Communicate and Fuse: How Filamentous Fungi Establish and Maintain an Interconnected Mycelial Network

Cell-to-cell communication and cell fusion are fundamental biological processes across the tree of life. Survival is often dependent upon being able to identify nearby individuals and respond appropriately. Communication between genetically different individuals allows for the identification of potential mating partners, symbionts, prey, or predators. In contrast, communication between genetically similar (or identical) individuals is important for mediating the development of multicellular organisms or for coordinating density-dependent behaviors (i.e., quorum sensing). This review describes the molecular and genetic mechanisms that mediate cell-to-cell communication and cell fusion between cells of Ascomycete filamentous fungi, with a focus on Neurospora crassa. Filamentous fungi exist as a multicellular, multinuclear network of hyphae, and communication-mediated cell fusion is an important aspect of colony development at each stage of the life cycle. Asexual spore germination occurs in a density-dependent manner. Germinated spores (germlings) avoid cells that are genetically different at specific loci, while chemotropically engaging with cells that share identity at these recognition loci. Germlings with genetic identity at recognition loci undergo cell fusion when in close proximity, a fitness attribute that contributes to more rapid colony establishment. Communication and cell fusion also occur between hyphae in a colony, which are important for reinforcing colony architecture and supporting the development of complex structures such as aerial hyphae and sexual reproductive structures. Over 70 genes have been identified in filamentous fungi (primarily N. crassa) that are involved in kind recognition, chemotropic interactions, and cell fusion. While the hypothetical signal(s) and receptor(s) remain to be described, a dynamic molecular signaling network that regulates cell-cell interactions has been revealed, including two conserved MAP-Kinase cascades, a conserved STRIPAK complex, transcription factors, a NOX complex involved in the generation of reactive oxygen species, cell-integrity sensors, actin, components of the secretory pathway, and several other proteins. Together these pathways facilitate the integration of extracellular signals, direct polarized growth, and initiate a transcriptional program that reinforces signaling and prepares cells for downstream processes, such as membrane merger, cell fusion and adaptation to heterokaryon formation.


CELL-TO-CELL COMMUNICATION IS A FUNDAMENTAL BIOLOGICAL PROCESS
Cells rarely exist alone, which has driven the evolution of diverse mechanisms for identifying and responding to the presence of other nearby cells. Animals have evolved complex organs (i.e., eyes and ears) for sensing and interpreting the surrounding environment. However, the functioning of these organs is dependent on the cells that make up each organ and their ability to accurately sense and communicate this information to other cells, resulting in a coordinated response. For example, when light enters the mammalian eye, rod and cone photoreceptors in retinal cells become activated, which triggers a change in the amount of glutamate that retinal cells release (Soucy et al., 1998). Nearby bipolar cells are sensitive to fluctuations in glutamate, which causes a flux in sodium and calcium cations, which activates nearby ganglion cells that initiate a signal transduction pathway that is ultimately received, integrated, and interpreted by multiple cells in the brain (Soucy et al., 1998;Cruz-Martín et al., 2014).
Cell-to-cell signaling is a fundamental biological process that has been well studied in many organisms besides mammals. For example, there is abundant research detailing the mechanisms of molecular warfare between plants and pathogenic fungi (Dodds and Rathjen, 2010), hormone signaling between plant cells (Jaillais and Chory, 2010), pheromone sensing between gametes during mating in yeast (Merlini et al., 2013), and quorum sensing that mediates coordinated behaviors (i.e., bioluminescence or biofilm formation) among bacteria (Waters and Bassler, 2005). The majority of cell-to-cell signaling research has focused on unraveling the mechanisms that mediate two general methods for cell-to-cell communication; either a density-dependent survey of how many other similar cells are nearby (i.e., quorum sensing), or signaling and sensing between cells that are intrinsically different from each other (i.e., host-pathogen interactions, or during mating). This review describes another type of cell-tocell communication in which communication occurs between genetically identical cells. While this form of communication is likely important for organisms across the tree of life, it is fundamentally important for the development and functioning of filamentous fungi.
Filamentous fungi exist as a multicellular, multinucleate interconnected network of tube-shaped cells called hyphae. In a hyphal colony, cells engage in cell-to-cell communication and undergo chemotropic growth toward each other until they make physical contact. Once contact is established, both cells will initiate the process of cell fusion, in which cell walls are remodeled, plasma membranes fuse, and ultimately the two cells become one with a shared cytoplasm. This process results in the hallmark interconnected colony associated with filamentous fungi (Buller, 1933;Hickey et al., 2002). Cell fusion between hyphae within an individual colony reinforces network architecture and influences the flow of resources throughout the colony (Hickey et al., 2002;Simonin et al., 2012;Roper et al., 2013Roper et al., , 2015. Hyphal fusion can also occur between fungal colonies, and can result in one of two outcomes. First, two genetically similar colonies can fuse together, resulting in shared resources and collaboration instead of competition (Bastiaans et al., 2015). Second, if two colonies that are genetically dissimilar at non-self recognition loci undergo fusion, the fused cells are compartmentalized and programmed cell death is initiated, which prevents shared cytoplasm from mixing throughout both colonies (Saupe, 2000;Glass and Kaneko, 2003;Gonçalves et al., 2017). In Neurospora crassa and most other fungi, there is no nuclear fusion or genetic recombination that occurs as a result of somatic cell fusion. However, some fungi (i.e., Aspergillus spp., Candida spp., and Ustilago spp.) are capable of a low frequency of somatic nuclear fusion following hyphal fusion, during which genetic recombination can occur via mitotic crossing-over events and chromosome loss, the so-called parasexual cycle (Garber and Ruddat, 1992;Bennett and Johnson, 2003;Schoustra et al., 2007). Unlike sexual reproduction, parasexual reproduction does not involve any specialized structures (i.e., gametes), meiosis does not occur, and no fruiting body or specialized spores are formed.
Communication-mediated cell fusion is important throughout the life cycle of most fungi, including Ascomycete fungi, such as N. crassa. Somatic cell fusion has been reported in fungi for over 100 years (Meyer, 1902). It is unclear whether species in the Mucoromycota undergo hyphal fusion, but hyphal fusion has been observed between colonies of the Glomeromycete species, Rhizophagus intraradices (Croll et al., 2009). Basidiomycete fungi depend on a unique form of hyphal fusion that forms small hyphal bridges around septa called 'clamp connections, ' which are important for facilitating nuclear movement during mitotic growth and maintaining a tightly regulated dikaryotic state within each cell (Buller, 1922;Read et al., 2010). Hyphal fusion is also important for building the tissue that ultimately defines the fruiting body of both Basidiomycete and Ascomycete fungi ( Van der Valk and Marchant, 1978;Read et al., 2010).

