The Fungal Type II Myosin in Penicillium marneffei, MyoB, Is Essential for Chitin Deposition at Nascent Septation Sites but Not Actin Localization

Genomic DNA was isolated as previously described (8). RNA was prepared by using the FastRNA red kit (BIO101). DNA-mediated transformation of P. marneffei has been previously described (8, 40). Reverse transcriptase PCR (RT-PCR) was performed by using SuperScript one-step RT-PCR with Platinum Taq (Invitrogen).

A primary clone of a myosin was obtained by PCR using degenerate primers MYO1 (5′-GGCGAGTCCGGCGCNGGNAARAC-3′) and MYO2 (5′-CGTTGGCGTAGTTGATGCAGADYTGYTCRAA-3′) directed to the motor domain and based on the CODEHOP protocol (45). Sequencing confirmed the cloning of a type II myosin-encoding gene. This PCR product was used to map the genomic locus by Southern blot analysis and as a probe to hybridize to a genomic DNA library constructed in λBlueSTAR (Novagen). A single clone of 5.8 kb was obtained, containing the 5′ end of the gene and spanning the ATG start codon and the promoter (plasmid p4699). Based on the genomic map, inverse PCR was used to clone the rest of the gene as follows. Genomic DNA was digested with EcoRI and then self-ligated and used as a template for PCR with the divergent primers myoB_inv_low (5′-GGTGGAAGTCATACGGTCG-3′) and myoB_for (5′-AAACGCCCAGACAGTGAGG-3′). The resulting PCR product was digested with different combinations of restriction enzymes and cloned into pBluescript II SK(−) (Stratagene), which was previously digested with the appropriate restriction enzymes.

The ΔmyoB deletion construct was generated by cloning an EcoRI/SmaI fragment containing the region upstream of the ATG of myoB from plasmid p4699 into the pyrG Blaster cassette plasmid pAB4626 (8) digested with EcoRI/EcoRV. The resulting plasmid was digested with XbaI/SmaI and ligated to a PCR product obtained from plasmid pDAP29 using primers M13 −21 and M13 reverse. Plasmid pDAP29 contained an EcoRI/HindIII fragment, from the inverse PCR product used to clone the 3′ half of the gene, in pBluescript II SK(−). The final construct, pDAP52, contained a deletion of most of the coding region, including the whole motor domain and part of the tail domain.

The RNA interference (RNAi) myoB construct was generated by PCR amplification of the 5′ end of the coding region with primers M13 −21 and myoB3BamHI (5′-AGGATCCCAGCCCTCATCAGTCAGTAG-3′) and cloning into pGEMTeasy (Promega). The resulting plasmid (pDAP31) was digested with BamHI/XbaI, and this fragment was cloned into the green fluorescent protein (GFP) gene-containing pALX196 digested with BglII/XbaI to produce pDAP30. The inducible promoter from the xylP gene was removed from pXYLNOM (59) by digestion with EcoRI/NcoI and cloned into pDAP30. The xylP(p)::GFP::myoBNt construct was digested with EcoRI/XbaI and cloned into the pyrG-containing pALX223 to give pDAP68. The 5′ end of the coding region of myoB (myoBNt) was PCR amplified with primers myoB3BamHI and myoBRNAiNcolow (5′-TCCATGGTTGCCGAAGTCGATGAATTGC-3′) and cloned into pGEMTeasy to give pDAP90. The myoBNt fragment from pDAP90 was removed by NcoI digestion and cloned into pDAP68 digested with NcoI, and clones were screened for the proper insert orientation. The resulting plasmid, called pDAP86, contains the RNAi-myoB construct under the control of the inducible promoter xylP(p) and the selection marker AnpyrG.

Strains were grown on Aspergillus nidulans medium (ANM) supplemented with 1% or 0.1% glucose as a carbon source and 10 mM γ-aminobutyric acid (GABA) as a nitrogen source at 25°C (15) or on either brain heart infusion (BHI) or synthetic dextrose (SD) medium supplemented with 10 mM (NH₄)₂SO₄ as a nitrogen source at 37°C. For induction of constructs under the control of the xylP promoter, strains were grown on carbon-free (CF) medium supplemented with 10 mM GABA as a nitrogen source at 25°C or on yeast nitrogen base (YNB) containing 10 mM (NH₄)₂SO₄ as a nitrogen source at 37°C and different concentrations of glucose and xylose (15, 59).

