Conidial swelling and germination.

Swelling, or isotropic growth of conidia, is the first step of biofilm initiation [18]. In Aspergillus spp., conidial swelling is triggered by suitable conditions that include, but are not limited to, nutrient availability, favorable temperatures, and sufficient oxygen [19–21]. Importantly, it has been shown that conidial density has a significant impact on the formation of a stable, drug resistant in vitro submerged biofilm [2]. While the exact mechanisms that trigger swelling in response to these environmental conditions are unknown in A. fumigatus, some of the sensors have been identified in A. nidulans. In A. nidulans, temperature is sensed by TcsB and FphA, where FphA modulates the expression of many downstream genes via the HOG signaling pathway [22]. RasA and cAMP mediated signaling initiate conidial swelling in response to carbon sources in A. nidulans [20,23]. However, RasA overactivation does not alter production of cAMP, suggesting that these signaling pathways act independently of each other [24]. Upstream of cAMP signaling, in A. nidulans surface G protein–coupled receptors (GPCRs), such as GprH, regulate initiation of germination in response to glucose [25].

In A. fumigatus, strains lacking the GPCR gprK had 80% germinated conidia compared to wild-type conidia, which had 35% germination at 8 hours of growth in the presence of carbon [26]. Without a carbon source, ΔgprK mutants had approximately 50% germination at 16 hours compared to wild-type conidia, which had less than 10% germination, indicating that GprK plays a role in regulating germination in response to carbon sensing. The GPCR subunits SfaD and GpgA also play a role in GPCR-mediated germination [27]. sfaD and gpgA null strain conidia begin germination between 12 and 14 hours of incubation, whereas wild-type and complemented strains germinate between 4 and 6 hours. In addition to glucose, inorganic phosphate, inorganic nitrogen, or magnesium sulfate are also required for initiation of germination in static Aspergillus niger cultures [28]. SakA, the A. fumigatus HOG1 homolog, is transcriptionally activated in response to nitrogen starvation, while the ΔsakA mutant strain has increased germination when grown in minimal media containing poor nitrogen sources (such as sodium nitrate) when compared to more preferred nitrogen sources (such as proline) [29].

Oxygen is also required for conidial activation and germination [21]. When incubated in anoxia, conidia stay dormant, and do not germinate until they are moved to 21% oxygen [30]. Additionally, a strain lacking srbA, a transcription factor essential for growth in low oxygen conditions, has hyphal growth in environments with 21% oxygen but not at 1% oxygen environments [31]. This result suggests that A. fumigatus conidia sense differences in oxygen tension to mediate germination. Currently, there are no known mechanisms of direct oxygen sensing in any Aspergillus species. It is possible that the activation of germination in response to oxygen is an indirect mechanism mediated by oxidative phosphorylation. Although this idea has not been directly tested, deletion of the kinases PkaC1 and PkaC2, which regulate the expression of genes involved in oxidative metabolism, results in reduced germination [32]. Thus, how oxygen impacts conidia activation remains an important area of A. fumigatus biofilm initiation research (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Summary of key stages of biofilm development.

Summary of each stage of biofilm development showing XZ view of a representative fluorescence confocal microscopy image of A. fumigatus CEA10 submerged biofilms. Here, we summarize features of each stage as well as key questions that remain to be answered. All images are 20× magnification. Scale bar is 100 um. ECM, extracellular matrix; eDNA, extracellular DNA; MIC, minimum inhibitory concentration.

https://doi.org/10.1371/journal.ppat.1009794.g002

Similarly, physical contact may also play a role in conidial germination via thigmotropism (response to touch). While there are no proteins annotated to contribute to thigmotropism in the Aspergillus genome, mechanisms utilized by other pathogenic fungi, such as the filamentous plant fungal pathogens, may be insightful to understand how A. fumigatus conidia coordinate germination and polarized growth in response to specific surfaces. For example, in plant pathogenic fungi such as Magnaporthe spp., the ends of hyphae form into an invasive structure termed the appressorium [33]. In Magnaporthe grisea, this process is coordinated via Pth11p-mediated GPCR signaling [34]. In the pth11 null mutant, germinated conidia have 10% to 15% appressorium formation when compared to wild type. It would be interesting to examine pth11 mutants biofilm initiation phenotypes in fungi such as A. fumigatus that do not generate appressoria.

