QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

Vol. 17, Issue 10, 4494-4512, October 2006

This Article Abstract Full Text (PDF) Supplemental Material All Versions of this Article: E06-03-0215v1 17/10/4494    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Helfer, H. Articles by Gladfelter, A. S. Search for Related Content PubMed PubMed Citation Articles by Helfer, H. Articles by Gladfelter, A. S.

AgSwe1p Regulates Mitosis in Response to Morphogenesis and Nutrients in Multinucleated Ashbya gossypii Cells

Hanspeter Helfer^*, and Amy S. Gladfelter^*,

*University of Basel Biozentrum, Molecular Microbiology, 4056 Basel, Switzerland; and Department of Biology, Dartmouth College, Hanover, NH 03755

Submitted March 21, 2006; Revised July 12, 2006; Accepted July 31, 2006
Monitoring Editor: Kerry Bloom

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Nuclei in the filamentous, multinucleated fungus Ashbya gossypii divide asynchronously. We have investigated what internal and external signals spatially direct mitosis within these hyphal cells. Mitoses are most common near cortical septin rings found at growing tips and branchpoints. In septin mutants, mitoses are no longer concentrated at branchpoints, suggesting that the septin rings function to locally promote mitosis near new branches. Similarly, cells lacking AgSwe1p kinase (a Wee1 homologue), AgHsl1p (a Nim1-related kinase), and AgMih1p phosphatase (the Cdc25 homologue that likely counteracts AgSwe1p activity) also have mitoses distributed randomly in the hyphae as opposed to at branchpoints. Surprisingly, however, no phosphorylation of the CDK tyrosine 18 residue, the conserved substrate of Swe1p kinases, was detected in normally growing cells. In contrast, abundant CDK tyrosine phosphorylation was apparent in starving cells, resulting in diminished nuclear density. This starvation-induced CDK phosphorylation is AgSwe1p dependent, and overexpressed AgSwe1p is sufficient to delay nuclei even in rich nutrient conditions. In starving cells lacking septins or AgSwe1p negative regulators, the nuclear density is further diminished compared with wild type. We have generated a model in which AgSwe1p may regulate mitosis in response to cell intrinsic morphogenesis cues and external nutrient availability in multinucleated cells.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Multinucleated cells are common throughout the biosphere. Filamentous, pathogenic fungi, metastasizing tumor cells, and the cells of the musculoskeletal and blood systems of mammals are examples of the diverse types of cells that contain many nuclei in one cytoplasm. In some cases these cells are produced from specialized cell cycles in which nuclear division occurs without cell division leading to large syncytia. For fungi, multinucleated cells may extend over hundreds of meters so that different regions of a single cell experience dramatically different microenvironments.

Mitosis in multinucleated cells can occur either in a coordinated, synchronous manner where all nuclei divide simultaneously or asynchronously where individual nuclei divide independently in time and space. Synchronous, parasynchronous, and asynchronous patterns of mitosis are observed in syncytial cells, and the different modes of nuclear division bring different potential advantages to cells (Nygaard et al., 1960; Clutterbuck, 1970; Gladfelter et al., 2006). In synchronous division, nuclei in distant regions of the cell can be coordinated, and a cell can globally respond to a signal or stimulus with a limited spatial distribution. Furthermore, synchronous mitoses such as those observed in Aspergillus nidulans tip compartments are coordinated with morphogenesis such as tip growth and septation (Harris and Momany, 2004). In asynchronous division, the cell can restrict mitosis to particular nuclei, providing a more local and spatially controlled response to a signal.

Asynchronous or local control of mitosis, which is observed in several filamentous fungi including Neurospora crassa and Ashbya gossypii, may enable cells to target both growth and nuclear division into nutrient-rich regions, while arresting these processes in areas of the cell lacking sufficient resources (Minke et al., 1999; Freitag et al., 2004; Gladfelter et al., 2006). Additionally, asynchronous control of mitosis may allow cells to directly link morphogenesis programs, such as branching patterns and septal positioning, to nuclear division. Thus external triggers such as nutrients and/or internal signals involving cell shape could spatially direct asynchronous mitosis in multinucleated cells.

In uninucleated Saccharomyces cerevisiae cells, cell shape (proper bud formation) is tightly coordinated with nuclear division through a set of cell cycle regulatory proteins that are anchored to the septin scaffold at the mother-bud neck (Kellogg, 2003; Lew, 2003). The septins are conserved filament forming proteins that assemble into heteromeric complexes with diverse functions in eukaryotes, including cytokinesis, morphogenesis, and secretion (Gladfelter et al., 2001; Longtine and Bi, 2003). When bud formation is delayed due to defects in the actin cytoskeleton or if the septin scaffold itself is perturbed, then the intrinsically unstable Swe1p kinase (Wee1 homologue) is stabilized and translocates to the nucleus where it can inhibit the cyclin-dependent kinase/Clb (CDK), Cdc28/Clb2p, through phosphorylation of the tyrosine 19 residue on Cdc28p (Sia et al., 1996; McMillan et al., 1998; Sia et al., 1998; McMillan et al., 1999a, 1999b). This inhibitory phosphorylation provides a G2-delay for cells to recover proper morphogenesis before mitosis, ensuring aberrant binucleate cells do not form in the absence of budding. On restoration of budding or adaptation, the Tyr19 phosphorylation is reversed by the phosphatase Mih1p (Cdc25 homologue), which promotes mitotic entry. Swe1p is then directed to the neck and phosphorylated by kinases, including the CDKs, Cla4p (PAK) and Cdc5p (Polo-like kinase; Asano et al., 2005; Sakchaisri et al., 2004; no. 6325). Neck localization of Swe1p is mediated by Hsl7p, which is anchored to the septins by Hsl1p (Barral et al., 1999; Shulewitz et al., 1999; Longtine et al., 2000; Cid et al., 2001). After phosphorylation, Swe1p is targeted for ubiquitin-mediated degradation in the proteasome (Sia et al., 1998). Thus the septin scaffold functions to coordinate bud morphogenesis and nuclear progression by regulating the abundance of Swe1p, and collectively these factors make up the budding yeast "morphogenesis checkpoint."

The homologues of all of these factors including the septins, Swe1p, Hsl1p, Hsl7p, and Cdc28p are present with varying degrees of homology in the genome of the multinucleated, filamentous fungus A. gossypii. This ascomycete diverged from a common ancestor of S. cerevisiae over 100 million years ago and shares many similar genomic features with this budding yeast yet never reproduces by budding and rather is exclusively found in a filamentous, multinucleated hyphal form (Dietrich et al., 2004). Given the absence of budding yet the conservation of all factors involved in the morphogenesis checkpoint, we speculated that septins may link morphogenesis to the nuclear division cycle in this filamentous fungus by directing where mitosis takes place in the cell. Thus here we have investigated if and how the septins may spatially direct mitoses in multinucleated hyphae.