The P. marneffei FRR2161, SPM4, and cflA^D120A strains have been previously described (8). The myoB deletion strain (ΔmyoB::pyrG⁺) was generated by transformation of SPM4 with the gel-purified deletion construct derived from pDAP52 and selection for pyrG⁺ transformants. Four transformants showing a distinct growth phenotype were examined by Southern blot analysis, and of these, one strain had a banding pattern consistent with replacement of the wild-type myoB allele with the deletion allele with no ectopic copies. Due to the poor aerial growth and the lack of conidia in the ΔmyoB strain, inoculation of the mutant strain, and of the wild-type strain when it was used as a control, was performed by excising a piece of agar medium containing vegetative growth, disintegrating this, and inoculating into SD liquid medium at 37°C. The liquid culture obtained was then used for inoculation of the corresponding media.

RNAi myoB strains were generated by transformation of SPM4 with the plasmid pDAP86 and selection for pyrG⁺. Transformants were further screened for phenotypes associated with induction on CF medium with 1% xylose.

P. marneffei strains were grown on slides covered with a thin layer of solid medium, inoculated with conidia (for RNAi experiments) or vegetative biomass grown at 37°C on SD medium (for deletion mutant experiments) from the appropriate strains, and incubated at the indicated temperature. All slides (except those stained with FM4-64) were fixed in 4% paraformaldehyde for 30 min. Immunofluorescence microscopy for the detection of actin was performed using mouse C4 monoclonal antiactin antibody (Chemicon International) as previously described (10). Plasma membrane staining was performed by immersing slides in 25 μM FM4-64 suspended in water for 15 min at room temperature, washing, and mounting in Tween 80 plus 1 μg μl⁻¹ calcofluor white (CAL) as previously described (17, 20). Single-dye control experiments were performed, and these showed no bleeding into the alternative filter set. Slides were examined using differential interference contrast (DIC) and staining with fluorescent brightener 28 (calcofluor white), 4,6-diamidino-2-phenylindole (DAPI), or Hoechst 33342 and visualized on a Reichart Jung Polyvar II microscope. Quantification was performed by counting a 100 cells or septa or 50 conidiophores in three independent experiments. The average number of septa in 100 μm was calculated by recording the number of septa in 10,000 μm in three independent experiments. Images were captured using a SPOT charge-coupled device (CCD) camera (Diagnostic Instruments) and processed in Adobe Photoshop 7.0.

Strains were grown on solid ANM for 8 days at 25°C. Excised cubes of agar containing fungal cells were fixed with 1% glutaraldehyde in phosphate-buffered saline (PBS) for 2 h at room temperature and then treated with 1% OsO₄ in PBS for 2 h at room temperature. Samples were slowly dehydrated on a rotating wheel using increasing concentrations of ethanol (six steps from 10% to 100% and then a repeat 100% step, each for 15 to 20 min). For scanning electron microscopy (SEM), samples were dried in a Balzers CPD 030 critical-point dryer and gold coated in an Edwards S150B gold sputter coater. Samples were examined with a Philips XL30 FEG field emission scanning electron microscope. For transmission electron microscopy (TEM), samples were embedded on a rotating wheel using increasing concentrations of LR white resin (four steps from 25% to 100% and then a repeat 100% step, each for 12 h), and thin sections were examined with a Philips CM120 BioTWIN transmission electron microscope.

BLAST searches were performed at the NCBI (http://www.ncbi.nlm.nih.gov/BLAST) (2). Alignments and phylogenetic analyses were performed with ClustalW 1.8 software (http://www.ebi.ac.uk/FTP/index.html). Trees were bootstrapped 1,000 times to assess the reliability of each branch point and drawn using TreeView X. Predictions of coiled-coil regions were performed at MIT (http://groups.csail.mit.edu/cb/paircoil/cgi-bin/paircoil.cgi) (5) and EMBnet (http://www.ch.embnet.org/software/COILS_form.html) (32).

