Guerrerostrongylus marginalis n. sp. (Trichostrongyloidea: Heligmonellidae) from the Guianan arboreal mouse (Oecomys auyantepui) from French Guiana

Based on the number and arrangement of cuticular ridges and configuration of the dorsal ray, nematode specimens collected from the small intestine of eight Guianan arboreal mice, Oecomys auyantepui (Rodentia: Sigmodontinae), in French Guiana are herein described and characterized. Guerrerostrongylus marginalis n. sp. (Heligmosomoidea: Heligmonellidae) shows a synlophe consisting of more than 40 ridges and a unique bursal arrangement with ray 8 (externo-dorsal) extending to the edge of the bursal margin, and appearing more prominent than the dorsal ray. This bursal arrangement is common in members of Hassalstrongylus Durette-Desset, 1971, but uncommon in the other four species in Guerrerostrongylus Sutton & Durette-Desset, 1991. The placement of the new species in Guerrerostrongylus is based on the number and nature of cuticular ridges and the ray arrangement and symmetry of the caudal bursa. Diagnostic characteristics of Guerrerostrongylus marginalis n. sp. include the length of ray 8 relative to bursal margin, the relative size of the spicules and vestibule, and the number of eggs in the uterus. We propose an amendment to the generic diagnosis of Guerrerostrongylus to modify the characters of the long rays 6 (postero-lateral), rays 8 (externo-dorsal), and dorsal ray as diagnostic, since at least ray 6 appears to be short in two different species in the genus, namely G. ulysi Digiani, Notarnicola & Navone, 2012 and G. marginalis n. sp.


Introduction
Trichostrongyloidea is the richest superfamily of nematodes both in the number of genera and species [5,6]. They infect the stomach and small intestine of all terrestrial vertebrates. Their classification and taxonomy are chiefly based on features of the caudal bursa and synlophe [6,9,10]. Trichostrongyles featuring a caudal bursa type of 2-2-1, oblique axis of orientation of ridges of synlophe, and tails devoid of a spine are typically assigned to Heligmonellidae. These nematodes are found in talpoid insectivores, lagomorphs, and rodents, and have a cosmopolitan distribution [6]. The combination of characters in the caudal bursa and the number and orientation of ridges in the synlophe are used in the identification of genera in this family. Among them, Guerrerostrongylus Sutton and Durette-Desset, 1991 was proposed to include species with a minimum of 40 longitudinal ridges (slender and slightly salient, less numerous toward anterior end); long dorsal ray and ray 6 (postero-lateral); and females with not bent tails, partially covered with an invaginated cuticle [20]. Species included in Guerrerostrongylus share several traits with species in Hassalstrongylus Durette-Desset, 1971: however, the larger number of ridges in the synlophe and the relatively long size of the dorsal ray of the former have acted as reliable characters [7,18,20]. Guerrerostrongylus includes four known species that infect sigmodontine and caviomorph rodents throughout the eastern half of South America. These include the type species G. uruguayensis Sutton and Durette-Desset, 1991, G. zetta (Travassos, 1937), G. gomesae Simões, dos Santos and Maldonado, 2012, and G. ulysi Digiani, Notarnicola, and Navone, 2012. Guerrerostrongylus uruguayensis is found in Oligoryzomys flavescens (Waterhouse) from Uruguay and Akodon simulator (Thomas) from Argentina [2,20]. Guerrerostrongylus zetta (Travassos, 1937) is found in Oligoryzomys nigripes (Olfers), Akodon cursor Winge, Cerradomys subflavus (Wagner), Euryoryzomys russatus (Wagner), Nectomys squamipes (Brants), Oligoryzomys eliurus (Wagner) and the caviomorphs Galea spixii (Wagler) and Thrichomys pachiurus Wagner from Argentina and Brazil [3,18,19,22]. Guerrerostrongylus gomesae Simões, dos Santos and Maldonado, 2012 is found in O. mamorae Thomas in southwestern Brazil [18], whereas G. ulysi infects Sooretamys angouya (Fischer) from northeastern Argentina [3].
Herein we present the description of a new species of nematode that combines characteristics of both Hassalstrongylus Durette-Desset, 1971 and Guerrerostrongylus, yet it is assigned to the latter based on the number of ridges in the synlophe and the relatively long dorsal ray featured by males. The species was collected from the small intestine of the Guianan arboreal mouse, Oecomys auyantepui Tate, an arboreal and graminivorous sigmodontine rodent that is found throughout the Guiana Shield; these rodents are considered medium-sized with an average adult body mass of 40 g [1,23]. In French Guiana, O. auyantepui is known through 60 preserved specimens from a dozen localities (unpublished data), within the large continuous track of primary rainforests; most collected animals have been caught in traps tied to lianas 1.0-1.5 m above the ground. The present findings constitute the first record of any endoparasite in this species of rodent. This new taxon is the fifth species in the genus and the uniqueness of its characters merits an emendation to the diagnosis of the genus that builds up in the variation in the number of ridges in the synlophe and the appearance of the bursal rays. . These mammals were collected using wire-mesh BTS traps and Sherman traps baited with peanut butter and local fruits and were placed in trees at different heights between 1 and 2 m as well as on the ground. The mammals were handled following the ethical chart of the American Society of Mammalogists [17]. Gastrointestinal contents were preserved in 70% ethanol and transported to the laboratory to be examined for helminths. Preservation, clearing, and mounting of parasites followed Pritchard and Kruse [14]. All helminths were preserved in 70% ethanol and kept under refrigeration.
Nematodes were cleared in lactophenol and mounted on temporary slides; all measurements are in micrometers. For each character, the range is given first, followed by the average, coefficient of variation, and sample size (when different from the number of specimens used in the description). All measurements of holotype, allotype, and paratypes are available at http://opensiuc.lib.siu.edu/zool_data/9/. Mammalian specimens used in the helminthological examinations are part of the holdings of the Muséum d'Histoire Naturelle de Genève, Switzerland (MHNG), and the Muséum National d'Histoire Naturelle, Paris, France (MNHN).
Genomic DNA was extracted, isolated, and purified from three vouchered nematodes following standard protocols [12,16]. These aliquots were used as a template to amplify a fragment of the mitochondrial gene coding for the large ribosomal subunit RNA (rrnL); the primers and thermal profile used to complete the reactions, as well as the postamplification processing of these fragments, are identical to those described elsewhere [11,16]. Published sequences of available herpetostrongyles, heligmosomoids, heligmonellids, and viannaids were downloaded from GenBank, aligned using Clustal Omega (http://www.ebi.ac.uk/Tools/msa/clustalo/) and analyzed for phylogenetic signal using Parsimony and Maximum Likelihood as optimality criteria in PAUP* v4.b10 [21]. For the latter, the GTR + G model of evolution -estimated with jModelTest [13] -was enforced. To test for branch support, 1,000 bootstrap replicates were performed using a heuristic search. The posterior probability of all branches was calculated using MrBayes v3.2.5 [15], which ran for 10 million generations with resampling every 1,000 iterations for a final burnin of 25%. The remaining trees were used to reconstruct the consensus. The matrix including the alignment and command lines used in both approaches is available at (http:// opensiuc.lib.siu.edu/zool_data/8/).