COMMUNICATION AND CELL FUSION IN Neurospora crassa
Neurospora crassa is a well-developed model organism for studying eukaryotic genetics and cell biology, including cellto-cell communication and cell fusion (Galagan et al., 2003;Colot et al., 2006;Leeder et al., 2011;Herzog et al., 2015). Communication and cell fusion are important at several points throughout the N. crassa lifecycle (Figure 1). N. crassa is a heterothallic, hermaphroditic species; each colony is capable of producing both male and female structures, but mating only occurs between colonies that encode opposite mating types, mat A or mat a. Sexual reproduction is initiated by a fertile hypha in the protoperithecium called the trichogyne ("female") that chemotropically grows toward a "male" cell (conidium or hypha) that is producing the opposite mating pheromone Borkovich, 2004, 2006). Upon cell fusion between the trichogyne and the male cell, the male nucleus is transported down the trichogyne to the protoperithecium, where it replicates with female nuclei to form ascogenous hyphae (Backus, 1939). Within ascogenous hyphae, nuclei of opposite mating types undergo FIGURE 1 | The life-cycle of Neurospora crassa. N. crassa is a heterothallic Ascomycete species with a distinct sexual cycle and asexual cycle. Conidia are clonal asexual propagules that can either generate a new colony on their own, or serve as a "male" partner to the "female" trichogyne during mating. The trichogyne is a specialized hypha that emerges from the protoperithecium and chemotropically grows toward a conidium of opposite mating type (shown here expressing H1-GFP). Ascospores are the result of meiosis, which occurs inside the perithecium. This lifecycle specifically highlights chemotropic interactions and cell fusion events. Stars indicate chemotropic interactions and fusion. Germlings, hyphae, and the trichogynes all undergo chemotropism and cell fusion. Protoperithecium image is from Lichius et al. (2012b), and the image showing many ascospores is from Raju (2009), with permission.
one mitotic division and septation to form the crozier. Opposite mating type nuclei in the penultimate cell of the crozier undergo karyogamy, meiosis, and ultimately becomes an ascus containing the eight meiotic ascospores (Raju, 1980;Fleißner et al., 2008b). This process is replicated many times, resulting in hundreds of asci within a single fruiting body.
Communication and fusion also occur during the vegetative phase of the N. crassa lifecycle. N. crassa, like many Ascomycete species, produces large quantities of clonal, asexual spores (conidia) that are easily dispersed by wind. When these genetically identical conidia land on a suitable substrate they will germinate and chemotropically grow toward each other to undergo cell fusion. Germinated conidia are called germlings. If non-clonal germlings are genetically different at loci termed "determinant of communication" or doc, chemotropic interactions are aborted and cell fusion frequency is drastically reduced (Heller et al., 2016). To prove that germling interactions are truly chemotropic, germlings were moved via optical tweezers; germlings re-oriented to maintain growth toward their communication partners (Wright et al., 2007). Lastly, new protein synthesis is not required for chemotropic interactions in N. crassa (Fleissner et al., 2009).
As a fungal colony continues to grow, hyphae within it undergo fusion to make the interconnected mycelial network via chemotropic interactions and cell fusion, which is mechanistically equivalent to the process of germling fusion. In fact, mutants that are defective at germling fusion are almost always also defective in hyphal fusion (Fleißner et al., 2008b;Fu et al., 2011). Fusion mutants are also almost always defective at producing extended aerial hyphae and developing female reproductive structures (protoperithecia), indicating that hyphal fusion may be important for supporting the development of these complex structures. However, there is no correlation between hyphal fusion and conidia production, which occurs on aerial hyphae (Li, 2005;Lichius et al., 2012a). Germling and hyphal fusion in N. crassa occurs between genetically similar cells and results in cytoplasmic mixing, but nuclei remain intact and there is no genetic recombination (Roper et al., 2011).
Since the discovery of the first hyphal anastomosis mutant, ham-1 (also called soft) (Wilson and Dempsey, 1999), more than 70 genes have been identified that are involved in the process of communication and/or fusion. Broadly speaking, the process of germling/hyphal communication and fusion is multifaceted. Germlings and fusion-hyphae produce, secrete, and sense a signaling molecule. To achieve chemotropic growth, the signal-sensing machinery is likely integrated with the machinery that directs polarized growth. Once interacting germlings make physical contact, they begin the process of cell fusion, which includes the breakdown of the cell wall, remodeling plasma membranes, and ultimately mixing of cytoplasm (Hickey et al., 2002;Goryachev et al., 2012). A pre-fusion checkpoint reinforces communication with genetically similar individuals and blocks communication with genetically dissimilar individuals (Heller et al., 2016). If two genetically dissimilar individuals do fuse, there is a post-fusion non-self recognition checkpoint mediated by heterokaryon incompatibility (het) loci . A recently fused cell will immediately induce cell death if fusion occurs between two cells that differ in specificity at het loci. While death is inherently costly, it may be an adaptive behavior to stop the spread of mycoviruses or cheater genotypes within a colony (Glass and Kaneko, 2003). Alternatively, in the plant pathogenic fungus Colletotrichum lindemuthianum, it has been hypothesized that temporary suppression of heterokaryon incompatibility may facilitate genetic diversification via horizontal gene transfer (Ishikawa et al., 2012).