Of the 18 classes of myosins in eukaryotes, four classes (types I, II, V, and XVII) are present in fungi (4). Type II myosins participate in cytokinesis and are important for contraction of the actomyosin ring at the site of division in yeast (13). A fragment of a myosin-encoding gene from P. marneffei was cloned by PCR using degenerate primers directed against sequences encoding the motor domain of fungal myosin heavy chains. Sequencing of the PCR fragment and BLAST searches against GenBank revealed high levels of similarity to the motor domains of other myosin heavy chains. This sequence was used to design specific primers that allowed the cloning of the entire gene by inverse PCR, as described in Materials and Methods. The open reading frame spans 7,215 bp and is predicted to have five exons which encode a 2,404-amino-acid protein. The protein shows very high levels of identity to type II myosins in BLAST searches (expect for values of 0.0 with human, S. cerevisiae, S. pombe, and other type II myosins), and relatedness analysis confirmed that the evolution of fungal type II myosins reflects the evolutionary history of these fungi rather than the different modes of cellular division (see Fig. S1 in the supplemental material). Based on the homology the gene was named myoB, following the A. nidulans nomenclature, in which the type I myosin was named myoA (34).

A search for conserved domains in the MyoB predicted polypeptide sequence using the Pfam databases (http://www.sanger.ac.uk/Software/Pfam/) identified a 680-amino-acid myosin motor domain, which contains a putative ATP binding site, an N-terminal SH3-like fold, and an IQ motif. The coiled-coil prediction algorithms Coils (32) and PAIRCOIL (5) predicted one major coiled-coil region at the N-terminal end of the tail and a second, smaller region toward the C-terminal end, which are separated by regions containing multiple proline residues (see Fig. S2 in the supplemental material). The tails of type II myosins in the filamentous fungi Aspergillus fumigatus and Neurospora crassa and the yeasts S. pombe (Myp2p) and S. cerevisiae (Myo1p) also contains several regions with a high probability of forming coiled-coil structures that are separated by proline residues, and this differs from the single coiled-coil region found in S. pombe Myo2p and chicken skeletal myosin II (31).

Reverse transcriptase PCR (RT-PCR) analysis showed that myoB is expressed during vegetative hyphal growth at 25°C, asexual development at 25°C, and vegetative yeast growth at 37°C (see Fig. S3 in the supplemental material). Expression was slightly higher during asexual development and yeast growth when standardized against expression from the β-tubulin-encoding benA gene, presumably reflecting a higher demand for MyoB protein under those conditions. The myoB transcript was not detectable by Northern blot analysis (data not shown).

The conventional type II myosin (Myo2) is essential in S. pombe (27), as is Myo1p in some genetic backgrounds of S. cerevisiae (50). In filamentous fungi vegetative cells do not separate from each other, and we hypothesized that deletion of the type II myosin would not be lethal. A deletion construct in which the entire motor domain and a portion of the tail domain of myoB was removed and replaced with the pyrG⁺ selectable marker was generated, and this construct was used to transform P. marneffei. PyrG⁺ transformant strains were screened by Southern blot analysis for replacement of the wild-type gene with the deletion construct and the absence of ectopic copies. Four transformants which showed striking growth defects at 25°C compared to the wild-type strain were isolated, and Southern blot analysis showed that these strains were deleted for myoB. One of these strains possessed no additional ectopic copies of the deletion construct and was used for further analysis. The ΔmyoB strain produced flat colonies (no aerial hyphae) with a waxy appearance and an intense red coloration compared to the wild-type strain (Fig. 1A). Closer examination of the colony also revealed an almost complete absence of mature conidiophores.

