The microbiota of the Lasius fuliginosus – Pella laticollis myrmecophilous interaction

Abstract

Myrmecophilous insects depend on their ant hosts during at least part of their life cycle. Lasius ants appear to be primarily involved as host in myrmecophilous interactions, especially with aphids and beetles, e.g. staphylinid. Pella laticollis is a rove beetle having a close ecological association with Lasius fuliginosus, which is correlated to its behavior adapted to avoid attack. Microorganisms can also play an important role in the maintenance of such relationships. We used 16S rRNA sequencing on the Illumina MiSeq platform to identify the bacterial communities associated with larvae and adults of both L. fuliginosus and P. laticollis. In addition, we determined the microbiota profiles of the nest-carton, an organic material lining the nest chambers. We obtained more than two million good quality reads. Taxonomic annotation showed that the profiles consisted of a total of 20 phyla, among which Proteobacteria was the most abundant. The samples were grouped according to the host’s developmental stage or the nest material, and the differences between those groups were significant. Only the microbiota of L. fuliginosus larvae and adults did not differ significantly. Analyses at the genus level indicated a heightened abundance of Pseudomonas in the insects’ profiles. The bacterial communities associated with the nest-carton included bacterial genera recorded from soil and dead wood. The profiles showed the presence of two well-known endosymbiotic bacteria, namely, Rickettsia and Wolbachia. According to our findings, the bacterial communities associated with larvae and adults of both L. fuliginosus and P. laticollis formed clusters according to the host’s identity and its developmental stage. The profiles determined for the nest-carton formed a separate group. Our study is in line with a new research trend that is focusing on microbiota associated not only with ants, but also with myrmecophiles and ant nests inhabited by those species.

Keywords:

• Myrmecophiles
• ants
• beetles
• nest-carton
• microbiota

Introduction

Insects have evolved non-pathogenic, persistent associations with various microorganisms. These microbes, particularly gut microbiota, have been shown to contribute to the host’s digestion, pathogen resistance, and physiology (Douglas 1998, 2009, 2015; Van Borm et al. 2002; Pinto-Tomas et al. 2009; Koch & Schmid-Hempel 2011; Raychoudhury et al. 2013; Engel & Moran 2013; Michalkova et al. 2014; Jing et al. 2020). The evolutionary success of insects has depended in part on these myriad relationships with beneficial microorganisms. Particular attention has been paid to the microbiota associated with social insects. Close interactions among host individuals, e.g. maternal care, food sharing, and grooming, facilitate the transmission of gut-associated bacteria, resulting in persistent associations, coevolution with hosts and the emergence of specialized symbiont lineages, as well as the competitive exclusion of unspecific microbes (Hongoh et al. 2005; Russell et al. 2009; Martinson et al. 2011; Anderson et al. 2012; Engel & Moran 2013; Kwong et al. 2017; Powell et al. 2018). Honeybees are important models in studies to determine the range of influences of microbial communities on their hosts and in understanding of the pathogenic and mutualistic components of the microbiota associated with these ecologically and economically important pollinators (Zheng et al. 2017; Jones et al. 2018; Wang et al. 2018; Raymann & Moran 2018; Kešnerová et al. 2020). The complex host-microbe relationships revealed by studies of honeybee microbiota have fuelled growing interest in the microbial communities associated with other social insects, such as ants.