Once activated, conidia break dormancy, and isotropic swelling initiates. Transcriptional analysis of swollen conidia in A. fumigatus revealed that mRNA from approximately half of all genes within the gene ontology (GO) category “RNA binding and translational regulation” were induced in conidia incubated in nutrient rich media for 30 minutes [35]. In contrast, only about 15% of genes involved in transcriptional regulation were observed in swollen conidia. Genes involved in amino acid metabolism, ribosome biogenesis, protein metabolism, and posttranslational modification are thus significantly enriched in swollen conidia. The coordination of increased translational activation and cell growth has been observed in both Saccharomyces cerevisiae and Candida albicans and is regulated via the HOG pathway [36,37]. These findings agree with previous observations that conidia contain prepackaged mRNA to expedite protein synthesis upon the initiation of swelling [38]. Finally, in swollen conidia, one-third of all genes in the stress response and oxidoreductase GO categories were down-regulated [35]. This highlights how conidia transition from a quiescent, stress resistant state in favor for a more metabolically and translationally active state necessary for active growth.

Germination of conidia begins around 4 to 8 hours, depending on the culture conditions and strain, and is the next step of A. fumigatus biofilm initiation. Germination is a 4-step process that begins with the aforementioned breaking of dormancy and conidial swelling followed by the initiation of polarized hyphal growth and emergence of a germ tube (Fig 1A) [39]. Active RasA is required for conidial swelling in response to favorable conditions but is inhibitory to germ tube development [40]. High levels of active RasA results in swollen conidia that fail to germinate until they are switched to conditions that favor intermediate RasA activity. The zinc finger transcription factor CrzA, a downstream mediator of calcineurin signaling, is also important at this stage of germination [41]. Strains lacking crzA had only 6% germinated conidia at 8 hours compared to a crzA complemented strain, which had 48% germinated conidia. The ΔcrzA strain had a defect in polarized growth and decreased glucan synthase in hyphae, suggesting that CrzA regulates germination and polarized growth through the biosynthesis of cell wall components. There are likely other signals that specifically mediate the polarized growth stage of germination, but conidial germination assays are often based only on the quantification of the percent of conidia, which have formed a germ tube—the last step in germination. This method lacks temporal precision and does not differentiate between signaling pathways necessary for polarized growth versus those required for the exit of dormancy. Therefore, there are still significant gaps in our knowledge of the pathways specifically required for the establishment of polarized growth and the formation of a germ tube. Cutting-edge flow cytometric analyses of conidia activation may be a high-throughput tool to identify key genes involved in the exit from dormancy.

As mentioned, conidia inoculum density has a significant impact on the formation of an A. fumigatus biofilm. The inoculum density requirement raises the potential for quorum sensing in A. fumigatus biofilm initiation. For bacterial biofilms, quorum sensing is a method of cell-to-cell communication that plays a major role in adherence, biofilm formation, colonization, and virulence [42–44]. In fungi, including species of Aspergillus, inoculum size has been observed to alter growth rate, morphology, and inhibitory concentrations of certain antifungals [39,45,46]. Thus, it is plausible that there are cell-to-cell communications utilized by A. fumigatus to mediate biofilm formation in response to cell density. While these mechanisms remain to be identified in A. fumigatus, quorum sensing mechanisms that mediate cell growth have been observed in other fungi. For example, C. albicans cultures grown in static conditions produce extracellular farnesol to mediate hyphal growth [47]. In Cryptococcus neoformans, an 11-amino acid quorum sensing–like protein, QSP1, facilitates growth in low density cultures in a tup1 null strain [48]. Given the importance of cell-to-cell communication in bacterial biofilm assembly and evidence for quorum sensing in other fungi, it would be beneficial to determine whether quorum sensing is utilized by A. fumigatus and how intercellular communication might play a role in biofilm formation. Here, one wonders about a potential role for the abundant secondary metabolites produced by A. fumigatus and other filamentous fungi [49].