We show here that the septin proteins assemble into rings in areas where growth and mitoses are most frequent, at branchpoints and growing tips in multinucleated A. gossypii cells. This spatial pattern of nuclear division requires septins and involves regulation of the AgSwe1p kinase in response to internal cues for branching. Additionally, AgSwe1p also regulates nuclear progression in response to external nutrient status through inhibitory phosphorylation of the CDK. We propose that this dual role for AgSwe1p may be used to facilitate a spatially controlled reaction to limited nutrient availability in the natural world.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
A. gossypii Methods, Media, and Growth Conditions
A. gossypii media, culturing, and transformation protocols are described in Wendland et al. (2000) and Ayad-Durieux et al. (2000). A. gossypii strains used here are listed in Table 1. To test the influence of high culture density, 10–20 ml of full medium was densely inoculated with spores (10x higher spore inoculum), and the cultures were grown until vacuoles were clearly observable by phase contrast microscopy. For all other conditions, 50–100 ml of full medium were inoculated at low density and regularly controlled for the absence of visible vacuoles, and fresh medium was added 1–2 h before taking samples or exposing the cultures to conditions of potential stress. Nocodazole stocks for the arrest and release experiment were 3 mg/ml in DMSO and were used at a final concentration of 15 µg/ml (Sigma, St. Louis, MO). Hydroxyurea was directly added to the cultures to a final concentration of 50 mM. Rapamycin stocks (LC Laboratories, Woburn, MA) were 1 mg/ml in 90% EtOH and 10% Tween-20 and were used at a final concentration of 200 nM (Loewith et al., 2002). To simulate the starvation response, low-density cultures were washed twice and resuspended in 40 mM MOPS, pH 7.0, 137 mM KCl. To test the lack of certain nutrients, different media based on Ashbya Synthetic Dextrose (ASD) medium were used: 1.70 g/l YNB without amino acids and without ammonium sulfate (Difco, Detroit, MI), 0.69 g/l CSM-LEU (BIO 101, Carlsbad, CA), 1.00 g/l Asn, 0.01 g/l adenine, 1.00 g/l myo-inositol, 20 g/l glucose, and 50 mM MOPS, pH 6.5.

A C-terminal fusion of GFP to AgSep7p was obtained by amplifying pGUG and pGUC with S7-G-S1 and S7-G-S2 or S7-G-S1, and S7-G-CS2, respectively. To generate AgSEP7-GFP-NAT1, the pGUC cassette was first constructed by digesting pGUG with BglII and BamHI and ligating the gel-purified 1026-base pair fragment into the dephosphorylated pUC19NATPS vector, which was linearized with BamHI.

The S7-G–S1 primer was designed with homology to the C-terminal region of AgSEP7 before the stop codon. The PCR products were each transformed together with pAG4401, containing AgSEP7, into the S. cerevisiae strain CEN.PK2 (Entian and Koetter, 1998) for in vivo recombination. The resulting plasmids pAGHPH002 and pAGHPH003 were purified and transformed into A. gossypii wt cells, leading to the strains AgHPH05 and AgHPH06. For integration into the genome, the two constructs were first subcloned into pUC19 to eliminate ARS activity. pUC19 was opened with SalI and XbaI, AgSEP7-GFP-GEN3 was excised from pAGHPH002 with EcoRI, and SphI and pAGHPH003 was cut with XhoI and SpeI to extract AgSEP7-GFP-NAT1. The fragments were gel-purified and separately ligated into the dephosphorylated pUC19 fragment to produce pUCHPH002 and pUCHPH003. The cassettes were excised with EcoRI and transformed into A. gossypii wt cells, giving rise to AgHPH07 and AgHPH08.

To generate ASG41 (AgHSL7-GFP-GEN3), the oligos Hsl7-S1 and Hsl7-S2 (with homology to the C-terminus of AgHSL7) were used with the template pGUG to generate the PCR product that was used to directly transform A. gossypii cells.

To make the nonphosphorylatable Y18F mutation of AgCDC28, the plasmid pAGHPH007 was created using in vivo gap repair in S. cerevisiae. First, AgCDC28 was amplified from genomic DNA using oligos Cdc28aIEcoR1 and Cdc28aIIBamH1, with homology upstream of AgCDC28 including the promoter region and limited downstream sequence. This PCR fragment was purified and digested with EcoRI and BamHI, sites present in the oligos, to clone it into pRS416 that was opened with these same enzymes. The entire AgCDC28 ORF was sequenced to confirm that PCR did not introduce point mutations. pRS416CDC28GEN3 was cut with XhoI at two sites, one being 23 base pairs upstream of Y18, the other being located in the pRS416 backbone. The resulting fragments were cotransformed into the S. cerevisiae strain DHD5 (Arvanitidis and Heinisch, 1994) together with the annealed oligonucleotides CDC28-Y18F-A/CDC28-Y18F-B and CDC28-glue-A/CDC28-glue-B for in vivo recombination. The pair CDC28-Y18F-A/CDC28-Y18F-B overlapped the first XhoI site and contained the Y18F replacement flanked by a total of 15 silent mutations to improve the fidelity of subsequent genomic integration into A. gossypii. The second XhoI site was overlapped by the pair CDC28-glue-A/CDC28-glue-B, which contained no mutations. The gap-repaired plasmid (pAGHPH007) was isolated and verified by sequencing of the altered regions. pAGHPH007 was digested with EcoRI, BamHI, and PstI and transformed into A. gossypii wt cells, resulting in AgHPH36. Correct integration and presence of the altered region was verified by analytical PCR with the oligonucleotides CDC28-IB/CDC28-Y18F-IA (binds to altered sequence, only) and G3/CDC28-G4.2 and CDC28-IB/CDC28-IA-Y18 (binds to wt sequence, only). For additional verification, a fragment containing the altered sequence was amplified using CDC28-G1.2/CDC28-IB and digested with KpnI, which cuts the mutated fragment at two places and the wt fragment at one site, only.

The plasmid carrying AgSWE1 (pAGHPH004) was generated by amplification of the gene from genomic A. gossypii DNA using the primers SWE1-B1 and SWE1-B2, which contain EcoRI and BamHI restriction sites. The fragments were purified, cut and ligated into pRS416. The resulting plasmid was verified to contain no mutations by sequencing.