THE CELL WALL INTEGRITY PATHWAY
The Cell Wall Integrity (CWI) pathway is composed of cell wall sensors that transduce information through a Mitogen Activated Protein Kinase (MAPK) cascade, which is a central signaling complex conserved across fungi. The ascomycete yeast, Saccharomyces cerevisiae, has five different cell wall sensors that activate the CWI pathway (Rodicio and Heinisch, 2010). A screen of mutants encoding cell wall proteins in N. crassa identified two genes (wsc-1 and ham-7) that encode GPI-anchored cell wall proteins that are required for maintenance of the cell wall, resistance to cell wall stress, and phosphorylation of MAK-1 (Maddi et al., 2012). Both the wsc-1 and ham-7 mutants have pleiotropic phenotypes, but only the ham-7 mutant is defective at germling and hyphal fusion (Maddi et al., 2012). There are likely more cell wall sensors that activate the MAK-1 pathway that have not yet been identified due to functional redundancy or mutant lethality.
The core conserved component of the CWI pathway is a MAPK cascade, which consists of MIK-1 (MAPKKK), MEK-1 (MAPKK), and MAK-1 (MAPK). Scaffold proteins often play an important role in regulating or buffering MAPK pathways, although a scaffold protein is not always necessary for pathway function. In S. cerevisiae, Spa2p acts as a scaffold that links the terminal CWI pathway MAPKK and MAPK to the rest of the polarisome, which coordinates polarized growth (van Drogen and Peter, 2002). In N. crassa, the spa-2 gene is dispensable for cell fusion, but necessary for polarized growth, and fluorescently labeled SPA-2 localizes to hyphal tips, germ-tube tips, septa, and sites of cell fusion (Araujo-Palomares et al., 2009;Lichius et al., 2012b). While the N. crassa SPA-2 protein is analogous to Spa2p in S. cerevisiae, it remains unclear if the N. crassa SPA-2 plays a direct roll in MAPK signaling or cell-to-cell communication. In contrast, the soft gene (also named ham-1) encodes a potential scaffold protein that physically interacts with MIK-1 and MEK-1 in both N. crassa and the related ascomycete species, Sordaria macrospora Weichert et al., 2016). During chemotropic interactions between genetically identical cells, the SOFT protein shows a dynamic localization pattern in which fluorescently tagged SOFT will assemble into distinct puncta focused at the cell periphery of the germ tube tip in communicating germlings [also called Conidial Anastomosis Tubes (CATs)] (Roca et al., 2005;Fleissner et al., 2009). These SOFT complexes remain assembled at the germ tube tip for roughly 4 min, after which they disassemble and become diffuse in the cytoplasm. Approximately 4 min later, SOFT complexes re-assemble at the germ tube tip. These roughly 4 min SOFT-complex oscillations of assembly and disassembly continue for the duration of chemotropic growth between two communicating germlings or hyphae undergoing chemotropic interactions (Fleissner et al., 2009). In contrast, the MAK-1 protein does not have a dynamic pattern of localization during chemotropic interactions, but localizes to germ tube tips at the site of fusion (Dettmann et al., 2013;Weichert et al., 2016). SOFT physically interacts with MIK-1 and MEK-1, but not MAK-1 Weichert et al., 2016). In S. macrospora, the soft-ortholog (PRO40) also directly interacts with an upstream activator, PKC1, and indirectly interacts with a Rho GTPase, RHO1. There are orthologs of rho1 and pkc1 in N. crassa, but there is no available rho-1 mutant, and the pkc-1 deletion is lethal (Park et al., 2011). Thus it remains unclear if pkc-1 and rho-1 are important for the CWI pathway in N. crassa or whether these proteins play a role in cell fusion.
Canonical MAPK cascades generally lead to a change in transcription. In S. cerevisiae, the CWI pathway regulates transcription via transcription factors Swi4, Swi6, and Rlm1 (Levin, 2005). In N. crassa, the MAK-1 protein localizes to both the cytoplasm and the nucleus, indicating that it may also play a role in regulating transcription (Dettmann et al., 2013). The circadian rhythmic pattern of expression for some mak-1-dependent genes mirrors the rhythmic expression pattern of genes dependent on the transcription factor ADV-1; both mak-1 and adv-1 are targets of the circadian clock (Bennett et al., 2013;Dekhang et al., 2017). Furthermore, MAK-1 is required for transcription of adv-1, and misexpression of adv-1 is sufficient to suppress the distinct colony morphology, slow growth rate, and short aerial hyphae phenotypes of the mak-1 mutant (Fischer et al., 2018).
In germlings and hyphae undergoing chemotropic interactions, the HAM-5, MAK-2, MEK-2, NRC-1, and STE-50 proteins form a complex that dynamically assembles and disassembles at roughly 4 min intervals at germ tube tips or hyphal fusion tips, mirroring the dynamic oscillations of SOFT (Fleissner et al., 2009;Dettmann et al., 2012;Jonkers et al., 2014). The MAK-2 complex assembles at the plasma membrane at germ tube tips, and when it disassembles, the components of the complex become diffuse in the cytoplasm (Fleissner et al., 2009). When the MAK-2 complex is assembled at the tip of one germ tube, at the tip of the opposing germ tube is a complex containing the CWI pathway scaffold protein, SOFT (Herzog et al., 2015;Fleißner and Herzog, 2016). Dynamics of SOFTcomplex assembly and disassembly is equivalent to, but perfectly out of phase with MAK-2-complex dynamics at germ tube tips and hyphal fusion tips (Fleissner et al., 2009). Oscillation of MAK-2 to germ tube tips is required for chemotropic interactions, as evidenced by tethering MAK-2 to the plasma membrane, which disrupts communication (Serrano et al., 2018). The observation of MAK-2 and SOFT dynamic oscillations led to a "ping-pong" model (Goryachev et al., 2012), in which SOFT is predicted to be involved in signal secretion, while the MAK-2 pathway is predicted to be involved in signal reception. By temporally separating signal sending and signal receiving, genetically identical germlings could to avoid self-stimulation and maintain communication and chemotropic growth toward a physiologically distinct, but genetically identical clone.
The upstream regulators of the MAK-2 pathway in N. crassa are largely unknown, except for STE-50 (a conserved kinase regulator), STE-20 (PAK kinase), and RAS-2 (GTPase). In N. crassa, these three proteins physically interact with members of the MAK-2 complex and are required for phosphorylation of MAK-2 . Additionally, the conserved Rho GTPases CDC-42 and RAC-1 are necessary for maintaining polarized growth and cell fusion in N. crassa; when RAC-1 was inhibited, oscillatory recruitment of MAK-2 was stopped (Lichius et al., 2014). In mammals, activation of the ERK1/2 pathway generally begins with signal reception via a tyrosine receptor kinase, which activates a Ras GTPase, which then activates the ERK1/2 cascade. Ccd42, PAK Kinases (STE-20like), and Rac GTPases also often act to modulate ERK1/2 signaling (Shaul and Seger, 2007). In S. cerevisiae, activation of the pheromone response pathway begins with signal reception via a G-Protein Coupled Receptor (GPCR; Ste2p or Ste3p), which leads to disassociation of the G-proteins, which activates Ste20p, Ste50p, and the Far1p/Cdc24p complex, and ultimately activates the Fus3p (mak-2 ortholog) MAPK cascade (Bardwell, 2005). Tyrosine receptor kinases are conserved in the Opisthokonts and Amoebozoa, but were lost in the Ascomycota (including N. crassa), Ustilaginomycotina, and Pucciniomycotina lineages (Zhao et al., 2014). Thus, based on homology to the S. cerevisiae pheromone response pathway, it is expected that a GPCR activates the MAK-2 pathway in N. crassa. However, the N. crassa pheromone GPCR receptors and G-proteins are completely dispensable for somatic communication and cell fusion, although they play an important role during mating cell fusion and sexual development Borkovich, 2004, 2006).
One intriguing component of the MAK-2 cascade in N. crassa is the NDR-kinase scaffold, HYM-1. NDR-kinases such as COT-1 regulate actin dynamics that define cell polarity (Yarden et al., 1992;Ziv et al., 2009), but are generally not essential for communication or fusion in N. crassa. HYM-1 is the scaffold protein for the COT-1 pathway, but HYM-1 also physically interacts the STE-20, and is required for the sequential phosphorylation of NRC-1, MEK-2, and MAK-2 (Dettmann et al., 2012). This observation suggests that HYM-1 is the bridge that connects MAK-2 pathway signaling with actin dynamics and cell polarity. However, the MAK-2 complex does not colocalize with actin markers, which indicates that the connection between MAK-2 and actin dynamics may be transient or indirect (Jonkers et al., 2016). In communicating germlings, both HYM-1 and COT-1 localize to the germ tube tip and do not oscillate dynamically, like MAK-2 or SOFT (Dettmann et al., 2012).
The transcription factor PP-1 (ortholog of Ste12p) is a conserved downstream target of MAK-2, and the pp-1 mutant phenocopies the mak-2 mutant (Li, 2005). A microarray expression study demonstrated that there is some overlap between MAK-2-dependent and PP-1-dependent gene expression . However misexpression of pp-1 in a mak-2 background is not sufficient to induce transcription of downstream genes, which suggests that MAK-2 may directly activate, or de-repress PP-1 (Fischer et al., 2018). Consistent with this hypothesis is the observation that the PP-1 protein has several phosphorylation sites that are phosphorylated in a MAK-2-dependent manner (Jonkers et al., 2016).

THE STRIPAK COMPLEX
The Striatin-Interacting protein Phosphatase And Kinase (STRIPAK) complex is a highly conserved multi-subunit protein complex that regulates diverse processes such as cytoskeletal organization, cell migration, and cell morphology in mammalian cells (Goudreault et al., 2009). In fungal cells, the STRIPAK complex is important for sexual development, growth, and cell fusion (Simonin et al., 2010;Dettmann et al., 2013;Kück et al., 2016). As evidence of its broad evolutionary conservation, the phenotype of the S. macrospora Striatin mutant was complemented with the mouse Striatin gene (Pöggeler and Kück, 2004). The STRIPAK complex in N. crassa consists of six proteins; PPG-1 (serine/threonine phosphatase), PP2A-A (phosphatase scaffold), HAM-3 (Striatin, phosphatase regulator), HAM-2 (membrane anchor), HAM-4 (membrane anchor), and MOB-3 (kinase activator) (Simonin et al., 2010;Dettmann et al., 2013). PP2A-A, HAM-3, and PPG-1 are the core members of a highly conserved phosphatase complex called PP2A that can function alone or in association with other complexes like the STRIPAK complex (Kabi and McDonald, 2014). The PP2A complex is considered to be the major kinase phosphatase in eukaryotic cells, and defines a central signaling hub for many eukaryotic signaling pathways. No kinases have been identified as components of the STRIPAK complex in N. crassa, however, in closely related S. macrospora, two kinases (KIN3 and KIN24) were shown to have a weak physical interaction with the Striatin ortholog in the STRIPAK complex (Frey et al., 2015); KIN3 is required for fruiting body formation, hyphal fusion, and septation (Radchenko et al., 2018). The homolog of KIN24 in N. crassa is MST-1 (NCU00772), which physically interacts with NRC-1 and MEK-2 , indicating that MST-1 may physically connect the MAK-2 pathway with the STRIPAK complex. However, the mst-1 gene is dispensable for germling communication and cell fusion. The MST-1 and the KIN3-ortholog STK-3 localize to septa and spindle pole-bodies, and were thus hypothesized to regulate septation and polarized growth in N. crassa (Frey et al., 2015;Heilig et al., 2014). Lastly, a SIKE-like small coiled-coil protein, SCI1, was recently identified as a component of the STRIPAK complex in S. macrospora (Reschka et al., 2018). The SCI1 protein physically interacts with the orthologs of HAM-3 and HAM-4, and the sci1 mutant displays defects in hyphal fusion, perithecial development, and growth rate (Reschka et al., 2018).

