Candida albicans hyphae have a Spitzenkörper that is distinct from the polarisome found in yeast and pseudohyphae

Throughout the fungal kingdom, there exists a broad range of cellular growth forms. Filamentous fungi form long, tube-like hyphae, which show extreme polarised growth from the tip. In contrast, cells of unicellular budding yeast are ovoid. In pseudohyphae, cellular compartments are more elongated and cells fail to separate after mitosis, forming branching chains of long cells. However, pseudohyphal cellular compartments are distinguished from those of hyphae by virtue of constrictions at the septal junctions, compared to the uninterrupted, parallel-sided walls of true hyphae (Sudbery et al., 2004). Although the mechanisms that drive polarised growth in both filamentous fungi and budding yeast have been studied extensively, it is not clear whether the mechanisms elaborated in yeast are applicable to filamentous fungi and vice versa. The human fungal pathogen C. albicans can grow in yeast, pseudohyphal and hyphal growth forms: a property that is thought to be important for its virulence. It thus provides the opportunity to compare, in the same organism, the mechanisms of polarised growth operating in these different growth forms.

S. cerevisiae has provided a model for polarised growth in budding yeast and pseudohyphal cells (Nelson, 2003; Pruyne and Bretscher, 2000; Pruyne, 2002; Rua et al., 2001). In yeast cells, growth is confined to the daughter buds, which initially grow in a polarised fashion and then, in G2, switch to isotropic growth (Kron and Gow, 1995). In pseudohyphal cells, polarised growth persists, resulting in longer buds (Kron et al., 1994). Polarised growth depends on the actin cytoskeleton, which consists of cortical actin patches and actin cables (Pruyne and Bretscher, 2000). Actin cables are oriented toward the site of polarised growth and are thought to form tracks along which the class V myosin, Myo2, and its regulatory light chain, Mlc1, transport secretory vesicles that contain the raw materials and enzymes for the synthesis of new cell walls and cell membranes in the growing bud (Johnston et al., 1991; Karpova et al., 2000; Schott et al., 1999). Cortical actin patches mediate endocytosis, which is also associated with polarised growth. Polarisation of the actin cytoskeleton is ultimately controlled by the Cdc42 GTPase, which localises to the incipient bud site and the bud tip.

Nucleation of actin cables is mediated by the polarisome, a protein complex that forms a cap covering the growing bud tip (Lew and Reed, 1995; Sheu et al., 1998; Lew and Reed, 1995). Components of the polarisome include Bud6, Spa2 and the formin Bni1 (Evangelista, 1997; Sheu et al., 1998). After the switch to isotropic growth, the polarisome disperses, later relocalising to the bud neck, to direct the secretory vesicles to the site of cytokinesis, when a contractile actomyosin ring guides the formation of the primary septum (Lippincott and Li, 1998). Like the polarisome proteins, Mlc1 localises to the bud tip where it interacts with Myo2 (Boyne et al., 2000). Later, Mlc1 interacts with the class II myosin, Myo1, in the cytokinetic ring (Boyne et al., 2000; Shannon and Li, 2000). Thus, the behaviour of S. cerevisiae Mlc1 is similar to that of the polarisome proteins, in that it localises to the bud tip during polarised growth and to the contractile ring during cytokinesis.

Polarised growth in filamentous fungi is much more extreme than in budding yeast: on an open Petri dish, Neurospora crassa hyphae extend at a rate of 38 μm/minute (Read and Hickey, 2001). A special structure, called the Spitzenkörper (apical body), which is located at or just behind the hyphal tip, is responsible for this dramatic polarisation of growth (Girbardt, 1957; Harris et al., 2005). The Spitzenkörper was originally recognised as a dark region in phase-contrast microscopy at the tip of actively growing hyphae (Girbardt, 1957). Subsequently, freeze-substitution electron microscopy revealed that it is rich in secretory vesicles (Grove and Bracker, 1970; Howard, 1981). More recently, it was shown that the amphiphilic styryl dye, FM4-64, labels the Spitzenkörper in numerous fungi (Fischer-Parton et al., 2000). This dye is incorporated into the plasma membrane, internalised by endocytosis and distributed to internal membranes. It is widely used as a marker of endocytosis and to visualise vacuolar membranes. However, when added to the growth medium it accumulates most rapidly at the Spitzenkörper, presumably through vesicle trafficking (Read and Hickey, 2001), so that after short exposures to FM4-64, the Spitzenkörper is preferentially labelled.

