BEM1 is dispensable for general hyphal polarity establishment and maintenance
Polarized growth is a hallmark of filamentous fungi. The life cycle of N. crassa includes numerous distinguishable polarization events, including spore germination, extension of vegetative hyphae, induction and tropic growth of fusion hyphae, conidia formation and tropic growth of trichogynes as mating structures (Riquelme et al., 2011).
Localization of BEM1 to the growing tips of different types of hyphae suggests that BEM1 is part of a conserved polarity machinery. However, our phenotypic characterization of Δbem1 mutants indicates that BEM1 is of different importance for the various polarization events in N. crassa. During spore germination or hyphal extension, its absence does not result in polarity defects. Similarly, in A. nidulans and E. festucae Δbem1 hyphae possess only minor growth defects, such as slight swelling of hyphal tips (Leeder and Turner, 2008; Takemoto et al., 2011). In A. nidulans, the formin SepA mislocalizes if BemA is absent, suggesting that BemA is required for Spitzenkörper organization (Leeder and Turner, 2008). However, when we localized the Spitzenkörper marker chitin synthase 1 (Sanchez-Leon et al., 2011) in the N. crassa Δbem1 mutant, no significant differences were observed (data not shown). Other proteins are likely to substitute BEM1 functions in general hyphal polarity, probably to a different extent in different species. Studies in yeast suggest that two positive feedback loops contribute in parallel to polarity establishment – one based on bem1p functions the other one on the actin cytoskeleton (Wedlich-Soldner et al., 2004). In the first one, active cdc42p locally recruits bem1p to the plasma membrane. bem1p interacts with cdc24p, which in turn activates more cdc42p. In the second one activated cdc42p leads to the formation of actin cables, resulting in the recruitment of vesicular cdc42p. In filamentous fungi, hyphal growth strongly depends on a highly polarized cytoskeleton. Thus, in general hyphal tip extension, positive cytoskeletal-based feedback loops might therefore be much more important than soluble signalling-based loops including BEM1.
In contrast to general hyphal growth, macroconidia formation and germling/hyphal fusion are severely affected in N. crassa Δbem1 mutants. Interestingly, macroconidiation and germling fusion in N. crassa share many similarities with yeast budding and shmoo formation respectively (Springer and Yanofsky, 1989; Glass and Fleißner, 2006; Read et al., 2009). Both processes of yeast development also depend highly on bem1p function (Chenevert et al., 1992; 1994; Kao et al., 1996; Lyons et al., 1996), suggesting that S. cerevisiae and N. crassa employ conserved polarity machineries in these related processes.
During macroconidiation in N. crassa the lack of BEM1 results in enlarged and misshaped spores, containing numerous nuclei. This phenotype hints to defects in the controlled re-initiation of apical budding after a spore has formed. These defects would then result in prolonged isotropic growth, given rise to the observed giant spores. In the absence of internal landmarks or external signals, S. cerevisiae cells spontaneously polarize. This phenomenon, known as symmetry breaking, led to a model, in which activated cdc42p-GTP, which is initially randomly distributed at the cell periphery, forms fluctuating clusters, which compete for rapidly diffusing bem1p. The winning polarity centre gets strongly reinforced by a positive feedback loop involving a bem1p–PAK–cdc24p complex, which interacts with active membrane-bound cdc42p-GTP. This competition/reinforcement leads eventually to a winner takes all situation resulting in the establishment of only one polarity axis (Atkins et al., 2008; Kozubowski et al., 2008). During macroconidiation of N. crassa lack of BEM1 potentially delays the successful establishment of one polarity centre, resulting in prolonged isotropic growth. In an alternative model, the increase in spore size would be a consequence of the higher number of nuclei and therefore of defective cytokinesis. However, this hypothesis appears less likely, since nuclear division and conidia formation are not clearly linked in N. crassa. Conidia are formed from multinucleated hyphae, which alternate between polar and apolar growth (Springer and Yanofsky, 1989). As a result a pearl necklace-like structure is formed. Then, secondary septa are introduced, resulting in the formation of individual spores with various numbers of nuclei (in wild type typically between one and four).
Formation of normal sized spores requires the PB-1 domain of the BEM1 protein. In contrast, this domain is dispensable for germling fusion. Similarly, expression of the 5′-extended bem1 open reading frame under control of the ccg-1 promoter resulted in strong defects in general hyphal polarity and establishment of a single polarity axis during germination. Strikingly, germling fusion appeared normal in these isolates. Together, these observations suggest that different parts of the BEM1 protein are of different importance for the various growth modes. BEM1 is probably part of different tip localized protein complexes, which have diverse molecular compositions and functions. These data also indicate that no correlation between spore size and fusion competence exist, thereby excluding the possibility that the fusion deficiency in Δbem1 might be a secondary effect of the conidial morphogenetic defects. The observation that expression of the extended construct has a strong dominant effect on hyphal polarity is best explained by a titration effect. The longer protein potentially binds certain interaction partners, but can then not fulfil its normal function, keeping the interacting factors captured. A similar observation was made in A. nidulans, when only a part of the BemA protein was expressed (Leeder and Turner, 2008).