At 37°C, the ΔmyoB strain showed poorer growth on standard defined (synthetic dextrose [SD]) medium than the wild type, and this was partially remediated by high concentrations of osmolytes such as sorbitol or NaCl (Fig. 1B). In contrast, the ΔmyoB strain was unable to grow on the standard complex undefined (brain heart infusion [BHI]) medium, even in the presence of osmotic stabilizers (Fig. 1B). At 25°C, the deletion strain was able grow on BHI medium (data not shown), suggesting that the growth defects of the ΔmyoB strain are temperature or cell type specific on this medium. Remediation of growth on defined medium at 37°C by high osmolarity suggested possible cell wall defects in the ΔmyoB strain. This was tested by growing the mutant and the wild-type strains in defined medium (ANM) containing different concentrations of the cell wall binding agent calcofluor white (CAL) at 25°C. The ΔmyoB strain showed greater sensitivity to calcofluor than the wild-type strain, supporting the hypothesis of cell wall defects (Fig. 2).

To investigate the nature of hyphal growth defects in the ΔmyoB strain, the wild-type and ΔmyoB strains were grown for 4 or 7 days at 25°C, stained with calcofluor to observe cell walls or with 4,6-diamidino-2-phenylindole (DAPI) for nuclei, and examined microscopically. At 25°C, the wild-type P. marneffei grows as septate, branched hyphae which elongate apically. Subapical hyphal cells are predominately uninucleate, whereas apical hyphal cells are multinucleate. After 4 days, only a small number of hyphae from the ΔmyoB strain exhibited morphological defects such as increased width, aberrant shape, apical branching, and multibranching (2.2% ± 0.2%, compared to 0.7% ± 0.7% for the wild type). However, unlike for the wild type, hyphae congregated as large longitudinally grouped hyphal bundles which ran along the surface of the agar (Fig. 3A and B; see Fig. S4 in the supplemental material). In contrast, after 7 days at 25°C, 30.1% ± 2.2% of ΔmyoB hyphae displayed aberrant morphology (compared to 0.7% ± 0.7% for the wild type). Defects ranged in severity across the mycelium and included less severe abnormalities such as thickened and bumpy subapical cells (Fig. 3C and D) and apical cells which were thickened, branched, and occasionally lysed (Fig. 4A) as well as severe defects such as hyphal cell collapse and excessive branching (Fig. 5A and B). ΔmyoB hyphae also displayed defects in nuclear number, size, and shape, and these defects correlated with morphological defects. The ΔmyoB hyphae displaying wild-type morphology possessed a low number of aberrant nuclei (6.3% ± 2.5%) compared to the wild type (4.3% ± 2.9%) (Fig. 4B). In contrast, ΔmyoB hyphae possessing aberrant hyphal morphology had both an elevated number of nuclei (Fig. 4A and 5A) and an increased number of nuclei (46.1% ± 2.9%), which were abnormally shaped and unevenly sized (Fig. 5C).

The phenotype of aberrant, collapsed, and multibranched apical cells is very similar to that previously observed in strains carrying dominant negative alleles of cflA (9). cflA encodes an orthologue of S. cerevisiae Cdc42p which, with the formins Bem1p and Bni1p, participates in a complex set of interactions with the various myosins (types I, II, and V) (23, 29, 57). The cflA^D120A dominant negative mutants produce hyphae which are swollen and misshapen and which have aberrant, multibranched, and fused apical cells (Fig. 5). However, in contrast to the case for the ΔmyoB strain, the aberrant apical cells of the cflA^D120A strain are also multiseptate, and, despite an increase in numbers, nuclear morphology appears to be normal (Fig. 5). This suggests that the ΔmyoB nucleation phenotype is independent of CflA.

In S. cerevisiae, the type II myosin is required for the formation of the actomyosin contractile ring, which participates in cytokinesis and cell separation (27, 50). Unlike in S. cerevisiae, in filamentous fungi cytokinesis (separation of the cytoplasm) is not accompanied by cell separation, but rather the cells remain attached, separated by cross walls called septa. To investigate whether myoB is required for septation in P. marneffei, septa were examined by calcofluor staining of cell walls in both the wild type and the ΔmyoB strain after 4 days of growth at 25°C. Wild-type septa occurred at regular intervals along hyphae, with an average of 1.42 ± 0.07 septa per 100 μm (Fig. 3A). In contrast, the ΔmyoB strain had substantially fewer septa which were unevenly distributed along the hyphae (0.10 ± 0.02 septa every 100 μm). In addition, almost all ΔmyoB septa (95.6% ± 2.0%) appeared very faint, malformed, or absent upon calcofluor staining compared to the wild type (Fig. 3B to F). At a low frequency, some hyphae also showed patches of calcofluor staining, suggesting random or delocalized chitin deposition (Fig. 3C to F).

