Introduction

Genome sequencing and phylogenetic inferences posit that the Microsporidian lineage is closely related to a recently erected, and highly diverse, phylum that sits at the base of the fungal tree, the Cryptomycota (James et al., 2013). Members of this lineage have lost or simplified several typical eukaryotic features through a spectacular course of evolution. This adaptation has fuelled the diversification of a widespread lineage of obligate and often opportunistic intracellular parasites, which now comprise over 1500 described microsporidian species distributed amongst more than 187 different genera (Vavra and Lukes, 2013). Species from this group infect a range of eukaryotes from protists to most animal lineages (Lee et al., 2008; Corradi and Selman, 2013; Vavra and Lukes, 2013) with increasingly evident medical, economic and environmental importance (Vavra and Lukes, 2013). Despite parasitising varied hosts, all microsporidian species present an original invasion strategy based on a highly specialised organelle: the polar tube. This represents one of the most sophisticated infection mechanisms among eukaryotes (Delbac and Polonais, 2008). After the germination process, the extruded polar tube penetrates the host’s cell membrane to release an infective sporoplasm within the target cell (Peyretaillade et al., 2011).

The key to the success of Microsporidia lies in the flexibility of their transmission strategies and life cycles. Microsporidia exploit different strategies to maintain their presence within host populations, which harness the host life cycles and make use of horizontal transmission (transfers between host individuals, including different species) and vertical inheritance (transmission of the parasite from parent to offspring(s)) (Dunn and Smith, 2001). In the latter process, some Microsporidia take advantage of the host reproduction system by distorting the sex ratio towards the transmitting sex (female), converting putative males into phenotypic functional females via a feminisation mechanism (Ironside et al., 2003). They can also specifically eliminate males (male-killing process) during the embryonic stage, resulting in the production of all-female offspring (Dunn and Smith, 2001). A purely vertical transmission mode is believed to be accompanied by asexuality in some taxa (Dunn and Smith, 2001). However, some data suggest that sex may be present in numerous if not all microsporidian species (Ironside, 2013; Krebes et al., 2014). The identification of gametes and the formation of diplokarya and mechanisms for haplosis via either meiosis or nuclear dissociation have been reported (Becnel et al., 2005), as has the presence of diploid polymorphisms (Cuomo et al., 2012; Selman et al., 2013). Furthermore, a sex-related locus and some meiosis genes suggest that Microsporidia have a genetically controlled sexual cycle (Katinka et al., 2001; Lee et al., 2008). Thus, some members of this lineage are highly adapted to vertical transmission, whereas others may supplement this mode of transmission with a sexual phase on occasion (Krebes et al., 2014).

Pathogens represent a large proportion of life on earth, but their ecological impact on ecosystems is not always appreciated (Marcogliese and Cone, 1997). Their role in population regulation is essential to maintaining ecosystem balance (Selakovic et al., 2014). High-density species become the prime target for pathogens, leading to their population decline and maintenance of high biodiversity in the environment. Conversely, the combination of several parameters (such as competition, predation or pathogenicity) can lead to a loss of balance in the ecosystem, reducing the biodiversity. Furthermore, because Microsporidia are often assumed to be a negligible biomass in ecosystems because of their small size and their intracellular status, they are often omitted from studies of ecosystem trophic networks. Nevertheless, their presence makes more complex food webs by increasing the connectance (the proportion of realised links among the possible ones) in ecosystems (Lafferty et al., 2006).

In summary, given their profound effects on trophic interactions through host population regulation, the overall function of the ecosystem may not be fully observed and understood unless Microsporidia are considered. To this end, metagenomics approaches (Handelsman et al., 1998) are well adapted to deciphering the role of microsporidian species within trophic networks because of substantial cost reductions in sequencing technologies that have led to the rapid growth of this scientific field in recent years. The diversity and functionality of microorganisms can thus be revealed from complex environmental samples using the above-mentioned tools (Segata et al., 2013), providing prime access to basic information about microbial traits such as physiology, epidemiology and evolutionary history that can be used to infer organism’s ecological role in a given environment (Kim et al., 2013).