The family Formicidae (Insecta, Hymenoptera) is an ecologically dominant group of invertebrates divided into 17 subfamilies and 333 genera. More than 13 000 known ant species occur worldwide (Bolton). The great diversity of ants is probably due to their ecological diversity, for example, as regards nesting, feeding preferences, social behaviour and the division of labor among castes (Wilson 1987; Hölldobler & Wilson 1990; Ramalho et al. 2017a). Analyses of the bacterial communities associated with ants have shown important aspects of host biology in structuring the diversity and abundance of these communities. Anderson et al. (2012) found similarity of the microbiota among ants species from the same trophic level, and identified differences between the microbiome profiles of herbivorous and predatory species. In the case of herbivorous ants, e.g. Cephalotes turtle ants, gut bacteria take part in nitrogen recycling and the synthesis of essential amino acids that are acquired by the host in substantial quantities (Hu et al. 2018). A nutritional role of associated bacteria has been identified for Camponotini ants (Sameshima et al. 1999; Wernegreen et al. 2009; Ramalho et al. 2017a). Blochmannia, a Proteobacteria specific to this tribe, not only assists in supplying its hosts with essential amino acids, but also harbours genes necessary for the metabolism of nitrogen, sulphur and lipids (Gil et al. 2003; Degnan et al. 2005; Feldhaar et al. 2007; Williams & Wernegreen 2010). It has also been shown that the number of Blochmannia increases strongly during host’s metamorphosis and that aposymbiotic larvae have much-reduced capacity to complete metamorphosis (Feldhaar et al. 2007; Stoll et al. 2010). This trend implies an important function of Blochmannia in the developmental phase during which the hosts are metabolically very active. Moreover, the bacteria associated with various ants species can produce antibiotics against a range of bacterial and fungal pathogens. One example is the genus Pseudonocardia, associated with ants from the Attini tribe, which produces antibiotics to suppress Escovopsis, a genus of virulent microfungal parasites that attacks the basidiomycetous fungi cultivated as the main food source by these ants (Currie et al. 1999, 2003; Cafaro & Currie 2005; Sen et al. 2009; Oh et al. 2009; Mueller et al. 2010; Zucchi et al. 2010). However, the associated bacteria do not only directly affect the fitness of their hosts. Microorganisms can also shape or interrupt close relationships among ants and other animals (e.g. myrmecophiles), plants or fungi (for ant interactions see (North et al. 1997; Pierce et al. 2002; Heil & McKey 2003; Chomicki & Renner 2017; Nelsen et al. 2018; von Beeren et al. 2018; Casacci et al. 2019; Tartally et al. 2019)). In a recent study, Di Salvo et al. (2019) identified two bacteria, Serratia mercescens and S. entomophila, which could be involved in the chemical cross-talk among microbiota, Maculinea larvae and Myrmica ants (Di Salvo et al. 2019). In turn, Fischer et al. (2017) have shown that bacteria may contribute to the recognition of distant aphid species by Lasius niger ants (Fischer et al. 2017). These results strengthen interest in studying the role of microorganisms in shaping relationships among closely associated species (Szenteczki et al. 2019; Kaczmarczyk-Ziemba et al. 2020).

Lasius ants appear to be primarily involved as host in myrmecophilous interactions with other insects, especially aphids (Fischer et al. 2017), but also beetles, such as staphylinids (Akino 2002; Lapeva-Gjonova 2013). The genus Pella is a rove beetle that has a close ecological association with Lasius ants (Hölldolber et al. 1981). For example, Lasius fuliginosus (Latreille, 1798) has been reported as the host of six Pella species (P. humeralis, P. funesta, P. cognata, P. similis, P. lugens, and P. laticollis) (Hölldolber et al. 1981). These myrmecophiles are found inside the carton-nests built by their hosts in tree hollows. The nests are divided into chambers by thin walls. The material used, usually macerated wood but occasionally soil particles, is intermixed and overgrown by fungus mycelium (Cladosporium myrmecophilum). Collected honeydews as well as supplied sugar-water are used as adhesives for sticking the wood and soil particles together, while at the same time providing a culture medium for the fungus (Fresenius 1852; Lagerheim 1900; Elliott 1915; Maschwitz & Hölldobler 1970). Further analyses of fungal communities associated with L. fuliginosus nests based on both morphological and molecular data have revealed the presence of a further three species of fungi cultured by ants. However, their identification using both standard keys and DNA sequence comparisons with GenBank entries has not been successful (Schlick-Steiner et al. 2008). More recently, Brinker et al. (2019) analyzed microbial communities associated with L. fuliginosus nests and found that the fungal communities were stable over years and distinct from surrounding soil (Brinker et al. 2019).

Larvae of Pella are observed in waste dumps inside those carton-nests, while adults are found in shelters during the day and in close proximity of hosts’ nests during the night. More detailed descriptions of Pella behaviour have been made by Hölldolber et al. (1981). These authors frequently observed the Pella larvae feeding on dead ants. In turn, when ants encountered the larvae, they usually attacked them. The attacks resulted in a typical defense behaviour with Pella larvae raising abdominal tip towards the head of the ants. This behavior apparently protected larvae against ant attacks although several such interactions ended with the death of the Pella larvae. On the other hand, when the temperature is low, the Pella larvae are able to come into close contact to L. fuliginosus workers without being attacked. Hölldobler et al. observed even the beetle larvae licking the cuticles of live ants, including the mandibles and mouthparts. However, experiments described by authors clearly showed that the Pella larvae do not solicit regurgitation in ants and dead ants are their main food source. In the case of P. laticollis adults, observations revealed that these insects predominantly live as scavengers feeding on dead or disabled ants, but are also able to act as very effective predators on ant adults and larvae, when starved (Hölldolber et al. 1981; Staniec et al. 2009).

