Correspondence: Manuel Mota, ICAAM – Instituto de Ciências Agrárias e Ambientais Mediterrânicas, Departamento de Biologia, Universidade de Évora, Évora, Portugal. Tel.: +351 266 760 881; fax: +351 266 760 914; e-mail: email@example.com
Pine wilt disease (PWD) has a tremendous impact on worldwide forestlands, both from the environmental and economical viewpoints. Monochamus sp., a xylophagous insect from the Cerambycidae family, plays an important role in dissemination of the pinewood nematode, Bursaphelenchus xylophilus, the primary pathogenic agent of PWD. This study investigates, for the first time, the bacterial communities of Monochamus galloprovincialis collected from Portuguese Pinus pinaster trees and B. xylophilus free, using a metagenomics approach. Overall, our results show that natural bacterial communities of M. galloprovincialis are mainly composed by γ-proteobacteria, Firmicutes and Bacteroidetes, which may be a reflection of insects' feeding diet and habitat characteristics. We also report different bacterial communities' composition in the thorax and abdomen of M. galloprovincialis, with high abundance of Serratia sp. in both. Our results encourage further studies in the possible relationship between bacteria from the insect vector and B. xylophilus.
Bacteria and insects have evolved together, developing specialized and interesting interactions, covering all types of biological interactions (Priest and Dewar, 2000). Wood-boring insects, such as cerambycids (Coleoptera: Cerambycidae), carry, in their gut, commensal bacteria intrinsically linked with specific roles in their lifestyle, mostly related with feeding and multitrophic interactions (plants/animals) (Dillon & Dillon, 2004). Other features such as vitamin biosynthesis, pheromone production, nitrogen fixation and degradation of tree compounds are also an essential bacterial contribution for wood-boring insect adaptation (Xu & Gordon, 2003; Raffa et al., 2005; Morales-Jiménez et al., 2009).
The pine sawyer beetle Monochamus spp. (Coleoptera: Cerambycidae) is a fundamental element in the epidemiology of pine wilt disease (PWD) (Mamiya, 1983; Linit, 1988), responsible for the dissemination of the pinewood nematode (PWN) Bursaphelenchus xylophilus (Mota & Vieira, 2008). Since the beginning of the twentieth century, PWD is considered one of the most severe diseases affecting coniferous trees worldwide (Jones et al., 2008; Vicente et al., 2012a). The first detection was in Japan in 1905, later spreading to China, Korea and Taiwan in the mid-80s (Mamiya, 1988). In Europe, PWD was first detected in continental Portugal, in 1999 (Mota et al., 1999), and more recently, in Spain (Robertson et al., 2011) and in the Portuguese island of Madeira, 1000 km SW of continental Europe (Fonseca et al., 2012). Different species of Monochamus have been described as vectors of the PWN, the most common being M. alternatus in East Asian forestlands (Mamiya, 1983) and M. galloprovincialis in Europe (Sousa et al., 2001; Naves et al., 2007). During spring season, adult insects are attracted by volatile organic components of stressed trees and fresh cut logs for mating and oviposition (Szmigielski et al., 2011). After hatching, the larvae of Monochamus begin constructing a gallery and feeding on woods' cells until reaching the cambial layer of the tree, where they begin pupation (Linit, 1988). PWN juveniles are chemically attracted to the pupae, where they aggregate and molt to a non-feeding state, the dauer larvae. At this stage, the PWN enters the insect vector (callow adult stage) and concentrates mainly in the tracheal system (Kobayashi et al., 1984). Low numbers of PWN have been recorded in the abdomen, appendages and head (Linit, 1988). The insect emerges from wood and flies to a new host tree for feeding. Upon arrival, the dauer juvenile exits the insect through the spiracles and infects the new host via insect-feeding wounds (Linit, 1990).
Although PWN is considered the primary pathogenic agent of PWD, its associated bacteria have been suggested to play an active role in disease development. Several studies demonstrate that PWN-associated bacteria could increase PWN virulence and induce PWD symptoms in various pine hosts (Oku et al., 1980; Han et al., 2003; Vicente et al., 2012b). Different bacterial species have been isolated from B. xylophilus (Table 1), mostly belonging to Enterobacteriaceae and Pseudomonadaceae families. The origin of these epiphytic bacteria harbored by the PWN is still unknown, possibly obtained from within the pine tree or carried by the insect vector. Furthermore, almost nothing is known about the diversity and ecology of microorganisms associated with Monochamus spp. (Park et al., 2007), and in the particular case of Monochamus galloprovincialis, the subject has never been investigated. The main aim of this study consisted in the characterization of bacterial communities associated with M. galloprovincialis collected from Portuguese Pinus pinaster trees, using a metagenomics approach. Moreover, we have analyzed the bacterial diversity and their relative abundances in the thorax and abdomen of M. galloprovincialis, which represent the most significant PWN location sites inside the insect vector (Kobayashi et al., 1984; Linit, 1988).