CROSSTALK BETWEEN THE CWI PATHWAY, THE MAK-2 PATHWAY, AND THE STRIPAK COMPLEX
There are several points of crosstalk between the CWI pathway, the MAK-2 pathway, and the STRIPAK complex, which is summarized in Figure 2. Both the MAK-1 and the MAK-2 pathways regulate gene transcription via the transcription factors PP-1 and ADV-1 Dekhang et al., 2017;Fischer et al., 2018). Briefly, MAK-2 is required for activation or derepression of PP-1, then PP-1 directly activates transcription of ADV-1, and ADV-1 directly activates transcription of many genes that are required for communication, fusion, growth, and cell wall stress response, including mek-1. Additionally, MAK-1 was shown to function upstream of ADV-1, independently of PP-1 FIGURE 2 | Diagram of crosstalk between the MAK-2 pathway, the CWI/MAK-1 pathway and the STRIPAK complex. A cyan "P" indicates canonical MAPK cascade phosphorylation, and a yellow "P" indicates MAK-2-dependent phosphorylation. MAPK pathway proteins are purple, the STRIPAK complex is blue, and pink transcription factors (PP-1 and ADV-1) are downstream of all three signaling components. MAK-2 is required for activation or de-repression of PP-1, and MAK-1 regulates ADV-1 in a PP-1-independent manner. (Fischer et al., 2018). The ham-7, mik-1, mek-1, and nrc-1 genes are all required for full phosphorylation of MAK-1 (Maddi et al., 2012;Leeder et al., 2013;Fu et al., 2014;Teichert et al., 2014) and both ham-7 and mak-1 have been shown to be required for full phosphorylation of MAK-2 (Maerz et al., 2008;Dettmann et al., 2012;Maddi et al., 2012;Leeder et al., 2013). These data suggest that HAM-7 either functions upstream of the MAK-1 cascade and influences MAK-2 phosphorylation indirectly via MAK-1, or HAM-7 functions directly upstream of both the CWI and MAK-2 pathways.
MAP kinases post-translationally regulate other proteins via phosphorylation, and MAK-2 is required for the phosphorylation of several proteins that are important for communication and cell fusion ( Table 1). Catalytically active MAK-2 is required for phosphorylation of kinases and scaffold proteins in both the MAK-2 and CWI pathways (Jonkers et al., 2014). MAK-2 also phosphorylates MOB-3, which affects accumulation of MAK-1 in the nucleus (Dettmann et al., 2013). Both mob-3 and ham-3 are required for full phosphorylation of MAK-1. In addition to MOB-3, three other STRIPAK proteins; HAM-3, HAM-4, and PP2A-A are also phosphorylated in a MAK-2-dependent manner. Additional evidence for STRIPAK-mediated crosstalk between the CWI and MAK-2 pathways was demonstrated via proteinprotein interaction experiments. An Affinity Purification-Mass Spectrometry (AP-MS) experiment identified a weak interaction between MAK-2 and the STRIPAK components PPG-1, PP2A-A, HAM-3, and MOB-3 . Furthermore, MST-1 was identified by both AP-MS and Y2H as interacting with NRC-1 and MEK-2 and, in S. macrospora, the ortholog of MST-1 was found to have a weak interaction with the HAM-3-homolog Frey et al., 2015).

PHOSPHATASES
The magnitude and duration of phosphorylation determines the signaling output of a kinase pathway, and phosphatases play a critical role in modulating protein phosphorylation (Behar et al., 2008;Caunt and Keyse, 2013). For example, constitutive activation of the Ras protein causes Schizosaccharomyces pombe cells to initiate fusion prior to making contact with a fusion partner (Merlini et al., 2018). The phosphatase genes csp-6 and ppg-1 are required for germling communication and cell fusion in N. crassa (Dettmann et al., 2013) and the ptp-2 gene is the ortholog of the phosphatase that negatively regulates the mak-2-ortholog in S. cerevisiae. The ptp-2 mutant is not defective at germling communication or fusion, but the PTP-2 protein is phosphorylated in a MAK-2-dependent manner, which may be indicative of a feedback loop in which MAK-2 regulates itself via PTP-2 (Jonkers et al., 2016). MAK-2 also regulates phosphorylation of CSP-6 and PP2A-A, which is the PPG-1 (phosphatase) regulatory subunit of the STRIPAK complex.
The csp-6 gene was initially identified in N. crassa in a screen of phosphatase mutants (Ghosh et al., 2014). The csp-6 mutant has a pleiotropic phenotype that is defective at conidial separation, growth rate, female sexual development, and circadian clock regulated developmental oscillations, and is sensitive to osmotic stress (Chinnici et al., 2014;Ghosh et al., 2014;Zhou et al., 2018). The CSP-6 protein is orthologous to Psr1p and Psr2p in S. cerevisiae, which physically interact with Whi2p to mediate the general stress response pathway (Kaida et al., 2002). Similarly, the CSP-6 protein physically interacts with the WHI-2 protein in N. crassa (Zhou et al., 2018). The CSP-6 protein was shown to dephosphorylate the main circadian clock regulator, WC-1, which affects ADV-1-dependent transcription. Furthermore, both CSP-6 and WHI-2 localize to the cytoplasm and the nucleus (Zhou et al., 2018). While phosphorylation of MAK-1 and MAK-2 in a csp-6 background has not been assessed, phosphorylation of both MAK-1 and MAK-2 is dramatically reduced in the whi-2 mutant , indicating that WHI-2, and likely CSP-6, function upstream of MAK-1 and MAK-2. Together with the STRIPAK complex, CSP-6 and WHI-2 may function as a signaling hub that contributes to the integration of multiple signaling pathways.

