Delineating species boundaries using an iterative taxonomic approach: The case of soldierless termites (Isoptera, Termitidae, Apicotermitinae)

https://doi.org/10.1016/j.ympev.2013.07.007Get rights and content

Highlights

  • Soldierless termites from Nouragues revealed differences in habitat preferences.

  • Species discrimination based on morphology and COI barcoding matched well.

  • Chemical fingerprints (GC, MALDI) produced groupings incongruent with other methods.

Abstract

Species boundaries are traditionally inferred using morphological characters, although morphology sometimes fails to correctly delineate species. To overcome this limitation, researchers have widely taken advantage of alternative methods such as DNA barcoding or analysis of cuticular hydrocarbons (CHs) profiles, but rarely use them simultaneously in an iterative taxonomic approach. Here, we follow such an approach using morphology, DNA barcoding and CHs profiles to precisely discriminate species of soldierless termites, a diversified clade constituting about one-third of the Neotropical termite species richness, but poorly resolved taxonomically due to the paucity of useful characters. We sampled soldierless termites in various forest types of the Nouragues Nature Reserve, French Guiana. Our results show that morphological species determination generally matches DNA barcoding, which only suggests the existence of three cryptic species in the 31 morphological species. Among them, Longustitermes manni is the only species whose splitting is corroborated by ecological data, other widely distributed species being supported by DNA barcoding. On the contrary, although CHs profiles provide a certain taxonomic signal, they often suggest inconsistent groupings which are not supported by other methods. Overall, our data support DNA barcoding and morphology as two efficient methods to distinguish soldierless termite species.

Introduction

The concept of species has been largely debated and more than 20 definitions, sometimes incompatible, have been proposed (Mayden 1997). Recently, de Queiroz (2007) attempted to reconcile those concepts emphasizing that they all share a common element, and proposed to define species as separately evolving metapopulation lineages. This unifying concept clearly separates the conceptual problem of defining the species category from the methodological problem of species delimitation, i.e. inferring species boundaries (de Queiroz 2007).

Traditional species delimitation is often based on morphological characters, but in many cases, morphology fails to discriminate among closely related species. This is especially true for species in which individuals do not recognize each other based on visual cues. In those taxa, differentiation between species may involve chemical or acoustic cues rather than visual ones, making species morphologically indistinguishable (Bickford et al., 2007, Henry, 1994). Such “cryptic” species have now been revealed in many animal taxa, generally by molecular data (Bickford et al., 2007), such as DNA barcoding.

DNA barcoding was proposed as an easy and cheap method to distinguish between species (Hebert et al., 2003). Species identification relies upon a COI tag sequence and a cut-off dissimilarity value, generally set at 2% for vertebrates and 3% for invertebrates, below which individuals are considered as conspecific and above which they are deemed to represent distinct species (Hebert et al., 2003). The simplicity of the concept is appealing and has attracted many researchers, quickly gathering a huge quantity of DNA barcodes (Hajibabaei et al., 2007). However, there are cases where DNA barcodes fail to identify species well characterized by other methods, e.g. where gene trees do not match species trees due to introgression or retention of ancestral polymorphism (Funk and Omland, 2003). For this reason and some other drawbacks, DNA barcoding has been heavily criticized by its detractors (e.g. Wheeler, 2005, Will et al., 2005), while its proponents underline its ease and advance that the method overcomes morphologically based taxonomy (e.g. Hebert et al., 2004, Packer et al., 2009).

Another commonly used non-morphological taxonomic characters are profiles of cuticular hydrocarbons (CHs), which have been proved useful to distinguish insect species, especially among social insects including termites (e.g. Haverty et al., 1992, Haverty et al., 2000, Haverty et al., 2005, Bagine et al., 1994). CHs, albeit in small amounts, form a substantial part of insect cuticle and are responsible for its waterproof properties. Insect hydrocarbons are in general synthesized de novo, often by modified epidermal secretory cells, the oenocytes (Chapman, 1998, Howard and Blomquist, 2005). At the same time, they are exploited by social insects as important identification tags allowing caste-, age-, or task-related recognition, as well as intra- and inter-specific discrimination (Blomquist et al., 1998, Singer, 1998, Martin et al., 2008). Using CHs profiles for species recognition thus allows comparing directly the chemical characters through which social insects recognize each other. A few studies focusing on termites have also shown the species-specificity of soldier frontal gland secretion (for review see Šobotník et al., 2010) or isoenzymatic composition (Lopes and Ruvolo-Takasusuki 2010) and their usefulness to infer species identity, but the DNA sequences and to a lesser extent the CHs profiles remain the two most common alternatives to morphology-based taxonomy.