Conidial adhesion.

Once conidia exit dormancy, they are capable of adhering to many surfaces ranging from plastic culture dishes to human epithelial and endothelial cells [50,51]. Adhesion of swollen and germinating conidia is likely reliant on surface adhesins as well as the production and secretion of a polysaccharide rich ECM. ECM-mediated adhesion of hyphae will be discussed in detail below for immature biofilms.

Initial adhesion of inert conidia to culture dishes is thought to be caused by weak, temporary electrostatic interactions [39]. However, once dormancy is broken, the hydrophobic outer layer of the conidia cell wall is broken down, exposing surface adhesins and cell wall α-glucans [52]. Approximately 80 genes in the A. fumigatus genome encode adhesin or adhesin-like proteins and about 30 of these genes have been classified as conidia specific [53]. Despite several putative adhesins being identified both bioinformatically and biochemically, the role of surface adhesins in biofilm formation, adherence, and general A. fumigatus physiology remains largely under explored. Biochemical studies have revealed that conidia can bind to host laminin and fibronectin in a protein mediated manner, but the responsible proteins remain to be determined [54]. The protein CalA was found to bind laminin, but its observed protein size does not match that of the original laminin binding study, suggesting that multiple proteins have the potential to bind laminin [55]. CalA also binds to integrin on host cells to induce endocytosis, therefore facilitating invasion of host tissue [56]. The role of CalA in invasive growth emphasizes the importance of identifying and characterizing genes encoding adhesin and adhesin-like proteins.

Hydrophobins on conidia are one group of proteins known to be involved in surface interactions. Hydrophobic rodlet proteins in the cell wall of conidia mediate binding to collagen, another host ECM protein abundant in the lung [57]. Additionally, biochemical studies have characterized a fucose binding lectin and a sialic acid binding lectin present on the surface of A. fumigatus that both have hemagglutination activity [58–60]. Alpha-glucans on the other hand have been shown to be relatively dispensable for vegetative growth on agar plates but are essential for adherence of swelling conidia to each other [61,62]. While adherence of conidia to other conidia is not a prerequisite for biofilm formation, it has potential implications for biofilms forming in environments where flow forces or stochastic disruptive forces, such as those caused by coughing, are present. Thus, while some biochemical and genetic groundwork has been laid, there is a need to further explore the players and mechanisms of fungal adherence in different contexts of A. fumigatus biofilm formation.

ECM-mediated adhesion.

The seminal discovery of A. fumigatus ECM production was a major step toward applying the biofilm concept to filamentous fungi [4,5,12]. ECM extracted from static grown A. fumigatus biofilms analyzed by nuclear magnetic resonance spectroscopy (NMR) was found to mostly consist of glycans and proteins but also contain small amounts of lipids, aromatics, and extracellular DNA (eDNA) [63]. While these experiments, conducted in Roswell Park Memorial Institute (RPMI) medium, give an example of ECM composition, they do not fully explore the complexity of the A. fumigatus ECM and how it changes under different environmental conditions. In C. albicans biofilms, matrix production is in part facilitated by endosomal sorting complexes required for transport (ESCRT)-mediated extracellular vesicles (EVs), and the individual deletion of 7 different ESCRT genes resulted in decreased ECM production [64]. Since A. fumigatus mycelia produce EVs future experiments assessing A. fumigatus, EV production within biofilms and their contribution to matrix production may be worthwhile [65].

In A. fumigatus mutants lacking uge3, a UDP-glucose epimerase required for galactosaminogalactan (GAG) synthesis, a major component of the ECM, germlings were unable to adhere to culture dishes and exhibited less than 10% adherence to epithelial cells when compared to wild-type germlings [66]. In an immunosuppressed mouse model of invasive aspergillosis, the Δuge3 mutant was found to have attenuated virulence with increased murine survival at 8 days post-fungal challenge compared to wild type [66]. GAG also functions to conceal surface β[1,3]-glucan, which may contribute to the decreased virulence observed for strains lacking uge3. While the exact kinetics of ECM secretion are not fully explored under different environmental and substrate conditions, ECM secretion likely begins shortly after swelling in order to facilitate surface adhesion. The signals and underlying mechanisms of ECM regulation at these early stages of filamentous fungal biofilm development remain to be fully defined.