Microsporidia in trophic networks

In pelagic aquatic environments, the flows of matter and energy are not organised only according to conventional linear trophic pathways based on photosynthetic assimilation where the primary producers (phytoplankton) represent a food source for zooplankton which in turn act as an intermediate trophic link to the top of the food chain (fish, mammals). These flows can also pass through a microbial loop (composed of prokaryotes, fungi, heterotrophic and mixotrophic nanoflagellates, ciliates, and viruses) to form a true food web that is directly or indirectly connected to the classic food web (Figure 1). Almost half of the identified microsporidian species parasitise aquatic organisms, such as fish, arthropods, non-arthropod invertebrates and protists (Stentiford et al., 2013; Vavra and Lukes, 2013), affecting the food web at various trophic levels (Figure 1). Specifically, the relationship between the phytoplankton, the prokaryotic biomasses and the zooplankton grazers can be modified by the presence of Microsporidia. Through the manipulation of their hosts, they alter the structure, stability and function of the entire ecosystem. For metazooplanktonic populations (for example, Daphnia), infection by Microsporidia leads to a reduced fecundity and longevity. An opaque appearance of tissues also makes them visually attractive to predators, such as zooplanktivorous fishes (Vizoso et al., 2005). Because Microsporidia can exert a strong control over the host zooplanktonic population by decreasing its abundance, it affects the primary producer population by limiting grazing by zooplanktonic populations (Figure 1).

Figure 1
figure 1

Simplified schematic representation of trophic networks integrating the microbial loop in pelagic aquatic environments (modified from (Amblard et al., 1998)). All components of the microbial loop are shown in grey. The red asterisks indicate known organisms that are susceptible to Microsporidia parasitism. DOM, dissolved organic matter.

Full size image

In some cases, infections from Microsporidia result in very large cysts packed with host spores and meronts, so it is possible that these could serve as an additional source of nutrients for the predators that inadvertently consume them. It is also noteworthy that, if the parasite is characterised by a sufficiently large host spectrum, the ingested Microsporidia may also impact the behaviour and the fitness of the predator population. In terrestrial ecosystems, the lifespan and egg production of the spined soldier bug (Podisus maculiventris) is significantly reduced after ingestion of prey infected with the Microsporidium Vairimorpha necatrix (Down et al., 2004). Microsporidia infections similarly affect intraguild predation, where predation was observed between potential competitors, that is, species exploiting the same resources in similar ways (Polis et al., 1989). The influence of microsporidian parasitism on predation hierarchy is also evident in freshwater environments. For example, microsporidian infection in the native amphipod Gammarus duebeni celticus results in an altered mobility because of abdominal musculature degeneration (MacNeil et al., 2003b), which ultimately results in a reduced capacity to prey on smaller species (G. tigrinus, Crangonyx pseudogracilis) and an increased vulnerability to predation by the largest invading species (G. pulex) (MacNeil et al., 2003a). Clearly, Microsporidia play hidden but crucial roles in predator–prey associations and competitor interactions.

Biodiversity is critical for an ecosystem, fuelling competition and natural selection, while reducing disease risk and vulnerability to pathogens. Losses in biodiversity can increase sensitivity to pathogens such as Microsporidia, which play a role in the success of biological invasions of new species threatening then ecosystem’s functions. A host that is tolerant to a parasite can increase in relative abundance, affecting the populations of less tolerant species. For instance, microsporidium (Nosema spp) infections give the invasive harlequin ladybird Harmonia axyridis a competitive advantage (Vilcinskas et al., 2013). In the North-American and European native ladybirds (Coccinella septempunctata, Adalia bipunctata), Microsporidia infection is lethal, but H. axyridis mounts an immune response, secreting large amounts of harmonin. This antimicrobial alkaloid may not only result in a greater resistance of H. axyridis to certain diseases but also act as a poison to native ladybirds that consume invasive ladybird eggs (Sloggett and Davis, 2010). Interestingly, such cases seem to reflect mutualism rather than parasitism as the relationship appears to be beneficial to both partners. Specifically, Microsporidia exploit host energetic resources for their growth and survival while the invasive harlequin ladybirds proliferate to new habitats following the activation of immune defences by the Microsporidia infection. Host H. axyridis has then adapted to resist Microsporidia, unlike other ladybird species. Other cases where Microsporidia may be redefined as mutualistic organisms include the amphipod G. roeseli, where infection by the microsporidian species N. granulosis results in increased survival (Haine et al., 2007).