To C-terminally tag AgSwe1p with GFP or 13-myc, the pGUC or pAG13-myc cassette was amplified with SWE1-G-S1 and SWE1-G-NS2 or SWE1-MS1 and SWE1-MS2, respectively. The resulting PCR products were cotransformed into the S. cerevisiae strain DHD5 together with pAGHPH004 for in vivo recombination, giving rise to pAGHPH005 and pAGHPH006. Correct fusion was verified by sequencing. Transformation of the plasmid pAGHPH006 into A. gossypii wt cells produced the strain AgHPH34. To obtain AgHPH32, pAGHPH005 was digested with BmgBI and BssHII before transforming it into A. gossypii wt cells, leading to genomic integration.

For overexpression of AgSWE1, the plasmid pAGHPH008 was constructed: the ScHIS3 promoter was amplified off pAGrPXC (kindly provided by Dominic Hpfner) using the primer pair MH1 and SWE1-SH2. The GEN3 cassette was amplified using the oligonucleotides MH2 and SWE1-SH2 with pGEN3 as template. MH1 and MH2 contained homology to both, the GEN3 cassette and the ScHIS3 promoter, generating an overlap between the two PCR products. SWE1-SH1 and SWE1-SH2 added 45-base pair homology to a region upstream of AgSWE1 START and 54-base pair homology to the very START of AgSWE1. The two PCR products were used together as template in a subsequent PCR reaction with the primers SWE1-SH1 and SWE1-SH2. The resulting GEN3-ScHIS3 cassette was fused in front of AgSWE1 by in vivo recombination with pAGHPH004 in the S. cerevisiae strain DHD5, leading to pAGHPH008. The plasmid was purified and verified by sequencing and transformed into A. gossypii wt cells to produce AgHPH37. To C-terminally tag GEN3-ScHIS3/SWE1 with HA, a 980-base pair fragment carrying the carboxy end of AgSwe1p fused to 6HA was cut out of pAgSWE1-HA-NAT1 using MscI and ScaI. To generate pAgSWE1-HA-NAT1, the Swe1F1 and Swe1F2 oligos, which have homology to the C-terminus of AgSWE1 and the universal tagging cassette, were used in a PCR reaction with the pN1–6HA (pAGT105-kindly provided by Andreas Kaufmann, University of Basel, Basel, Switzerland) template, and the resulting product was cotransformed into budding yeast cells with the plasmid pAGHPH004. Recombination of the plasmid with the linear PCR fragment resulted in yeast cells resistant to CloNAT. Plasmids were rescued from these resistant yeast strains and correct fusion of 6HA to AgSWE1 was verified by restriction digest. pAGHPH008 was opened with MscI and XbaI. T4 DNA polymerase was used to fill the 5' protruding end generated by XbaI. The dephosphorylated 8910-base pair vector fragment was gel-purified and ligated with the 980-base pair insert to produce pAGHPH009. In frame fusion of the tag was verified by sequencing. Transformation of the plasmid pAGHPH009 into A. gossypii wt cells produced the strain AgHPH38.

Protein Extraction and Western Blotting
Cells grown under the desired conditions were collected, washed once with ice-cold PBS containing 1 mM Na₃VO₄ and 1 mM -glyerolphosphate, and suspended in ice-cold lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% NP-40, 2 mM sodium pyrophosphate, 1 mM Na₃VO₄, 1 mM -glycerolphosphate, and Roche complete protease inhibitor cocktail). An equal volume of 0.5-mm glass beads was added to this suspension, and cells were broken by vigorous vortexing during 6 x 30-s intervals in a bead beater at 4°C. Total cellular proteins were separated by SDS-PAGE (12%) and transferred to nitrocellulose membranes (Amersham Biosciences), using standard conditions. Blocking was performed with 5% milk in TBS-Tween-20 (0.1%). For detection of phosphorylated AgCdc28p, rabbit anti-phospho-cdc2(Tyr15) (Cell Signaling, no. 9111) was used at 1:1000 in TBS, 5% BSA, 0.1% Tween-20, and HRP-anti-rabbit (Cell Signaling, no. 7074) was used at 1:2000 in TBS, 5% milk, 0.1% Tween-20. As positive control for phosphorylated CDK, total cellular proteins were extracted from the S. cerevisiae strain DLY5544 (cdc12-6, GAL/SWE1), grown in YP 2% sucrose, and induced by adding 2% galactose. For detection of HA-tagged proteins, mouse anti-HA was used at 1:1000 in TBS, 5% BSA, 0.1% Tween-20, and HRP-anti-mouse (Jackson ImmunoResearch, West Grove, PA) was used at 1:1000 in TBS, 5% milk, 0.1% Tween-20. Detection of labeled proteins was performed by using the enhanced chemiluminescence (ECL) detection system (Amersham Biosciences). Either nonspecific cross-reacting bands or bands from Ponceau red staining of the membrane was used for load controls.

Immunofluorescence, Image Acquisition, and Processing
DNA stainings (Hoechst 33342) and immunofluorescence stainings were performed as described previously (Gladfelter et al., 2006). Rabbit anti-ScCdc11p (Santa Cruz Biotechnology, Santa Cruz, CA) was diluted in PBS and used at a 1/10 dilution. Rabbit anti-phospho-cdc2(Tyr15) was used at 1/50 for immunofluorescence. For in vivo observation of mycelium containing GFP-tagged proteins, the cells were grown for 10–12 h in full medium and then washed and resuspended in ASD or MOPS/KCl for observation under the microscope.

The microscopy setup used was the same as described in Hpfner et al. (2000), Knechtle et al. (2003), and Gladfelter et al. (2006). The following filter sets were used for the different fluorophores: no. 02 for Hoechst, no. 15 for Rhodamine 568 (Carl Zeiss, Thornwood, NY), and no. 4108 for GFP (Chroma Technology, Brattleboro, VT).

For still images, multiple planes with a distance between 0.3 and 0.5 µm in the Z-axis were taken. MetaMorph 4.6r9 software (Universal Imaging, West Chester, PA) was used for ("no-neighbors") 2D-deconvolution. The stacks were flattened into a single plane using stack arithmetic "maximize" command. Outlines of the cells were obtained by doing phase-contrast Z-series. The stacks were passed through the Metamorph filter "detect edges: Laplace 2" and flattened into a single plane using the stack arithmetic "sum" command. Fluorescent and phase-contrast images were scaled and overlaid using "color align."