TRANSCRIPTION FACTORS
Two transcription factors, adv-1 and pp-1 are necessary for germling communication and fusion in N. crassa Dekhang et al., 2017). PP-1 is a C2H2-Zn 2+ transcription factor that is homologous to the pheromone response pathway transcription factor, Ste12p in S. cerevisiae . In S. cerevisiae, the pheromone response pathway is orthologous to the MAK-2 pathway and results in de-repression of the Ste12p transcription factor (Bardwell, 2005;Merlini et al., 2013). Ste12-like proteins, including PP-1, have two C2H2-Zn 2+ motifs and a homeobox-like STE domain that is involved in binding DNA (Errede and Ammerer, 1989). In N. crassa the STE domain is essential for function, but the C2H2-Zn 2+ motifs are dispensable . Ste12-like transcription factors are regulated by direct phosphorylation and phosphorylation of associated regulatory proteins (Blackwell et al., 2007). Several phosphorylation sites have been identified on PP-1 in N. crassa, but the biological significance of these sites remains unknown Jonkers et al., 2014;Xiong et al., 2014). Additionally, there are no known regulatory proteins that bind PP-1 in a way that is analogous to how Dig1p/Dig2p bind and repress Ste12p in S. cerevisiae (McCullagh et al., 2010). However, it has been suggested that a nuclear protein called NsiA may physically bind and positively regulate the ortholog of pp-1 in the plant endophyte, Epichloë festucae (Green, 2016).
Ste12-like proteins have been well studied and they regulate diverse functions throughout the fungal kingdom, such as growth, virulence, sex, dimorphism, and asexual development (Hoi and Dumas, 2010). Ste12p and its homologs in the yeasts S. cerevisiae and S. pombe are essential for sexual reproduction, as they respond to pheromone stimulus and activate transcription of mating-specific genes (Bardwell, 2005;Merlini et al., 2013). Ste12p also regulates the morphological switch from yeast to pseudohyphal growth in response to starvation (Roberts and Fink, 1994). In addition to regulating sexual reproduction and morphological switching, a Ste12 homolog also regulates virulence in the animal-pathogenic yeasts Candida glabrata and Cryptococcus gattii (Calcagno et al., 2003;. In the dimorphic opportunistic human pathogen Penicillium marneffei, the Ste12-homolog StlA is dispensable for the switch from the yeast to hyphal morphology and has no other known function. However, the identification of StlA in the genome suggests that cryptic sex exists in P. marneffei; stlA from P. marneffei is sufficient to complement the sexual defects in the Aspergillus nidulans mutant, steA (Borneman et al., 2001). In A. nidulans, SteA is necessary for sexual reproduction, but dispensable for asexual growth and development (Vallim et al., 2000). In contrast, other filamentous fungi depend on Ste12 to regulate growth, asexual development, and pathogenicity. The Ste12 homolog in N. crassa and S. macrospora is not essential for mating cell fusion, but it is essential for wild-typelike growth, somatic cell fusion, protoperithecium development (N. crassa only), ascus development, and ascospore germination (Colot et al., 2006;Nolting and Pöggeler, 2006;Leeder et al., 2013). In S. macrospora, the NADPH-oxidase nox2 is specifically required for ascospore development, and there is overlap in the differentially expressed genes in the nox2 and ste12 mutants as compared to wild-type cells (Dirschnabel et al., 2014). Filamentous plant pathogenic fungi (i.e., Magnaporthe oryzae, Alternaria brassicola, Verticillium dahlia, Fusarium oxysporum, and Colletotrichum lagenarium) require the Ste12-homolog for either appressorium formation or plant penetration, and subsequent pathogenesis on the plant host. Furthermore, the Ste12-homolog in these plant pathogens is dispensable for sexual reproduction where the sexual cycle is observable (Park et al., 2002;Tsuji et al., 2003;Cho et al., 2009;Rispail and Di Pietro, 2010;Sarmiento-Villamil et al., 2018). In contrast, the plant pathogen Cryphonectria parasitica requires a Ste12-homolog to regulate both female fertility and virulence (Sun et al., 2009).
ADV-1 is a Zn(II) 2 Cys 6 transcription factor that, like PP-1, regulates growth, sex, virulence, sexual development and cell fusion (Colot et al., 2006;Dekhang et al., 2017). ADV-1 regulated transcription is a target of the CWI and MAK-2 pathways (Fischer et al., 2018). In contrast to PP-1, ADV-1 has also been shown to be a target of the circadian clock master regulator, the White Collar Complex (WCC) (Dekhang et al., 2017). ADV-1 is essential for developmental oscillations (i.e., conidiation) and the quantity of adv-1 mRNA and protein correlates well with the circadian clock, though adv-1 expression is never fully off . The expression of ADV-1 target genes in a mature colony also occurs rhythmically and in concert with the circadian clock. These data led to the hypothesis that hyphal fusion within a fungal colony is a clock-regulated developmental process (Dekhang et al., 2017).
ADV-1 is not as broadly conserved as PP-1, as clear adv-1 homologs are restricted to the filamentous Ascomycetes (Pezizomycotina). In Aspergillus spp., the adv-1 homolog RosA represses fruiting body development whereas a second adv-1 homolog, NosA, induces fruiting body development. nosA and rosA share 43% sequence similarity, and nosA is a target of the metabolism and the developmental regulator, LaeA (Vienken et al., 2005;Vienken and Fischer, 2006;Zhao et al., 2017). In S. macrospora, the pro1 gene is the homolog of adv-1, and the N. crassa adv-1 gene is sufficient to complement the S. macrospora pro1 mutant. A pro1-like gene (named hpro1A) was identified in both N. crassa and S. macrospora that shares 40% similarity with pro1 and adv-1, however, this potential paralog does not complement the function a S. macrospora pro1 mutant (Masloff et al., 2002). It remains unclear if hpro1A plays a role in communication, fusion, or related processes in either species. PRO1, like other Zn(II) 2 Cys 6 transcription factors, has a Gal4-like DNA-binding domain and a transcription-activation domain. However, unlike Gal4p, PRO1 lacks a coiled-coil dimerization domain, indicating that PRO1 likely works alone (Masloff et al., 2002). The S. macrospora pro1 mutant and the N. crassa adv-1 mutant both produce protoperithecia, but are defective in the early stages of perithecial development (Masloff et al., 1999;Fischer et al., 2018). In plantpathogenic A. brassicicola, the adv-1 homolog is required for growth and may also be involved in virulence (Cho et al., 2009). Another plant pathogen, C. parasitica, does not require the adv-1 homolog for virulence, but does require it for female fertility and maintaining stable mycovirus infection (Sun et al., 2009). The adv-1 homolog in the plant endophyte E. festucae is required for hyphal fusion and maintenance of mutualism. E. festucae is normally a mutualistic grass endophyte, but certain environmental conditions or the deletion of specific genes, such as the adv-1 homolog, cause E. festucae to switch to a pathogenic lifestyle, resulting in stunted growth of the host plant (Tanaka et al., 2013).

THE NOX COMPLEX AND REACTIVE OXYGEN SPECIES
While Reactive Oxygen Species (ROS) can be destructive, ROS are also known to be an important aspect of several eukaryotic signaling pathways including cell differentiation, development, and cytoskeletal remodeling ( Hansberg et al., 1993;Finkel, 2003;Aguirre et al., 2005). NADPH-oxidases (NOX) are evolutionarily conserved and produce superoxide by oxidizing NADPH and reducing molecular oxygen. Three different nox gene subfamilies have been identified in Aspergillus spp.: noxA, noxB, and noxC (Aguirre et al., 2005). The genome of N. crassa has nox-1 (noxA) and nox-2 (noxB) homologs, but not a noxC homolog (Galagan et al., 2003). The noxA and noxB homologs have diverse functions in fungi, but in general, noxA is important for a variety of processes including fruiting body development, cell fusion, growth, sclerotial development, and virulence of plant pathogens (Lara-Ortíz et al., 2003;Malagnac et al., 2004;Egan et al., 2007;Cano-Dominguez et al., 2008;Segmüller et al., 2008;Dirschnabel et al., 2014). The noxB homologs often have a more specific function such as ascospore development or mediating appressorium penetration of a host plant (Egan et al., 2007;Cano-Dominguez et al., 2008;Segmüller et al., 2008;Tanaka et al., 2008;Dirschnabel et al., 2014). In S. macrospora, many of the genes required for communication, cell fusion, and sexual development are differentially expressed in a nox-1 mutant as compared to wild-type hyphae/protoperithecia (Dirschnabel et al., 2014). These data suggest that the NOX-1 complex likely functions upstream of a transcriptional response during communication and cell fusion.
The nor-1 gene is a conserved regulatory subunit of both nox-1 and nox-2, however, only nox-1 and nor-1 are required for germling communication, cell fusion, growth, and female sexual development in N. crassa (Cano-Dominguez et al., 2008). NoxR (nor-1 ortholog) was first described in the plant endophyte E. festucae, and was shown to activate Nox1 in a complex with the conserved regulatory proteins, BemA, Cdc24, and RacA (Takemoto et al., , 2011Tanaka et al., 2006Tanaka et al., , 2008. Recently, Cdc42 was identified as a negative regulator of Nox1 in E. festucae, and which also physically interacts with BemA (Kayano et al., 2018). Orthologs of BemA, RacA, Cdc24, and Cdc42 are conserved general regulators of polarized growth and cytoskeletal dynamics in fungi and have been particularly well characterized in S. cerevisiae (Drees et al., 2001;Martin and Arkowitz, 2014). Strains containing mutations in orthologs of RacA, Cdc24, and Cdc42 in N. crassa and other filamentous fungi have a strong pleiotropic phenotype, while strains carrying mutations in the BemA ortholog are less pleiotropic and are not severely compromised for polarized growth (Araujo-Palomares et al., 2011;Takemoto et al., 2011;Schürg et al., 2012). However, these four proteins may represent a link between cytoskeletal dynamics and the NOX complex.
The HAM-6 protein is a component of the NOX-1 complex in filamentous fungi. The ham-6 gene was initially identified in S. macrospora and N. crassa in screens for mutants defective at sexual development and germling fusion (Nowrousian et al., 2007;Fu et al., 2011). This gene encodes a transmembrane domain protein that is conserved among the filamentous Ascomycetes (Pezizomycotina). Recent research demonstrated that HAM-6 is a necessary component of the NOX-1 complex in filamentous fungi. The ham-6 ortholog in Botrytis cinerea and Podospora anserina (a close relative of N. crassa) physically and genetically interacts with the nox-1 and nor-1 orthologs (Lacaze et al., 2015;Siegmund et al., 2015;. The ham-6 gene is differentially expressed in nox-1 cells compared to wild-type, and ham-6 is epispastic to soft in S. macrospora (Nowrousian et al., 2007;Dirschnabel et al., 2014). Interestingly, ham-6 is required for full phosphorylation of MAK-1 , indicating that HAM-6, and potentially the NOX-1 complex, function upstream of the CWI pathway. Additional evidence of a connection between the NOX-1 complex and the CWI pathway was demonstrated in the mycoparasitic Dothidiomycete, Coniothyrium minitans, in which the MAK-1 ortholog physically interacts with the NOX-1 ortholog (Wei et al., 2016). Furthermore, over-expression of the mak-1ortholog was sufficient to partially suppress the phenotype of the rac-1 mutant (Wei et al., 2016). In B. cinerea, Nox1 was also recently shown to physically interact with Mak1 via the scaffold protein Iqg1 . The Iqg1 protein is a conserved scaffold that links multiple signaling pathways, including the NOX and CWI pathways .
The superoxide produced by NOX enzymes is highly reactive, short-lived, and unable to cross biological membranes because of the negative charge. It is therefore highly unlikely that ROS functions as a signal mediating germling communication. However, ROS could play an indirect roll in extracellular signaling. Recent evidence in B. cinerea indicates that ER localization and the N-terminus of NOX-1 are both essential for mediating plant pathogenicity, while plasma membrane localization and the C-terminus of NOX-1 are essential for mediating germling communication and fusion .
These data indicate that the NOX-1 complex regulates germling communication and fusion from the plasma membrane. However, it is unclear if ROS are directly important for signaling, or if the NOX-1 complex is primarily important for coordinating cytoskeletal dynamics with other signaling pathways.