The staphylinid, Pella laticollis (Maerkel, 1845) (Staphylinidae, Aleocharinae) is associated only with L. fuliginosus. Thus, the association between P. laticollis and its host seems to be species-specific on the part of Pella. Moreover, previous studies have suggested that P. laticollis exhibits behavioural pre-adaptation towards a closer relationship with its host better than other Pella representatives (Hölldolber et al. 1981; Stoeffler et al. 2011; Staniec et al. 2018). Evidence for this statement is provided by the defensive behaviour observed in P. laticollis larvae and adults. On encountering an ant, a beetle discharges a secretion from its tergal glands, while simultaneously flexing its abdomen (Staniec et al. 2009). This “duck down” appeasing behaviour and presenting the abdominal tip, observed only in P. laticollis, appears to mollify the hosts’ aggressive reaction: the ants focus on the tip of the beetle’s abdomen, which allows the latter to escape (Stoeffler et al. 2011). These results can bring us nearer to understanding of the mechanisms that govern the close relationships between P. laticollis and its host L. fuliginosus.

Undoubtedly, one cannot gainsay the possible role of microbiota in shaping the interaction between these two species. On the other hand, bacterial communities associated with both P. laticollis larvae and adults may be similar to those associated to their host’s adults. Both developmental stages of this staphylinid feed on dead and disabled L. fuliginosus ants (and ants’ larvae, if available) and thus, they may acquire fractions of microbiota associated with their host. To test this hypothesis, we investigated the bacterial communities associated with the larvae and adults of both P. laticollis and its host L. fuliginosus. We used 16S rRNA sequencing to define whether the bacterial communities varied among two closely associated species. Lastly, we analyzed the microbiome profile of the nest-carton – the material which lines the inside of L. fuliginosus nests and may mediate the horizontal transfer of bacteria.

Materials and methods

Sample collection

Samples were collected during the spring of 2018. Permission for the field studies was granted by the Polish Ministry of the Environment (Permit No. DOP-WPN.286.214.2018.MŚ). Larvae and adults of both L. fuliginosus and P. laticollis were obtained from a single, large carton-nest found inside an old birch tree trunk (Betula sp.) near Lake Długie (Polesie National Park; N 51º27ʹ04.12” E 23º10ʹ14.17”). We also sampled the nest-carton. The number of collected samples was limited (5 samples of L. fuliginosus (2 for larvae and 3 for adults), 6 for P. laticollis (3 for larvae and 3 for adults) and 3 for nest material). All larvae and adults were starved prior to the subsequent analyses, then placed separately into Eppendorf or Falcon tubes and stored at −80°C until DNA extraction.

DNA extraction and Illumina Miseq sequencing

The protocol for the sterilization of the ants cuticles was invoked to eliminate surface microbes and exogenous DNA prior to the investigation of internal microbe diversity. Individuals were processed separately with 70% ethanol for 30 s and then rinsed for 1 min in three successive baths of DNA-free water. The DNA was then extracted from whole specimens using a Sherlock AX Purification Kit (A&A Biotechnology, Poland). The DNA from the nest material (carton) was in turn extracted using a DNeasy PowerSoil Kit (Qiagen, UK) preceded by a homogenization step with garnet beads. All the extractions were performed in accordance with the manufacturer’s protocols. Negative controls were not included in the present study, but all the procedures were performed in such a way as to prevent DNA contamination.

After extraction, the quantity and quality of the DNA were evaluated with a Nano Drop ND-1000 spectrophotometer (Nano Drop Technologies, USA). The extracted DNA was sent to the Genoplast Laboratory (Poland) for library preparation and 16S Miseq Illumina sequencing as described in (Kaczmarczyk-Ziemba et al. 2019). Sequences of the V3-V4 region of the 16S rRNA gene were amplified using 341 F and 785 R primers. Amplification was performed using NEBNext® Hotstart DNA polymerase with high fidelity using the following protocol: 1 cycle at 98°C for 30 s, 25 cycles at 98°C for 10 s, at 55°C for 30 s and at 72°C for 20 s, and finally 1 cycle at 72°C for 2 min.

The DNA extracted from 14 samples was sequenced separately: 5 samples of L. fuliginosus (2 of larvae and 3 of adults), 6 samples of P. laticollis (3 of larvae and 3 of adults) and 3 samples of nest material. For the subsequent analyses, the samples were designated as follows: Lf-L and Lf-A – L. fuliginosus larvae and adults, Pl-L and Pl-A – P. laticollis larvae and adults, and K – carton. The raw sequential reads were submitted to the Sequence Read Archive (SRA) database under accession number PRJNA612990.