Time-lapse images were taken as described in Gladfelter et al. (2006) and assembled into movies using QuickTimePro 6.5. Mitotic events in Supplementary Movie 1 were labeled manually using Adobe Photoshop CS.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Mitoses Are Enriched at Branching Points
To investigate if there is spatial control of asynchronous mitosis in multinucleated A. gossypii cells, we measured the mitotic index in tips, branchpoints, and regions between these growth areas using several methods. First, we observed cells growing on agar expressing GFP-labeled Histone H4 (AgHHF1-GFP) using time-lapse microscopy to follow nuclear dynamics and morphogenesis over multiple hours in vivo (Supplementary Movie 1). Mitoses were frequently observed near newly emerging branches at the junction between the new branch and the "mother" hypha (Figure 1A). In the frames of the movie shown in Figure 1, mitoses occurred preferentially at the branching site at 171, 183, 201, and 213 min. Mitoses observed near a newly emerging branch are labeled with a box just before division, and new daughter nuclei are then outlined with circles. In 84 total mitoses captured by time lapse, 12% were at hyphal tips (which represent 11% of total hyphal length), 51% were at current or future branching sites (which represent 30% of total hyphal length), and 37% were in the middle of hyphae (a region not near the tip or branching sites that represents 59% of total hyphal length). Thus, there were 1.7 times the number of mitoses at branches than would be expected if mitoses were distributed without respect to location. Additionally, to assay mitotic frequency and position using an alternative method, wt cells were grown in liquid medium and then processed for anti-tubulin immunofluorescence to visualize mitotic spindles. In this case, branchpoints were similarly enriched for mitoses with 45% of mitotic nuclei at these sites, whereas tips hosted 30% of the dividing nuclei and 25% were present in the interregion (N >200 nuclei, presented in graphical form in Figure 4C).

View larger version (46K): [in this window] [in a new window]   Figure 1. Mitoses are enriched at branching points. (A) AgH4-GFP cells were grown as described in Supplementary Movie 1, and images were collected at 3-min intervals. Cell outlines are in red, nuclei are in green, and mitoses are false-colored blue and labeled with a box just before division and new daughter nuclei are outlined with circles. Bar, 10 µm. (B) Schematic illustrates the method used for scoring the positions of mitoses within the developing mycelium. The closest 10 µm (shaded blue circles) to growing tips were considered as tip region T. B and FB indicate zones of 10 µm (shaded blue circles) around sites of branch formation or future sites of branch formation, respectively. The regions between tips, and branches were scored as interregions ‘I'.

Septins Localize to Branches and Tips and Influence the Position of Mitoses
We speculated that septin proteins may function as a marker directing where nuclei divide and thus localized septin proteins in A. gossypii. The A. gossypii septins were visualized in living cells using a GFP-tagged septin, AgSep7p-GFP, and by immunofluorescence using an antibody generated against ScCdc11p. In living and in fixed cells, the septins assembled into discontinuous rings at the cortex at the tips of hyphae, along the main hyphae between tips and branches and at branch sites. Rings were composed of discrete "bars" or filaments that were long and faint when present at the tip and short and bright when at the bases of branches or along the main hypha away from the tip (Figure 2A). The rings at the base of branches often appeared asymmetric with the apical half of the ring composed of short, bright bars and the basal half made up of fainter less well-organized bars (Figure 2A, inset). Overall, 18% of septin rings were found at tips, 50% were found at bases of branches, and 32% on the main hyphae (n = 127 septin rings). Fifty percent of the septin rings found on the main hypha were located just beside the spot where a branch emerged. On the basis of this localization of septins and the mitotic index near branches, we would predict that there is a higher proportion of mitotic nuclei in the vicinity of septin rings. In young cells in which septins and tubulin have been visualized by immunofluorescence, 71% of nuclei with mitotic spindles were adjacent to either a tip, main hyphal or branch septin ring and 52% of nuclei with duplicated SPBs were adjacent to septin rings (n = 230 nuclei, Figure 2B). Further support of this observation can be seen in Supplementary Movie 2, in which H4-GFP and SEP7-GFP are both monitored in living cells and mitoses are observed near septin rings.

View larger version (20K): [in this window] [in a new window]   Figure 2. Septins localize to branches and tips. (A) AgSEP7-GFP (AgHPH07) cells were grown overnight for 12 h at 30°C before live imaging. Tip rings (1 and inset, left), branch rings (2 and inset, right) and main hyphal rings (3) are visible. (B) Wild-type cells (leu2 thr4) were grown overnight for 16 h at 30°C and processed for anti-tubulin and anti-Cdc11p immunofluorescence. In overlay, DNA is blue, tubulin is red, and septins are green. (C) Agcdc12SEP7-GFP (AgHPH15) cells were grown for 12 h at 30°C and then imaged live. Bars, 10 µm.

How might the septins locally promote mitosis? Given their role as scaffolds for a multitude of factors at the mother- bud neck in budding yeast, it is conceivable that the septins are similarly recruiting either mitosis-promoting or -inhibiting factors to branching spots in A. gossypii cells. In budding yeast, the proteins involved in the morphogenesis checkpoint are recruited to septins for the inhibition of the CDK inhibitor Swe1p. We identified the A. gossypii homologues of the key components of the S. cerevisiae morphogenesis checkpoint as a basis to begin evaluating if these factors may be used to direct where mitosis is taking place in multinucleated cells that do not bud. Homology and length of these factors vary tremendously; however, core domains are conserved and they are present in syntenic positions of the genome, suggesting that protein functions such as kinase (AgSwe1p) and phosphatase (AgMih1p) activity are conserved (Table 4).

View larger version (55K): [in this window] [in a new window]   Figure 3. AgHsl7p localizes to septin-dependent cortical rings. (A–D) AgHSL7-GFP (ASG41) cells were grown for 12 h at 30°C and then imaged live. (B) 90° 3D reconstruction of the discontinuous ring shown in A. (E and F) Agcdc12HSL7-GFP (AgHPH19) and Aghsl1HSL7-GFP (AgHPH20) were grown for 12 h at 30°C and then imaged live. Phase contrast, fluorescence and merged images are ordered from left to right (A) or from top to bottom (C–F). Bars, 10 µm.

