Compost is a dynamic and complex habitat wherein microorganisms play fundamental roles. Bacterial communities in these systems have been extensively studied (Tiquia, 2005; Takaku et al., 2006; Thummes et al., 2007; Partanen et al., 2010; De Gannes et al., 2012, 2013). However, fungi are another major group of compost microbes that contribute essential decomposition processes, particularly the initial degradation of plant polymers (Sharma, 1989; Floudas et al., 2012). Relatively little is known, however, about the composition, diversity, and succession patterns of fungal communities in compost.

While compost is valued as a soil amendment, potential health hazards have also been recognized, as some saprotrophic fungi that colonize these materials are also opportunistic human pathogens. The primary risk is exposure to bio-aerosols containing compost-derived fungi, which in humans can cause mycoses of the lungs, skin, ears, sinuses, or bone marrow (Bunger et al., 2000). Most attention has focussed on Aspergillus fumigatus, the most commonly identified fungal pathogen in composts (Ryckeboer et al., 2003; Anastasi et al., 2005; Dehghani et al., 2012). But, other pathogens have also been reported, including Candida tropicalis, Candida krusei (Bonito et al., 2010), Scytalidium lignicola, and Alternaria alternata (Anastasi et al., 2005). The spectrum of saprotrophic fungi that may also be opportunistic human pathogens is much greater than the foregoing list, but their occurrence in composts has not been extensively explored.

Molecular analyses have the potential to provide new insights into the composition of fungal communities in compost. However, to date, such analyses have indicated that fungal diversity in these systems is low. For example, Tiquia (2005) applied 18S rDNA terminal-restriction fragment length analysis for profiling of manure composts, and identified a total of 18 unique fragments, presumably indicative of 18 different phylotypes. Also, 10 genera were identified from libraries of 18S rRNA genes that were cloned from a model compost of dog food-wood chips (Hansgate et al., 2005). Hultman et al. (2010) created libraries of internal transcribed spacer (ITS) region sequences from municiple waste compost, and identified 166 OTUs (≥99% similarity) from ca. 2900 clones. It's unclear if the relatively low diversity indicated by these studies accurately reflects the communities, or if it results from a limited depth of analysis. Similarly, the dynamics of fungal communities as a function of compost phase are as yet ill-defined. In principle, since most fungi are mesophiles, a reduction in their diversity might be expected as compost heat to thermophilic temperatures. But, to date, results from molecular analyses do not consistently support that concept (Tiquia, 2005; Bonito et al., 2010) and it's unclear as to the extent to which the divergent finding reflects variation in compost properties (e.g., level and duration of heating) vs. limitations of the molecular analyses applied.

Recently, knowledge of fungal communities in soils has been expanded with the advent of the use of high-throughput sequencing technologies such as 454-pyrosequencing. Compared to traditional approaches based on cloning and Sanger sequencing in terms of cost, time, and number of sequences obtained, 454-pyrosequencing has proven to be a powerful alternative for the identification of a greater depth of taxa (Tedersoo et al., 2010). For example, Lentendu et al. (2011) demonstrated an unprecedented depth of fungal diversity in forest soil, and illustrated the strong spatial heterogeneity of fungal communities in these soils. Buee et al. (2009) were able to identify rare individuals (OTUs), which could be correlated with ecological parameters. Pyrosequencing of the nuclear ribosomal ITS region revealed fungal communities of a cork oak forest soil were distinct from those described in other soils (Orgiazzi et al., 2012). Pyrosequencing also proved useful in the examination of fungal communities related to plant disease in agriculture soils (Xu et al., 2012).

While high through-put sequencing has greatly advanced our understanding of fungal communities in soil, its application to assess these organisms in compost has not been reported. Pyrosequencing analysis could provide new insights into the composition of fungal communities in compost, and how their evolution compares to the chemical evolution of the composts. Furthermore, deep sequencing could expand the diversity of human pathogens present in compost. Thus, the goal of our study was to obtain in-depth sequence information to assess diversity and dynamics of fungal communities in three composts of lignocellulosic agricultural wastes. Fungal composition was assessed via high throughput 454-pyrosequencing of ITS1 libraries derived from each of the three major phases of the compost process. We hypothesized that community structure of each system is unique throughout the composting process, and that these systems do not converge on a common composition in the finished product. We also hypothesized that deep sequencing of these communities would reveal a greater diversity of potential pathogens than has been indicated in the literature to date.