Expression of the GAG biosynthetic gene cluster containing uge3 is regulated by the 3 transcription factors, MedA, StuA, and SomA [66–68]. Loss of MedA or StuA transcriptional regulators results in reduced expression of uge3, little to no detectable GAG, and reduced adhesion [66,67]. SomA regulates GAG synthesis through the regulation of MedA and StuA and by direct binding to the promoter regions of the GAG biosynthesis genes agd3 and ega3 [68,69]. In the medA null mutant, the expression of 8 putative adhesins was compared to wild type and one, Afu3g00880, was found to be regulated by MedA [67]. To date, this gene has yet to be validated as an adhesin and further highlights a key gap in knowledge regarding protein-based adhesins important for biofilm formation and structure. The interplay between protein-based adhesins and the ECM-mediated adherence remains to be fully defined. It is unclear how ECM composition may change under different conditions and under what conditions and in what contexts protein-based adhesins may play an important role in surface adherence (Fig 2). A complete catalog of transcription factors required for adhesion to diverse surfaces under a variety of conditions is now possible with the release of an A. fumigatus whole genome transcription factor deletion collection [70]. Broadening our understanding of the key transcriptional regulatory circuits involved in adhesion is expected to yield new insights into fungal biofilm formation.

To this end, SakA and MpkA are 2 MAP kinases involved in A. fumigatus surface adhesion. sakA and mpkA null mutants had less adhesion when grown in liquid minimal media in fibronectin treated plates as quantified by a crystal violet (CV) assay [71]. Germlings of an MpkA null mutant have 50% less GAG, while SakA null germlings have 50% more GAG when compared to WT strains when quantified using the Soy Bean Agglutinin (SBA) lectin staining [71]. SBA preferentially binds oligosaccharides with terminal alpha or beta-linked N-acetylgalactosamine residues and galactose residues to a lesser extent. The MpkA-regulating phosphatase SitA also plays a key role in regulating cell wall composition and surface adhesion [72]. Upstream mediators of SitA as well as transcription factors downstream of SakA and MpkA that mediate this adherence phenotype are currently unknown. However, the SakA regulated transcription factor AtfA is critical for conidia stress resistance and germination [73]. In C. albicans, Mkc1p, the MpkA homolog, is phosphorylated in a contact-dependent manner, suggesting that this mechanism may also occur in A. fumigatus but has yet to be fully explored [74]. The CV assays with ΔmpkA and ΔsakA were performed using mature biofilms so future experiments assaying the adhesion of ΔmpkA and ΔsakA germlings are needed to better understand the role of these kinases in adhesion during germination and immature biofilm development. Conceptually, investigating the signals responsible for the activation of these MAP kinases in the context of surface adhesion will give us greater insight into mechanisms of A. fumigatus biofilm formation.

Antifungal resistance of immature biofilms.