Life history traits are not only genotype-specific but also environmentally mediated. Environmental parameters may affect the coevolution of Microsporidia and their hosts. Environmental variation may influence the host population density, altering the transmission efficiency of the parasite. Thus, it has been suggested that large aggregations of harlequin ladybirds in winter can facilitate the transfer of Nosema spp. between individuals (Reynolds, 2013).

Parasitism can be particularly important when the hosts are keystone species with crucial functions in the environment, as it is the case of the honey bee (Apis mellifera) and the bumble bee (Bombus terrestris) pollinator species. Indeed, fecundity of these pollinators is affected by various Nosema species (N. apis, N. ceranae and N. bombi). Infection by Microsporidia also modifies honey bee behaviours, such as earlier and more intense foraging (Lin et al., 2009) and higher sucrose consumption (Alaux et al., 2010). These host–parasite interactions, combined with environmental parameters may have potentially contributed to the recent and steep global decline in populations of honey bee (Bromenshenk et al., 2010) and bumble bees (Meeus et al., 2011). The heterogeneity of the ecosystems, due in part to intensification of agriculture and landscape alterations, affects foraging areas by creating different floral resources. Consequently, different pollen diets (low in pollen or high in monofloral pollens) lead to a reduced lifespan of the honey bees parasitised by N. ceranae (Porrini et al., 2011; Di Pasquale et al., 2013). Similarly, pollutants (for example, pesticides) may act synergistically with N. ceranae infections to further increase the mortality of A. mellifera populations worldwide (Vidau et al., 2011; Pettis et al., 2012). The diminution of this microsporidian host in turn inhibits lower trophic levels, namely the plants that have to be pollinated. Consequently, Microsporidia can be associated with biodiversity losses in both pollinator host populations and pollinated plants. It is clear that environmental factors, which can fluctuate through time and space, play an essential role in host–parasite interactions and coevolution. Such factors have to be taken into account in investigations of the exact role of Microsporidia in ecosystems.

Microsporidia, like pathogens in general, play a tremendous role in ecosystems worldwide, at all levels, and for this reason, they should be better integrated into studies aimed at understanding processes involved in the ecological networks. Hopefully, the recent acquisition of large scale genome data from several microsporidian species, coupled with advances in the field of metagenomics, will help reveal the true extent of microsporidian diversity in the field and the extent of their functionality in terrestrial and aquatic ecosystems.

Structure and content of microsporidian genomes—the key for environmental success of Microsporidia

Both molecular karyotype analyses and genome sequencing of several microsporidian species indicate that their nuclear genomes range in size from only 2.3 Mbp for the human pathogen Encephalitozoon intestinalis (Corradi et al., 2010) to 24 Mbp for the Daphnia magna pathogen Hamiltosporidium tvaerminnensis, formerly named Octosporea bayeri (Corradi et al., 2009). Figure 2 describes the general features of sequenced microsporidian genomes and their phylogenetic relationships.

Figure 2
figure 2

General features and phylogenetic relationships of sequenced microsporidian genomes. Phylogenetic relationships were established based on several analyses (Wang et al., 2006; Corradi et al., 2009; Cuomo et al., 2012; Suankratay et al., 2012).