The Spitzenkörper is thought to drive hyphal growth because changes in the direction of hyphal growth are anticipated by changes in the position of the Spitzenkörper (Reynaga Pena et al., 1997; Lopez-Franco, 1996). The vesicle supply centre (VSC) model of polarised growth in filamentous fungi posits that the Spitzenkörper is the repository for secretory vesicles that are transported along hyphae towards the tip (Bartnicki-Garcia et al., 1989). Vesicles radiate from the Spitzenkörper and travel to the cell surface, where they fuse with the plasma membrane and release their cargo. Because of the proximity of the Spitzenkörper to the growing tip, a greater concentration of vesicles per unit area arrives at the tip than in more distant parts of the hypha. Computer modelling shows that a Spitzenkörper generates a concentration gradient of vesicles in the form of a hyphoid curve at the cell surface, and this curve closely predicts the actual shape of the hyphal tip in the 34 different fungal species examined (Bartnicki-Garcia et al., 1995). Although the behaviour of the Spitzenkörper has been well documented, less is known about its molecular composition and the mechanisms by which it drives polarised growth in filamentous fungi. In Aspergillus nidulans, the formin SepA localises to the hyphal tip and to sites of septation. Interestingly, SepA localises to a crescent at the tip surface and to a spot just behind the tip (Sharpless and Harris, 2002). As the SepA spot colocalises with FM4-64 (Harris et al., 2005), it is likely to be a component of the Spitzenkörper.

Studies of polarised growth in yeast and filamentous fungi have largely proceeded independently and the relationship between these structures has not been widely addressed, although recently it was suggested that the polarisome complex is a sub-component of the Spitzenkörper (Harris et al., 2005). C. albicans provides the opportunity to use the same organism to compare the molecular mechanisms that drive polarised growth in yeast, pseudohyphae and hyphae. We demonstrate here that C. albicans hyphae contain a Spitzenkörper-like structure by localisation of Mlc1-YFP and FM4-64 to a discrete spot at, or just behind, the growing tip. In contrast, we find that polarisome components and YFP-CDC42 localise predominantly to a surface crescent or cap. Furthermore, we identify genetic and cytoskeletal requirements for each structure as well as organisational differences during the cell cycle that distinguish the Spitzenkörper from the polarisome. Taken together, our results demonstrate that these structures are indeed distinct and that different molecular mechanisms drive polarised growth in hyphae as compared to yeast and pseudohyphae. Interestingly, we observed both Spitzenkörper and polarisome structures in hyphal tips, suggesting that both of these mechanisms may be used, possibly at the same time, in C. albicans hyphae.

Heterozygous fusions to green fluorescent protein (GFP), yellow fluorescent protein (YFP) and cyan fluorescent protein (CFP), and generation of promoter fusions were constructed as described previously (Gerami-Nejad et al., 2001; Gerami-Nejad et al., 2004). Candida albicans disruption strains were also constructed as described previously (Wilson et al., 1999). Full details of the primers used are presented in supplementary material Table S1. The genotypes of the Candida albicans strains used are listed in Table 1.

YEPD medium consists of 2% Difco peptone, 1% Difco yeast extract (both from BD Diagnostics, Sparks, MD), 2% glucose, and 20 μg ml^–1 uridine. SDC medium consists of 0.67% yeast nitrogen base and 2% filter sterilised glucose added after autoclaving. When used, serum was added after autoclaving.

For all but the time-lapse videos experiments, yeast, pseudohyphae and hyphae were cultured as follows: cells were grown to saturation overnight in YEPD at 30°C to produce a culture of unbudded yeast cells. For yeast cells, cultures were reinoculated at 10⁶ cells ml^–1 into YEPD, pH 6.0, and incubated at 30°C. For pseudohyphal cells, cultures were reinoculated at 10⁶ cells ml^–1 into YEPD, pH 6.0, and incubated at 36°C; in these conditions 10-20% of cells develop as hyphae (Sudbery, 2001). For hyphae, unbudded cells were reinoculated into YEPD, pH 6.0, plus 20% serum and cultured at 37°C. For microscopy, cells were washed once in distilled water before examination. Most images were of unfixed cells. However, for time-course experiments, the sample was removed from the culture and cells were fixed immediately in 2% formaldehyde for 10 minutes. The GFP-YFP signal was preserved in this procedure.

For time-lapse videos of live cells, 15 μl of the cell suspension was placed on a glass slide and the cells were allowed to settle for 10 minutes before the liquid was carefully removed by aspiration. Two microliters of SDC containing 2% molten SeaPlaque ultra pure low melting point agarose (Cambrex, Wokingham, UK) were added to cells on the slide. For induction of hyphae, the SDC/low melting point agarose mixture was supplemented with 20% serum. A cover slip was immediately placed over the cells and sealed with a 1:1:1 mixture of Vaseline^®, lanolin, and paraffin wax.

Yeast, pseudohyphae and hyphae were incubated at the appropriate temperature in an environmental chamber fitted to the microscope (Solent Scientific, Portsmouth, UK). Under these conditions, cells grew for several cell cycles, although the mean cell-cycle time was extended to approximately 140 minutes compared with approximately 100 minutes for pseudohyphal cells growing in liquid YEPD.

Expression of YFP-CDC42 from the PCK1 promoter was carried out as follows. Cells from a freshly streaked plate were grown in YEPD-liquid culture overnight. They were reinoculated into YEPS (YEP plus 2% succinate) and grown to stationary phase at 30°C. These stationary phase cells were reinoculated to a density of 10⁶ cells ml^–1 in YEPS plus 5% serum and cultured at 37°C for 90 minutes.