Data analysis

The samples were analysed using the QIIME 1.9.1 pipeline (Caporaso et al. 2010b). Paired-end reads from MiSeq sequencing were quality trimmed and joined with PANDAseq version 2.8 (Masella et al. 2012) with a quality threshold of 0.9. The sequences that did not meet the quality criteria were removed from further analysis (mean quality >20). Chimeric reads detection was performed with VSEARCH, version 1.7.0, an open-source replacement of USEARCH software. The clustering of operational taxonomic units (OTUs) at 97% similarity was performed using the uclust method, version 1.2.22q (Edgar 2010). Sample reads were rarefied to 52 233 reads. Taxonomic assignment was performed at 97% against the SILVA v.132 database (Quast et al. 2013), with the taxonomy assignment tool PyNAST (Caporaso et al. 2010a). OTU saturation was evaluated with rarefaction curves using the Chao1 richness estimate. The BIOM table (The Biological Observation Matrix) was used as the core data for downstream analyses (McDonald et al. 2012). Any sequences that were classified as Mitochondria or Chloroplast, as well as the singletons, were filtered out of the dataset. The diversity indices, including Shannon and observed OTUs were estimated for each sample. Further statistical analyses were preceded by data distribution analysis using the Shapiro-Wilk test. Differences in alpha diversity indices among the sample groups were statistically assessed with the Kruskal-Wallis test. The bacterial community structure was compared using UniFrac (Lozupone & Knight 2005) and Emperor (Vázquez-Baeza et al. 2013). A two-dimensional principal coordinate analysis (PCoA) was conducted from weighted UniFrac distances obtained from the core diversity analyses.

A permutation-based multivariate analysis of variance (PERMANOVA) was performed to assess whether there were differences among the predefined groups of samples, i.e. Pella larvae, Pella adults, Lasius larvae, Lasius adults, and nest material. Similarity percentage (SIMPER) analysis was performed to calculate the average dissimilarities in the bacterial community structures between samples and to ascertain which family was responsible for the observed differences. The relative abundances of those bacterial families that were primarily responsible for the differences between samples were then used for a proximity analysis. To illustrate the most abundant bacterial families and microbiome relationships across the tested samples, a heatmap and dendrogram were generated with the Bray-Curtis dissimilarity index. A similarity profile (SIMPROF) test was used to identify well-defined groups of samples (Clarke et al. 2008). All the statistical analyses were performed using Primer version 7 software (Clarke & Gorley 2015).

Results and discussion

To date, studies using high throughput sequencing techniques for microbiota profiling have focused primarily on identifying and analysing the bacterial communities associated with different ant species (Kautz et al. 2013b; Vieira et al. 2017; Lucas et al. 2017; Ramalho et al. 2017b, 2017c, 2019; Hu et al. 2018; Chua et al. 2018; Segers et al. 2019). Although myrmecophiles interact closely with ants and are often found inside ants’ nests, no exhaustive analyses of the bacterial communities associated with those species have been carried out (but see (Liberti et al. 2015; Fischer et al. 2017; Ivens et al. 2018; Di Salvo et al. 2019; Szenteczki et al. 2019; Kaczmarczyk-Ziemba et al. 2020)). In the present study, we have extended the state of the art regarding the microbiota of ants and their associates by investigating the potential similarities among the bacterial communities associated with two closely interacting species, L. fuliginosus and P. laticollis (both larvae and adults). We also determined the microbiota profile for the nest-carton, an organic material present inside the nest chambers. We collected only a limited number of samples, performing the analyses on 5 samples of L. fuliginosus (2 of larvae and 3 of adults), 6 of P. laticollis (3 of larvae and 3 of adults) and 3 of nest material. Thus, presented results and statistics should be recognized rather as the first attempt to identify trend characteristics in L. fuliginosus – P. laticollis interaction and more comprehensive analyses involving additional biological replicates should be performed in the future to test the stability of the described patterns.

Following demultiplexing, quality filtering and chimera removal, 2 006 036 reads were retained for the 5 samples of L. fuliginosus (2 of larvae and 3 of adults), the 6 samples of P. laticollis (3 of larvae and 3 of adults) and the 3 samples of nest material (52 233–169 028 reads per sample). The calculated alpha diversity metrics differed significantly among the samples grouped according to the source (Kruskal-Wallis test (H): 12.13, p < 0.05 for OTU and 11.32, p < 0.05 for the Shannon indices).

Taxonomic classification yielded 20 phyla present in the bacterial communities analysed (Figure 1, Table S1). All the samples contained high abundances of Proteobacteria (up to 99.84% for the Pl-L1 profile), Actinobacteria (up to 29.19% for Lf-L3) and Firmicutes (up to 23.58% for Lf-L3) (Figure S1). These three phyla have frequently been listed among the most abundant groups of bacteria present in the microbiota of various insect species (Jones et al. 2013; Yun et al. 2014), including ants (Kellner et al. 2015; Sapountzis et al. 2015; Lanan et al. 2016; Meirelles et al. 2016; Vieira et al. 2017; Łukasik et al. 2017; Ramalho et al. 2017a) and their associates (Liberti et al. 2015; Ivens et al. 2018; Szenteczki et al. 2019; Kaczmarczyk-Ziemba et al. 2020).