Full size image

Potential impacts of transposable elements on microsporidian genome structure

Transposable elements (TEs) and repeated sequences contribute to genome plasticity by enabling large-scale chromosomal rearrangements and by acting as buffer regions where transpositions and recombination events can occur without disrupting vital functions (Kidwell and Lisch, 2000; Biemont, 2009). In genomes without intergenic regions, paralogs and TEs, opportunities for chromosomal shuffling are limited. The smallest microsporidian genomes are found in the Encephalitozoonidae. The genomes of the Encephalitozoon spp. E. cuniculi (2.9 Mbp; (Katinka et al., 2001)), E. intestinalis, (2.3 Mbp; (Corradi et al., 2010)), E. hellem and E. romaleae (2.5 Mbp each; (Pombert et al., 2012)) show an almost perfect content and synteny conservation (Peyret et al., 2001; Corradi et al., 2010; Pombert et al., 2012). Their cores are extremely compact, with median intergenic lengths ranging from 75 to 82 bp (unpublished data, Figure 3a) and a few overlapping Coding DNA Sequences (CDSs) in head-to-head, tail-to-head or tail-to-tail configuration (Peyretaillade et al., 2009). This dense arrangement means few intergenic regions are available for potential recombination events. Furthermore, Encephalitozoon genomes lack simple sequence repeats, minisatellite arrays and TEs, features that all result from their extreme reductive evolution. The absence of RNA interference machinery in these genomes, whose function is involved in the repression of TE expression (Chung et al., 2008), could in part explain their lack of TEs. Indeed, an asexual genome that would not be able to control the TE activity would be submitted to too many deleterious TE insertions, which would irremediably decrease its fitness (Dolgin and Charlesworth, 2006; Kraaijeveld et al., 2012). Alternatively, the elimination of the RNA interference pathway could instead have occurred subsequently to the loss of TEs, following which its machinery would have become superfluous.

Figure 3
figure 3

Impact of TEs on microsporidian genome architectures. (a) Schematic representation of three genome structures observed in the microsporidian species Encephalitozoon cuniculi, Nosema ceranae and Anncaliia algerae. Genes are indicated in yellow boxes, and TEs are indicated in blue boxes. (b) Overview of TEs described in the microsporidian genomes. Information was obtained from both the literature and personal analysis with TransposonPSI (http://transposonpsi.sourceforge.net/).

Full size image

In contrast, intermediate-sized microsporidian genomes display large variations in the numbers and types of TEs that they contain. Only a few TEs were initially reported in N. apis (Chen et al., 2013) whereas the genomes of N. ceranae, N. antheraeae and N. bombycis were shown to display around 1%, 6% and 20% of TEs, respectively (Pan et al., 2013). However, BLASTP (Altschul et al., 1997) and TransposonPSI (http://transposonpsi.sourceforge.net/) analyses allowed us to determine that many TEs were wrongly annotated as host genes in the four species (unpublished data), such that the real TE content is higher in these genomes and, respectively, closer to 2%, 7%, 10% and 25% in N. apis, N. ceranae, N. antheraeae and N. bombycis. In any case, the differences in TE content between N. bombycis and the other Nosema species are substantial and correlated with genome size variations, a trend also observed between Nematocida species (Cuomo et al., 2012). The unusually large expansion in N. bombycis has been attributed to the propagation of TEs, to the acquisition of genes by horizontal transfer and to large-scale and small-scale gene duplication events (Pan et al., 2013). All of these species feature LTR retrotransposons but differ in non-LTR retrotransposon, DNA transposon and helitron contents (Figure 3b; (Cornman et al., 2009; Peyretaillade et al., 2012; Chen et al., 2013; Pan et al., 2013)). For example, N apis lacks Merlin and Mariner DNA transposon families whereas N. ceranae lacks piggyBac DNA transposons. The distribution of TEs is not uniform in the N. ceranae genome and core genes are often clustered in islands that are either free of or rarely interrupted by TEs (Figure 3a). Although the intra-generic synteny observed between genomes from the Nosema and Nematocida lineages is relatively high, it is lower than that observed within the Encephalitozoonidae, which suggests that the presence of TEs could be responsible for numerous rearrangements.