Cells were stained with Calcofluor and DAPI as described previously (Sudbery, 2001). For filipin staining, cells were grown in YEPD under the desired inducing conditions, a 1 ml sample of cells was taken and 1 μl Filipin (8 μg μl^–1 stock, Sigma) was added and incubated at room temperature for 20 minutes. The sample was then centrifuged briefly and resuspended in 1 ×PBS. For FM4-64 staining, 5 μl of stock solution (1.64 mM in DMSO) was added directly to 20 ml of culture and the culture was then incubated for a further 10 minutes. Cells were harvested by brief centrifugation and washed with 1 ×PBS. Microscopic imaging was performed immediately, as FM4-64 staining of the Spitzenkörper dissipates quickly when hyphae are not actively growing.

Fluorescence microscopy was performed with a Delta Vision Spectris 4.0 microscope running Softworx^™ version 3.2.2 (Applied Precision Instruments, Seattle). The standard DAPI filter set was used to visualise DAPI, Calcofluor and Filipin fluorescence, the standard FITC filter set was used for GFP and YFP, the standard TRITC filter set was used for FM4-64, and a Chroma YFP/CFP filter set (Chroma Technology Corp., Rockingham, part number 86002) was used for the simultaneous visualisation of CFP and YFP. Preliminary experiments confirmed that there was no bleed-through of either fluorophore into the inappropriate channel. Z-stack images were collected with step sizes of 0.2 μm and deconvolved using Softworx™. Where quantitation of fluorescence was carried out, the exposure time was kept constant at one second. Except where stated otherwise, images are projections of the deconvolved Z-stacks. Where enlargements are shown, the images were depixellated using the interpolated zoom facility. Images were exported as tif files and image size was adjusted to 300 dpi and cropped using Adobe Photoshop 5.5. Multipart figures were assembled in Microsoft PowerPoint 9.0 and saved as encapsulated postscript (.eps) files. Where stated, Z-stacks of images generated with the Delta Vision microscope were further processed by Volocity software (Improvision, Warwick UK). All quantitative measurements and model building were carried out using deconvolved image files. Quantitation of fluorescence signals was carried out using the data inspector in the Softworx suite.

To generate models, we used the modelling module in Softworx™ or the Volocity software suite (Improvision, Warwick, UK). The Softworx™ module operates by identifying polygons in each layer where the pixels record fluorescence from each fluorophore that is above a user-prescribed threshold. It then synthesises a 3D model from the polygons identified in each layer. We set the threshold so that the polygons corresponded as closely as possible to that visible in the image presented alongside each model. A limitation of the Softworx module is that the models show the boundaries of all fluorescence above the threshold, so there is no way of representing variations in the intensity. However, an advantage is that where one fluorophore is enclosed by a second fluorophore, as was the case with Spa2 and Mlc1 or FM4-64, the outer fluorophore can be represented by a wire frame to allow the boundary of the inner fluorophore to be visualised. Volocity analyses the Z-stack to identify volumes of equal intensity, called voxels, from which it generates a 3D model. Like Softworx, the models represent intensity above a user-prescribed threshold.

Mlc1-YFP localisation was examined in hyphae, pseudohyphae and yeast using a Delta Vision Spectris 4.0 3D restoration microscope, which applies iterative deconvolution to a Z-stack of epifluorescence images (Swedlow et al., 2002). In hyphal germ tubes, Mlc1 appeared as a bright spot of fluorescence at the tip of 100% of germ tubes. This spot colocalised with FM4-64, which has been shown to visualise the Spitzenkörper at the hyphal tip of filamentous fungi (Fischer-Parton et al., 2000) (Fig. 1A). Close examination of individual germ tube tips revealed that, in some cases, the Mlc1-YFP spot localised at the tip (Fig. 1B), whereas in others it was located just behind the tip (Fig. 1B,F). In addition to the spot, 60% of the germ tubes exhibited a crescent of Mlc1-YFP at the tip (Fig. 1B,C). Examination of 200 nm optical sections through these tips revealed that the crescent is in a different focal plane from the spot (Fig. 1C). We found that the crescent is always a surface structure, whereas the spot is sometimes internal to the germ tube tip. The proportion of cells having a crescent as well as a spot showed little variation through the cell cycle (data not shown).

Deconvolved images in a Z-stack contain information concerning the distribution of fluorescence in the Z-plane as well as the X-Y planes. Two-dimensional (2D) projections of the Z-stack can summarise this information, but fail to recapitulate the full three-dimensional (3D) relationship of the fluorescence in different planes. Computer models utilise the information in the Z-planes to represent a 3D structure (see Materials and Methods for more detail). To determine whether the spot of Mlc1-YFP was a 3D object projected onto a 2D plane, we generated a 3D model of the Mlc1 structure using the Volocity program. Fig. 1D shows such a model in a cell co-stained with filipin to define the position of the plasma membrane at the tip. Mlc1-YFP localises to a 3D ball enclosed by the plasma membrane.