Figure 1. Bacterial phyla present in nest material and in profiles of Pella laticollis and Lasius fuliginosus individuals

The average similarities of the profiles calculated on the basis of the Bray-Curtis index at the bacterial family level varied among the predefined groups. The bacterial communities associated with P. laticollis adults resembled each other the most (99.84% of the average similarity within the group). The profiles of the other samples showed the following average similarities at the bacterial family level: 59.71% for P. laticollis larvae, 55.90% for L. fuliginosus larvae, 47.77% for L. fuliginosus adults and 82.61% for the nest-carton. SIMPER analysis indicated that 23 families were primarily responsible for the differences between the samples. The relative abundances of these families were used to generate a heatmap for all the samples (Figure 2(a)). The bacterial communities were grouped according to their origin, this approach being supported by the results of the SIMPROF and PERMANOVA (Pseudo-F = 13.4, p = 0.001) analyses. Only sample Lf-A5 was grouped with the L. fuliginosus larvae samples. This pattern was also discernible in the bar plot at the phylum level and in the PCoA plot (Figures 1 and 2(b)). The relationships among the microbiome profiles on the PCoA plot and the heatmap were congruent. The microbiota associated with the carton and L. fuliginosus ants (both larvae and adults) displayed the greatest similarity, whereas the bacterial communities associated with P. laticollis adults were the most highly differentiated from the other communities (Figure 2(a)). In turn, PERMANOVA calculated for all pairs of samples showed that only the microbiota associated with Lasius larvae and adults were not significantly different (Table I). However, some of these significant differences may have resulted from our use of two different methods of extracting DNA from insects and the nest-carton (e.g. (Ketchum et al. 2018; Mallott et al. 2019)).

Table I. Results of PERMANOVA analysis after 999 permutations. Values above diagonal indicate p values. Groups were determined according to the sample origin

	K	Lf-L	Lf-A	Pl-L	Pl-A
K		0.011	0.011	0.001	0.001
Lf-L	3.542		0.129	0.041	0.006
Lf-A	2.985	1.706		0.019	0.003
Pl-L	4.374	2.359	2.657		0.002
Pl-A	12.773	5.914	3.939	5.275

Figure 2. The heatmap showing relationships among tested profiles of bacterial communities and the Principal Coordinate Analysis (PCoA) plot. (a) – The heatmap showing bacterial families distributed across tested samples. Only those families which were primarily responsible for the observed differences among samples were considered. Both dendrograms were estimated with the Bray-Curtis dissimilarity index; (b) – PCoA of bacterial communities associated with tested specimens based on weighted UniFrac distances

More detailed examination of the L. fuliginosus microbiota showed that Acetobacteraceae (Proteobacteria) was the most abundant in the bacterial communities associated with adults. Previous studies had found that members of this family were characteristic of the microbiota of ants consuming sugar-rich substances like honeydew, which is the principal dietary component of L. fuliginosus adults (Hu et al., 2017; Russell et al. 2009; Brown & Wernegreen 2016; Ivens et al. 2018). At the genus level, all the ant microbiota profiles were characterized by a relatively high percentage of certain bacterial genera: Pseudomonas and Acinetobacter (Proteobacteria), Mycobacterium, Cutibacterium and Nocardioides (Actinobacteria), Bacillus, Staphylococcus and Streptococcus (Firmicutes). Moreover, a relatively high abundance of Turicella (Actinobacteria) was identified in the microbiota of adults (Table S1). These bacteria may play key roles in different digestive and defensive processes. For example, Pseudomonas species engage in cellulolytic, lipolytic, esterase, amylolytic and xylanolytic activities (Briones-Roblero et al. 2017b), which appear to be necessary for honeydew digestion and wood maceration during nest building. In turn, Mycobacterium may impair the development of fungal infections in insects (Kabaluk et al. 2017), an aspect especially crucial in the case of L. fuliginosus, as living in a social group may increase the risk of an epidemic outbreak. Entomopathogenic fungi, as well as their toxins, are natural threats to social insect colonies. To combat them, social insects have evolved a series of unique disease defences at the colony level, which consist of behavioural and physiological adaptations, i.e. social immunity (Liu et al. 2019). We cannot rule out the possibility that the Mycobacterium found in bacterial communities associated with ants is involved in these defensive mechanisms. Bacteria-derived antifungal compounds have recently been described from the head and body of L. fuliginosus (Liu et al. 2016a, 2016b; Ye et al. 2017; Shen et al. 2017; Brinker et al. 2019). The higher abundances of the genera listed above are in accordance with the profiling of ant microbiota performed to date. Pseudomonas and Mycobacterium, as well as other genera listed above, have been found in communities associated with ant species belonging to the genera Acromyrmex (Zhukova et al. 2017), Allomerus and Tetraponera (Seipke et al. 2013), Camponotus and Colobopsis (Ramalho et al. 2017b), Cephalotes (Kautz et al. 2013a; Lanan et al. 2016), Formica (Kaczmarczyk-Ziemba et al. 2020), Oecophylla (Chua et al. 2018), Solenopsis (Li et al. 2005; Ishak et al. 2011), Brachymyrmex and other Lasius species (Ivens et al. 2018).