Guerrerostrongylus Sutton and Durette-Desset, 1991
Heligmonellidae. Medium-sized worms, with females reaching or exceeding 8 mm. Synlophe with at least 40 continuous cuticular ridges at midbody, sporadically 35 in males. Height of ridges in anterior half of the body unequal, height of ridges of similar size in posterior half. Caudal bursa subsymmetrical, with ample dorsal lobe; ray 6 (postero-lateral) projected posteriad, dorsal ray long; ray 8 (externo-dorsal) usually shorter than dorsal ray. Bursal pattern of type 2-2-1 or 2-2-1 tending to 1-3-1. Genital cone not enlarged. Posterior end of female not bent; vulva opens near posterior end, tail tapers to a blunt end. Etymology: The species name, marginalis, refers to the extension of ray 8 (externo-dorsal), which reaches the posterior margin of the bursa.
Other characters that assist in the discrimination of G. marginalis from other species in the genus include a combination of the relative size of the spicules, size of genital cone, length of the uterus, and size of eggs (Table 1). A comparison against each species follows. First, G. ulysi features a proportionally longer dorsal ray that causes rays 9 and 10 to extend farther posteriad than rays 6 and 8; in G. ulysi the length of the dorsal ray represents 60% of the length of the caudal bursa. Second, G. marginalis can be discriminated from G. zetta in the relative length of rays 6 and 8, in addition, the dorsal ray is 70% of the length of the caudal bursa. Regarding traits in females, the vulva in G. zetta appears to be closer to the posterior end than the vulva of G. marginalis. Third, the dorsal ray in G. uruguayensis is 65% the length of the caudal bursa; in contrast, the genital cone is very small in G. uruguayensis (14 · 9 versus 71 · 72 in G. marginalis). Interestingly, both uterus and vestibule are longer in G. uruguayensis (2,500 and 350, respectively) than the homologous structures in G. marginalis (2,020 and 205, respectively). Finally, the most similar species to G. marginalis is G. gomesae, yet both can be discriminated because ray 5 of G. gomesae appears to be relatively longer than ray 6. In contrast, the spicules as well as the eggs of G. marginalis tend to be larger. The range for spicules is 544-829 (average 717) for G. marginalis and 310-560 for G. gomesae, whereas the range for their eggs is 31-59 · 25-35 and 50-72 · 30-60, respectively. The number of eggs in the uterus of G. marginalis is greater than the number of eggs in G. gomesae. Another notable difference is the length of the vestibule, which is reportedly shorter in G. gomesae than the homologous structure in G. marginalis (Table 1).