Table 1. List of bacterial species found worldwide (China, Japan, Korea and Portugal) associated with Bursaphelenchus xylophilus
Insects were collected from Atalhada, a PWD affected area in Central Portugal (Penacova, north of Coimbra). The area, a 25-year-old pine stand, is almost exclusively composed of P. pinaster (95%). Despite PWD, no other phytosanitary issues were recorded from the collection site. During the summer season of 2011 (August/September), multifunnel traps were placed in the extremities of a few tree canopies. Besides M. galloprovincialis, other insects inhabiting P. pinaster such as Ips sexdentatus, Tomicus minor, T. piniperda and many unidentified species were also collected. After 15 days, the content of each trap was collected, placed inside sterile plastic containers and stored at 4 °C. Adult M. galloprovincialis (n = 8) were surface sterilized with 70% ethanol for 1 min, rinsed in sterile distilled water and dissected in sterile 1X phosphate-buffered saline (PBS) solution. The insect body, slightly crushed, was dipped in sterile PBS solution, remaining 24 h in a Baermann funnel (Viglierchio & Schmitt, 1983). Prior to insect DNA extraction, insects were checked for the presence of the PWN. For such purpose, the obtained suspension was screened under a binocular scope (Olympus SZX-12; Olympus Corporation, Tokyo, Japan). Afterward, the insect was divided into three parts: head, thorax and abdomen. The thorax and abdomen were selected for bacterial characterization due to its significance in PWD (PWN location spots in the insect vector) (Kobayashi et al., 1984; Naves et al., 2006). The head was discarded. Separately, the insect thorax and abdomen were squeezed with a sterile plastic pestle in 1X PBS solution and the exoskeleton removed. The remaining suspension was transferred to a 1.5-mL sterile tube, centrifuged at low speed (1000 r.p.m., 1 min) to pellet insect tissues, and the supernatant collected for bacterial DNA extraction. All the above procedures were conducted under aseptic conditions (The Baker Company, Sanford) to avoid possible airborne contaminations.
DNA extraction and PCR amplification
DNA extraction was performed using Purelink Genomic DNA kit according to the manufacturer's instructions (Invitrogen, Life technologies). The 16S rRNA gene was PCR amplified using the following set of primers: w024 (5′-GCRAACVGGATTAGATAC-3′) and w018 (5′-GAGTTTGATCMTGGCTCAG-3′) as forward primers and w002 (5′-GNTACCTTGTTACGACTT-3′) as reverse primer, for partial and complete 16S rRNA gene of bacteria (Godon et al., 1997). PCRs (50 μL) were prepared as follows: 10–100 ng total DNA, 0.1 μM of each primer, 0.2 mM dNTPs, 2.5 mM MgCl2, 1 U Taq polymerase (Fermentas, Thermo Fisher Scientific Inc.) and 1X Taq buffer with KCl. A ‘touchdown’ (TD) PCR was performed to increase specificity of amplification and to reduce formation of by-products (Korbie & Mattick, 2008). The reaction conditions were as follows: 95 °C for 3 min; 1st phase TD: 10 cycles (95 °C for 1 min, 58 °C for 1 min and decreasing 1 °C per cycle until 48 and 72 °C for 1 min); 2nd phase TD: 20 cycles (95 °C for 1 min, 48 °C for 1 min, 72 °C for 1 min); and final extension at 72 °C for 10 min. PCR products were purified using illustra GFX™ PCR DNA and gel Band purification kit, following the manufacturer's instructions (GE Healthcare, Life Sciences, Piscataway).
Gene library construction, screening and sequencing
Purified PCR products were ligated to pCR®4TOPO®vector (Invitrogen, Life Technologies) and transformed in competent Escherichia coli Top-10 cells (Invitrogen, Life Technologies) according to the manufacturer's instructions. For each 16S library, successfully recombinant clones were selected for further analysis. Plasmids were extracted using NZY Miniprep kit (NYZtech Lda, Portugal) and analyzed for insertion using EcoRI (NYZtech Lda) restriction. Sequencing was conducted at Macrogen (Europe) by capillary sequencing using M13 primers.
