bodies were still colocalized with peroxisomes during this movement (Fig.  3E). Additionally, the overall Woronin body and peroxisome co-occurrence (SspA/PTS1) in pxdAΔ hyphae was similar to what was observed in wild type ( Fig. 2E-G, . CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for this this version posted January 21, 2023. ; https://doi.org/10.1101/2023.01.20.524968 doi: bioRxiv preprint S1F, and S1J). These findings demonstrate that the majority of Woronin body movement occurs via association with hitchhiking peroxisomes and that PxdA is not important for the assembly of Woronin bodies.

Essential Woronin body proteins are not required for peroxisome hitchhiking
As Woronin bodies and peroxisomes comigrated during long-distance runs, and most motile peroxisomes colocalized with Woronin bodies while moving (Fig. 2E), we wondered if the presence of Woronin body proteins was required for peroxisome motility. We tested this using mutants lacking either HexA or SspA (Fig. S2A). Loss of HexA or SspA did not significantly affect the average number of moving peroxisomes per cell (Fig. S2B). In hexAΔ hyphae, peroxisomes showed a slight accumulation at hyphal tips (Fig. S2C), but to a lesser extent than the . CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for this this version posted January 21, 2023. ; https://doi.org/10.1101/2023.01.20.524968 doi: bioRxiv preprint accumulation observed in pxdAΔ hyphae (Fig. S2D). Despite nearly all moving peroxisomes possessing Woronin body proteins, these data show that the ability to form Woronin bodies is not required for peroxisome hitchhiking.

Woronin bodies localize at septa and plug damaged hyphae independently of hitchhiking
We next wanted to test whether peroxisome-Woronin body hitchhiking is required for Woronin bodies to localize to septa in undamaged hyphae. First, we quantified the cytoplasmic localization of Woronin bodies in strains lacking peroxisome movement (pxdAΔ and hookAΔ) as well as in a strain with non-functional Woronin bodies (hexAΔ). We included hookAΔ because HookA is the adaptor protein that links microtubule motors to early endosomes in A. nidulans 30 , and loss of HookA therefore disrupts both early endosome and peroxisome distribution and motility 7 . As predicted, loss of peroxisome movement in both pxdAΔ and hookAΔ hyphae results in an accumulation of SspA-labeled Woronin bodies at the hyphal tip ( Fig. 4A-B). There were occasionally very bright Woronin bodies at the apex of wild type hyphal tips, corresponding to a slight peak in intensity around 0.5 μm from the tip (Fig. 4B, black line) that was abolished in hexAΔ hyphae (Fig. 4B, red line). This peak in SspA intensity in wild type tips could correspond to higher levels of SspA import into peroxisomes in this region, or the presence of larger Woronin body/peroxisome organelles that have not yet undergone fission.
We next examined Woronin body localization at tip apical septa in undamaged, wild type hyphae as well as pxdAΔ, hookAΔ, and hexAΔ mutants. We chose tip apical septa because these septa were most recently formed 31 and are still capable of dynamically opening and closing via septal plugging 14 . Transmission electron microscopy studies in A. nidulans have previously reported between four and nine Woronin bodies at septa 18 . In assessing Woronin body localization at septa via fluorescence microscopy, three classes of Woronin body localization were identified: (i) septa with Woronin body signal on either side likely representing more than one Woronin body, (ii) septa with Woronin body signal on only one side and (iii) no Woronin body signal (Fig. S3A-B).
Manual scoring of these localization patterns revealed that most tip apical septa had Woronin bodies on one or both sides in wild type hyphae (Fig. 4C-D). Furthermore, approximately one third of tip apical septa had Woronin bodies on both sides in wild type (Fig. 4D). As expected, most hexAΔ hyphae lacked Woronin bodies at apical septa. Both pxdAΔ and hookAΔ hyphae had an intermediate phenotype, with a slight reduction in the number of septa with Woronin bodies on both sides, but similar numbers of septa with Woronin bodies on one side compared to wild . CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for this this version posted January 21, 2023. ; https://doi.org/10.1101/2023.01.20.524968 doi: bioRxiv preprint type. This finding suggests that PxdA-mediated hitchhiking could contribute to achieving Woronin bodies on both sides of nascent septa. However, fluorescence measurements show there is no significant difference in SspA signal at septa in wild type versus pxdAΔ or hookAΔ hyphae (Fig.  S3C). Overall, our data suggest that Woronin body localization at septa does not depend on Woronin body hitchhiking. Though Woronin body hitchhiking may have a minor contribution to Woronin body recruitment at septa, it is not the primary mechanism by which these organelles localize at septa. Woronin bodies in pxdAΔ hyphae could be potentially recruited to tip-apical septa by other mechanisms, such as cytoplasmic movement via 'poleward drift' 32 or passive association with septal assembly sites that are marked as the hyphal tip grows 31 .
Finally, we wanted to test if Woronin body hitchhiking was important for Woronin body-based sealing of septal pores following hyphal damage. Growth on sorbose media induced tip lysis in just under half of cells, regardless of genotype (Fig. S3D). Tip lysis and cytoplasmic puddling was clearly visible on sorbose media for all strains (Fig. 4E). The amount of leaked protein was higher in lysed hexAΔ than lysed wild type, pxdAΔ, or hookAΔ cells (Fig. 4F), despite leakage areas appearing similar in size (Fig. S3E). This pattern of cytoplasmic leakage was recapitulated when dried mycelia were exposed to hypotonic shock via distilled water (Fig. S3F). Together, these data demonstrate that Woronin body hitchhiking is not required for proper septal plugging after hyphal damage.