Transmission electron microscopy (TEM) was performed on the wild-type and ΔmyoB strains to examine the septation defect. Wild-type septa appeared as a complete and distinct layer separating two cellular compartments in the hyphal cells (Fig. 6). In the ΔmyoB strain some septa appeared complete; however, many incomplete septa which failed to span the width of the hyphal cell and which displayed a serpentine shape were noted (Fig. 6). This suggests that the myoB mutant is partially impaired in the formation of septa.

The cortical cytokinetic ring in eukaryotes is composed of actin and myosin. While it has been shown that in fungi such as S. pombe and A. nidulans cytokinesis is actin dependent, it is also clear that in some fungi, such as S. cerevisiae, type II myosins play an important but nonessential role (19, 50). To assess whether actin localization at the septation site is MyoB dependent, actin was visualized by immunofluorescence in both wild-type and ΔmyoB strains (Fig. 7). In the wild type, both actin and chitin are readily simultaneously detectable at nascent septation sites (Fig. 7A), and this is followed by loss of actin staining in mature septa (10). Surprisingly, actin correctly localized to presumptive septation sites in hyphae of the ΔmyoB strain, which displayed relatively normal morphology despite the lack of any subsequent chitin deposition and septation (Fig. 7B). In two independent experiments, 100% of the wild-type apical cells with actin staining at nascent septation sites also showed chitin staining (n = 31), while 0% of the ΔmyoB apical cells with actin staining at presumptive septation sites also had chitin staining (n = 22). No transverse actin localization, indicative of sites of septation, was observed in ΔmyoB hyphae which displayed extremely aberrant morphology. Localization of actin at the hyphal apex and cortical actin patches was indistinguishable in the wild-type and ΔmyoB strains (Fig. 7C and D).

During cytokinesis, the actomyosin ring contracts, which leads to membrane invagination. The formation of the primary septum separates the two cells and splits the plasma membrane. As the ΔmyoB strain has normal actin localization at nascent septation sites but lacks chitin at septa, we assessed whether this mutant was capable of cytokinesis during division. The wild type and ΔmyoB strains were grown for 3 days at 25°C and costained with calcofluor and the lipophilic membrane dye FM4-64 (see Materials and Methods). In the wild type, FM4-64 staining was observed around the cell periphery, surrounding vesicles at the hyphal apex, as a crescent at the presumptive Spitzenkorper, and colocalized with chitin at septation sites as transverse membranes partitioning the hyphae into separate cellular compartments (Fig. 8A). Plasma membrane staining by FM4-64 in the ΔmyoB strain was also noted at the cell periphery, as vesicles at the hyphal apex, and at the presumptive Spitzenkorper (data not shown). However, in contrast to the case for the wild type, transverse membranes were not readily observed, suggesting that membrane invagination during cytokinesis was defective, thus failing to produce distinct cellular compartments (Fig. 8A). Of the small number of transverse membranes which were observed in the hyphae, some but not all were associated with weakly calcofluor-stained septa. In addition, a large number of circular membranes accumulated along the hyphae of the ΔmyoB strain (Fig. 8A), and a proportion of these membranes colocalized with the delocalized chitin spots (Fig. 8B and C). These membranes could be either vesicles or abnormal plasma membrane invaginations.