Sequences were checked for chimera artifacts using decipher (Wright et al., 2012) and compared with nonredundant GenBank library using blast search (Johnson et al., 2008) and Ribosomal Database Project 10.0 (Cole et al., 2005). Sequences were aligned using NAST_ALIGN tool from greengenes (DeSantis et al., 2006). A distance matrix was created using dnaml (DNA maximum likelihood) option of dnadist from greengenes (McDonald et al., 2012). Operational taxonomic units (OTUs) were calculated using mothur (Schloss et al., 2009) and as input file the distance matrix previously generated. To infer adequacy of clone sampling, rarefaction curves (RC) were constructed at distance levels of 20%, 10%, 5%, 3% and 1%, considered to represent the phylum, class, genus, species and strain levels (Schloss et al., 2004). Nonparametric richness estimator Chao 1 was calculated to infer the diversity of the community (Chao et al., 2004). Chao 1 estimator uses the number of singletons (OTUs represented by only one individual in a sample) and doubletons (OTUs represented by two individuals in a sample) to estimate the diversity of a given environment (Bohannan & Hughes, 2003).
Analyzed sequences and bacteria type strains sequences were aligned using clustalw in bioedit 7.0 (Hall, 1999). Maximum likelihood analyses and selection of the appropriate model of sequence evolution were performed using mega 5.0 (Tamura et al., 2011). The model K2+G (Kimura 2-parameter + gamma distribution) was selected for tree construction. The Archaea Haloquadratum walsbyi DSM16790 was used as outgroup. The stability of the tree branches was assessed by the bootstrap method using 1000 replicates.
Nucleotide sequence accession numbers
Nucleotide sequences were deposited at GenBank (NCBI) under the accession numbers KC253403-KC253895.
A total of 492 partial sequences (750–1200 bp) of 16S rRNA gene were considered in this study. RC at 97% (species level) and 95% (genus level) were calculated for insect total bacterial content and in the thorax sampling (Fig. 1a). RC for insect abdomen could not be computed, probably due to the similarity of sequences and dominance of a single phylotype. At 95–97% distance level, RC slope for total sampling and thorax clones is nearly reaching the flattened phase indicating sufficiency in the number of clones sampled. Likewise, the collector's curve for Chao 1 estimator at 3% and 5% difference between sequences had leveled off, indicating that most species had been sampled (Fig. 1b). In addition, the RC analyses showed 18 OTUs observed at species-genus level.
Bacterial communities associated with M. galloprovinciallis are mainly composed by Proteobacteria (78.5%), followed by Firmicutes (20.8%) and Bacteroidetes (< 1%). From the phylum Proteobacteria, three classes were present: γ-proteobacteria (87.9%), β-proteobacteria (11.6%) and α-proteobacteria (0.5%). The most abundant genera of Proteobacteria were Serratia (76.4%), followed by Janthinobacterium (11.6%), Rahnella (5.0%), Pseudomonas (3.6%) and Nevskia (2.1%). In < 1%, other genera were found: Klebsiella, Morganella, Providencia, Stenotrophomonas and Sphingomonas. Among Firmicutes, the genera found were as follows: Bacillus (95%), Paenibacillus (3%), Lactococcus (1%) and Lysinibacillus (1%). From the phylum Bacteroidetes, representatives were identified as Sphingobacterium (n = 1), Sediminibacterium (n = 1) and uncultured Bacteroidetes (n = 2).
In the thorax of M. galloprovincialis, the most abundant genera were as follows: Serratia (44%), Bacillus (29%), Janthinobacterium (13%), Pseudomonas and Providencia (4%), and Nevskia (2%) (Fig. 2). Other genera were found in lesser relative abundance (< 2%): Bacteroidetes, Lactococcus, Lysinibacillus, Morganella, Paenibacillus, Photorhabdus, Rahnella, Sphingomonas, Sphingobacterium and Stenotrophomonas. Although with a lower number of clones in the insect abdomen, the most abundant genus was Serratia (95%), followed by Bacillus (2.5%), Klebsiella (< 1%), Janthinobacterium (1%), Rahnella (< 1%) and Sediminibacterium (< 1%). Three main genera were found to be common in the thorax and abdomen of M. galloprovincialis mainly Serratia, Bacillus and Janthinobacterium, although with different relative abundances (Fig. 2).
The phylogenetic affiliations of the most dominant bacteria found in M. galloprovincialis are presented in Fig. 3. In Proteobacteria, the sequences of the genus Serratia were predominantly from S. marcescens, Pseudomonas sequences were identified as P. fluorescens, Rahnella sequences were assigned to R. aquatilis, Nevskia sequences belonged to N. soli and sequences of Janthinobacterium genus were J. agaricidamnosum. In relation to Firmicutes, Bacillus sequences belonged to B. weihenstephanensis, the only representative of Lactococcus was L. garvieae, and Paenibacillus sequences were identified as P. kribbensis. From the Bacteroidetes, sequences from Sediminobacterium are described as S. salmoneum, the one species described so far in this genus. An uncultured Bacteroidetes T26-51 could not be assigned to a specific species, yet the respective sequence clustered together with S. salmoneum.