Concluding thoughts
This study shows that Woronin bodies move dynamically and bidirectionally in the hyphal tip by hitchhiking on PxdA-labeled early endosomes in A. nidulans. Though Woronin bodies have been described to move in some fungi (Z. tritici 33 ; A. fumigatus 19 ; A. nidulans 26 ), their motility has not been well characterized and the mechanism by which they move was previously unknown. We showed that Woronin bodies and peroxisomes often coexist as connected organelles in the cytoplasm, where they hitchhike together on early endosomes via PxdA with no directional bias (Fig. 4G). We have identified that virtually all peroxisomes colocalize with a Woronin body while hitchhiking, despite only two thirds of peroxisomes colocalizing with Woronin bodies while immotile in the cytoplasm. Despite this intriguing finding, we show that Woronin body proteins are not required for the identification of peroxisomes as hitchhiking cargo.
It was previously hypothesized that Woronin body biogenesis begins in the hyphal tip, after which they are transported retrograde towards tethering sites at nascent septa [17][18][19] . We found that Woronin body movement in hyphal tips is bidirectional and requires the hitchhiking protein PxdA. Regardless, we ultimately determined that Woronin bodies do not require PxdA-mediated hitchhiking to reach nor associate with septa, thus disproving the hypothesis that they are actively transported retrograde towards septa.

Acknowledgements
LDS is funded by an NSF-GRFP and the UCSD Quantitative Integrative Biology T32 from the NIH (1T32GM127235). JRC is funded by a MOSAIC award (K99/R00) from the NIH (K99GM140269). SLRP is supported by HHMI and NIH (1R35GM141825). We thank John Salogiannis and Valentin Wernet for ideas and insight, Kaeling Tan for cloning RPB564, James Holcomb for cloning RPB2190, and Xin Xiang for the mGFP5 plasmid. This paper was typeset with the bioRxiv word template by @Chrelli: www.github.com/chrelli/bi-oRxiv-word-template

Author contributions
LS, JRC, and SLR-P devised the experiments. LS and DB performed the experiments. JRC and SLR-P supervised the research. LS, JRC, and SLR-P wrote and edited the manuscript.

Competing interest statement
Authors declare that they have no competing interests.

Strain construction
A. nidulans strains used in this study are listed in Table S2. New strains were generated by transforming PCR-amplified targeting DNA into A. nidulans protoplasts 34 or via genetic crosses 35 . The genotypes of novel strains were confirmed by PCR amplification from isolated genomic DNA 36 and Sanger sequencing.