From 8 to 12 hours, the voriconazole minimum inhibitory concentration (MIC₉₀; the concentration of voriconazole needed to reduce cell viability by 90%) increases from 0.5 mg/L to 8 mg/L [15,75]. How does this remarkable increase in MIC for antifungal drugs arise at this stage of fungal biofilm development? These data also raise an important consideration for in vivo drug treatments; even an immature filamentous fungal biofilm displays antifungal drug resistance through an unknown mechanism(s). A potential explanation for decreased drug susceptibility throughout biofilm formation might be the increased expression of multidrug resistance (MDR) efflux pumps. MDR efflux pumps are ATP-dependent transporters, which belong to the larger ATP-binding cassette (ABC) transporter family. ABC transporters as well as major facilitator superfamily (MFS) transporters, the second major class of transporters found in A. fumigatus, actively transport a variety of molecules, such as the drug voriconazole, to remove them from the cell [76]. The A. fumigatus genome contains 49 predicted ABC transporters as well as 278 predicted MFS proteins, of which 35 are putative multidrug permeases [77,78]. Transcripts for the ABC-type transporter encoding gene mdr4 are significantly increased at 12 hours compared to 8 hours when grown in vitro in RPMI media [79]. This increase in mdr4 transcription in 12-hour biofilms compared to 8-hour biofilms also correlated with a significant increase in drug resistance. Additionally, random mutagenesis of A. fumigatus conidia identified mutants that were highly resistant to the antifungal itraconazole, which had increased basal expression mdr3 and mdr4 or increased expression in the presence of itraconazole in shaking cultures when compared to wild type [80]. Future experiments assessing the azole susceptibility of an mdr4 mutant strain are crucial to better understanding the relationship between mdr4 expression and antifungal resistance within biofilms. Induction of transporters may indicate a complex interplay between individual hyphal cell metabolism and microenvironmental niches that remain to be explored. As efflux pumps are important for removing a variety of compounds, it is possible that these pumps are up-regulated in order to remove toxic molecules that might become concentrated in biofilms and serendipitously aid in antifungal resistance as a result. For example, biofilms grown on an agar surface down-regulate glycolysis in favor of alternate metabolic pathways that may lead to the buildup of potentially toxic metabolic byproducts [79]. Metabolite profiling of the extracellular biofilm environment at different stages of development may yield new insights into fungal biofilm metabolism and subsequent stress resistance phenotypes.

Further studies defining regulators of drug efflux pumps in biofilms will provide much needed insight into biofilm mediated antifungal resistance at this early stage of biofilm development (Fig 2). Transcriptional or posttranslational MDR pump regulation remains under-explored in A. fumigatus. The putative drug efflux pump Cdr1B has been shown to be regulated by the transcription factor AtrR [81]. In a ΔatrR strain, expression of cdr1B was reduced to less than 10% of wild-type expression. However, the role of AtrR and Cdr1B in A. fumigatus biofilms remains to be defined. Direct examination and genetic analyses of additional specific transporters are needed, but these studies may be confounded by the substantial redundancy of pump function. An overexpression-based approach may reveal new key efflux pump-based mediators of biofilm antifungal drug sensitivity, but the power of transcription factor and protein kinase mutants to identify candidate effector genes should not be overlooked.

Physical barriers, such as the aforementioned ECM or cell wall, are a likely contributor to fungal biofilm antifungal drug resistance and tolerance, as observed with cell wall glucan-dependent drug resistance of C. albicans biofilms [82]. Cell wall changes are dynamic, robust, and occur rapidly as conidia swell, germinate, and transition to hyphae. Cell wall differences between strains has been linked to virulence, showing how the cell wall plays an important role in infection and disease outcome [83]. Transcriptional analysis of colony biofilm (on agar surface) cultures suggest that 16-hour biofilms have significantly more protein glycosylation than hyphae in planktonic (free floating in liquid) cultures [7]. This increased glycosylation results in “sticky” proteins, which may glue hyphae together within the biofilm and could conceivably make biofilms less permeable to drugs.

Regarding the ECM, studies have intriguingly revealed that A. fumigatus Δuge3 biofilms, lacking the ECM component GAG, do not have decreased antifungal resistance at 18 hours when compared to wild-type biofilms [14]. However, antifungal treatment of 9-hour A. fumigatus biofilms with the addition of hydrolases capable of degrading GAG showed significantly decreased MIC₅₀ for all 3 classes of antifungals [84]. This highlights the importance of GAG during biofilm initiation while also illustrating how other factors besides GAG contribute to antifungal resistance at later stages of biofilm development. How other ECM components contribute to antifungal resistance will be covered later in this review when we discuss antifungal resistance within mature biofilms.