Beside the debate surrounding DNA barcoding and its oversimplified species concept, several authors have argued about the necessity of switching over an integrative taxonomy approach. The term “integrative taxonomy” was coined simultaneously by Dayrat (2005) and Will et al. (2005) and refers to the use of independent sets of characters to precisely assess species boundaries. In this study, we followed a multisource approach to distinguish species, combining the results of three disciplines, i.e. morphology, DNA barcoding, and CHs profiling, analyzed separately and compared for concordance (Schlick-Steiner et al., 2010). This comparative approach, testing species boundaries against several lines of evidence, was termed “iterative taxonomy” by Yeates et al. (2011), who defined “integrative taxonomy” as the use of different data contributing together to the delimitation of species boundaries. Here, we will follow Yeates et al. (2011) nomenclature and relate to the term “iterative taxonomy”.

Termites are abundant and diversified in tropical and subtropical ecosystems. Several studies have already pointed out the presence of cryptic species in termites, highlighting that as in many other animal taxa, distinct termite species are sometimes morphologically indistinguishable (e.g. Bagine et al., 1994, Cheng et al., 2011, Davison et al., 2001, Roy et al., 2006, Watson et al., 1989). Additionally, genetic characters have also been used as evidence to synonymize introduced species characterized by a disjunct distribution (Austin et al., 2005, Scheffrahn et al., 2005a, Scheffrahn et al., 2005b). Here, we focus on soldierless termites, mostly represented by soil-dwellers, feeding on degraded organic matter (e.g. Bourguignon et al., 2011a). Soldierless termites constitute about one-third of the termite diversity in African and South American rainforests (Eggleton 2000). Because they lack the soldier caste, species identification must be based on alate imagoes, when collected, and on workers, and requires the dissection and close examination of their digestive tube (Noirot 2001). Such a complicated way to discriminate species, coupled with the cryptic life-style of most species, explains the paucity of studies devoted to their taxonomy (Sands, 1972, Sands, 1999, Fontes, 1986, Bourguignon et al., 2010), leaving a very large gap in termite knowledge.

This study aims at two goals:

  • (1)

    We applied an iterative taxonomic approach, taking advantage of the multitude of methods currently available, to provide an accurate image of species boundaries (Schlick-Steiner et al., 2010, Yeates et al., 2011). We combined morphological examination, DNA barcoding, and CHs profiling using two analytical approaches in order to understand the species diversity and boundaries of the soldierless termite fauna in the rainforest of the Nouragues Nature Reserve in French Guiana. Our first aim was to compare the effectiveness of morphological, molecular and chemical characters for species delimitation.

  • (2)

    The Nouragues reserve is composed of a wide variety of habitats from which we sampled a continuum from the forest growing on poorly drained soil beside a creek, to the low forest growing on the foothills of the Nouragues inselberg (Poncy et al., 2001). We previously showed that different forest habitats differ in their termite fauna, although some species are ubiquitously found throughout a wide range of habitats (Bourguignon et al., 2011b). A recent study demonstrated that generalist springtail soil species often form complexes of cryptic species, more specialized than previously expected (Emerson et al., 2011). Our second aim was therefore to find out whether the ubiquitously distributed soil-feeding termite species are true species, or if they actually are clusters of ecologically specialized cryptic species.

Section snippets

Study site

Termite sampling took place in the Nouragues Nature Reserve (04°05′N, 52°41′W). The site is covered by primary tropical rainforest characteristic of the Guyana Shield. Although the Nouragues research station is dominated by the inselberg, culminating at 430 m, the altitude of sampling sites was between 80 m and 200 m. We sampled termites in 10 parcels around the station to cover the diversity of available habitats (Fig. 1). All the sites but three (H, I and J) consisted in squares of 100 × 100 m

Morphological taxonomy

Altogether, we collected 153 samples representing 31 morphological species (Fig. 2). Most samples were unambiguously assigned to one morphological species, but in a few cases assignment was eased by genetic identification (see below). One sample, G587, was assigned to the Anoplotermes-group sp E species-group without a precise determination at the species level and was later separated as a distinct species based on COI barcode.

Molecular taxonomy

We visually inspected all the sequences and did not find any

Are CH profiles efficient at delimiting species boundaries?