Plasmid cloning
Plasmids and primers generated for this study are listed in Table S3 and Table S4, respectively. All DNA constructs were designed for homologous recombination at the endogenous locus with the endogenous promoter unless otherwise stated. Novel plasmids were cloned using Gibson isothermal assembly 37,38 . Briefly, fragments were amplified using PCR (NEB Q5 highfidelity DNA polymerase, M0491) from other plasmids or A. nidulans genomic DNA. Then, fragments were assembled into the Blue Heron Biotechnology pUC vector at the 5′EcoRI and 3′HindIII restriction sites. The cloning of mTagGFP2-rabA::AfpyrG, mCherry-FLAG-PTS1::AfpyroA, and 2xBFP-PTS1::AfriboB were described previously 39 . The cloning of pxdA-mKate2::AfpyroA and pxdAΔ::AfriboB were also previously published 7,12 . All fluorescently tagged proteins had a GA5 linker. The sequence for mGFP5 was optimized for expression in A. thalania plants 40 . To grow adult hyphae for imaging, spores were inoculated into MM agar with supplements and incubated for 16-20hr at 37°C. Individual colonies were cut from the agar and inverted onto 35 mm #1.5 glass-bottom imaging dishes (Cellvis D35C4201.5N). To grow germlings for imaging, spores were collected into 0.01% Tween-80 (Sigma-Aldrich P1754) and their density was measured using a hemocytometer. The spores were then inoculated into liquid MM with supplements at a density of 5x10 5 spores/mL in a four-chamber 35 mm dish and incubated for 16-20 hr at 30°C in an unsealed humid plastic container.

Phylogenetic analysis of peroxisome hitchhiking proteins
Pezizomycotina fungi included in this analysis were previously published to have Woronin bodies and/or have been referenced as representatives of different clades 42,43 . Species were only included if they had their whole genome sequenced and proteome annotated and/or drafted at the time of this analysis (30 August 2022). The phylogenetic tree for fungi was drawn manually based on the taxonomic rankings of each species on NCBI.
Homologous genes for PxdA (AN1156) and DipA (AN10946) were identified using BLASTp available at NCBI (https://blast.ncbi.nlm.nih.gov/) (parameters: E = 1.0x10 -5 , word size of 6, BLOSUM62 matrix, gap costs existence 11 and extension 1, filtering low complexity regions, minimum percent coverage 20%). Reciprocal BLASTs were used to confirm all candidate homologs. The reference proteins database (refseq_proteins) was used to reduce redundancy of search results. For species where the refseq_proteins database was unavailable, the nonredundant protein sequences database (nr) was used instead. The E value, query coverage, and percent identity of the top hits for a reference strain of each species was recorded (Table S1).
. CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for this this version posted January 21, 2023. ; https://doi.org/10.1101/2023.01.20.524968 doi: bioRxiv preprint To identify PxdA homologs, three query sequences were used: the PxdA coiled-coil region, previously defined as CC1-3 (aa 1466-2115) 7 ; the PxdA uncharacterized N-terminus (aa 1-1465); and the full-length PxdA protein (aa 1-2236). The conserved stretch of amino acids in the PxdA CC3 domain are aa 2033-2058. The coiled-coil region of all PxdA homologs was individually predicted by querying each sequence in the Waggawagga coiled-coil prediction software (https://waggawagga.motorprotein.de/) with the following settings: Marcoil, Scorer 2.0 oligomerization, and window length of 21. PxdA possesses a predicted coiled-coil region that is necessary and sufficient for peroxisome hitchhiking 7 . The coiled-coil prediction software Waggawagga estimated between three to six distinct coiledcoils in each PxdA homolog, with the total coiled-coil region varying from 167 to 307 aa long (Table S1). All PxdA homologs possessed an N-terminal extension of variable length with no discernable conserved domains, based on NCBI and PFAM databases.
For identifying DipA homologs, query sequences included both the full-length DipA protein (aa 1-704) and the DipA metallophosphatase domain alone (aa 45-96). DipA homologs were defined to have both a metallophosphatase domain followed by a conserved domain of unknown function (DUF2433). The amino acid boundaries of both the metallophosphatase domains and the DUF2433 were identified for all DipA homologs using NCBI domain prediction and MUSCLE multiple protein sequence alignments to the A. nidulans DipA reference sequence in SnapGene.