CYTOSKELETON AND SECRETION
The actin cytoskeleton in fungi is important for mediating cell polarity, septation, exocytosis, endocytosis, organelle movement, and chemotropic growth (Huckaba et al., 2004;Moseley and Goode, 2006;Roca et al., 2010;Berepiki et al., 2010). In N. crassa germlings undergoing chemotropic interactions, actin cables and actin patches are dramatically enriched at CAT tips . Actin patches are associated with endocytic vesicles, while actin cables are associated with exocytosis and sites of growth and organelle movement. When two communicating germlings make contact and begin the process of cell fusion, the actin cables slowly disappear, while the patches remain throughout the fusion process (Berepiki et al., 2010;Roca et al., 2010). When germlings were treated with the microtubule-inhibitor Benomyl, no effect on communication or cell fusion was observed. In contrast, when germlings were treated with a sub-lethal dose of the actin-polymerization-inhibitor Lactrunculin B, germling communication and fusion was significantly reduced . Together, these data demonstrate that microtubules are dispensable for germling communication, while actin is critically important.
The Arp2/3 complex is a highly conserved actin-regulatory complex in all Eukaryotes (Robinson et al., 2001). This complex has been well characterized in vertebrates, S. cerevisiae, and Arabidopsis thaliana, but is largely uninvestigated in filamentous fungi. The Arp2/3 complex consists of seven subunits, the names of which are not consistent between organisms ( Table 2). These genes are generally considered to be essential in N. crassa, however, mutants of arp-6, arp-14, and arp-13 were shown to be defective at germling communication and fusion . Both ARP-2 and ARP-3 co-localize with actin in N. crassa and these proteins were developed as actin reporters (Delgado-Álvarez et al., 2010).
Proteins that are peripherally associated with the MAK-2 pathway in N. crassa are involved in regulating actin dynamics. The HYM-1 protein is the scaffold protein for COT-1 and POD-6, which are involved in regulating actin dynamics and polarized growth in both fungi and animals (Baas et al., 2004;Maerz et al., 2008). The HYM-1 protein is required for the sequential phosphorylation of the MAK-2 MAPK cascade, and HYM-1 and STE-20 physically interact in N. crassa (Dettmann et al., 2012). The ste-20 gene encodes a conserved upstream Pak1-like kinase that connects multiple signaling pathways that are involved in polarized growth and signal response. In S. cerevisiae, Bem1p interacts with both Ste20p and actin to regulate polarized growth (Leeuw et al., 1995). In mammalian cells, the ste-20-ortholog phosphorylates members of the Arp2/3complex to directly regulate actin dynamics (Vadlamudi et al., 2004). An adenylate cyclase (CR-1) and its associated adenylate cyclase capping protein (CAP-1) physically interact with the MAK-2 pathway in N. crassa via the conserved MAP-kinase regulator, STE-50 . The CAP-1/CR-1 complex binds actin monomers, is regulated by RAS-2, and has been shown to be involved in regulating actin dynamics and cell polarity in diverse fungi (Rocha et al., 2001;Bertling et al., 2007;Zhou et al., 2012). Lastly, a myosin mutant, myo-2 was shown to be defective at germling communication and fusion in N. crassa . These data combined demonstrate the central role of actin and its associated regulatory proteins during germling communication, which involves polarized growth.
Actin is important for mediating endocytosis, exocytosis, and vesicle tracking, which are all necessary for polarized growth and fusion. Amphiphysins regulate membrane curvature, vesicle trafficking, endocytosis, exocytosis, and actin organization (Sivadon et al., 1995;Ren G. et al., 2006;Toume and Tani, 2016). The amph-1 gene in N. crassa encodes an amphiphysinlike protein with a Bin-Amphiphysin-Rvs (BAR) domain; an amph-1 mutant in N. crassa is defective at germling communication, cell fusion, growth, and sexual development (Fu et al., 2011. Proteins containing a BAR domain are involved in protein-protein interactions and create a scaffold for regulating membrane curvature and connecting membrane dynamics with the cytoskeleton (Sivadon et al., 1995;Ren G. et al., 2006;Toume and Tani, 2016). The amph-1 homologs in S. cerevisiae are RVS161 and RVS167, which are required for polarization of the actin cytoskeleton, polarized growth, cell fusion, and endocytosis (Youn et al., 2010). Amphiphysins regulate exocytosis in neuronal synaptic cells in Drosophila melanogaster (Mathew et al., 2003) and have been well studied in animals where these proteins interact with actin-regulatory proteins (i.e., Arp2/3) and Rho GTPases (i.e., Cdc42) to mediate endo-and exocytosis of synaptic vesicles (Aspenström, 2014). A genome wide association study of N. crassa wild isolates identified several additional genes that are required for communication and are involved in the secretory pathway, endocytosis, or exocytosis (Palma-Guerrero et al., 2013). The GTPase activating protein Gyp5p physically interacts with amphyphysins to recruit them to sites of polarized growth and endocytosis or exocytosis in S. cerevisiae (Prigent et al., 2011). The gyp-5 mutant in N. crassa is a hyper-fusion mutant that fuses significantly more than wild-type cells (Palma-Guerrero et al., 2013). The Sec15 gene encodes an essential member of the exocyst complex in S. cerevisiae, and the sec-15 mutant in N. crassa is defective at germling communication (Palma-Guerrero et al., 2013). The sec-22 gene encodes a SNARE protein that is also involved in vesicle trafficking, particularly between the ER and Golgi (Liu and Barlowe, 2002). The sec-22 mutant in N. crassa has a reduced frequency of germling communication (Palma-Guerrero et al., 2013), and in S. macrospora, sec22 is necessary for the later stages of sexual development (Traeger and Nowrousian, 2015). In mammalian cells, a biochemical link has been established between the exocyst complex (including sec-15), GTPase activating proteins (i.e., gyp-5), Rac1, and BAR-domain containing proteins (i.e., amph-1), which all work together to regulate cell motility (Parrini et al., 2011). It is possible that a similar regulatory network exists in fungi for mediating polarized growth, including chemotropic interactions.