To confirm that the localisation of Mlc1-YFP was unaffected by the YFP moiety, we investigated the localisation of Mlc1 in germ tubes of a wild-type strain by immunocytofluorescence using an antibody to S. cerevisiae Mlc1 (kindly provided by J. Boyne and C. Price, Lancaster, University, Lancaster, UK). As with the Mlc1-YFP strains, a discrete spot at the hyphal tip is clearly visible by immunofluorescence, despite the higher background observed with this method (Fig. 1E). Therefore, the localisation of Mlc1-YFP to a spot at the hyphal tip is not an artefact of the YFP fusion. Furthermore, this localisation is evident in the hyphae of several independently derived MLC1-YFP strains and in an MLC1-CFP strain, so it is not peculiar to the particular strain used.

The colocalisation of Mlc1-YFP and FM4-64 to a spot at the hyphal tip suggests that a Spitzenkörper is present at the tip of C. albicans hyphae. If this is the case, then hyphal dimensions should conform to a hyphoid curve predicted by the VSC model and should be predictable from the position of the Spitzenkörper (Bartnicki-Garcia et al., 1989). These predictions were tested using the Fungus Simulator program version 4 (Bartnicki-Garcia et al., 1989). The shape of hyphae predicted by this program, according to the distance of the Spitzenkörper core from the hyphal tip, was in good agreement with the observed dimensions of hyphae. An example is shown in Fig. 1F, where we visualised the Spitzenkörper with Mlc1-YFP and the cell wall with Concanavalin A conjugated to Texas Red (Fig. 1F). Similar agreement was seen in ten additional hyphae analysed in this manner. The mean width measured in these hyphae 2 μm from the tip was 2.13±0.09 μm (mean±s.d.) compared to an average predicted width at this point of 2.26±0.19 μm. Interestingly, in the experiment shown in Fig. 1B, the Spitzenkörper was closer to the tip and the hyphae were narrower (1.59±0.27 μm) (Fig. 1B), which is consistent with the prediction from the VSC model (1.51±0.27 μm). Thus, the distance of the Spitzenkörper from the tip correlates with hyphal width, in good agreement with the VSC model. The difference in hyphal widths observed in the two experiments is probably due to the relative growth rates: the culture in Fig. 1B had been incubated for 70 minutes after hyphal induction, whereas the Concanavalin A-treated culture had been cultured for 120 minutes. Taken together, these results are consistent with the presence of a Spitzenkörper at the tip of C. albicans hyphae and thus we will refer to the spot of Mlc1 and FM4-64 fluorescence as the Spitzenkörper.

We next investigated the pattern of localisation in hyphal tips of Bni1-YFP, Bud6-GFP and Spa2-YFP, orthologs of polarisome components in S. cerevisiae. Like Mlc1, Bni1 localised to a bright spot just behind the hyphal tip (Fig. 2A). Apical localisation of Spa2 has been reported recently (Zheng et al., 2003), but in contrast to Mlc1 and Bni1, we observed that both Spa2 and Bud6 localised predominantly to a crescent or cap at the tip of all hyphae (Fig. 2B,C). However, closer examination revealed that areas of more intense staining were often visible along with the cap (Fig. 2C, arrow; Fig. 2D). The extent of Spa2-YFP colocalisation with FM4-64 and Mlc1-CFP was determined in colocalisation experiments (Fig. 2D,E). Spa2-YFP formed a crescent or a cap located slightly closer to the tip than the spots of Mlc1-CFP and FM4-64. The 3D model of Spa2 and FM4-64 localisation (Fig. 2E) shows that the Spa2 cap covers the ball of FM4-64-stained material, consistent with the idea that both a polarisome and a Spitzenkörper are present at the hyphal tip. Thus, Spa2 and Bud6 localise predominantly to the polarisome, whereas Bni1, Mlc1 and FM4-64 localise predominantly to the Spitzenkörper. The more intensely fluorescing patches of Spa2 and Bud6 may indicate that some of these proteins are present in the Spitzenkörper. However, close examination reveals that the overlap between Spa2 and FM4-64 or Mlc1 is limited (Fig. 2D,E).

Cdc42 plays many roles in coordinating polarised growth in S. cerevisiae, and in C. albicans it has been localised to the hyphal tip (Hazan and Liu, 2002). We visualised Cdc42 with an N-terminal fusion to YFP (YFP-Cdc42). As found by others, no signal was detected when YFP-Cdc42 was expressed from its own promoter (Hazan and Liu, 2002). We therefore overexpressed YFP-Cdc42 by placing it under the control of the PCK1 promoter, which is induced by growth on succinate (Leuker et al., 1997), using a cassette designed for the PCR-mediated construction of such alleles (Gerami-Nejad et al., 2004). Expression of the YFP-Int1, the C. albicans homolog of S. cerevisiae Bud4, was elevated eightfold when expressed from the induced PCK1 promoter compared to the native promoter (Gerami-Nejad et al., 2004). Thus, use of the PCK1 promoter results in a moderate degree of overexpression. Strikingly, YFP-Cdc42 localised to two structures: a bright crescent at the surface of the hyphal tip and a fainter spot just behind the crescent (Fig. 3A,D). In these experiments, FM4-64 stained a discrete spot at the tip (Fig. 3B,E) that colocalised with the subapical spot of Cdc42 (Fig. 3C,F). We conclude that, like Spa2 and Bud6, but in contrast to Mlc1 and Bni1, the majority of Cdc42 localises to the polarisome, whereas a smaller amount is located within the Spitzenkörper.