Many tools are now available to assess species boundaries (Seifert, 2009, Sites and Marshall, 2003) and allow taxonomists to delineate species not only based upon morphology but also using genetic, acoustic or chemical characters (Walker, 1964, Hebert et al., 2003, Hebert et al., 2004, Soisook et al., 2008, Howard and Blomquist, 2005, Blomquist and Bagnères, 2010). In this study, we followed an iterative taxonomic approach and used four distinct datasets (one morphological, one genetic and two

Acknowledgments

We are indebted to Kateřina Nováková for assistance with the analysis of cuticular hydrocarbons. We thank the Canadian Center for DNA Barcoding for DNA barcodes acquisition. Financial support was provided by the Belgian FRIA (fellowship to TB), the French CNRS (Nouragues grant), the Czech Science Foundation (P206/12/1093), and the Institute of Organic Chemistry and Biochemistry (RVO: 61388963). JŠ thanks to project 20124364 of Internal Grant Agency of Faculty of Forestry and Wood Sciences

References (59)

  • D. Baracchi et al.

    Medium molecular weight polar substances of the cuticle as tools in the study of the taxonomy, systematics and chemical ecology of tropical hover wasps (Hymenoptera: Stenogastrinae)

    J. Zool. Syst. Evol. Res.

    (2010)
  • J. Bergsten et al.

    The effect of geographical scale of sampling on DNA barcoding

    Syst. Biol.

    (2012)
  • G.J. Blomquist et al.

    Nestmate recognition in termites

  • T. Bourguignon et al.

    Towards a revision of the Neotropical soldierless termites (Isoptera: Termitidae): redescription of the genus Anoplotermes and description of Longustitermes, gen. nov

    Invertebr. Syst.

    (2010)
  • T. Bourguignon et al.

    Feeding ecology and phylogenetic structure of a complex neotropical termite assemblage, revealed by nitrogen stable isotope ratios

    Ecol. Entomol.

    (2011)
  • T. Bourguignon et al.

    Beta-diversity of termite assemblages among primary French Guiana rain forests

    Biotropica

    (2011)
  • R.F. Chapman

    The Insects. Structure and Function

    (1998)
  • S. Cheng et al.

    Evidence for a higher number of species of Odontotermes (Isoptera) than currently known from peninsular Malaysia from mitochondrial DNA phylogenies

    PLoS ONE

    (2011)
  • D. Darriba et al.

    JModelTest 2: more models, new heuristics and parallel computing

    Nat. Methods

    (2012)
  • D. Davison et al.

    Species-level systematics of some Kenyan termites of the genus Odontotermes (Termitidae, Macrotermitinae) using mitochondrial DNA, morphology, and behaviour

    Insectes Soc.

    (2001)
  • B. Dayrat

    Towards integrative taxonomy

    Biol. J. Linn. Soc.

    (2005)
  • K. de Queiroz

    Species concepts and species delimitation

    Syst. Biol.

    (2007)
  • P. Eggleton

    Global patterns of termite diversity

  • B.C. Emerson et al.

    Phylogeny, phylogeography, phylobetadiversity and the molecular analysis of biological communities

    Phil. Trans. R. Soc. B: Biol. Sci.

    (2011)
  • L.R. Fontes

    Two new genera of soldierless Apicotermitinae from the Neotropical region (Isoptera, Termitidae)

    Sociobiology

    (1986)
  • D.J. Funk et al.

    Species-level paraphyly and polyphyly: Frequency, causes, and consequences, with insights from animal mitochondrial DNA

    Annu. Rev. Ecol. Syst.

    (2003)
  • Haverty, M.I., Nelson, L.J., Thorne, B.L., Collins, M.S., Darlington, J.P.E.C., Page, M., 1992. Cuticular hydrocarbons...
  • M.I. Haverty et al.

    Cuticular hydrocarbons of termites of the Hawaiian Islands

    J. Chem. Ecol.

    (2000)
  • Cited by (22)

    • Efficient capture of natural history data reveals prey conservatism of cryptic termite predators

      2016, Molecular Phylogenetics and Evolution
      Citation Excerpt :

      Sequences of COI or COII from reduviid gut extracts with ⩾98% identity to annotated termite sequences were considered to be of conspecific origin. We consider this a conservative approach as other studies using COI or COII sequences in a general mixed Yule-coalescent model for species delimitation of various termites found mean distances between species of ∼9–13% divergent, although in one case, a purported interspecific distance can as similar as 98.5% identical (Bourguignon et al., 2013; Monaghan et al., 2009; Roy et al., 2014). Separate gene regions were aligned using MAFFT (E-INS-i strategy) and concatenated using SEQUENCEMATRIX v1.7.8.

    View all citing articles on Scopus
    View full text