Introduction

The genus Nesticus Thorell, 1869 is distributed worldwide except for south-eastern Asia and Australia and comprises 125 species and 8 subspecies (Platnick 2011). In Europe Nesticus is represented by 23 species, of which five are known from Iberia. Unlike Nesticus cellulanus (Clerck, 1757), a species with a holarctic distribution, the four others are endemic to Iberia, being cavernicolous species with more or less evident troglomorphic features.

The first species described from Iberia was Nesticus obcaecatus Simon, 1907, found only from a single locality: Cueva del Molino de Aso (Huesca), on the southern slopes of the Central Pyrenees (Simon 1907). The species description was originally based on a single female specimen. The male was described several decades later (Ribera 1979), from a series of both sexes collected from the same cave. The second Iberian endemic is Nesticus lusitanicus Fage, 1931, a native species of the karst landscape in central Portugal. This species was described on the basis of females (Fage 1931), its male was found half of a century later (Ribera 1988). The third species, which was found in several caves in Asturias and Cantabria (north-western Iberian Peninsula) (Ribera and Guerao, 1995), is Nesticus luquei Ribera & Guerao, 1995. Finally, Nesticus murgis Ribera & De Mas, 2003 was described from a cave in the province of Almeria (Ribera and De Mas 2003). Overall, ranges of these species are rather small, in two species (Nesticus obcaecatus and Nesticus murgis) being restricted to a single cave.

This work describes a new cavernicolous species whose distribution includes several karst landscapes of the different mountain formations that make up the Sistema Bético, the ridge in southern Iberian Peninsula.

Materials and methods Taxonomy

Abbreviations used in the text are as follows: PL = prosoma length (from posterior edge of carapace to front edge of clypeus, measured at midline); PW = maximum prosoma width; OL = opisthosoma length (excluding the pedicel); OW = maximum opisthosoma width; MA = median apophysis; Mt = metatarsus; Tb = tibia; TTA = theridioid tegular apophysis; TTA p1 = process 1 of TTA; TTA p2 = process 2 of TTA; TL = total length (excluding the pedicel). Eyes: AM = anterior median; AL = anterior lateral; PM = posterior median; PL = posterior lateral; i = immature; sub = subadult; GEV = Grupo de Espeleología de Villacarrillo.

Female vulva was removed and treated with 30% KOH. After observation and drawing, the vulva was washed in distilled water and stored in 70% ethanol. Left male palps were illustrated in all cases. We follow Coyle and McGarity (1992) for describing the paracymbium, and Huber (1993) and Agnarsson et al. (2007) for other parts of male and female copulatory organs. Holotype and paratypes have been deposited in the Arachnida Collection of the CRBA (Centre de Recursos de Biodiversitat Animal) at the University of Barcelona. Catalogue numbers are given in brackets.

A Nikon Coolpix 4500 digital camera attached to a stereomicroscope was used to capture images. Following Coleman (2003) and with the aid of a Trust Scroll Tablet TB-4200 the samples were drawn with repeated reference to the specimen under the microscope. The specimens used for SEM studies were dehydrated with alcohol gradient dehydration and ultrasonically cleaned. They were then critical-point dried and were mounted and covered with gold and examined using a HITACHI S-2300 Scanning Electron Microscope (SEM) (SCT, Universitat de Barcelona, Spain).

Phylogeny

Taxonomic sampling. Taxa analyzed in the present study are listed in Appendix 1. All the Iberian species are included except Nesticus murgis due to impossibility to obtain fresh material for DNA analysis. Nesticus eremita Simon, 1879 from Croatia and Nesticus ionescui Dumitrescu, 1979 from Romania are also included to test the monophyly of Iberian species. Sequence from Nesticus sp. from China (Arnedo et al. 2004) was also included in the analysis as a more distantly related Nesticus species that was used to root the tree.