The rank-abundance data showed that, in the coffee system, diversity decreased between the mesophilic and thermophilic phases, and then increased to its highest levels in mature phase. The rice compost showed a contrasting pattern with diversity increasing to its maximum level in the thermophilic phase, and then declining in the mature phase. The bagasse system, which lacked a thermophilic phase, showed the greatest expansion in fungal diversity from the mesophilic to mature stage, and thus achieved the highest level of diversity overall. These findings supported the concept of fungal community successions occurring across compost phases (Ryckeboer et al., 2003). But, the data did not consistently support prior findings that the occurrence or extent of a thermophilic phase to be the most important factor affecting the levels of diversity in the finished product (mature phase).

Previous studies have shown that fungal diversity decreased at thermophilic temperatures (Ryckeboer et al., 2003; Tiquia, 2005; Bonito et al., 2010; Hultman et al., 2010). But for the composts which attained thermophilic temperatures in the present study (rice and coffee), only the coffee was consistent with this finding. A possible explanation for the divergent behavior of fungal communities in these composts could have been the level of heat generated and duration of the thermophilic phase. In this regard, the coffee compost reached and maintained 64°C for 5 d, while the rice system heated to 57°C for 3–4 d. Notably the bagasse system, which never attained a thermophilic temperature exhibited the greatest expansion in fungal diversity. This could reflect the fact that most terrestrial fungi are mesophiles (Dix and Webster, 1995), and maintenance of mesophilic conditions in the bagasse could have allowed their undisrupted growth, and consequently development of communities more diverse than those of the other composts. Previously we reported analysis of prokaryotic diversity in the same bagasse, coffee, and rice composts (De Gannes et al., 2013), and comparison of that data with the present study shows that prokaryotic diversity was far greater than that of fungi. This finding was congruent with those of other investigators (Ryckeboer et al., 2003; Tiquia, 2005).

The prevalence of Fusarium might be anticipated given its ubiquity both in soil and association with plants. But, the relative abundance of Fusarium may to some extent have been an indicator of compost maturity, as its sequences tended to be greatest in meso- and thermo-philic phases (Table S2). Thus, mature phase compost was characterized by a reduction in Fusarium. All Fusarium sequences were matched to Fusarium oxysporum (Table S2), strains of which exhibit a variety of activities that range from lignin degradation (Kadarmoidheen et al., 2012) to causing plant (Beckman, 2000), animal (Odds et al., 1998), and human diseases (Boutati and Anaissie, 1997). Scytalidium was also prevalent in the composts, and the two species identified, Scytalidium thermophilum and Scytalidium lignicola, were divergently distributed (Table S2). The former of these was predominantly associated with the coffee and rice composts, and in the coffee system showed a striking increase in abundance in the mature phase (Table S2). Prior investigators have noted that this thermophilic fungus often exhibits profuse growth in mature phase compost from spores produced during thermophilic phase growth (Wiegant, 1992). In contrast, the majority of Scytalidium lignicola sequences were identified in the mature phase of the bagasse compost (Table S2), which lacked a thermophilic phase. The two species of Chaetomium identified, Chaetomium funicola and Chaetomium globosum, showed a similar bipartite distribution with the thermophilic organism, Chaetomium funicola, occurring primarily in rice and coffee composts (especially the coffee mature phase) while abundance of the mesophilic Chaetomium globosum was heavily weighted toward the mature phase bagasse (Table S2).