The bacterial communities associated with P. laticollis larvae and adults exhibited lower alpha diversity indices than those calculated for the ant profiles. At the phylum level, all these communities were dominated by Proteobacteria, but Actinobacteria were also present in the larval profiles. These differences were reflected in the profile clustering (congruent to the host’s developmental stage) and the significant PERMANOVA value. Analyses of lower taxonomic ranks indicated relatively high abundances of the families Pseudomonadaceae in P. laticollis larvae and Rickettsiaceae in adults; this was directly related to higher percentages of two genera, i.e. Pseudomonas and Rickettsia, respectively. As mentioned above, members of Pseudomonas participate in many different enzymatic activities. Moreover, members of Pseudomonas associated with females from the genus Paederus (staphylinids like Pella laticollis) can synthesize the polyketide pederin, the most complex non-proteinaceous insect defensive secretion known (Kellner & Dettner 1996; Kador et al. 2011; Krinsky 2019). In rove beetles, pederin acts as a substance for chemical defence against potential predators (Kellner & Dettner 1996). Furthermore, both Pseudomonas and Serratia found in the larval profiles possess antimicrobial activity and regulate mechanisms within bacterial populations (quorum sensing) (Venturi 2006; Liu et al. 2007; Charyulu et al. 2009). Briones-Roblero et al. (2017a) hypothesized that those bacteria, along with an insect’s immune system, could regulate the gut bacterial communities associated with the bark beetle Dendroctonus rhizophagus (Briones-Roblero et al. 2017a). Apart from host-species, Serratia has also been found in the microbiota of other curculionid beetles, as well as scarabaeid and staphylinid species (Kolasa et al. 2019). Here, we found a heightened abundance of S. proteamaculans (3.34% for Pl-L2 and 5.90% for Pl-L3), which is responsible for amber disease in the New Zealand beetles Costelytra zealandica, Pyronota festiva and P. setosa (Sanchez-Contreras & Vlisidou 2008; Hurst et al. 2018). Its pathogenicity is directly correlated with the presence of the pADAP plasmid, so not all strains of S. proteamaculans are pathogenic (Koppenhöfer et al. 2012). Moreover, the pADAP plasmid has only been found in New Zealand bacterial isolates (Jurat-Fuentes & Jackson 2012). Members of our team have been taking part in long-term studies of the biology of myrmecophiles, P. laticollis in particular (Staniec et al. 2009), and their investigations have not revealed any potential symptoms of amber disease in P. laticollis larvae. Although S. proteamaculans does not appear to affect the biology of P. laticollis, endosymbiotic Rickettsia (over 99.00% in adult profiles) can induce diverse effects on the host, from influencing host fitness to manipulating reproduction. Rickettsia strains have been reported to affect host fitness negatively, for example, by reducing in body weight, fecundity and longevity, and by increasing susceptibility to insecticides (Chen et al. 2000; Sakurai et al. 2005; Simon et al. 2007; Kontsedalov et al. 2008). There is also evidence that this endosymbiont has positive effects on host fitness, for example, via a possible role in oogenesis and the promotion of more offspring (Zchori-Fein et al. 2006; Himler et al. 2011). Finally, Rickettsia may manipulate the host’s reproductive processes in such a way as to bias the host sex ratio towards females, thus favoring the spread of this transovarially inherited endosymbiont in the infected population (Perlman et al. 2006; Hurst & Frost 2015). Infections with Rickettsia have been identified in various groups of insects (Weinert et al. 2009), including ants (Hu et al., 2017; Lester et al. 2017), and also myrmecophilous species like the caterpillars of the Large Blue butterfly Maculinea alcon butterfly (Di Salvo et al. 2019) and adults of the beetles Thiasophila angulata, Monotoma angusticollis, Dendrophilus pygmaeus, Leptacinus formicetorum, Myrmechixenus subterraneus and Ptenidium formicetorum (Kaczmarczyk-Ziemba et al. 2020). Nevertheless, the potential effects of such an infection with Rickettsia have not been detected in those hosts. Similarly, long-term studies of the biology and morphology of P. laticollis have not revealed any female-biased sex ratio in its natural populations (authors’ observations). It is highly advisable, however, that the prevalence of Rickettsia in P. laticollis populations be analysed in further studies. Here, the sample size was limited and only three profiles of adult individuals collected from a single nest were determined. Thus, we cannot state definitely that all P. laticollis adults are heavily infected with Rickettsia. Neither can we rule out the possibility that immature individuals, as well as their host L. fuliginosus, are also infected with these endosymbiotic bacteria. In the present study, Rickettsia was found in microbiota of a single P. laticollis larva (0.05% for Pl-L3) and a single L. fuliginosus adult (0.55% for Lf-A6).