Sample Storage and DNA Extraction. Specimens were preserved in 95% or absolute ethanol and stored at 4°C. Total genomic DNA was extracted from legs of a single specimen using the QIamp® DNA Mini Kit (QIAGEN) following the manufacturer’s protocols. The approximate concentration and purity of the DNA obtained were verified using 1% agarose/TBE gel electrophoresis.

PCR Amplification and Sequencing. Two regions of the mitochondrial DNA corresponding to a fragment of the cytochrome oxidase I gene (cox1) and 16S rRNA (rrnL) were selectively amplified using PCR with the following primer pairs: for cox1 C1-J-1718 (5' GGAGGATTTGGAAATTGATTAGTTCC 3') with C1-N-2191 (5' CCCGGTAAAATTAAAATATAAACTTC 3') (Simon et al. 1994); for rrnL LR-N-13398 (5' CGCCTGTTTATCAAAAACAT 3') (Simon et al. 1994) with LR-J-12864 (5' CTCCGGTTTGAACTCAGATCA 3') (Arnedo and Gillespie 2006). The PCR reaction mixture contained a final concentration of 0.2 μM of each primer, 0.2 mM of each dNTPs, 0.5 U Taq polymerase (Promega), with the supplied buffer, and 1.5–2.5 mM Mg Cl2 in a final volume of 25 μL.

A Perkin-ElmerCetus Moldel 480 thermocycler was used to perform 35 iterations of the following cycle: 30s at 95°C, 45s at 45°C, and 1 min at 72°C, beginning with an additional step of 3 min at 95°C, and ending with another step of 10 min at 72°C. PCR results were visualized by means of a 1% agarose/TBE gel. Amplified products were purified using Microcon PCR columns following the manufacturer’s specifications. Purified products were directly cycle-sequenced from both strands using ABI BigDye (Applied Biosystems) chemistry, precipitated in DyeEx Spin kit (Qiagen, Chatsworth, CA) columns, and run out on ABI Prism 377 (Applied Biosystems) automated sequencers. Sequencing reactions were performed in our lab with the forward and reverse PCR primers. Resulting product were run and analyzed at the Serveis Científico-Tècnics of the Universitat de Barcelona.

Alignment. Raw sequences were compared against chromatograms and complementary contigs built and edited using the Geneious Pro 3.6.2 software (http://www.genious.com). Sequences were manipulated and preliminary manual alignments constructed using BioEdit V.7.0.5.3 (Hall 1999). Alignment of the cox1 gene fragments was trivial due to the absence of length polymorphism. However, there were some length differences among the rrnL fragments, suggesting the occurrence of insertion/deletion events during the evolution of these sequences. Automatic alignment algorithms have been considered as superior to manual protocols due to their objectivity and repeatability (Giribet et al. 2002). Automatic alignments for the rrnL data set were constructed with the program MAFFT v 6.240 (Katoh and Toh 2007) The alignment was constructed using the manual strategy option set Q-INS-I, the most accurate multiple sequence alignment, whit default options. All analyses were performed by coding gaps as absence/presence character following Simmons and Ochoterena’s simple coding method (Simmons and Ochoterena 2000), as implemented in the software GAPCODER (Young and Healy 2002). This method allows the inclusion of gap information in phylogenetic inference, minimizing the effect of increasing the weight of overlapping multiple non-homologous gaps that results from scoring gaps as 5th state (Pons and Vogler 2006).

Phylogenetic analyses. Parsimony analyses of the combined data matrices were conducted with the program Winclada v.1.00.08 (Nixon 2002) using the following heuristic tree search strategy: 1000 iterations of 10 Wagner trees constructed with random addition taxa and subsequent TBR branch swapping, holding a total maximum of 10000 trees. This program facilitated combination of the different gene fragments in a single data set for simultaneous analyses and also provided additional statistics for those trees (CI and RI values). Clade support was assessed via Bootstrap (Felsenstein 1985) as implemented in Winclada, based on 1000 bootstrap replicates with 20 iterations and 10 starting trees per replica. Uncorrected genetic distances between taxa of cox 1 gene from terminal taxa were assessed with the program MEGA v.3.0 (Kumar et al. 2004).