Deletion of SPA2 results in germ tubes that are broader and less polarised than the wild type (Zheng et al., 2003) (Fig. 4C,E). To investigate whether SPA2 affects the integrity of the Spitzenkörper, we constructed an Mlc1-YFP fusion in a spa2Δ/spa2Δ strain. Loss of SPA2 resulted in polarisome-like localisation of Mlc1-YFP rather than a Spitzenkörper-like localisation in 70 out of 74 hyphal tips examined (Fig. 4A). This mislocalisation of Mlc1-YFP and change in morphology is consistent with the idea that Spa2 is required for Spitzenkörper formation and, consequently, for hyphal growth.

To determine whether the integrity of the Spitzenkörper was specifically dependent upon SPA2 or if it required other components of the polarisome as well, we introduced Mlc1-YFP into a bud6Δ/bud6Δ strain. The absence of BUD6 resulted in a phenotype similar to that of spa2Δ/spa2Δ strains. When grown under hyphal-inducing conditions, the daughter cells were wider and less polarised (Fig. 4D). Consistent with this change in morphology, Mlc1-YFP no longer localised to a Spitzenkörper, but instead was distributed in a surface crescent in all 20 hyphal tips examined (Fig. 4B). Thus, the formation of the C. albicans Spitzenkörper requires at least two polarisome components, Spa2 and Bud6.

The formation of the Spitzenkörper in Fusarium acuminatum is disrupted by methyl benzimidazole-2-ylcarbamate (MBC), an inhibitor of tubulin polymerisation (Bartnicki-Garcia et al., 1995; Howard, 1981). There are conflicting reports of whether hyphal growth in C. albicans is sensitive to MBC (Akashi et al., 1994; Yokoyama et al., 1990). We found that 0.1 mg/ml MBC blocked hyphal extension within 10 minutes (data not shown). To determine whether microtubules would be disrupted by MBC, we treated a Tub1-YFP-expressing strain with MBC. In untreated cells, long microtubules extend along the long axis of germ tubes, consistent with previous reports (Fig. 5A) (Barton and Gull, 1988; Hazan et al., 2002). Microtubules were disrupted within 10 minutes of exposure to 0.1 mg/ml MBC. In MBC-treated hyphae, the organisation of FM4-64 (Fig. 5C) and Mlc1 (Fig. 5D) was disrupted in 61 out of 62 hyphal tips examined. Thus, the integrity of the C. albicans Spitzenkörper requires microtubules. As we found for spa2Δ/spa2Δ and bud6Δ/bud6Δ strains, the residual organisation of Mlc1-YFP in MBC-treated hyphae resembled that of a polarisome, although the structure was more discontinuous than it was in wild-type pseudohyphae or yeast (see next section). This suggests that localisation of Mlc1 to a polarisome-like structure requires neither microtubules nor the polarisome components Spa2 and Bud6.

Polarised growth in S. cerevisiae yeast cells is not dependent on microtubules, but does require actin cables. In C. albicans, actin cables are specifically disrupted with Cytochalasin A (CA), which has no effect on actin cortical patches and microtubules (Akashi et al., 1994). When growing hyphae are treated with CA, the hyphal tips swell (Akashi et al., 1994), as they do in other filamentous fungi treated with inhibitors of the actin cytoskeleton. We used CA to test the dependency of the Spitzenkörper on actin cables in cells expressing Mlc1-YFP. We also wanted to investigate whether the tip swelling caused by CA is due to new cell wall material added through isotropic growth, or whether the swelling arises from expansion of a pre-existing wall weakened by the CA treatment. To do this, we exploited the cell wall-binding lectin Concanavalin A (ConA) conjugated to different fluorophores to distinguish between old and new cell wall material. ConA-Alexa 488 (green fluorescence) was used prior to CA treatment and ConA-Texas Red (red fluorescence) was used to stain cell walls after the CA treatment (Fig. 6). Because the fluorescence of YFP is compatible with the filter set used to visualise Alexa 488, the green channel shows Mlc1-YFP fluorescence along with cell wall material present prior to CA addition (Fig. 6A,D). As reported previously, CA treatment resulted in uniform swelling of all hyphal tips (Fig. 6B,C). The swollen tip exhibited no green fluorescence, but did stain with ConA-Texas Red (Fig. 6A,B). Thus the swollen tip arises from the isotropic deposition of new cell wall material and is not due to expansion of a weakened cell wall. In CA-treated cells, the Spitzenkörper completely disappeared (Fig. 6A,C), whereas in untreated controls the Spitzenkörper remained visible (Fig. 6F inset). Thus, when actin cables are disrupted with CA, the Spitzenkörper disappears and the growth mode switches from polarised to isotropic. Interestingly, Mlc1-YFP localisation to the cytokinetic ring is not disrupted after CA treatment (arrow, Fig. 6C), thus this inhibitor is specific to actin cables that deliver cargo to the hyphal tip.