Based on 16S rRNA sequencing analysis, M. galloprovincialis bacterial communities show a predominance of the genera Serratia, Bacillus and Janthinobacterium in both thorax and abdomen. Previous culture-dependent studies have also reported isolation of Serratia marcescens in M. alternatus (Ma et al., 2009), Aerobacter aerogenes and Bacillus cereus var. mycoides in M. scutellatus, M. notatus and M. marmorator (Soper & Olsen, 1963). Other bacteria genera were found as well, in less abundance, in M. galloprovincialis namely: Bacteroidetes, Lactococcus, Lysinibacillus, Morganella, Paenibacillus, Rahnella, Sphingomonas, Sphingobacterium and Stenotrophomonas in the thorax; in the abdomen Sediminibacterium, Klebsiella and Rahnella. To check whether clone sampling was enough to infer the bacterial diversity in M. galloprovincialis, RC analysis was conducted. In general, the results showed sampling sufficiency. The low percentage of some species could be a reflection of low abundance in the community or bias introduced by DNA extraction and/or PCR amplification (Schloss et al., 2006), although TD PCR was chosen to overtake these situations as reported by Cottrell et al. (2005).
Most studies of cerambycids microbiota are related with gut-bacterial communities due to their importance in the insect's biology and ecology. Essentially, the research in this field showed that insect microbial communities are limited by specific niche characteristics. Grünwald et al. (2010) compared wood-boring beetles with different diet/feeding area preferences (rotten softwood; bark of hardwoods; bark of softwoods) and showed that microbial communities were significantly distinct. Our results show a predominance of γ-proteobacteria in M. galloprovincialis, which might be intrinsically related with their feeding diet and habitat characteristics. The presence of Serratia in insects is well documented, both as symbiont as well as pathogen, and it is related with their fitness ability to resist antibacterial substances ingested by the insect, as well as the powerful enzymatic cocktail produced (chitinases, lecithinases, and proteinases) (Grimmont & Grimmont, 2006). Our results indicate a high-density population of Serratia spp. within the M. galloprovincialis bacterial community. In contrast to these findings is the fact that Serratia spp. (S. marcescens) has been presented as a natural enemy of M. alternatus (Nakamura-Matori, 2008). Further studies are needed to understand their functional contribution to the bacterial community structure of M. galloprovincialis.
Many studies have suggested a role for PWN-associated bacteria in PWD (Han et al., 2003; Proença et al., 2010; Vicente et al., 2011), but not many have paid attention to the bacterial communities within the insect vector, how it may influence insect lifestyle and development as well as the interactions with other organisms such as the PWN. In this work, we have analyzed PWN-free M. galloprovincialis, indicating that the bacterial communities obtained are natural to this insect. In light of all knowledge regarding bacterial communities of B. xylophilus (Table 1) and the results here presented, it is tempting to establish the hypothesis that perhaps the PWN can harbor bacteria from the insect. An example is the predominance of Serratia in M. galloprovincialis and also B. xylophilus, as reported by our previous study (Vicente et al., 2011). Vicente et al. (2012b) described that some PWN-associated bacteria, including Serratia spp., were able to degrade cellulose, an advantage in the adaptation and colonization of wood tissues (Harakava & Gabriel, 2003). Cheng et al. (2013) showed the presence of enzymes involved xenobiotics metabolism (α-pinene) in the genome of Serratia sp. M24T3 isolated from B. xylophilus (Proença et al., 2012). Although with this study, it is not possible to establish a comparison between insect vector and PWN bacterial communities, the results presented are useful and encourage future work in this subject. Understanding the role of bacterial transmission in the PWD complex will bring important knowledge for future prospects in the disease management and control.
This work was supported by the Portuguese national scientific Portuguese national scientific agency FCT (Fundação para a Ciência e Tecnologia)/project PTDC/BIA-MIC/3768/2012 (FCOMP-01-0124-FEDER-028368), the European Project REPHRAME – Development of improved methods for detection, control and eradication of pine wood nematode in support of EU Plant Health policy, European Union Seventh Framework Programme FP7-KBBE-2010-4; by the Chubu Science and Technology Center fellowship to Cláudia Sofia Leite Vicente; and by FEDER Funds through the Operational Programme for Competitiveness Factors – COMPETE and National Funds through FCT – Foundation for Science and Technology under the Strategic Project PEst-C/AGR/UI0115/2011.