Description Discussion

Specimens and sequences, with corresponding Genbank accession numbers, analyzed in the present study, are listed in Appendix 1. Alignments of two mitochondrial genes and gap scores as presence/absence were merged resulting in a combined data matrix of 950 characters (cox1=472, rrnL=456 and 22 gap characters). Uncorrected cox1 genetic divergences among terminal taxa are summarized in Appendix 2. Parsimony analyses of the combined data matrix yielded a single most-parsimonious tree of 705 steps (CI = 73 and RI = 62) (Fig. 1).

The results show that the Iberian species do not constitute a monophyletic group. Nesticus luquei, Nesticus lusitanicus and Nesticus baeticus sp. n. form a clade with high bootstrap support (99%), while Nesticus cellulanus is nested within a clade that also includes Nesticus ionescui, from Romania, and Nesticus eremita, from Croatia. Nesticus obcaecatus is the sister group of the remaining species of the ingroup. This topology, along the high genetic divergences observed between Nesticus obcaecatus, Nesticus cellulanus and the remaining Iberian species suggest the existence of tree independent colonization to the Iberian Peninsula. Preliminary results of a more extensive phylogenetic analysis, including almost all the Mediterranean species of Nesticidae (our work in progress), support this hypothesis.

The morphology of both the male and female copulatory organs of this Iberian group of species (Nesticus luquei, Nesticus lusitanicus, Nesticus murgis and Nesticus baeticus sp. n.) shows important differences as compared to Nesticus cellulanus, the type species of the genus Nesticus, as well as to the Carpathonesticus species. Thus, the absence of paracymbial ramification, the shape and size of the median apophysis, plus clear differences in size and arrangement of the TTA processes constitute the major differences in the males. The number of spermathecae is the most noticeable character in the females.

With regards to Nesticus obcaecatus, significant differences in the shape and structure of the paracymbium, the median apophysis and the TTA (see Ribera 1988), as well as in the shape of the epigynum and the arrangement of the spermathecae (see Fage 1931), indicate that this species is most distantly-related to all the Iberian endemics known to date. On the basis of the aforementioned characters Nesticus obcaecatus seems to be more closely-related to Nesticus idriacus Roewer, 1931 known from the eastern area of the Alps and to Nesticus morisii Brignoli, 1975 known from Italy. Besides, the conformation of the copulatory organs of these three species are very similar to Typhlonesticus absoloni (Kratochvil, 1933) known from Montenegro. Thus, all three species are likely to belong with Typhlonesticus as well. Yet, in order to prove or reject this assumption, a molecular phylogeny of all the Mediterranean Nesticidae is to be performed. Our future research is focused on resolving the phylogeny of the Mediterranean Nesticidae (work in progress) and will include most of the Mediterranean species to test the monophyletic status of current genera.

Nesticus obcaecatus shows highly troglomorphic characters, such as complete depigmentation, reduction of the eye size and their number (only six eyes), and is known from a single cave. These data alongside its phylogenetic uniqueness (basal position and a deep genetic distance from other Iberian congeners) suggest that this species may be considered a relict representative of an old colonization to Iberia, and should be a candidate for protection.

Climatic relict hypothesis assume that adaptation and speciation to caves are mainly driven by climatic factors. The Pleistocene glacial cycles has been identified as de main driver of the evolution of cave-dwelling fauna in the Paleartic (Barr 1968; Vandel 1958, 1964). The uncorrected genetic distances between Nesticus baeticus, Nesticus lusitanicus and Nesticus luquei range between 14.3 and 15.5%. Assuming an average substitution rate for arthropod mitochondrial genes between 2% (DeSalle et al. 1987) to 2.3% (Brower 1994) we can conclude that the origin of these species preceded the Pleistocene glacial cycles and, hence that other climatic or environmental factors may have been responsible for the evolution of these taxa.