We also investigated the localisation of Mlc1-YFP in pseudohyphae and yeast. In some newly evaginated pseudohyphal buds, Mlc1-YFP localised to a spot similar to that seen in hyphae (Fig. 7A, cell 1). However, this quickly changed to a crescent or patch (Fig. 7A, cells 2-4; Fig. 7B, cell 1) which, when analysed by 3D modelling, was located around the surface of the tip (Fig. 7D). In elongated buds, Mlc1-YFP fluorescence at the tip was low or undetectable (Fig. 7B, cell 2) and other cells with elongated buds contained a ring of Mlc1-YFP at the bud neck (Fig. 7C). Quantification of the different patterns of Mlc1 localisation, categorised according to bud length confirms that the Mlc1-YFP crescent or spot present at the tip of young buds disappears as cells enter mitosis to be replaced by a ring at the bud neck. Spa2 and FM4-64 also localised to a crescent at the tips of pseudohyphal buds (Fig. 7E,F). Similarly, 3D modelling showed that both Spa2 and FM4-64 were located in a surface crescent (data not shown). As we found for pseudohyphal buds, newly evaginated yeast buds had a spot of MIc1 fluorescence (Fig. 8A) that quickly changed to a crescent as the bud enlarged. Again, like pseudohyphae, Mlc1 was reduced at the tip of large yeast buds compared to small buds and eventually a ring of Mlc1 formed at the bud neck (Fig. 8A). This pattern of localisation was observed in the localisation of Spa2 in yeast (Fig. 8B). Thus, Mlc1, Spa2 and FM4-64 localise to a surface cap or crescent and are not found in a spot at the tips of pseudohyphal or yeast cells. We conclude that only a polarisome is present in pseudohyphae and yeast.

The reduction in signal intensity for polarity determinants at the tips of large pseudohyphal and yeast buds suggested that the polarisome is regulated in a cell cycle-dependent fashion. To investigate this further, we recorded time-lapse videos of pseudohyphal cells in a strain coexpressing Nop1-CFP and Mlc1-YFP. Nop1 is an abundant nucleolar protein, thus the fluorescent fusion protein allows the nucleus to be visualised in a living cell (Fig. 9A and supplementary material, Movie 1). During or just after mitosis, an Mlc1 ring appears at the bud neck and then contracts (Fig. 9A) as it does during cytokinesis in yeast cells (Lippincott and Li, 1998). In 13/13 videos, the polarisome disappeared late in the cell cycle and did not coexist with the cytokinetic ring. Quantification of random populations of cells showed that Mlc1 is not present at the bud tip in 80% of cells displaying a ring of Mlc1 at the septation site. In the majority of the remaining cells, the residual fluorescence of Mlc1 at the bud tip is faint. Therefore, we conclude that, in pseudohyphae, the polarisome disappears at or before mitosis and rarely coexists with the cytokinetic ring.

In S. cerevisiae, cells switch from polarised to isotropic growth at a specific point in the cell cycle after phosphorylation of Cdc2/Cdc28 (the cyclin-dependent kinase) by the Swe1 kinase (Lew and Reed, 1995). Prior to this switch, cells exhibit polarised growth with length increasing faster than cell width and, as a result, the axial ratio (length/width) of the growing bud is greater than one and continues to increase as the bud length increases. After the switch to isotropic growth, length and width increase at a similar rate and thus the axial ratio no longer increases relative to bud length. Plotting bud length versus the axial ratio reveals the average cell length at which polarised growth (curve has a positive slope) switches to isotropic growth (slope of the curve becomes shallower). In C. albicans yeast cells, the switch to isotropic growth occurs at a cell length of ∼1.75 μm and an axial ratio of 1.25 (Fig. 9B). In contrast, pseudohyphal cells switch to isotropic growth at a cell length of 7.5 μm and an axial ratio of 4.5 (Fig. 9C). In general, cells with Mlc1 staining at the bud tip fall into the polarised growth part of the curve and cells without Mlc1 staining fall into the isotropic part of the curve. Thus, the presence of the polarisome is correlated with polarised growth and this polarisome-associated growth persists for a longer period in pseudohyphal compared to yeast cells.

In contrast to both yeast and pseudohyphae, hyphal cells do not exhibit a cell cycle-dependent localisation of Mlc1. Rather, Mlc1-YFP localises to the Spitzenkörper in all hyphal cells, even when it localises to the cytokinetic ring (Fig. 9D,E). Thus, the Spitzenkörper is continuously present at the hyphal tip during all stages of the cell cycle, including septation.