In addition to the microbiota profiles determined for L. fuliginosus and P. laticollis, we analysed the bacterial communities associated with the nest-carton. Ant colonies build extensive nests to live in, produce waste, and interact in ways that produce distinct microbial communities (Hölldobler & Wilson 1990; Hölldobler & Wilson 2009). Interest in analyses of the microbiota associated with the nest material has been increasing in recent years, and one such analysis has been conducted for L. fuliginosus (Brinker et al. 2019). In that study, however, the composition of the bacterial community was characterized by automated ribosomal intergenic spacer analysis (ARISA), an approach enabling the rapid generation of whole-community “fingerprints” of microbiota assemblages (Fisher & Triplett 1999; Ranjard et al. 2001). Although ARISA-based approaches are suitable for investigating microbiota patterns, they prevent direct comparison with available data because comparable taxonomic information is missing (Brinker et al. 2019). In the context of those results, therefore, we cannot evaluate the structure of the nest-carton bacterial community identified here. Nevertheless, reports based on 16S sequencing have been published for nests of the Neotropical ants Azteca trigona and A. alfari, and the Palearctic species Formica exsecta (Lucas et al. 2017, 2019; Lindström et al. 2019). Three phyla (Proteobacteria, Actinobacteria and Acidobacteria) were the most abundant in the microbiota profiles determined in those studies; the same pattern emerged in the present study. Nine genera within these phyla (Acidipila, Granulicella, Candidatus Koribacter, Mycobacterium, Acidothermus, Roseiarcus, Afipia, Burkholderia-Caballeronia-Paraburkholderia and Rhodanobacter) were more abundant than in the other samples tested and accounted jointly for 32.21–39.93% of those communities. These bacteria are frequently listed as members of microbiota present in soil (Rawat et al. 2012; Ritpitakphong et al. 2016; Siles & Margesin 2017; Haas et al. 2018; Karimi et al. 2018; Jiao et al. 2019; Lee et al. 2019; Domeignoz-Horta et al. 2019; Bartelme et al. 2020) and dead wood (Tláskal et al. 2017; Bani et al. 2018). Inside L. fuliginosus nests, however, soil particles as well as shredded wood are bound into a cardboard-like building material (carton) (Maschwitz & Hölldobler 1970; Brinker et al. 2019). Thus, the presence of bacteria characteristic of soil and decomposing dead wood in microbiota associated with the nest-carton is perfectly possible. In addition, some genera identified in carton-associated communities may give an indication of the acidity of the soil surrounding the nest. The genera Acidipila and Roseiarcus, found in these profiles, were previously more frequently identified in bacterial communities associated with acidic soil taken from a coniferous forest (Siles & Margesin 2017). The L. fuliginosus nest examined in the present study was located inside a birch trunk growing between oak-hornbeam and alder carr forests, in close proximity to a dystrophic lake with a littoral zone of floating sphagnum mats. Although the pH of the soil surrounding the nest in question was not measured, the soil incorporated into the nest is likely to have been of higher acidity.