Molecular results
The phylogenetic reconstruction based on the mitochondrial gene rrnL is shown in Figure 13. This tree is the consensus resulting from the estimation of the posterior probabilities of the branches. The analysis of the dataset using parsimony and Maximum Likelihood results in six and three trees, respectively. The trees obtained using Maximum Likelihood are essentially the same, since the only difference is the reciprocal position of the specimens identified as G. marginalis. The six trees generated with parsimony as optimality criterion have a length of 1,411 steps and a consistency index of 0.43, resulting Nevertheless, the monophyly of G. marginalis is supported in all three analyses (Fig. 12). This species appears to be clustered with the heligmonellid Hassalstrongylus sp. and Stilestrongylus sp., in a clade that shows a strong support of 100% and a posterior probability of 1.

Discussion
The configuration of the caudal bursa of G. marginalis resembles the homologous structure in some species of Hassalstrongylus. This is because the extension of the dorsal ray appears to be 50% the length of the caudal bursa, ray 8 extends more posteriad than rays 9 and 10, and the extension of rays 4 through 6 gives the caudal bursa the appearance of an irregular trapezoid. The perception of the overall shape of the caudal bursa of G. marginalis seems to differ from the caudal bursa of other members of Guerrerostrongylus, which was described as ellipsoidal, rectangular, or heart-shaped [3,18]. Irrespective of the interpretation of the shape of the bursa, the overall symmetry in all five species is sub-symmetrical as described in Durette-Desset and Digiani [7]. Additionally, the number of ridges in the synlophe, the size variation of these ridges, and the posterior end of the females are typical of Guerrerostrongylus.
The original diagnosis of the genus was based on two species that bear striking morphological resemblances, namely G. uruguayensis and G. zetta. Since its original description [20], the diagnosis has been translated into English [9], yet this diagnosis predates the description of three more species (G. gomesae, G. ulysi, and G. marginalis) that show more variability in some of the characters used for the diagnosis, including the size of the worms, the number of ridges in the synlophe, and the relative size of rays 6, 8, and dorsal (including rays 9 and 10). For example, ray 6 in G. ulysi is not as long as the homologous structure in G. uruguayensis and G. zetta. Furthermore, the number of ridges at midbody in the synlophe of males of G. gomesae can be 36 [18], which is also the case for G. marginalis. Although the proposed changes are minor, the emended diagnosis we present accounts for the variability observed in the number of ridges and the sub-symmetrical shape of the caudal bursa. The direct observation of paratypes Figure 13. Placement of Guerrerostrongylus marginalis n. sp., relative to available heligmonellid nematodes. The phylogenetic tree (based on a fragment of the mitochondrial gene coding for the large ribosomal subunit RNA -rrnL-) represents the consensus used to calculate the posterior probability of the branches. Posterior probability is to the right of each node. Bootstrap support values are indicated to the left of the node, with values for Maximum Likelihood support over those obtained by parsimony. of G. gomesae allows the detection of minor inconsistencies in the measurements of the vagina, vestibule, and sphincters. The range for these measurements is noted in parentheses in Table 1, and it also includes the number of eggs counted in the uteri of two paratypes. Digiani et al. [4] have shown that this value, as well as the length of the uterus, are reliable characters to assist in the discrimination of syntopic species of Hassalstrongylus. This suggests that the statistical analyses of meristic data may yield unexpected useful characters in species discrimination. With the expectation that other scientists can complete these tests, we have made the measurements for the type specimens universally available (http:// opensiuc.lib.siu.edu/zool_data/9).
For the completion of the present work, specimens of G. zetta collected from Oligoryzomys nigripes (Olfers) in Argentina were kindly provided by Dr. Mike Kinsella. Unfortunately, attempts to amplify DNA from these individuals failed, perhaps as a result of their previous contact with clearing reagents. As a consequence, the relationship of G. marginalis with the rest of the species, as well as their placement in Heligmonellidae, remains to be tested.