Assessing the chromosomal architecture and TE content in large microsporidian genomes is currently limited by the dearth of sequenced representatives. None of the lineages investigated so far featured more than one species per genus (this should be rectified soon with the incoming Anncalia algerae genomes) and those that were investigated were carried out at the draft level, focusing on core genes rather than structure. Both the Hamiltosporidium tvaerminnensis (Corradi et al., 2009) and A. algerae (Peyretaillade et al., 2012) drafts, respectively, covered 13.3 and 13.8 Mbp of their estimated ∼24 Mbp genome size and corresponded to numerous small-to-mid-size contigs, whose assemblies were hindered by the presence of repeated elements. Indeed, with an estimated 34.2–37.2 and 4.7 coverage of H. tvaerminnensis and A. algerae genomes, respectively, more than 50% of these genomes may correspond to TEs or repeated sequences artefactually assembled. This is especially true for H. tvaerminnensis, whose draft was performed in the early stages of Illumina sequencing with 35-bp long short reads. TEs are found in both the H. tvaerminnensis and A. algerae genomes (Figure 3b), with at least one-third of the A. algerae predicted CDSs found on contigs that contain one or more TE (Figure 3a, unpublished results), and we infer that more will be found in the missing regions upon further completion of the drafts.

Microsporidian genomes containing TEs are expected to be flexible and may thus be able to adapt faster to varying conditions. This in turn could occasionally yield genotypes that can jump to new host populations or even to new host species, as suggested for filamentous plant pathogens (Raffaele and Kamoun, 2012). Furthermore, recent study also suggests that A. algerae could use a TE as a lure strategy to escape the host innate immune system (Panek et al., 2014). Therefore, TEs may play an important role not solely in modulating the microsporidian genome architecture but also in the capacity of these parasites to infest multiple hosts and especially to adapt to new hosts.

Gene content and common transcriptional and translational microsporidian features

Microsporidian species are minimalist parasites with an extremely reduced metabolic potential and fewer or derived cellular components compared with free-living fungi. They do not possess peroxisomes, their mitochondria have been reduced to mitosomes that are incapable of oxygenic respiration and their Golgi apparatus is atypical and unstacked (Vavra and Lukes, 2013), the sum of which makes them highly host-dependent (Corradi and Selman, 2013; Corradi and Slamovits, 2011). Despite a 10-fold variation in genome size, Microsporidia seem to have conserved a similar gene repertoire ranging from ca. 1800 to 2600 protein coding genes. This array contains an irreducible set of core genes preserved throughout their evolution (Corradi and Slamovits, 2011; Vavra and Lukes, 2013) and variant subsets likely reflecting their adaptation to different hosts and environments (Vavra and Lukes, 2013). For instance, comparative genomics of four microsporidian species highlight a core genome of 932 CDSs (Peyretaillade et al., 2012), from which 141 appear to be found exclusively in the microsporidium phylum (for example, polar tube proteins involved in the infestation process). Comparative analysis of 11 genomes also allowed to define a microsporidian-specific core including especially 32 gene families specific to Microsporidia phylum (Nakjang et al., 2013). However, the high sequence divergence of microsporidian genes makes comparison with other organisms difficult. This high divergence also complicates the identification of microsporidian sequences from complex environments using only sequence similarity and base composition, leading to undetected contamination (Heinz et al., 2012; Peyretaillade et al., 2012; Vavra and Lukes, 2013).