In yeast mating, bem1p functions as a molecular adaptor connecting the MAP kinase module of the pheromone response pathway with the core polarity machinery, resulting in recruitment of the MAP kinase fus3p to shmoo tips (Leeuw et al., 1995; Lyons et al., 1996). In germling fusion in N. crassa, the fus3p homologous MAP kinase MAK-2 gets also recruited to fusion tips, however, in a highly dynamic oscillating fashion, in which recruitment alternates between the two fusion partners (Fleißner et al., 2009b). This highly orchestrated process likely involves positive and negative feedback loops, which partially might depend on BEM1. In S. cerevisiae the absence of bem1p leads to a substantially reduced sensitivity to α-factor, indicating that bem1p facilitates efficient pheromone signalling (Kao et al., 1996). Our finding that in Δbem1 germling fusion is reduced, delayed and takes place only over very short distances, suggests a similar promoting function in N. crassa. Long distance interactions of conidial germlings require the polarity axis to be quickly adjustable to incoming signals. If one assumes additive actions of a cytoskeletal-based and a soluble BEM1-dependent polarity mechanism, the latter one might be much more suited to allow highly dynamic tropic interactions. Thus, in contrast to general polarized hyphal growth, germling fusion might much stronger depend on BEM1 functions. Consistent with this hypothesis, general hyphal growth ceases after disintegration of the microtubular cytoskeleton, while tropic growth involved in germling fusion continues (Roca et al., 2010). Lack of BEM1 has the exact opposite outcome.
When we tried to analyse the subcellular dynamics of SO and MAK-2 in the Δbem1 mutant, the tropic interaction of germlings was frequently and quickly aborted, and SO became mislocalized in both fusion partners. In an earlier study, we found the exact same characteristic pattern of SO mislocalization after specific inhibition of MAK-2 activity by a kinase inhibitor (Fleißner et al., 2009b). Together these observations suggest that in Δbem1 MAK-2 activation is highly instable and that BEM1 is essential for robust signalling via the MAP kinase module. So far, it remains unclear where and how BEM1 might stabilize cell communication since no clear indications for colocalization of the proteins were found. Similarly, testing potential physical interactions of MAK-2 and BEM1 via yeast two-hybrid analysis rendered only negative results (data not shown). In yeast, bem1p functions as a scaffold connecting cdc42p with its activator cdc24p and the cdc42p effector ste20p, a p21-activated kinase (Elion, 2000). In addition to Cdc42, filamentous fungi contain the closely related Rac1. In Ustilago maydis interaction of this GTPase with Cdc24 is also promoted by Bem1 (Frieser et al., 2011). In addition, in a screen for hyphal fusion-deficient mutants in N. crassa, a rac-1 mutant was isolated (Fu et al., 2011). In mature hyphae, CDC42 and RAC1 both localize to hyphal tips. The localization pattern of CDC42 is highly reminiscent of BEM1. In germlings, both GTPases localize overall to the plasma membrane but are enriched at hyphal tips (Araujo-Palomares et al., 2011). Together, these observations support a model in which RAC1 and/or CDC42 are involved in activation of MAK-2, and this activation is promoted by BEM1. The clear function and relationship of these factors during germling fusion of N. crassa have to be determined in future studies.
BEM1 as a potential linker of ROS signalling and polar growth?
A recent publication reported the BEM1 homologue of E. festucae (BemA) as being physically interacting with NoxR, a regulatory component for NADPH oxidases, which produce ROS (Takemoto et al., 2011). Stable NoxR complex formation at hyphal tips is dependent on the presence of BemA, suggesting a direct link between ROS generating systems and polarity machineries. Interestingly, we recently showed that NoxR and the NADPH oxidase Nox1 are essential for germling fusion in the grey mould Botrytis cinerea (Roca et al., 2012). A comparable observation was reported for N. crassa (Read et al., 2012), suggesting a potential role of ROS signalling in germling fusion. Based on these observations, we hypothesize that BEM1 might function as a linker connecting ROS and MAP kinase signalling with cell polarity. Consistent with this hypothesis Δbem1, Δmak-2 and Δnor1 (Δnoxr) of N. crassa share other characteristic phenotypes, such as inviable sexual progeny (ascospores) (Cano-Dominguez et al., 2008). Ascospore germination is thought to include the controlled production of ROS spikes, resulting in establishment of a polarity axis and germ tube formation. To detect potential differences in ROS production between wild type and Δbem1 we stained germlings with the ROS-sensitive Nitro blue tetrazolium salt (NBT). Similar to what we have reported for B. cinerea (Roca et al., 2012), ROS were detected at the tips of germ tubes, at fusion tips and at fusion points after cell–cell contact. However, no differences in these patterns were found in wild type and mutant (data not shown). Similarly, ΔbemA hyphae in E. festucae exhibited similar ROS patterns as wild-type ones (Takemoto et al., 2011). The general detection of ROS, independent of their origin, appears to be not specific enough to detect subtle, but potentially very significant differences in their spatial and temporal production in mutants and wild type.
Based on the presented data we propose a model for germling fusion in N. crassa, in which a fusion signal activates the MAP kinase MAK-2 probably through the production of ROS. Activated MAK-2 is recruited to the site of polar growth potentially promoted by BEM1. In addition, BEM1 promotes also the localized accumulation of the ROS producing systems and the polarity machinery. By linking these different activities, BEM1 enhances positive feedback loops, which result in robust but highly dynamic cell-to-cell signalling and tropic growth. Further analysis of the composition and activity of BEM1 complexes in different polarization events in N. crassa will extend our understanding of the molecular networks controlling polarized growth, tropic responses and cellular differentiation.