CALCIUM-DEPENDENT PROTEINS
Calcium is important for mediating polarized hyphal growth, and extracellular calcium is necessary for germling communication in N. crassa (Palma-Guerrero et al., 2013;Takeshita et al., 2017). Furthermore, four different calmodulin-dependent kinases (CaMK) in N. crassa are required for growth, sexual development, and stress response (Kumar and Tamuli, 2014). The germling fusion phenotype of these CaMK mutants is unknown. Calcium plays a critical role during the process of membrane fusion, and some of the genes involved in communication are predicted either to bind calcium or function in a calcium-dependent manner. The HAM-3 (Striatin) protein in the STRIPAK complex has a binding motif specific for the calcium-signaling regulator, Calmodulin, and Striatins in general are well known to integrate with calcium signaling (Bartoli et al., 1998;Simonin et al., 2010). The cch-1 gene encodes an integral membrane calcium channel, and the cch-1 mutant has a reduced frequency of germling communication (Chu, 2013).
The ham-10 gene is essential for germling communication and it encodes a putative C2-domain protein (Fu et al., 2011. The C2-domain exists across Eukarya and is defined by two 4-stranded β-sheets that bind phospholipids in a calciumdependent manner. C2-domain proteins have diverse functions that are involved in various aspects of membrane repair, vesicle trafficking, anchoring cytoskeletal elements to membranes, and lipid-based signal transduction (Zhang and Aravind, 2010).
Homologs of three N. crassa genes (cse-1, pik-1, and nfh-1) are important for calcium-dependent exocytosis in mammals and S. cerevisiae. In N. crassa, the cse-1, pik-1, and nfh-1 mutants all have a significantly reduced frequency of germling communication, and the CSE-1 protein co-localizes with Golgi markers (Palma-Guerrero et al., 2013). The CSE-1 protein is predicted to bind calcium ions, which directly impacts Calmodulin-dependent calcium signaling (Tamuli et al., 2011). In S. cerevisiae, the Pik1p physically interacts with Frq1p (CSE-1 homolog) and promotes its association with the Golgi, which is necessary for regulating calcium-dependent exocytosis. Additionally, in S. cerevisiae Pik1p is required for activation of the mak-2-ortholog, Fus3p (Cappell and Dohlman, 2011). The S. cerevisiae nfh-1-ortholog is a regulator that also binds Pik1p and shuttles it between the nucleus and cytoplasm (Demmel et al., 2008).

MEMBRANE FUSION
The result of cell-to-cell communication and chemotropic growth is cell fusion, and five genes have been identified that are specifically involved in membrane merger during cell fusion in N. crassa. An ergosterol mutant ( erg-2) was recently identified that fails to arrest growth after two germlings make contact (Weichert et al., 2016). Thus, the sterol composition of the cell membrane may play an important role in mediating the switch from communication to cell fusion. The deletion mutants, prm-1, lfd-1, lfd-2, fig-1, and NCU01697 engage in germling chemotropic growth, but are defective in the process of cell fusion and generally have increased levels of cell lysis, likely due to a compromised plasma membrane. The prm-1 gene encodes a conserved integral membrane protein that localizes to the plasma membrane and is necessary for mediating membrane fusion after cell wall breakdown (Fleißner et al., 2008a). The lfd-1 gene is unique to the filamentous Ascomycetes (Pezizomycotina) and encodes a protein with a single transmembrane domain that localizes to the plasma membrane. The phenotype of lfd-1 mutant is remarkably similar to the prm-1 mutant, and both mutants are abnormally sensitive to calcium-stress (Palma-Guerrero et al., 2014). The lfd-2, fig-1, and NCU01697 mutants also have a calcium-dependent cell-lysis phenotype (Palma-Guerrero et al., 2015). The NCU01697 locus encodes an uncharacterized protein that is predicted to have an acyltransferase domain, which could be important for phospholipid biosynthesis. The LFD-2 protein resides in the membrane of the Golgi and ER, while the FIG-1 protein is a plasma membrane lowaffinity calcium ion channel, and both proteins are important for regulating membrane homeostasis and membrane repair. Epistasis experiments demonstrated that LFD-2 and FIG-1 likely function in separate pathways (Palma-Guerrero et al., 2015).

NON-SELF RECOGNITION BEFORE AND AFTER CELL FUSION
Non-self recognition regulates chemotropic interactions between cells that have different allelic specificities at the doc loci (Heller et al., 2016). If cells have identical allelic specificity at the doc loci, robust chemotropic communication between cells ensues. However, if cells have different allelic specificities at the doc loci, chemotropic interactions are not reinforced and cell fusion frequencies are greatly reduced. Wild populations of N. crassa fall into five different communication groups that are correlated with five different doc haplotypes. Thus, the doc genes function as "greenbeard" genes (Heller et al., 2016), similar to genes in other organisms that regulate cooperation and cell identity. Greenbeard genes are involved in long-distance kind discrimination, leading to cooperation between non-genealogical relatives with similar greenbeard genes. Post-fusion non-self recognition mechanisms also exist in filamentous fungi whereby if two cells undergo communication and fusion, but carry different allelic specificities at heterokaryon incompatibility (het) loci, fusion cells undergo a type of programmed cell death (Saupe, 2000;Glass and Kaneko, 2003;Glass and Dementhon, 2006;Gonçalves et al., 2017). Some non-self recognition mechanisms mediated by genetic differences at het loci are only active in hyphae, because death is not immediately activated if two incompatible germlings undergo fusion (Ishikawa et al., 2012). In contrast, the incompatibility loci sec-9 and plp-1 mediate post-fusion programmed cell death between both germlings and in hyphae carrying incompatible sec-9/plp-1 alleles (Heller et al., 2018).

MISCELLANEOUS AND HYPOTHETICAL COMMUNICATION GENES
Several genes have been identified as necessary for germling communication and fusion, but it is unclear how these genes fit into the regulatory network that mediates germling communication. Many of these loci were named "hyphal anastomosis mutant" (ham) genes, which reflected their uncharacterized nature. Below is a brief description of what is known about these genes.

ham Genes
The ham-8 mutant is defective at communication, cell fusion, growth rate, and sexual development (Fu et al., 2011). The ham-8 gene encodes a predicted transmembrane domain protein with four transmembrane domains characteristic of the mammalian MARVEL proteins, which regulate tight-junctions between mammalian cells and vesicle transport (Sánchez-Pulido et al., 2002). The ham-8 mutant in N. crassa also has reduced levels of phosphorylated MAK-1, and to a lesser extent, MAK-2 , which indicates that ham-8 may function upstream of the CWI/MAK-1 pathway. The HAM-8 protein is phosphorylated in a MAK-2-dependent manner (Jonkers et al., 2014).
The ham-9 mutant is defective at communication, cell fusion, growth rate, and sexual development (Fu et al., 2011). The ham-9 gene encodes a protein with a predicted pleckstrin domain, which transiently associates with lipid bilayers. Pleckstrin domains are common in a variety of proteins that are generally associated with signal transduction (Lemmon, 2007). The ham-9 mutant has wild-type-like levels of phosphorylated MAK-1 and MAK-2, indicating that the ham-9 gene is not required for activation of these two MAPK pathways.
The ham-11 mutant has a unique phenotype that is unlike any other communication mutant . The ham-11 mutant is indistinguishable from wild-type in terms of hyphal cell fusion, growth rate, aerial hyphae production, and sexual development, but ham-11 germlings fail to normally undergo chemotropic interactions or cell fusion. However, ham-11 germlings undergo chemotropic interactions and fusion with otherwise isogenic wild-type germlings . Furthermore, SOFT and MAK-2 dynamic oscillations are restored in ham-11 germlings when communicating with wild-type germlings. The ham-11 gene is restricted to the Sordariaceae, has no characterized orthologs, and encodes a plasma membrane protein with either one or two predicted N-terminal transmembrane domains and a long cytoplasmic C-terminal tail (Fischer et al., 2019). HAM-11 is phosphorylated in a MAK-2-dependent manner (Jonkers et al., 2014). Recently, a genetic interaction between ham-11 and doc-1/doc-2 was identified that regulates reinforcement of MAK-2 signaling during chemotropic interactions (Fischer et al., 2019).
The ham-12 mutant has a reduced frequency of germling communication and fusion compared to wild-type cells, and ham-12 germlings communicate with wild-type germlings at an equivalently reduced frequency . The ham-12 gene is conserved among the filamentous Ascomycetes (Pezizomycotina), and is predicted to encode a protein of unknown function with one transmembrane domain.
The ham-13 mutant has reduced frequency of germling communication and fusion as compared to wild-type cells. The HAM-13 protein was discovered in an Affinity Purification Mass Spectrometry (AP-MS) experiment to identify proteins that physically interact with MAK-2 . The HAM-13 protein likely interacts indirectly with MAK-2 via other members of the MAK-2 pathway complex. HAM-13 is also required for full phosphorylation of MAK-2.
The HAM-14 protein was identified as a phosphorylation target of MAK-2 and it is necessary for germling fusion, but not hyphal fusion (Jonkers et al., 2014(Jonkers et al., , 2016. The HAM-14 protein physically interacts with MAK-2 and MEK-2 via co-immunoprecipitation, but was not observed to co-localize with fluorescently tagged MAK-2 (Jonkers et al., 2016). These observations led to the hypothesis that HAM-14 may facilitate the formation of MAK-2-pathway complexes, but HAM-14 is not involved in the oscillatory recruitment of MAK-2-pathway complexes during chemotropic interactions (Jonkers et al., 2016).