Prior studies have provided insights about the presence of opportunistic pathogens associated with composting systems, viz. Candida tropicalis and Candida krusei (Bonito et al., 2010), Aspergillus fumigatus (Dehghani et al., 2012), Scytalidium lignicola, and Alternaria alternata (Anastasi et al., 2005). However, the deep sequencing presented here has revealed the broadest spectrum of potentially pathogenic fungal species yet reported for a composting study. The group of eight species detected in all composts includes all but one of the potential pathogens identified by prior investigators, and thus could be representative of the main types of potential pathogens that might be encountered in composts. If so, bio-monitoring efforts could be expanded beyond Aspergillus fumigatus to include these organisms as well.

Fungal pathogens are an increasing threat to human health, and those listed above are known to cause several types of mycoses, with immune-compromised individuals being particularly susceptible. Allergic broncho-pulmonary aspergillosis, caused primarily by Aspergillus fumigatus (Latge, 1999; Chakrabarti et al., 2002; Hope et al., 2005; Agarwal et al., 2009) is a wide-spread concern (Howard et al., 2010; Denning et al., 2011). Skin diseases are also common and include phaeohyphomycosis caused by Scytalidium lignicola, Candida tropicalis (Wingard et al., 1979), Scedosporium prolificans (Revankar et al., 2002), and Bipolaris spicifera (McGinnis et al., 1992) as well as chromoblastomycosis, which have involved Fonsecaea pedrosoi and Chaetomium funicola (Pipenbring et al., 2007). The latter affliction is common in tropical and subtropical regions (Silva et al., 2002).

The clinical spectrum for most of these fungal pathogens is diverse. For example, Bipolaris spicifera is a source of nosocomial infections and has also caused a type of meningitis (Ogden et al., 1992) as well as cutaneous and pulmonary infections (McGinnis et al., 1992). Scedosporium prolificans have been shown to cause ocular infections (keratouveitis; Arthur et al., 2001) as well as cases of pneumonia (Berenguer et al., 1997); treatment of Scedosporium prolificans infections are difficult because of its resistance to many commonly used anti-fungal agents. Scedosporium apiospermum causes a range of infections such as bullous and necrotic purpura (Miyamoto et al., 1998), osteomyelitis (German et al., 2004) while infection by Scedosporium aurantiacum can result in malignant otitis externa, osteomyelitis, invasive sinusitis, keratitis, and pneumonia (Alastruey-Izquierdo et al., 2007). Concerns with Fusarium oxysporium include production of a range of mycotoxins (trichothecenes, zearalenon, and fumonisins) in addition to infections ranging from skin lesions to allergies (Boutati and Anaissie, 1997). Alternaria alternata also produces mycotoxins that caused allergies and asthma in children (Halonen et al., 1997). Infections caused by Cladophialophora arxii are chronic and appear as spreading mycoses of the skin, subcutaneous tissues, and nail (Barde and Singh, 1984).

According to the American Biological Safety Association, fourteen of the fungal pathogens identified in this study are categorized as Biosafety Level 2 (Alternaria alternate, Aspergillus fumigatus, Candida tropicalis, Chaetomium funicola, Cladosporium cladosporioides, Fusarium oxysporum, Scytalidium lignicola, Bipolaris spicifera, Fonsecaea pedrosoi, Metarhizium anisopliae, Retroconis fusiformis, Scedosporium apiospermum, Scedosporium aurantiacum, Scedosporium prolificans) and one (Cladophialophora arxii) belonged to Biosafety Level 3. Hence, it is recommended that personal protective equipment be worn when handling these composts. These include face protection (goggles, face masks), disposable gloves, and protective coats. It should be noted that, for all of the foregoing species, there can exist substantial intra-specific (strain level) variation in virulence (Ben-Ami et al., 2010). Thus, molecular detection such as that of the present study can provide an indicator of the potential health threat, but additional bioassays of isolates are needed to assess virulence.

Results of the study supported both hypotheses. First, the community structure of each system was unique throughout the composting process, and did not converge on a common composition in the finished product. Thus, there was no single fungal community structure that could be considered typical. Second, the diversity of potential pathogens was greater than previously reported. Molecular identification of these indicated that safety precautions would be warranted for both compost producers and users. Overall, the study demonstrated that deep sequencing can effectively elucidate fungal community diversity in compost and that such information can have important implications for compost use and human health.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.