A comparison of shaking planktonic and static colony biofilm cultures found increased expression of ergosterol biosynthesis genes in colony biofilms. Deletion of the gene cyp51A, which encodes for a 14-α sterol demethylase, has been shown to increase sensitivity to azoles [85]. This observation raises an intriguing question regarding the mechanism for enhanced ergosterol gene expression in biofilms and whether this transcriptional increase corresponds to alterations in sterol levels within the fungal cells at this point in biofilm development. Sterol biosynthesis is directly impacted by iron and oxygen levels and reductions in these key molecules stimulates ergosterol synthesis gene expression through SrbA-dependent mechanisms [86]. In submerged biofilm cultures, oxygen levels are depleted over the course of biofilm development and correspond with an increase in transcription of the SrbA dependent gene erg25A [14]. Thus, as biofilm cell density increases, oxygen levels within the immature biofilm become depleted leading to both an increase in expression of ergosterol biosynthetic genes and a decrease in metabolism. These 2 factors may contribute to the localized increases in antifungal resistance within the developing biofilm. This hypothesis remains to be tested.

Polarized growth and morphology.

Lastly, a mechanism important for the morphology of immature biofilms is cell polarity, which serves to guide hyphae as they grow to form an interconnected mycelium. Filamentous fungi grow in a polarized manner, directing growth at the tip via formation of a vesicle rich structure called the Spitzenkörper [87]. Consequently, regulators of fungal cell polarity machinery such as endocytosis and cytoskeletal elements are likely important for biofilm form and function but remain largely unstudied in the context of submerged biofilms. Seminal discoveries of genes involved in these processes in model organisms such as N. crassa and A. nidulans can serve as a foundation for studies in A. fumigatus. Here, the question of developmental programs comes into play. Are these vertical growing hyphae under static liquid submerged culture conditions a result of asexual development initiation? Or do they remain in a vegetative state consuming and probing for new nutrient sources such as oxygen (Fig 2)?

In static liquid cultures with adequate nutrients and oxygen, A. fumigatus hyphae consistently form a basal mycelium followed by vertical hyphal growth (Fig 1A) [3,14]. This vertical polar growth toward the surface is seen in immature biofilms by 12 hours [14]. Intriguingly, SrbA, which is required for growth in low oxygen conditions in A. fumigatus, has been shown to be involved in cell polarity and microtubule dynamics in Aspergillus spp. [31,88]. Consequently, it is interesting to note that oxygen availability clearly alters biofilm architecture. In static liquid cultures, the morphology of biofilms changes with oxygen availability [14]. Hyphae within A. fumigatus submerged biofilms grown in 21% oxygen are mostly vertical, while hyphae grown under a 5% oxygen atmosphere or less intriguingly grow more laterally or horizontally. A. fumigatus strains that are more fit in low oxygen conditions also exhibit altered biofilm morphology at higher oxygen concentrations with an increase in the percentage of filaments that grow horizontally [83]. These data suggest connections between cell polarity, biofilm morphology, and oxygen availability and pose important questions. For example, while oxygen availability alters biofilm morphology, the signaling pathways that mediate this process are unknown. Furthermore, the connection between oxygen availability and cell polarity remains to be defined. These questions highlight the importance of understanding and better characterizing the immature, transitional stage of biofilm development.

Complexity and antifungal resistance.

The buildup of ECM material peaks at 36 hours in static cultures of A. niger grown in potato dextrose broth on polystyrene dishes, encasing the fungus in a protective layer that glues hyphae together [89]. As mentioned previously, in C. albicans, ECM serves as a physical barrier to antifungals through glucan binding and sequestration of antifungals [82]. It has also been shown that the ECM of C. albicans also acts as a physical barrier to host leukocytes and is capable of inhibiting extracellular trap formation by neutrophils (NETs) [90]. In A. nidulans, increased production of GalNAc-rich GAG through the overexpression of ugeB or heterologous expression of A. fumigatus uge3, increased resistance to neutrophil NET formation in vitro [91]. How mature A. fumigatus biofilms impact immune effector cell function remains to be fully defined and is an important area of investigation moving forward.

In addition to polysaccharides, the presence of eDNA in the ECM has been reported at 24 to 48 hours in A. fumigatus biofilms [92]. eDNA not only supports fungal biofilm architecture but also aids in antifungal resistance [92]. In static cultures, the concentration of amphotericin B and caspofungin needed to inhibit fungal growth decreased as the biofilms were treated with increasing concentrations of DNAse, suggesting a functional role for eDNA release in biofilms. Previous studies in both bacteria and Aspergillus spp. suggests that eDNA is released via autolysis, a process that likely occurs within growing biofilms as they adapt to their environment [93,94]. Still, the exact mechanism of eDNA secretion into the ECM is unknown, and the extent to which this occurs under conditions relevant in the infection environment is unclear (Fig 2).