Spinning disk confocal microscopy
Live cell imaging was performed at room temperature using a Nikon Ti2 microscope mounted with a Yokogawa W1 spinning-disk confocal scan head and two Prime95B cameras (Photometrics). All images were acquired in 16-bit format. The NIS Elements Advanced Research software (Nikon) was used to run the microscope using the 488 nm and 561 nm lasers of a six-line LUN-F-XL laser engine (405 nm, 445 nm, 488 nm, 515 nm, 561 nm, and 640 nm). The stage xy position was controlled by a ProScan linear motor stage controller (Prior) and its z position was controlled by a piezo Nano-Z stage positioner (Mad City Labs). An Apo TIRF 100x 1.49 NA objective (Nikon) was used for all experiments except for the imaging of apical septa for hyphae grown in glucose and sorbose agar, which required an S Fluor 40x 1.30 NA objective (Nikon) to capture a larger field of view.
Simultaneous dual-color imaging was used to assess colocalization of Woronin body and peroxisome proteins in germlings. Z-stacks were collected to capture the complete volume of germling cells in the field of view (0.2 µm step size, 8-10 µm total depth). The 488 nm and 561 nm lines were used to excite mGFP5-and mCherry-tagged fusion proteins, respectively. The emission was split with a Cairn TwinCam with a 580LP filter. The GFP/488 emission was reflected and passed through a 514/30 bandpass filter onto camera two. The mCherry/561 fluorescence was passed through a 617/73 and an additional 600/50 bandpass filter to camera one. The cameras were aligned manually in NIS Elements using 0.1 µm Tetra-Speck beads (Thermofisher T14792). Image acquisition settings, such as exposure time and laser power, were determined at the beginning of each imaging session to eliminate fluorescent bleed through between each channel.
Triggered acquisition of 488/561 was used when imaging Woronin body association at apical septa with the 40x objective and when imaging SspA-mGFP5 comigration with PxdA-mKate with the 100x objective. The firing of the 488 nm and 561 nm lasers was synchronized by the Prime95B camera trigger signal, which was integrated into a Nikon BB connected to a 6723 DAQ board housed in an external Pxi chassis (National Instruments). A quad bandpass filter (Chroma ZET405/488/561/640mv2) was placed in the emission path of the W1 scan head.

Quantitative image analysis
All image analysis was performed using ImageJ/FIJI (National Institutes of Health, Bethesda, MD). All microscopy images were blinded prior to analysis using a custom script in R 44 . Statistical tests were performed using GraphPad Prism (Windows version 9.4.1). Code availability: All macros (ImageJ, Windows version 1.53q) and data processing scripts (R, Windows version 4.1.0; Rstudio, Windows version 2022.12.0 Build 353) are available on GitHub (https://github.com/LiviaSongster).

Colocalization and co-occurrence
To prepare Z-stacks for colocalization analysis, whole germlings were cropped, max-Z projected, and saved as separate files. For both the 488 and 561 channel, a Gaussian Blur (sigma = 1) and background subtraction (rolling = 50) were applied prior to thresholding (Yen dark) to generate a binary puncta mask. The threshold for the green channel was set at 15 to 65535, and for the red channel at 20 to 65535. All gray values outside of that channel's mask were cleared to eliminate background. Colocalization was then quantified using Pearson's coefficient of correlation and Mander's coefficient of co-occurrence algorithms via the JACoP plugin in ImageJ 45,46 .
The fraction of SspA puncta with PTS1 and the fraction of SspA puncta with PTS1 was manually scored for four separate experiments containing at least ten germlings each. Germlings were imaged for 60 sec (300 ms interval) with a single z-plane using simultaneous dual-color settings. Puncta in the cytoplasm or at septa were manually segmented as an ROI and scored as colocalized if there was a punctum from both channels that overlapped by >50% and if the mean gray value of the puncta of the opposite channel was at least 10% brighter than the background. Motile puncta were assessed by first generating kymographs, and for directed runs where a punctum traveled >2 µm, the co-occurrence of puncta in both channels was manually scored.
To measure the fluorescence intensity ratio of different proteins, puncta from sum Z projections of germlings were manually circled as regions of interest (ROIs). The gray value of all puncta was measured in both the 488 and 561 channel and the intensity ratio for each ROI was calculated as the mean gray value of the 488 channels divided by the mean gray value of the 561 channels.

Organelle flux
Total puncta movement was estimated from single-plane, triggered-acquisition movies (30 sec total length, 300 ms interval, 100 ms exposure per laser). The number of puncta crossing a perpendicular line approximately 10 μm from the hyphal tip was manually counted.

Intensity profile during movement
The intensity along a line (1 pixel width) was measured during a directed run for both the 488 and 561 channel. The line scans from multiple kymographs were compiled in R. For each line scan, the maximum intensity in the 561 channel was set as x position 0 μm and then applied to the 488 channel.