Mature conidiophores were not readily observed on the colony surface of the ΔmyoB strain (Fig. 1A). To examine this defect further, the wild-type and the ΔmyoB strains were incubated in defined medium with low (0.1%) glucose (carbon-free [CF] medium), which strongly induces asexual development, for 7 days. These cells were stained with calcofluor to observe cell walls or Hoechst 33258 for nuclei and viewed microscopically. The wild type undergoes asexual development by producing stalks from the vegetative hyphal cells, followed by the differentiation of sterigmata (metula and phialide) cells by sequential budding from the tip of the stalk. Metulae are produced from the stalk and these bud phialides, which in turn repeatedly bud to produce long chains of conidia (asexual spores) (Fig. 9; see Fig. S4 in the supplemental material). All budded cell types of the conidiophore are uninucleate, and each cell type is separated by a septum, reflecting the tight coupling of nuclear division and cell division. Cell-cell adhesion between mature conidia is easily disrupted to liberate free conidia. In contrast to those of the wild type, the conidiophores of the ΔmyoB strain lacked clearly defined cell types. Typically stalks produced a “phialide-like” cell from which a single terminal conidium was observed. Occasionally one additional malformed phialide-like cell was also noted, which often failed to produce conidia (Fig. 9A; see Fig. S4 in the supplemental material). In the wild type, the conidium-phialide boundary is clearly marked by a septum. Either the conidiophore cells of the ΔmyoB strain lacked this septum or it was malformed (Fig. 9B). In addition, septa separating the metulae and phialides were also absent or partially formed (Fig. 9A and C). DAPI staining showed that all conidiophores of the ΔmyoB strain also possessed cells with nuclear abnormalities, and in contrast to the wild type, in which each budded cell type of the conidiophore contains a single nucleus, most of the conidiophore cells (95.7% ± 0.9%) of the ΔmyoB strain contained nuclei with an aberrant morphology (Fig. 9D). The nuclei appeared fragmented, clumped, and poorly defined, suggesting division defects (Fig. 9D). Unlike the compact nucleus observed in the terminal conidium of the wild type, the nuclei in the terminal conidia of the ΔmyoB conidiophores were unevenly shaped and string-like (Fig. 9E and F).

P. marneffei is unique among dimorphic fungi in that yeast cells divide by fission rather than budding. In the fission yeast S. pombe, which unlike most other fungi has two type II myosins, the myo2 gene is essential while the myp2 gene is dispensable (6, 27). Yeast morphogenesis in the wild-type P. marneffei is evident after 4 days at 37°C, when cellular division and nuclear division become coupled in cells known as arthroconidiating hyphal cells, and this is followed by the deposition of double septa at sites of cell division and cell separation at these sites to liberate uninucleate yeast cells. This morphogenetic process is termed arthroconidiation, and the resultant yeast cells subsequently divide by fission. Microscopic examination of the ΔmyoB strain after 4 days at 37°C showed that it cannot undergo arthroconidiation (data not shown). Instead, filamentous growth continues, in which septum staining by calcofluor was hardly discernible. The hyphae showed different degrees of branching, and aberrant branching occurred in some apical cells. Similar to the case at 25°C, some cells collapse and show uncontrolled nuclear division. These data suggest that myoB type II myosin is important for cell division during yeast morphogenesis and growth in P. marneffei.

As a consequence of the cellular defects exhibited by the ΔmyoB strain, it was not possible to genetically complement the mutation. To show that the observed phenotypes are due to loss of myoB, an RNAi strategy in which a hairpin construct was under the control of an inducible promoter was adopted (see Fig. S5 in the supplemental material). This SPM4 recipient strain was transformed with the RNAi construct, and transformants were selected for uracil prototrophy (PyrG⁺) under noninducing conditions. Transfer of the transformants to medium containing increasing concentrations of inducer led to increasing levels of growth inhibition at 25°C (see Fig. S5 in the supplemental material). The strongest phenotype was equivalent to that noted for the deletion strain, namely, an absence of mature conidiophores and a waxy colony appearance with dark red coloration. Microscopic examination of the RNAi myoB strains showed wild-type morphology on noninducing medium with septum formation at regular intervals, production of conidiophores under the appropriate conditions at 25°C, and yeast morphogenesis (see Fig. S5 in the supplemental material). Similar to the case for the deletion mutant, when the RNAi strains were incubated in the presence of inducer, very faint or no calcofluor staining of septa was observed. Thus, all of the phenotypes of the deletion strain were recapitulated by the RNAi strains.