Proteins That Physically Interact With the MAK-2 Complex
The YPK-1, CKA, CKB-1, and CDC-25 proteins were all identified as having a weak physical interaction with the MAK-2 complex via AP-MS, and these proteins were also shown to be required for full phosphorylation of MAK-2 . The ypk-1, cka, ckb-1, and cdc-25 mutants all show a reduced frequency of communication and reduced levels of phosphorylated MAK-2 as compared to wild-type cells . ypk1 encodes a serine/threonine protein kinase in S. cerevisiae that phosphorylates regulators of sphingolipid-biosynthesis, which is an essential structural component of cellular membranes (Roelants et al., 2011). The cka and ckb-1 genes encode the two subunits of the caseinkinase-2 (CK2) heterodimer; CK2 is conserved across eukaryotes and mediates a broad diversity of biological processes. In N. crassa, CK2 is known to be an important component of the circadian clock, specifically in response to temperature (He et al., 2006;Mehra et al., 2009). The cdc-25 gene encodes a conserved guanine nucleotide exchange factor. In addition to CDC-25 affecting the phosphorylation of MAK-2, CDC-25 is phosphorylated in a MAK-2-dependent manner (Jonkers et al., 2014). In S. cerevisiae, Cdc25p is required for progression through the G1 phase of the mitotic cell cycle and indirectly regulates adenylate cyclase via Ras1p and Ras2p (Broek et al., 1987).

Transcriptional Targets of ADV-1
The genes ncu04645, ncu04487, ncu05836, and ncu05916 were identified as targets of a communication-activated transcription factor protein, ADV-1 (Fischer et al., 2018). The NCU04645 protein is orthologous to IDC4 in P. anserina (Gautier et al., 2018) and has a putative C-terminal AIM24-like domain, which is involved in coordinating the MICOS protein complex on the mitochondrial membrane. However, neither NCU04645 nor IDC4 contain a predicted mitochondrial targeting signal sequence. Consistent with this prediction is the observation that IDC4-mCherry localizes to the cytosol FIGURE 3 | All currently known proteins required for germling communication and fusion in N. crassa. Proteins are color-coded by function. Red proteins are components of the NOX complex that produces superoxide by reducing NADPH and oxidizing molecular oxygen. Purple proteins make up two different MAPK pathways; the Cell Wall Integrity/MAK-1 pathway, or the MAK-2 signal response pathway. Yellow proteins are involved in actin dynamics, vesicle trafficking, endocytosis/exocytosis, or secretion. Green proteins are involved in long-distance non-self recognition. Blue proteins compose the STRIPAK complex. Pink proteins are transcription factors. Brown proteins are involved in membrane fusion. Orange proteins are uncharacterized or have an unknown function. Gray proteins are associated and relevant, but are dispensable for communication or fusion. Proteins with a cyan-P are phosphorylated as part of a MAPK cascade, and proteins with a yellow-P with a green outline are phosphorylated in a MAK-2-dependent manner. (Gautier et al., 2018). The N. crassa NCU04645 mutant is completely defective at germling communication, while the idc4 mutant in P. anserina is defective at fruiting body development and suppresses the crippled growth phenotype (Gautier et al., 2018). The NCU04487, NCU05836, and NCU05916 mutants all showed a significantly reduced frequency of germling communication as compared to wildtype (Fischer et al., 2018). NCU04487 encodes a protein with a predicted C-terminal transmembrane domain, NCU05836 is orthologous to the putative ER mannosidase Mnl2p in S. cerevisiae (Nakatsukasa et al., 2001), and NCU05916 is a putative alpha-1,3-mannosyltransferase with homology to the virulence protein CMT1 in Cryptococcus neoformans (Sommer et al., 2003).

pkr-1
The pkr-1 gene was identified in a screen for mutants defective at germling communication and female sexual development (Fu et al., 2011). The pkr-1 gene is conserved among fungi and encodes a putative ER membrane protein. In S. cerevisiae, Pkr1p is an important assembly factor involved in the formation of the V-ATPase in the membrane of the ER (Davis-Kaplan et al., 2006).

idc-3
The idc-3 gene was identified in a screen for P. anserina mutants that suppressed a crippled growth phenotype (Lalucque et al., 2016). In addition to crippled growth suppression, the P. anserina idc-3 mutant is defective at cell fusion and female sexual development. The idc-3 homologous gene from the plant endophyte E. festucae was sufficient to complement the phenotype of the P. anserina idc-3 mutant (Lalucque et al., 2016), indicating that idc-3 function is likely to be broadly conserved among the filamentous Ascomycete species. The IDC-3 protein also has two conserved cysteine residues that may participate in NOX-related ROS signaling (Lalucque et al., 2016).

lao-1
The lao-1 gene encodes a L-ascorbate-oxidase-like laccase, and the lao-1 mutant was shown to be completely defective at germling fusion (Lichius, 2010). The LAO-1 protein has multiple predicted copper-oxidase domains and there are no biochemically characterized orthologs of lao-1.
arg-15, spr-7, and nik-2 The arg-15, spr-7, and nik-2 genes were identified in a genome wide association screen for genes associated with communication frequency across a genetic population of N. crassa wild isolate strains (Palma-Guerrero et al., 2013). The arg-15 mutant has a reduced frequency of germling communication compared to wild-type, while the spr-7 and nik-2 mutants have an increased frequency of germling communication compared to wild-type (Palma-Guerrero et al., 2013). The arg-15 gene encodes a putative acetylornithine-glutamate transacetylase that is orthologous to Dug2p in S. cerevisiae, which is involved in degrading glutathione and other peptides containing a gammaglu-X (Kaur et al., 2012). Glutathione is an antioxidant that reacts with ROS for detoxification in cells experiencing oxidative stress (Pócsi et al., 2004). The spr-7 gene encodes a putative secreted subtilisin-like serine protease, with no clear orthologs that have been characterized. Given the hyper-fusion phenotype of the spr-7 mutant it is tempting to speculate that the SPR-7 may negatively regulate an extracellular signaling peptide. However, subtilisin-like serine proteases encompass a broad diversity of enzymes (Siezen and Leunissen, 2008), and it is unclear what type of subtilisin-like protease is SPR-7. The nik-2 gene encodes a histidine kinase with no clear orthologs that have been characterized. However, in S. cerevisiae and other fungi, histidine kinases function upstream of the osmotic stress MAPK pathway (Zhao et al., 2007).

CONCLUSION AND SUMMARY
The mechanisms that mediate communication and fusion in N. crassa are complex and integrate components of several pathways, which are depicted in Figure 3. Several factors influence crosstalk between the MAK-2 pathway, the MAK-1 pathway, and the STRIPAK complex. The broadly conserved regulatory proteins RAC-1, BEM-1, CDC-24, and CDC-42 have been directly implicated as part of the NOX complex in the filamentous Ascomycetes, but it is likely that these proteins also regulate chemotropic growth by influencing actin dynamics and the MAK-2 pathway. Several genes represent various aspects of the canonical ER-Golgi secretion pathway, for which actin dynamics are critically important. Calcium plays an important role in mediating both communication and cell fusion. Lastly, there are several genes that remain uncharacterized, except for their necessity during cell communication in N. crassa. Future work will elucidate the functions of these uncharacterized genes and further illuminate the complex signaling network that mediates somatic cell-to-cell communication in filamentous fungi.

AUTHOR CONTRIBUTIONS
MF and NLG wrote and edited the review. MF produced the figures.

FUNDING
This work was funded by a grant from the National Science Foundation to NLG (MCB1412411).