Macroscopic colony morphology and virulence.

Exposure to low oxygen also results in changes in macroscopic colony biofilm morphology, such as increased colony furrows and a white perimeter of vegetative growth, on solid media [83]. Some strains have a hypoxia-locked morphology, termed “H-MORPH,” where colony biofilm morphology resembles low oxygen culture conditions even in oxygen-replete conditions. H-MORPH has been associated with increased mortality in a murine model of invasive aspergillosis [83]. A single cryptic gene, bafA, is sufficient to induce H-MORPH in both the A. fumigatus reference strains AF293 and CEA10 and the A. niger reference strain A1144 [95]. Overexpression of bafA in these strains alters both macroscopic and microscopic biofilm morphology and decreases adhesion of submerged biofilms. Interestingly, bafA has several putative orthologs, including a pseudogene, which display copy number variation across strains of A. fumigatus. Moreover, bafA was discovered in the sub-telomeric region of chromosome 5 of strain AF293, a region of novel and rapidly evolving genes [96]. Sub-telomeric region genes’ mRNA levels were significantly induced in the agar colony biofilm model [79]. However, the molecular function of bafA remains unknown and it is unclear how its expression impacts biofilm morphology. Yet, the example of bafA also highlights that many uncharacterized fungal specific genes found in sub-telomeric regions have important and uncharacterized roles in fungal biology.

Heterogeneity within biofilms.

Evidence of metabolic heterogeneity within a mature biofilm has been observed, but it is not yet understood how and to what extent these metabolic changes contribute to biofilm form and function. For example, fungal respiration reduces available oxygen within the biofilm, resulting in zones of oxygen depletion by 18 hours of growth in the submerged biofilm model (Fig 3) [14]. Hyphae at the base and toward the interior of the biofilm are thus likely in an altered metabolic state driven in part by reduced oxygen availability compared to peripheral cells. The interplay between localized oxygen depletion and alterations in metabolism throughout the biofilm remains to be defined experimentally. Importantly, hyphae in regions of oxygen limitation are strikingly resistant to antifungal drugs. Azoles, in particular, such as voriconazole, require metabolic activity to effectively target ergosterol synthesis (Fig 3). Curiously, planktonic grown hyphae or conidia are not more resistant to antifungals in an atmosphere with reduced oxygen availability [14]. This raises intriguing questions about the hypoxia response in planktonic versus submerged biofilm conditions in fungi that remain to be defined. Presumably, carbon and nitrogen sources available to the fungus in the interior of the biofilm are also altered compared to exterior and planktonic cells, further impacting the metabolism of these oxygen limited regions. Yet, studies addressing these questions are currently in their infancy with filamentous fungal biofilms. In addition, the low oxygen environment inside the biofilm might inhibit the generation of reactive oxygen species (ROS) by neutrophils and confer resistance to leukocyte mediated killing but remains to be experimentally tested for filamentous fungal biofilms.

As A. fumigatus biofilms mature and oxygen gradients are established, the complexity of these biofilms increases. As we better understand biofilm growth, the need for spatial resolution becomes more apparent. For example, the decreased expression of efflux pumps at 24 hours may not be a global change within biofilms, but rather a result of altered metabolic activity in different regions of the biofilm. Oxygen gradients are just one example of important macro and micronutrient gradients that may become established over time in biofilms. Additional factors such as other gases, nutrients, secondary metabolites, host-generated molecules, and host cell infiltration may also occur regionally within biofilms to impact their function and maturation (Fig 3). The use of reporter strains is one tool that can be used to investigate spatial heterogeneity in these filamentous fungal biofilms. The development and application of other tools to further spatially dissect fungal biofilms and focus on one region at a time will greatly increase our understanding of biofilm form and function.