The number of predicted genes may also drastically fluctuate depending on the inclusion of small CDSs (<=300 bp) defined by ab initio prediction software, and software-induced over-predictions produced a number of false positives in microsporidian annotations (Cheng et al., 2011). To ensure identification of such small genes, a re-evaluation of previous annotations of Encephalitozoon genomes performed using upstream transcriptional signals and synteny resulted in better delineated intergenic regions. Following these analyses, we propose that E. cuniculi, E. intestinalis, E. romaleae and E. hellem contain 2126, 1927, 1904 and 1955 CDSs, respectively (unpublished data). For these Encephalitozoon species, the mean CDS length is 991, 999, 996 and 1004 bp, respectively after reanalysis (unpublished data). Moreover, taking into account the new predictions, between 8 and 11% of the proteins are encoded by small CDSs for these four species (unpublished data). A high gene number has been identified for N. antheraeae, N. bombycis and T. hominis with mean CDS length 775, 741 and 871 bp, respectively (Heinz et al., 2012; Pan et al., 2013) indicating a potential over-prediction of small genes. For example in T. hominis, approximately 23% (736) of the predicted genes encode proteins smaller than 100 amino acids. Protein homology has only been established for 147 of these 736 CDSs and many of the remaining CDSs may correspond to falsely predicted genes.

Genome sequencing has also revealed a variable proportion of duplicate genes from different species (Nassonova et al., 2005; Akiyoshi et al., 2009; Pan et al., 2013). Nevertheless, in some cases, such gene duplication may in fact be related to polyploidy of the microsporidian genome as suggested for N. ceranae and A. algerae (Peyretaillade et al., 2012; Roudel et al., 2013)

In microsporidian genomes, transcriptional and translational regulation signals are highly conserved. Regardless of the base composition, CCC-like or GGG-like signals located in close proximity to the start codon appear to play a key role in transcription initiation (Cornman et al., 2009; Peyretaillade et al., 2009; Peyretaillade et al., 2012; Chen et al., 2013) (Supplementary Figure S1). 5′ RACE-PCR and RNAseq show that 5′UTRs may be highly reduced and in some cases absent in Microsporidia (Williams et al., 2005; Corradi et al., 2008; Peyretaillade et al., 2009; Gill et al., 2010; Heinz et al., 2012). Even in these cases, CCC-like or GGG-like signals are located upstream of the translational start site. In various genomes with a low G+C content, 5′UTRs are absent and the base composition upstream of the initiation start codon is strongly A+T-biased (that is, over 80%) (Peyretaillade et al., 2012). The analysis of highly expressed genes, such as those encoding ribosomal proteins, highlights a complementary AAAATTTT-like signal in close proximity to the CCC-like or GGG-like signals which seem universal in the microsporidian phylum (Peyretaillade et al., 2009; Peyretaillade et al., 2011; Peyretaillade et al., 2012) (Supplementary Figure S1). For the 3′-end mRNA processing, the consensus sequence AAUAAA or some other slightly degenerate signal (that is, only differing by a single nucleotide) is near the stop codon for all annotated genes in E. cuniculi, E. intestinalis, N. ceranae and E. bieneusi species (Peyretaillade et al., 2009; Peyretaillade et al., 2012). Finally, concerning the translational signals, no bias is detected upstream of the AUG codon, which is consistent with the existence of a highly reduced 5′UTR. Nevertheless, a strong bias towards adenine or guanine (∼78%) in the +4 position and a significant bias for adenine (∼45%) in the +5 position have been identified in these species (Peyretaillade et al., 2009).

Genomic features to explore the microsporidian life cycle

Although Microsporidia are present in many environments, the first difficulty in exploiting genomic data from these species is to isolate their cells. For this, cell culture and single-cell approaches can be considered. Nevertheless, such techniques require to know culture conditions especially host cells and to be able to discriminate the microsporidian cells from other microbial populations (Figure 4). To study microbial diversity from complex environments, direct DNA extraction and sequencing remain the most promising and easy-to-use strategies to implement. However, current DNA procedures are not adapted to the different microsporidian forms, which are present as highly resistant spores in outside environments or as fragile meronts when inside their hosts (Vavra and Lukes, 2013). The release of DNA from the spores is difficult because of insufficient current lysis approaches. The methods applied to the infected host organisms to recover parasite DNA involve heating steps (Izquierdo et al., 2011; Wang et al., 2013; Sapir et al., 2014) that may cause damage to A+T-rich microsporidian genomes (Figure 2).

Figure 4
figure 4

Schematic representation of the major approaches to appraise Microsporidia diversity in complex environments using microsporidian genomic features.