View larger version (36K): [in this window] [in a new window]   Figure 4. AgSwe1p regulation influences the position of mitoses. (A) Radial growth on solid medium of wild type (leu2 thr4) and deletion mutants Agmih1 (AgHPH23), Aghsl1 (hsl1NAT1), Aghsl1mih1 (AgHPH28), Agcdc12 (AgHPH15), Agcdc12mih1 (AgHPH27), and Agswe1 (ASG35) compared with wild type (100%; top). A small piece of homokaryotic mycelium (1 mm3) from each strain was spotted on full medium plates and grown for 4 d. Nuclei visualized in different strains (bottom). Cells (same strains as above) were grown overnight at 30°C in selective conditions and fixed and processed for Hoechst dye staining. Numbers indicate average distance between two nuclei. (B) Percentage of nuclei in different nuclear division cycle stages evaluated by Hoechst and anti-tubulin staining (Ana: anaphase; Meta: metaphase; 2 SPB: 2 spindle pole bodies; 1 SPB: 1 spindle pole body). (C) Percentage of mitoses found at sites of branch formation, at tips and in the interregion of hyphae evaluated by Hoechst and anti-tubulin staining. Bars, SE. (D) Wild-type cells (leu2 thr4) grown under low-density conditions were stained with Hoechst for DNA visualization (left) and assayed for phosphorylated CDK by immunofluorescence using the anti-phospho-cdc2(Tyr15) antibody (right). (E) Protein extracts from wt cells (leu2 thr4) grown under low-density conditions and then arrested and released from nocodazole were assayed for phosphorylated CDK on a Western blot, using the anti-phospho-cdc2(Tyr15) antibody. A nonspecific band from the Ponceau red treated membrane was used as loading control. Extracts from the S. cerevisiae strain DLY5544 (cdc12-6, PGAL1-SWE1) served as positive (+) control for this and all following Western blots. The numbers indicate the molecular weight in kDa. Bars, 10 µm.

To determine if AgSwe1p spatially influences mitosis by phosphorylating the tyrosine 18 residue of the CDK (the analogous tyrosine that has been shown in a variety of systems to be a substrate of Swe1/wee1 kinases), we applied a phospho-specific anti-cdc2(Tyr15) antibody to cells and cell extracts. First, phosphorylated CDK (AgCdc28p) was assayed for in cells by immunofluorescence to see if a subset of nuclei, such as those in the interregion far from branchpoints, may be enriched for modified AgCdc28p (Figure 4D). No phosphorylated protein was detected by this method, but potentially this could be due to technical problems resolving the protein with this antibody by immunofluorescence. Thus, whole cell extracts were also generated to evaluate if the antibody can recognize the A. gossypii CDK by Western blot. Minimal phosphorylation on AgCdc28p, however, was apparent in either asynchronously growing cells or cells that had been artificially synchronized (70% synchrony) in G2/M with nocodazole and released for progression through mitosis (Figure 4E). Thus, although AgSwe1p and septins both seem to contribute to the spatial control of mitosis, it is still unclear if they function through tyrosine phosphorylation of the CDK.

AgCdc28 Is Phosphorylated on Tyrosine 18 when Cells Are Starved for Nutrients
The absence of detectable CDK tyrosine phosphorylation led us to ask if we could ever observe such modifications during the A. gossypii nuclear division cycle, perhaps only under certain environmental conditions. Given the common role of Wee1 kinases in different checkpoint responses across eukaryotes, we evaluated if AgCdc28p was phosphorylated when cells were exposed to potential checkpoint triggers and environmental stresses including hydroxyurea (to impair DNA replication), nocodazole (to impede spindle assembly), starvation, osmotic shock, and high temperature (42°C). No or limited phosphorylation was detected using the anti-phospho-cdc2(Tyr15) antibody on whole cell extracts from cultures treated with hydroxyurea or nocodazole (Figure 5A). High-temperature stress and osmotic shock resulted in moderate phosphorylation on a protein of the predicted size of AgCdc28p. Surprisingly, starvation that was induced by high-density growth (evaluation described in Materials and Methods) resulted in a level of phosphorylation that was markedly stronger than in any other condition tested (Figure 5A).

View larger version (42K): [in this window] [in a new window]   Figure 5. AgCdc28 is phosphorylated on tyrosine 18 when cells are starved for nutrients. (A) Wild-type cells (leu2 thr4) were grown for 12 h at 30°C before exposure to checkpoint stimuli or environmental stresses followed by lysis and Western blotting of whole cell extracts using the anti-phospho-cdc2(Tyr15) antibody. When not indicated otherwise, the cells were grown at low density before applying the various stresses. Nocodazole was used at 15 µg/ml and hydroxyurea at 50 mM. As loading control, a nonspecific cross-reacting band was used for the blot at the left, and a nonspecific band from the Ponceau red treated membrane was used for the blot at the right. (B) Wild-type cells (leu2 thr4) grown under high-density conditions for 16 h were stained with Hoechst for DNA visualization (left) and assayed for phosphorylated CDK by immunofluorescence using the anti-phospho-cdc2(Tyr15) antibody (right). Bar, 10 µm. (C) Wild-type cells (leu2 thr4) were grown under high-density conditions in 20 ml of AFM until prominent vacuole formation was observed (18 h). The cells were washed and resuspended in 200 ml of fresh AFM to recover. Samples were taken after 3 and 5 h, cells were lysed, and the levels of CDK phosphorylation in whole cell extracts were detected by Western blotting using the anti-phospho-cdc2(Tyr15) antibody (left). A nonspecific cross-reacting band was used as loading control. The degree of vacuolization was observed by phase-contrast microscopy (right). (D) Wild-type (leu2 thr4) and Agcdc28Y18F (AgHPH36) cells were grown to the indicated densities, lysed, and processed by Western blot using the anti-phospho-cdc2(Tyr15) antibody. A nonspecific cross-reacting band was used as loading control.

To confirm that the phosphorylation observed is as predicted on the tyrosine 18 residue of AgCdc28p, this residue was mutated to phenylalanine, which should be nonphosphorylatable. The phosphorylated AgCdc28p recognized in wt, high-density cultures is not present in lysates generated from Agcdc28Y18F cells grown to similar high density (Figure 5D). This indicates that the anti-phospho-cdc2(Tyr15) antibody is recognizing the conserved tyrosine phosphorylation on the A. gossypii CDK.

This phosphorylated AgCdc28p observed in densely grown cultures could be formed in response to nutrient deprivation or could be a density-dependent reaction to a quorum-sensing molecule that accumulates at high density. Several approaches were taken to attempt to distinguish between these possibilities. First, cells were treated with rapamycin which likely inhibits the kinase AgTor1/2p, which has a conserved role in nutrient sensing in a variety of eukaryotes. Inhibition of Tor kinase mimics starvation in many cells and in budding yeast simulates nitrogen starvation (Martin and Hall, 2005). AgCdc28Y18 phosphorylation clearly accumulated with time in the presence of rapamycin even in cells grown at low density in rich-nutrient conditions (Figure 6A). Thus, the appearance of the phosphorylated form did not depend on high density here but did appear during this mock starvation. Furthermore, when cells grown at low density are transferred to MOPS/KCl buffer (osmotically stable but without nutrients) to induce rapid low-density starvation, CDK phosphorylation appeared within 30 min and accumulated to levels comparable to high-density starvation by 105 min (Figure 6A). Finally, the response was independent of the type of nutrient that was restricted and was observed when either carbon (AFM with 0.1% glucose) or nitrogen (ASD-Asn or with one quarter amino acid concentration) was the limiting resource even when cells were grown to low density (Figure 6B). Thus, AgCdc28Y18 phosphorylation appears in response to nutrient deprivation rather than other possible signals that may accumulate at high-density growth.