To compare the amount of Mlc1 in the Spitzenkörper and the polarisome, we quantified the mean of the peak values of Mlc1 intensity at the tips of pseudohyphal buds and hyphal germ tubes and plotted these values against germ tube or pseudohyphal bud length (Fig. 10A,B). In hyphae, the mean peak intensity increased from ∼1000 units immediately after evagination to ∼3000 units when the germ tube length had reached 35-50 μm (Fig. 10A). In contrast, the mean peak intensity in pseudohyphae was ∼600 units after evagination, and declined thereafter, consistent with its disappearance from the bud tip later in the cell cycle (Fig. 10B). Furthermore, the 3D graph of Mlc1-YFP intensity in hyphae (Fig. 10A) reveals a ridge, corresponding to the crescent, on either side of the main peak. The intensity of this ridge of fluorescence is similar to the intensity of the polarisome in pseudohyphae, consistent with the hypothesis that the crescent at hyphal tips represents a polarisome and the ball at hyphal tips represents a Spitzenkörper. Peak fluorescence intensity in yeast was similar to that of pseudohyphae (Fig. 10C).

Like Mlc1, Spa2 is also distributed as a surface ridge in hyphae (Fig. 10D), in contrast to the cone observed in the 3D graph of Mlc1 fluorescence (Fig. 10A) and consistent with the model shown in Fig. 3D. Interestingly, when we compared the relative fluorescence intensity between the growth forms, we found that the fluorescence of Mlc1-YFP increased approximately fourfold in hyphae compared to yeast and pseudohyphae, whereas the fluorescence intensity of Spa2-YFP did not show much difference between the different morphologies (Fig. 10C), supporting the idea that Spa2 is predominantly in the polarisome, whereas Mlc1 is predominantly in the Spitzenkörper.

The primary aim of this work was to test the hypothesis that there are different properties of polarised growth that account for the distinctive shapes of hyphae relative to yeast and pseudohyphae. To this end, we found molecular, genetic and temporal evidence that support this hypothesis. First, we found two informative patterns of protein localisation at hyphal tips. The localisation of Mlc1, which is required to transport vesicles to the growing tip along actin cables, and the formin, Bni1, which nucleates actin cables, differs from that of Spa2 and Bud6, which are components of the S. cerevisiae polarisome, and of Cdc42, which orchestrates many of the processes required for polarised growth. Mlc1 and the Spitzenkörper marker FM4-64 colocalised predominantly to a spot (Fig. 1A). The spot of Mlc1-YFP fluorescence was revealed to be a ball when rendered into a 3D model (Fig. 1D). In some cells, the spot was located a short distance from the tip and the plasma membrane (Fig. 1B,F). In contrast, Spa2, Bud6 and Cdc42 localised mostly to a crescent that showed limited colocalisation with FM4-64 (Fig. 2E, Fig. 3F). When Spa2-YFP and FM4-64 were rendered into three dimensions, Spa2 formed a surface cap at the tip and was more apical than the FM4-64 stained material (Fig. 2E). Thus, we conclude that there are two structures present at the hyphal tip: a polarisome and a Spitzenkörper.

Although Mlc1 localised mainly to the Spitzenkörper, it was also detected in the polarisome of many hyphae (Fig. 1B,C and Fig. 10A). However, whereas the Mlc1 spot was sometimes observed without the crescent (Fig. 1B,D), we never observed the crescent alone: all hyphae displayed a spot of Mlc1 at the tip (Fig. 1A). The simultaneous localisation Mlc1-YFP to a spot and a crescent in the hyphal tip is strikingly similar to the localisation of the A. nidulans formin, SepA, which localises to two clearly separate structures: a spot a small distance away from the tip and a crescent at the tip (Sharpless and Harris, 2002). A small amount of Cdc42 was present in the Spitzenkörper of most cells (Fig. 3). This observation is subject to the caveat that we could only visualise Cdc42 when it was overexpressed, thus the localisation of Cdc42 to the Spitzenkörper could be artefactual. Spa2 and Bud6 were also observed to localise to a spot in some cells (Fig. 2C,D). Thus Cdc42, as well as Spa2 and Bud6, may also be components of the Spitzenkörper. The observation that Bni1 and Mlc1 were apparently located predominantly in the Spitzenkörper, whereas Cdc42, Spa2 and Bud6 were found predominantly in the polarisome, may reflect a genuine difference in the proportions of each protein in the Spitzenkörper and polarisome. However, the apparent differences could also be due to technical issues such as different degrees of competition with the wild-type protein in the two structures. However, note that the indirect immunofluorescence with anti-Mlc1 antibodies (Fig. 1E) shows that the localisation of Mlc1-YFP to a spot is not artefactual.

Second, we investigated the genetic requirements for the Spitzenkörper and found that the polarisome components Spa2 and Bud6 are required for Spitzenkörper formation (Fig. 4). In mutants lacking either of these proteins, the Spitzenkörper was disrupted and Mlc1 localised only to the polarisome. Moreover, there was an accompanying change in the morphology of the mutants, such that they resembled pseudohyphae rather than hyphae. This is consistent with the Spitzenkörper driving the hyphal growth of C. albicans. Thus, polarisome components are required for hyphal growth, but are not required for yeast or pseudohyphal growth. Similarly, we found that the Spitzenkörper was disrupted by treatment with the microtubule inhibitor MBC (Fig. 5). This dependence on microtubule function is consistent with the observation that MBC disrupts the Spitzenkörper in the filamentous fungus, Fusarium acuminatum (Howard and Aist, 1981), whereas microtubules are not required for bud growth and secretory vesicle transport in S. cerevisiae (Huffaker et al., 1988).