The division of cellular cytoplasm during mitosis (cytokinesis) is an essential process during the proliferative growth and development of all organisms. This study has shown that the type II myosin in P. marneffei, encoded by myoB, is required for the formation of the primary septum during all modes of division. At 25°C, ΔmyoB hyphae and conidiophores elaborate substantially fewer septa, which are often malformed. Similarly, at 37°C, the arthroconidiating hyphae display very few septa, and consequently no uninucleate yeast cells are produced. Thus, both budding and fission modes of division are affected by deletion of myoB. Calcofluor staining showed that chitin was absent from nascent sites of septation and poorly deposited in the few mature septa, which were often incomplete and failed to span the width of the hypha.

Actin localization at nascent septation sites is unaffected in the ΔmyoB strain despite the deletion strain exhibiting cytokinesis defects and a lack of chitin deposition at these nascent septation sites. This is unexpected, as myosin is essential for actin ring formation in S. cerevisiae, S. pombe, and Ashbya gossypii (7, 25, 31). The simplest explanation is that MyoB is not required for the localization of actin at nascent septation sites in P. marneffei but is essential for actin ring contraction during cytokinesis. Another possible explanation for the lack of MyoB-dependent actin ring formation is that P. marneffei can use a second, actomyosin-independent pathway to compensate for the loss of myoB and allows the formation of actin rings. This may also account for the partially formed and weakly staining septa. In S. cerevisiae an independent pathway involving Cyk3p, Hof1p, and Inn1p facilitates actin ring formation and cytokinesis in the absence of Myo1p (7, 22, 28, 44, 52, 53). P. marneffei has a conserved Cyk3p homologue (required for actin ring formation), but homologues of either Hof1p (regulates actomyosin ring dynamics and interacts with formins and septins) or Inn1p (required for the ingression of the plasma membrane) are poorly conserved, with BLAST E values of e⁻¹⁷ and e⁻⁵, respectively (data not shown). It is possible that like in S. cerevisiae, the P. marneffei Cyk3p homologue allows formation of actin rings in the absence of MyoB; however, unlike in S. cerevisiae, cytokinesis cannot proceed due to a lack of Hof1p and Inn1p activity. Searches for homologues of Hof1p and Inn1p in all annotated fungal genome databases showed that they are poorly conserved outside the Saccharomycotina.

Although actin localization is unaffected when myoB is deleted, the ΔmyoB strain lacks chitin deposition at forming septa and accumulates vesicles in hyphae which colocalize with the abnormal chitin deposits. Based on FM4-64 staining, these vesicles are derived from endocytic recycling. It is currently unclear whether the endosomes contain the chitin or are merely located at the same cellular position, which may be locations of failed or incomplete septation. This raises the possibility that MyoB may be playing a role in the transport of chitin-containing endosomes to the nascent septation sites during cytokinesis. In mammalian cells it has been shown that membrane insertion at the cell center (midbody) is required for abscission of the narrow bridge formed as the cleavage furrow ingresses during cytokinesis (reviewed in reference 37). The membrane vesicles inserted at the midbody are derived from endocytic recycling, but how these endosomes are transported to the midbody remains unclear. The microtubule motor dynein is responsible for endosome clustering during cytokinesis; however, none of the kinesins characterized have been shown to control membrane transport to the bridge (37). This theory is further supported by previous studies which have shown that myosin types V and I are required for transporting vesicles along actin filaments in fungi. Type V myosins are required during exocytosis to deliver secretory vesicles to the growth region in S. cerevisiae and to the hyphal apex (and Spitzenkorper) in the plant pathogen Ustilago maydis (18, 24, 46, 47). In contrast, type I myosins play a role during endocytosis and have been shown to mediate the endocytic uptake of the marker dye FM4-64 into the vacuole in A. nidulans and Candida albicans (41, 58). This study suggests that type II myosins may be required for the endocytic transport of vesicles to nascent septation sites during cytokinesis.