Wolbachia was also found in the nest-carton profiles, but its relative abundance was low, not exceeding 1% (Table S1). Like Rickettsia, this well-known endosymbiotic bacteria, can induce mechanisms manipulating its host’s reproductive processes, causing cytoplasmic incompatibility, parthenogenesis, male killing, or male feminization (Stouthamer et al. 1999; Charlat et al. 2003; Correa & Ballard 2016; Pietri et al. 2016). Wolbachia potentially infects more than 65% of insect species (Hilgenboecker et al. 2008; Lewis & Lizé 2015), including a wide range of ant genera, such as Atta and Acromyrmex (Van Borm et al. 2003; Frost et al. 2010, 2014), Camponotus (Ramalho et al., 2018; Ramalho et al. 2017), Cephalotes (Kelly et al. 2019; Reeves et al. 2020), Colobopsis (Ramalho et al. 2017b), Formica (Keller et al. 2001; Viljakainen et al. 2008; Kaczmarczyk-Ziemba et al. 2020), Polyrhachis (Ramalho et al. 2017a), Solenopsis (Ishak et al. 2011; Martins et al. 2012; de Souza et al. 2014), and others (Wenseleers & Ito 1998; Russell et al. 2012; Pontieri et al. 2017). This endosymbiont has also been identified with a high relative abundance in bacterial communities associated with myrmecophilous insects, such as caterpillars or adults of Maculinea butterfly (Patricelli et al. 2013; Di Salvo et al. 2019) and adults of the beetles Dendrophilus pygmaeus, Leptacinus formicetorum, Monotoma angusticollis, Myrmechixenus subterraneus, Ptenidium formicetorum and Thiasophila angulata (Kaczmarczyk-Ziemba et al. 2020). Here, we confirmed the presence of Wolbachia only in the microbiota from the nest material, which consists of dead wood particles chewed by ants. Thus, in view of the limited number of samples analysed, we cannot unreservedly rule out the possible horizontal transmission of Wolbachia and the potential infection of L. fuliginosus ants. This phenomenon has recently been described in insects sharing the same habitats and/or host plants. Li et al. (2017) described plant-mediated horizontal transmission of Wolbachia between infected and uninfected Bamisia tabaci whiteflies. When uninfected individuals fed on the infected cotton leaves (after infected whiteflies fed on leaves, Wolbachia was visualized in the phloem vessels), the majority of them became infected with the symbiont. Moreover, Wolbachia persisted in cotton leaves for at least 50 days. The role of host plants in endosymbiont spread has been also described for two species of Crioceris leaf beetles (Kolasa et al. 2017) and revised by Chrostek et al. (2017). Moreover, there are known groups of insects that inhabit the same environments and share the same or very similar Wolbachia strains (Kawasaki et al. 2016; Kajtoch & Kotásková 2018). For example, bacteria horizontal transfer was observed when infected and uninfected larvae of parasitoid Trichogramma wasps shared the same host egg (Huigens et al. 2004). In turn, Stahlhut et al. (2010) found that ecological associations can facilitate horizontal of Wolbachia within mycophagous Diptera species. On the other hand, not all Wolbachia strains have the same genetic potential for horizontal transmission (Tolley et al. 2019). Our analyses were based on 16S rDNA sequencing, which cannot distinguish closely-related Wolbachia strains (Andersen et al. 2012). For further strain typing, the diversity of the wsp gene, as well as five genes described in the MLST protocol, will need to be analysed (Baldo et al. 2005, 2006; Bordenstein et al. 2009).

Conclusion

This paper presents the results of a study of the bacterial communities associated with the larvae and adults of both L. fuliginosus and the myrmecophilous beetle P. laticollis, as well as the nest-carton used by L. fuliginosus to build chambers inside its nests. The close relationships between these insects, as well as the shared habitat may have been seen as potential factors shaping the similarities in their microbiota. However, our findings show that the bacterial communities associated with both the larvae and adults of both L. fuliginosus and P. laticollis were clustered according to the host’s identity and its developmental stage. The profiles determined for the nest-carton comprised a separate group. The differences between all pairs of groups but one were significant. Only the bacterial communities associated with L. fuliginosus larvae and adults were not significantly different. In these profiles, we identified the presence of two well-known endosymbiotic bacteria – Rickettsia and Wolbachia although the latter was found only in the nest-carton profiles. Our study is in line with a new research trend that is focusing on the microbiota associated not only with ants but also with myrmecophiles and the nests inhabited by those species. There is still a great deal to discover about these systems. Some important questions that need to be addressed are: What factors shape the myrmecophiles’ microbiota profiles? What functional benefits are derived from the associated bacteria? Do close interactions among myrmecophiles and their hosts permit horizontal transmission of microbiota? Can microorganisms associated with nests protect the ant colony and its inhabitants against pathogens? The broad nature of these questions means that multidisciplinary collaborations will be required to address them in the future.

Acknowledgements

The raw sequential reads generated during the current study are available in the Sequence Read Archive (SRA) repository under accession number PRJNA612990 (https://www.ncbi.nlm.nih.gov/sra/). The authors declare that they have no competing interests.

We are very grateful to the staff of the OMICRON Centre for Medical Genomics for their advices.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by grants from the State Forests National Forest Holding, Poland, and the Polesie National Park, Poland, respectively numbered EZ.0290.1.25.2018 and NB 520-2/2018.