View larger version (43K): [in this window] [in a new window]   Figure 6. CDK phosphorylation produces nuclear delay in nutrient deprivation. (A) Wild-type cells (leu2 thr4) were grown under low-density conditions before adding 200 nM rapamycin or transferring the cells to MOPS/KCl (40 mM MOPS, pH 7.0, 137 mM KCl) except for the left lane where cells were grown to high density for a positive control. CDK phosphorylation was detected by Western blotting as described above. (B) Low-density wt cells (leu2 thr4) were grown in full medium (AFM) for 12 h before shifting them into media limited for different resources: AFM Glc 0.1% (20 times less glucose than standard growth media), ASD (synthetic media without yeast extract), ASD-Asn (synthetic media lacking asparagine), and ASD with 0.25 normal amino acid concentration. Samples were taken 5 h after the media shift and assayed for CDK phosphorylation as described above. (C) Wild-type (leu2 thr4) and Agcdc28Y18F (AgHPH36) cells were grown to the noted densities, fixed, and stained with Hoechst to visualize nuclei. Bars, 10 µm.

If such a delay is stochastic such that limited nutrients can block nuclei in any stage of the nuclear division cycle, then high-density cultures would be expected to have similar proportions of nuclei in each nuclear division cycle stage as those found in low-density cultures. If, however, the delay is regulated such that it occurs in a specific window of time, the proportions of nuclei in each nuclear division cycle stage should vary between the different growth conditions. Under high-density growth conditions, more nuclei accumulated in cells with duplicated SPBs and metaphase spindles compared with low-density cultures. Cells in high-density cultures contained 41% nuclei with duplicated SPBs compared with 28% in low density and 16% metaphase nuclei compared with 4% in low-density cultures (N >200 nuclei). Thus the deprivation of nutrients leads to accumulation of nuclei in specific stages of division, both just before and during metaphase (summarized as part of Figure 7B).

View larger version (38K): [in this window] [in a new window]   Figure 7. AgSwe1p is the kinase responsible for AgCdc28Y18 phosphorylation. (A) Wild-type (leu2 thr4), Agswe1 (AgHPH24), and Agmih1 (AgHPH23) cells were grown to high density, and whole cell extracts were assayed for CDK phosphorylation as described above (left). Wild-type (leu2 thr4) and Agswe1 (AgHPH24) cells were grown overnight to low density and then incubated with rapamycin (200 nM). A nonspecific cross-reacting band was used as loading control. (B) Wild-type (leu2 thr4), Agswe1 (AgHPH24), and PScHIS3-AgSWE1 (AgHPH37) cells were grown at 30°C to high/low density and then fixed and processed for anti-tubulin immunofluorescence to evaluate the percentage of nuclei in different nuclear division cycle stages. (C) Nuclei of cells grown for (B) visualized by Hoechst staining. Bar, 10 µm. (D) Whole cell extracts from PScHIS3-AgSWE1 (AgHPH37, lanes 1 and 2), PScHIS3-AgSWE1-6HA (AgHPH38, lanes 3 and 4), and AgSWE1-6HA (AgHPH39, lanes 5 and 6) were generated from cells grown to either high or low density. Extracts were then probed by Western blot using either the anti-phospho-cdc2(Tyr15) or anti-HA antibody. A nonspecific cross-reacting band was used as loading control. (E) Wild-type (leu2 thr4), Agmih1 (AgHPH23), Agswe1 (AgHPH24), Agcdc28Y18F (AgHPH36), PScHIS3-AgSWE1 (AgHPH37), Aghsl1 (hsl1NAT1), Aghsl1mih1 (AgHPH28), Agcdc12 (AgHPH15), and Agcdc12mih1 (AgHPH27) cells were grown to either high or low density and fixed, and nuclei were visualized by Hoechst staining to calculate the average distance between two nuclei for each strain and condition. Error bars reflect weighted SEs.

Are the altered nuclear density and nuclear division cycle proportions observed under high-density conditions due to the action of AgSwe1p? Agswe1 cells grown to high density failed to respond as strongly to these conditions, but still showed some increase in nuclear density (8.0 µm [N >500 nuclei] compared with 7.9 µm in Agcdc28Y18F and more than 10 µm in wild type, Figure 7C). Agswe1 mutant cells had fewer nuclei with duplicated SPBs than wt cells at high density but retained a similar proportion mitotic nuclei (Figure 7B). These results suggest that AgSwe1p induced AgCdc28Y18 phosphorylation is responsible for a delay in nuclear division in G2 and that some other factors are responsible for the altered proportion of metaphase nuclei in high-density growth.

To determine if AgSwe1p activity was sufficient for AgCdc28p phosphorylation and altered nuclear division even in the absence of high-density starvation, AgSwe1p was overexpressed in cells grown in low-density conditions. For this experiment the AgSwe1p promoter was replaced with the S. cerevisiae HIS3 promoter, which leads to high, constitutive expression of proteins in A. gossypii (Dominic Hpfner, personal communication). Protein levels were assayed by Western blot against a 6HA-epitope tag fused to the C-terminus of AgSwe1p to confirm that the HIS3 promoter leads to overexpression of the AgSwe1p protein (Figure 7D). Additionally, AgSWE1 expression under its own promoter was evaluated for regulation and to confirm that the HA fusion protein was functional during high- and low-density growth. Under control of the native promoter, AgSwe1p-6HA was barely detected in low density cultures, whereas in high-density growth, it was somewhat more abundant. Unlike in wt cells, phosphorylated AgCdc28p was readily detected in low-density conditions when AgSwe1p was overexpressed from the ScHIS3 promoter (lanes 2 and 4 compared with lane 6, Figure 7D).

To determine if this Y18 phosphorylation of AgCdc28p could alter nuclear cycle progression even in low-density cultures, nuclear density was evaluated in these cells overexpressing AgSwe1p. The overexpressed AgSwe1p and presumably the subsequent AgCdc28p phosphorylation led to a 50% decrease in nuclear density with an average distance between nuclei of 8.4 µm compared with 4.6 µm wild type at low density (Figure 7C). Additionally, the overexpressed AgSwe1p led to a dramatic delay in the nuclear division cycle leading to a population in which almost 60% of nuclei had duplicated SPBs compared with only 28% in wt, low-density cultures, further suggesting that AgSwe1p-induced AgCdc28p phosphorylation acts before metaphase (Figure 7B).