Speed and distance travelled of migrating puncta
Adult hyphae were imaged for 60 sec (300 ms interval) with a single zplane using simultaneous dual-color settings. The line tool was used to trace puncta trajectories over time and movies were resliced along these lines to generate kymographs. The instantaneous speed and run lengths of individual moving puncta were calculated from the inverse of the slopes of manual kymograph traces. A punctum was scored as moving if it exhibited a directed run (no pauses) > 3 μm during its movement.

Hyphal tip line scans
Whole adult hyphae were imaged live as a Z-stack (0.2 µm step size, 8-10 µm total depth). Max-Z intensity projections of fluorescence micrographs were obtained and the brightfield channel was used to trace from the hyphal tip inwards using the segmented line tool (pixel width 20). The traces were superimposed on the fluorescence micrographs to project the average fluorescence intensity along the line. The mean background intensity was measured for each cell and subtracted from that cell's respective line scan prior to binning.

Woronin body localization at tip apical septa
Cells expressing mCherry-SsoA and SspA-mGFP5 were grown overnight with supplements in either regular MM (1% glucose) or sorbose MM (2% sorbose/0.1% glucose). Adult hyphae were imaged using the 40x objective to collect complete volume Z-stacks (0.2 µm step size, 8-10 µm total depth). Images were max-Z projected and the apical septa were manually identified as ROIs and cropped. Crops of single septa were blinded and Woronin body localization was manually scored as the fraction of apical septa with an associated Woronin body per experiment. The mean fluorescence intensity of SspA-mGFP5 from these max-Z projections was measured within a 0.5 µm-wide by 1 µm-long rectangle centered at an apical septum. The mean background intensity adjacent to each septum was measured and subtracted from the mean septal intensity. Background-corrected measurements were then normalized to the average septal intensity of wild type.

Cytoplasmic leakage
Cytoplasmic leakage was measured for burst hyphal tips in sorbose agar. First, complete Z stacks (0.2 µm step size, 4 µm total depth) were imaged, sum-Z projected, and blinded. The integrated density (mean intensity of puddle times puddle area) for each cytoplasmic puddle was measured in ImageJ and corrected by subtracting the product of the mean background intensity and puddle area. This background-corrected integrated density was then normalized to the mean observation for wild type to calculate the normalized amount of leaked cytoplasm.
As a complimentary approach, mycelia were also subjected to hypotonic shock to induce global tip lysis, using an adapted protocol for A. nidulans 47 . Spores from strains expressing a cytoplasmic GFP were inoculated . CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for this this version posted January 21, 2023. ; https://doi.org/10.1101/2023.01.20.524968 doi: bioRxiv preprint into liquid YG media and grown overnight at 37°C. Mycelia was collected in Miracloth (EMD Millipore 475855) and dried with paper towels. Approximately 300 ± 10 mg dried mycelia were submerged in 1 mL sterile ddH2O and vortexed to induce hypotonic damage at hyphal tips. Samples were spun at 15,000 x g for 1 min in a tabletop centrifuge to pellet cell debris and the supernatant was diluted 1:10 in sterile ddH2O for a 96-well microplate Bradford microassay (Bio-Rad 5000201) to estimate the amount of leaked protein. The absorbance at 595 nm was measured using a Cytation 5 plate reader (BioTek) and Gen5 data analysis software (Windows version 3.05).
. CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for this this version posted January 21, 2023. ; https://doi.org/10.1101/2023.01.20.524968 doi: bioRxiv preprint Movie S1 (separate file). Woronin body (SspA-mGFP5) and peroxisome (mCherry-PTS1) comigration in a wild type hyphal tip. A single Z plane image was taken every 300 ms via simultaneous dual-color spinning disk confocal microscopy (Nikon). The white arrow indicates a comigrating, bidirectional run. Scale bar = 10 µm.
Table S1 (separate file). PxdA (sheet 1) and DipA (sheet 2) BLASTp results for all 33 fungal species tested. The amino acid sequences used for each BLASTP search are shown in row 1 on both sheets.
. CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for this this version posted January 21, 2023. ; https://doi.org/10.1101/2023.01.20.524968 doi: bioRxiv preprint Table S2. Strains used in this study.
. CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for this this version posted January 21, 2023. ; https://doi.org/10.1101/2023.01.20.524968 doi: bioRxiv preprint Table S3. Plasmids generated for this study.
. CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in . CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in  Table S4. Primers used to make plasmids for this study.
. CC-BY 4.0 International license perpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for this this version posted January 21, 2023. ; https://doi.org/10.1101/2023.01.20.524968 doi: bioRxiv preprint