Full size image

Microsporidian taxa detection is also limited by the molecular identification methods used, that is, the sequencing of PCR products targeting small subunit ribosomal (rRNA) genes. This leads to an underestimation of their presence and diversity in different environments, because microsporidian rRNA genes are shorter than those of other eukaryotic organisms and feature extremely reduced variable regions (Peyretaillade et al., 1998). Thus, microsporidian diversity is investigated with available non-degenerate specific primers that target only few organisms (Fournier et al., 2002; Ardila-Garcia and Fast, 2012; Ardila-Garcia et al., 2013; Wang et al., 2013; Sapir et al., 2014) or a weakly degenerate primer set designed from the few known Microsporidia with available sequences in public databases (Vossbrinck et al., 2004). Several studies show the presence of some known human microsporidian species in various environments (Fournier et al., 2000; Kahler and Thurston-Enriquez, 2007; Izquierdo et al., 2011). However, to our knowledge, only two studies highlight the microsporidian diversity in different ecosystems: soil, sand and compost (Ardila-Garcia et al., 2013) or water from a mosquito larval habitat (Avery and Undeen, 1987). Thus, to unveil the microsporidian community structure as defined in microbial ecology (that is, richness, diversity and composition), new analytical tools must be developed. The definition of explorative primers to target either phylogenetic markers (small subunit rRNA) or functional genes (Delbac and Polonais, 2008; Dugat-Bony et al., 2012; Peyretaillade et al., 2012) can be accomplished with recently developed design software such as HiSpOD (Dugat-Bony et al., 2011) or KASpOD (Parisot et al., 2012), and will be crucial for identifying uncharacterised microsporidian sequences and anticipating genetic variations (Figure 4).

Advances in high-throughput sequencing have led to a rapid growth of metagenomics studies to monitor complex ecosystems (Segata et al., 2013). Binning, which represents the taxonomic classification of reads, is the most basic step in the characterisation of microbial communities from metagenomic samples. Classification can be performed either by de novo binning using intrinsic sequence features or by using extrinsic information provided in sequenced microbial genome databases (Segata et al., 2013). Therefore, the identification of microsporidian sequences relies on a detailed knowledge of the genomic characteristics of these pathogens including specific gene repertoires and the particular organisation of their genomes (Figure 4). The unique transcriptional and translational regulation signals described herein could also confirm the presence of Microsporidia. Finally, identification of TEs specific to microsporidian species (unpublished data) provides sequence information to validate the presence of these organisms in complex environments. Encompassing all this genomic information into a comprehensive database dedicated to Microsporidia could improve both extrinsic and intrinsic approaches. Nevertheless, some horizontal gene transfers from different eukaryotic and prokaryotic donors can interfere with identification and should then be taken into account (Cuomo et al., 2012; Heinz et al., 2012; Pombert et al., 2012).

By including metatranscriptomic data, a peculiar feature observed in some microsporidian species can also be utilised. For some species, the most gene-dense regions have increased levels of ‘overlapping transcription’. This phenomenon is an atypical transcriptional process that produces mRNA encompassing multiple CDSs from different strands (Williams et al., 2005; Corradi et al., 2008; Peyretaillade et al., 2009). The presence of such mRNAs coupled with previously described genomic features can strengthen the taxonomic assignment to the Microsporidia phylum (Figure 4).

‘Omics’ studies require, however, a very large effort in both sequencing and analyses to assess the microsporidian species. Interesting alternatives rely on the reduction of sample complexity through methods such as single-cell genomics (Stepanauskas, 2012) or gene capture to specifically enrich for the nucleic acid sequences of interest. Gene capture targeting specific microsporidian genes can then be coupled to high-throughput sequencing for accurate and efficient identification of species (Denonfoux et al., 2013) (Figure 4). Using these new tools will highlight the high diversity and distribution of environmental Microsporidia and will increase the understanding of host interactions and the roles of these organisms in the functioning of ecosystems.

Data archiving

There were no data to deposit.