Nuclear Delay in Response to Starvation Is Exacerbated in Septin Mutants
Is AgSwe1p-induced AgCdc28p phosphorylation in response to low nutrients regulated by septins and AgHsl1p? If these factors do negatively regulate AgSwe1p in high-density growth, then cells lacking these controls may show an exacerbated response to starvation with even larger distances  between nuclei. To begin to evaluate if these factors contribute to AgSwe1p regulation during starvation, nuclear density was evaluated in Aghsl1, Agcdc12, Aghsl1, Agmih1, and Agcdc12mih1 mutants in high-density cultures. Agcdc12 (13.3 µm), Aghsl1mih1 (14.7 µm), and Agcdc12mih1 (15.5 µm) mutants all had in average higher nuclear densities than wild type (10.2 µm) or even cells overexpressing AgSwe1p (11.0 µm). One caveat to these experiments however is that the slower growth rate of septin mutants necessitates a longer period of overall growth to achieve high density, and thus the growth time is not identical between septin mutant strains and other strains. Nonetheless these data raise the possibility that the septins and AgHsl1p contribute some form of negative regulation to AgSwe1p during high-density growth (Figure 7E).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
In this work, we have sought to identify mechanisms for how mitosis may be spatially regulated in multinucleated hyphal cells. We have shown that septins promote mitosis in the vicinity of branchpoints and that this is likely regulated through AgSwe1p activity. This represents a novel role for a Wee1 kinase in spatially directing mitosis in multinucleated cells. AgSwe1p also promotes AgCdc28Y18 phosphorylation to limit mitosis when these cells are starved for nutrients. Based on this data, it appears that Swe1p in A. gossypii responds both to intrinsic morphogenesis cues and external environmental conditions. Here we discuss these data and propose a hypothetical model in which activity of AgSwe1p and assembly of septins may enable these multinucleated cells to react to a spatially limited pool of nutrients and restrict mitoses to the area of the nutrient source (Figure 8).

View larger version (20K): [in this window] [in a new window]   Figure 8. Model of spatial control of mitosis by septins and nutrient availability. (A) Wild-type cells grown in the laboratory have mitoses enriched at branch sites near assembled septin rings. (B) As branch sites mature, new septin rings in developing branches acquire mitosis promoting capacity, leading to mitotic events in interregions. (C) Mitoses are no longer localized preferentially to branch sites in cells lacking septins. (D) In low-nutrient conditions, nuclei are arrested or delayed with duplicated SPBs and high CDK-Tyr 18 phosphorylation levels. (E). Hypothesis for how mitoses may be spatially controlled in response to heterogeneous nutrient supplies in the natural world, where a single hypha may simultaneously experience "feast and famine" conditions. Local pools of nutrients may trigger assembly of a septin ring, which in turn promotes local mitosis through the down-regulation of AgSwe1p, enabling the cell to exploit nutrients without duplicating nuclei distant to nutrient source.

Complex patterns of septin organization have also been observed in a variety of other multinucleated and filamentous fungi (Douglas et al., 2005; Gladfelter, 2006). In Candida albicans hyphae, the septins establish a nuclear division plane in hyphal cells. In Aspergillus nidulans, septin ring assembly at the bases of branches coincides with the reentry of nuclei into mitosis, but the link between septins and the nuclear division machinery is still unknown in this system (Westfall and Momany, 2002).

What are possible ways the septin rings at branchpoints could influence the mitotic machinery in A. gossypii? Based on the wiring of the morphogenesis checkpoint in S. cerevisiae and conservation of these factors, it is easy to hypothesize that septins direct mitosis through the recruitment and local inactivation of AgSwe1p. Potentially in unbranched areas of cells, some nuclei are delayed in division by active AgSwe1p. This delay then is relieved upon septin ring formation at branchpoints, which leads to a local depletion/inhibition of AgSwe1p and then nuclear division at such sites. The observation that deletion of AgSWE1, AgHSL1, AgMIH1, or AgCDC12 similarly shifts mitoses away from branches to random positions supports the possibility that the balance of AgSwe1p activity controls the spatial pattern of mitosis. Cells lacking AgSwe1p would no longer have the means to limit mitoses in unbranched areas, and cells lacking septins would not be able to promote mitoses at branches, both of which result in random division phenotypes. Furthermore, the localization of AgHsl7p, whose homologue functions in S. cerevisiae as an adaptor that helps bring Swe1p to the septins where it is targeted for degradation, is suggestive that the links in this regulatory system may be preserved in A. gossypii. Thus, in response to some cell intrinsic patterning signals that direct branching in rich nutrient conditions, there is evidence that the septins may direct mitosis spatially through AgSwe1p activity.

There are some problems, however, in overlaying the details of the S. cerevisiae morphogenesis checkpoint system onto the A. gossypii spatial control system that functions in response to cell intrinsic cues. AgHsl7p is not observed at all branch septin sites and rather seems to be more common at septin rings on the main hyphae, although this could just be due to detection problems. Additionally, the AgHsl7p has a key residue differences in its sequence compared with the ScHsl7p at phenylalanine 242. When ScHSL7F242 is mutated to leucine, the same substitution observed in the A. gossypii sequence, ScHsl7p no longer can interact with ScHsl1p, and ScSwe1p is no longer targeted for degradation in budding yeast cells (Cid et al., 2001). This raises the question as to whether AgHsl1p and AgHsl7p form a stable complex that could promote AgSwe1p recruitment. More notably, no AgCdc28Y18 phosphorylation is detectable in cells grown in high nutrient conditions, suggesting that the septins and AgSwe1p could function through other means to regulate the nuclear cycle machinery and influence the spatial balance of mitoses in this condition.

What are alternative modes by which the septins and AgSwe1p may function here? Certainly septins recruit a plethora (>50) of proteins to the mother-bud neck in budding yeast so conceivably the A. gossypii septins are recruiting and either inhibiting/activating additional cell cycle factors locally at branches (Gladfelter et al., 2001). Alternatively, the septins may interact more directly with nuclei at branches by capturing the ends of astral microtubules. Such associations have been observed by tubulin immunofluorescence of A. gossypii cells (A. Gladfelter, unpublished results) and potentially such physical connections, which may generate tension on the astrals, could be a trigger for mitosis. AgSwe1p may itself have other targets than AgCdc28p that are relevant for nuclear cycle progression, and potentially AgSwe1p-dependent phosphorylation impacts their activity and in turn progression. Or AgSwe1p may be able to directly inhibit AgCdc28p independent of Y18 phosphorylation, and there is actually precedent for this alternative inhibition of the CDK in the morphogenesis checkpoint in budding yeast (McMillan et al., 1999b).