Interestingly, the ΔmyoB mutant displays a phenotype which bears a striking similarity to that of the A. nidulans ΔcsmA mutant. Deletion of csmA, which encodes a class V chitin synthase with a myosin motor-like domain (class XVII myosin), results in hyphal lysis and bundling, poorly developed conidiophores, increased sensitivity to calcofluor, and structurally abnormal septa (21). This might indicate either that the vesicles transported to septa by MyoB contain CsmA or that MyoB and CsmA have overlapping roles at the nascent septa.

In contrast to hyphal cells, all cell types of the conidiophore are uninucleate and each cell type is separated by a septum, reflecting a tight coupling of nuclear and cellular division. The ΔmyoB conidiophores lacked clearly defined cell types and contained multiple nuclei with aberrant morphology. Interestingly, the absence of clearly defined sterigmata cell types may indicate that the sequential production of metulae, phialides, and conidia by budding requires septation to be successfully completed. This hypothesis is further supported by the phenotype of the P. marneffei ΔpakB strain (CLA4 homologue) in which although all conidiophore cell types were observed, multiple chains of conidia were not produced due to defects in septation at the phialide-to-conidium cell boundaries (12). Likewise, mutation of the A. nidulans septin gene, aspB, also resulted in conidiophores which were arrested at the vesicle stage of development (54). The presence of multiple nuclei in the ΔmyoB conidiophores indicates that lack of septation does not block nuclear division in the P. marneffei conidiophore. For A. nidulans it is thought that passage through mitosis is required to activate septation (56), and although this is likely to be so for P. marneffei, the myoB mutation leads to a failure to produce wild-type septa.

The nuclei in the ΔmyoB conidiophores also appeared clumped. This is likely to be a secondary effect arising from the septation defects exhibited in ΔmyoB conidiophores. This result suggests that the conidiophore septa may be playing a spatial role in the placement of nuclei in the conidiophore. It is unlikely that MyoB is playing a role in nuclear distribution, as it is generally accepted that microtubules and their associated motors control nuclear migration in filamentous fungi (48). However, there is a small possibility that MyoB may play a role in nuclear positioning in conidiophores, as there is some evidence that in S. cerevisiae actin and actin-dependent motors might be required for certain steps in nuclear migration, specifically during the orientation of the spindle pole body (nuclear envelope-embedded microtubule-organizing center), which is required to orient the nucleus with respect to the growth axis of the cell during division (35, 49). This might be crucial during processes which require a switch from a syncytial stage to one with a tight coupling of nuclear and cellular division such as conidiation. In addition, the S. cerevisiae MYO1 mutant, in addition to displaying an increase in nuclei numbers and a change in morphology, also displays aberrant nuclear positioning during division (7, 53). During division in S. cerevisiae, the nucleus has to migrate only a small distance from a random position in the mother cell to the budding neck, where it undergoes mitosis to provide each cell with a nucleus; however, this is clearly a myosin-regulated process.

MyoB clearly does not play an important role during nuclear migration in hyphae. The ΔmyoB aberrant apical hyphal cells also possessed an elevated number of nuclei, while subapical hyphae with relatively normal morphology showed a normal nuclear distribution, suggesting that MyoB is not required for proper nuclear distribution in hyphae. The elevated numbers of nuclei observed in apical cells are likely to be a secondary effect of disrupting septation due to the mode of duplication in filamentous fungi. In A. nidulans nuclear division in apical cells is initiated in waves which extend basally, and this mitotic wave is concluded by septum formation in the apical cell (14, 26, 54). Defects in apical septum formation may result in the loss of the signal to halt the wave of nuclear division.

In summary, this study has shown that the type II myosin in P. marneffei is required for the formation of the primary septum in all cell types (vegetative hyphal, yeast, and differentiating conidiophore cells) and modes of division (budding and fission). This indicates that although fungi have evolved different modes of division to suit their life cycle or environment, the molecular mechanisms they utilize have been conserved and are modified to suit the particular application. Understanding how fungi regulate syncytial versus uninucleate growth remains an exciting area for exploration, and filamentous and dimorphic fungi have much to contribute.