Third, we tested the prediction that there would be no Spitzenkörper in yeast and pseudohyphae. Indeed, we found that Mlc1, Spa2 and FM4-64 localised to a surface crescent rather than a Spitzenkörper-like structure (Fig. 7A-F and Fig. 8). Therefore, we conclude that the Spitzenkörper is specific to hyphae and is not involved in yeast or pseudohyphal growth.

Finally, we observed two differences between the regulation of the polarisome in yeast and pseudohyphae relative to the Spitzenkörper in hyphae. First, polarisome components disappear from the tip of yeast and pseudohyphal buds at or before mitosis and reappear at the cytokinetic ring after mitosis. Mlc1 at the tip rarely coexisted with the cytokinetic ring (Fig. 9A). In contrast, in hyphae, the Spitzenkörper was continually present at the tip and coexisted with the contractile Mlc1 ring at the site of septation (Fig. 9D,E). Second, the intensity of Mlc1 fluorescence was approximately fourfold greater in the Spitzenkörper as compared to the polarisome. This was true even in newly evaginated yeast and pseudohyphal buds, where the pattern of Mlc1 localisation resembled that of hyphae in some buds (Fig. 9). These results support the existence of two complexes that are subject to different modes of regulation.

Maintenance of the Spitzenkörper requires both actin cables and microtubules: the Spitzenkörper disappeared on treatment with cytochalasin A and was disrupted upon treatment with MBC. However, there were interesting differences between the effects of these inhibitors on growth patterns. Growth completely ceased upon MBC treatment. In contrast, growth continued upon Cytochalisin A treatment, but switched from a polarised to an isotropic mode, resulting in a swelling at the hyphal tip (Fig. 6). These observations are consistent with the idea that microtubules mediate long-distance transport of vesicles and that actin cables mediate short distance distribution of vesicles from the Spitzenkörper to the growing tip. According to this scenario, growth ceases upon MBC treatment because secretory vesicles generated in the body of the hypha cannot be transported to the tip region. In the absence of the continued arrival of vesicles, the Spitzenkörper disperses. The remaining Mlc1-YFP in a crescent at the surface may reflect the continued operation of the polarisome-mediated transport of vesicles along actin cables. Upon disruption of actin cables, vesicles would continue to arrive at the tip region, transported along microtubules. Without a Spitzenkörper to focus these vesicles to the tip, they would disperse randomly in all directions, resulting in isotropic growth.

As the same components are present in both the Spitzenkörper and the polarisome, albeit in different proportions, it may be argued that the structure described here as a Spitzenkörper is a hyperactive polarisome. Three lines of evidence suggest that the difference is not simply semantic. First, in hyphae, the presence of the Spitzenkörper is independent of the cell cycle, whereas in yeast and pseudohyphae it is cell cycle dependent, disappearing before mitosis. Second, the Spitzenkörper is clearly a 3D object, which in some hyphae is clearly separated from the hyphal tip, whereas the polarisome consistently forms a surface cap at the growing tip. Third, loss of Spa2 or Bud6 or disruption of the microtubules with MBC results in loss of the Spitzenkörper, whereas a polarisome-like structure appears to persist and growth assumes the pseudohyphal morphology. In S. cerevisiae, microtubules are not required for polarised growth, and so presumably are also not required for the integrity of the polarisome. Thus, not only is the Spitzenkörper regulated differently from the polarisome, it has different genetic and cytoskeletal requirements to the polarisome.

Taken together, our data show that there are considerable differences in the properties of the Spitzenkörper, present only in hyphae, and the polarisome that is present in hyphae, pseudohyphae and yeast. Importantly, although pseudohyphae may superficially resemble hyphae, the underlying mechanism of polarised growth in pseudohyphae is more similar to that in yeast, supporting the view that hyphae and pseudohyphae in C. albicans are qualitatively different states (Sudbery et al., 2004). It will be of great interest to understand the molecular mechanisms responsible for organising these different modes of polarised growth and to determine how signal transduction pathways regulate the structures that give rise to them and thereby mediate morphologic changes.

This work was supported by the Wellcome Trust (Grant No: 060862/Z/00/Z) and BBSRC grant to P. E.S., the National Institutes of Health (R01 AI/DE 14666) to J.B. and a University of MN Graduate School grant to C.G. Helen Crampin was in receipt of a BBSRC Research Training Studentship, Helen Court is in receipt of an MRC Research Training Studentship and K.F. was supported by NIH Biotechnology training grant GM08347. We thank Yue Wang for providing a spa2Δ/spa2Δ strain, Alan Tilley of Improvision for generating the 3D models using the Volocity Software suite and Jim Boyne and Clive Price for antisera to S. cerevisiae Mlc1.