Despite a change in the position of mitoses, there was no considerable change in nuclear density in any of the single and double deletion mutants analyzed here under low-density conditions. We would expect that elimination of local control of mitosis might be globally reflected in the average nuclear distance. The fact that this is not the case suggests the existence of redundant mechanisms to regulate nuclear division rate. Loss of local stimulation of mitosis at branching sites could be compensated for by yet unidentified ways to decrease AgSwe1p activity in a more global manner, coupled with efficient migration and distribution of nuclei along microtubules (Alberti-Segui et al., 2001). Notably, such potential mechanisms of compensation can be bypassed or suppressed by overexpressing AgSWE1 from the ScHIS3 promoter, which results in decreased nuclear density even under rich nutrient conditions. One way to explain this is that AgSwe1p activity could globally be restricted on a transcriptional level in addition to being locally controlled by protein inhibition/degradation. This transcriptional control would no longer be functional when AgSwe1p is expressed from an exogenous promoter.

Although it is still unclear how AgSwe1p influences the position of mitotic progression in the presence of abundant nutrients, we present evidence that AgSwe1p does phosphorylate and inhibit AgCdc28p and pause nuclear division under starvation signals, even when hyphal growth continues. We observed that AgCdc28p is phosphorylated on the Y18 residue in high-density growth or conditions that mimic starvation, and this is correlated with an accumulation of nuclei with duplicated SPBs and metaphase spindles and a decreased nuclear density (Figure 8D). Overexpression of AgSwe1p is sufficient to induce some changes to the nuclear division cycle, the delay in G2, even without nutrient deprivation. Interestingly, the AgSwe1p-dependent response to low nutrients was exacerbated in cells lacking septins or cells in which the septins have been deleted in combination with the phosphatase, AgMih1p, which likely removes the Y18 phosphorylation. This suggests that the septins may regulate the AgSwe1p-dependent response to nutrient status. Alternatively or additionally, this AgSwe1p response could also be spatially controlled by clustering of receptors/signaling factors for sensing nutrient status on the cell surface, which then inactivate a pool of AgSwe1p to generate mitoses locally.

Surprisingly, a complete nuclear arrest was never observed even when both the putative AgSwe1p inhibitor AgHsl1p and the phosphatase AgMih1p are deleted. This is in clear contrast to budding yeast, where double deletion of ScMih1p and ScHsl1p leads to a lethal arrest at the G2-M transition (McMillan et al., 1999a). Conceivably, some portion of the CDK pool is shielded from interaction with AgSwe1p so that a uniform arrest does not take place. For example, some nuclei could block nuclear import of AgSwe1p and AgCdc28p in these nuclei would no longer be accessible for AgSwe1p-mediated inhibition, enabling such nuclei to divide independently of the levels of active AgSwe1p. Alternatively, Y18 phosphorylation of AgCdc28p could only partially inhibit activity of the CDK in A. gossypii.

How might AgSwe1p activity be enhanced by nutrient limitation? In budding yeast cells, ScSwe1p participates in the filamentous differentiation response to nutrient limitation through inhibiting CDK/Clb2 complexes, which extends G2 and promotes hyperpolarized bud growth (Edgington et al., 1999; La Valle and Wittenberg, 2001). It is unclear, however, how the status of environmental conditions is transmitted to ScSwe1p and its regulators that are partially responsible for sustaining this differentiation program. Both the STE MAPK and the RAS/cAMP pathways have known "anchors" at the plasma membrane sensor level and the nuclear transcriptional response level; however, ScSwe1p can act independently of these pathways in filamentation (Ahn et al., 1999). How Swe1p senses nutrient status in budding yeast is unknown. In fission yeast, nutrient limitation signals through a MAP kinase called Spc1 and triggers a Wee1-mediated G2 delay potentially through the regulation of nim1 (Shiozaki and Russell, 1995; Belenguer et al., 1997). A. gossypii is constitutively filamentous, and unlike in S. cerevisiae strains, filamentous morphology changes are not observed under nutrient deprivation. Nevertheless, AgSwe1p activity makes dramatic contributions to the nuclear division cycle kinetics in starving A. gossypii cells. It was remarkable that inhibition of AgTor1/2p by rapamycin could activate AgSwe1p similarly to nutrient deprivation, suggesting that AgSwe1p activity in response to starvation may be regulated downstream of this master regulator of cell growth. This Tor1/2p path and/or a MAPK/RAS/cAMP path could communicate nutrient status and influence AgSwe1p in a variety of ways either by influencing its stability, its localization, its transcription, or its intrinsic kinase activity or affinity for AgCdc28p. Future work will be aimed at understanding how this conserved kinase senses and responds to nutrient fluctuations.

Filamentous fungi inhabit heterogenous environments in the natural world. There is spatial and temporal irregularity in many factors such as water, temperature, minerals, and pH, in addition to carbon and nitrogen sources. A. gossypii is found outside the laboratory as a tropical plant pathogen where it is transmitted by sucking insects and can infect cotton and citrus fruits leading to dry rot (Ashby, 1926; Batra, 1973). Given the large potential size of a filamentous fungal mycelium in which many interconnected hyphae share a common cytoplasm, a single cell may simultaneously experience "feast and famine" conditions. In some fungi, branching is induced by exposure to pockets of nutrients, allowing the cell to forage and exploit that presumably limited source (Ritz, 1995). A cell would ideally target these resources locally, producing more nuclei to support the exploring hyphae, while the rest of the nuclei and cell remain quiescent. In a syncytium, however, it is unclear how a cell may delineate the active and inactive regions. Our current laboratory culturing conditions of A. gossypii make it difficult to simulate the nutrient heterogeneity such cells would encounter in the natural world. However, we have attempted to integrate the data we have generated on AgSwe1p spatially directing mitosis in the presence of abundant nutrients and AgSwe1p inhibiting mitosis during starvation into a model. We envision testing this model once more heterogeneous environmental conditions can be replicated in the laboratory. We speculate that septins and AgSwe1p may function to locally generate nuclei in a region of the cell experiencing a limited surplus of nutrients. In this model, sudden increases in the pools of nutrients would lead to local assembly of a septin ring, branch initiation, and perhaps activation of the TOR pathway, which may then promote mitoses in this limited area through the inactivation of AgSwe1p locally (Figure 8E). Future studies will examine this novel role of AgSwe1p and septins for influencing the position of mitoses in multinucleated cells.

	ACKNOWLEDGMENTS

 
We thank Dann