An Example of How Barcodes Can Clarify Cryptic Species: The Case of the Calanoid Copepod Mastigodiaptomus albuquerquensis (Herrick)

Background The freshwater calanoid Mastigodiaptomus is a genus with high richness in the Americas and is composed of nine species, seven recorded in Mexico and four that are apparently endemic to small areas. Mastigodiaptomus albuquerquensis is a common, widely distributed species ranging from the southern USA to Central America. This species can be easily identified by a notable butterfly-like sclerotization on the basis of the right fifth leg of males. Nevertheless, morphological differences observed among populations throughout this species distributional range have led to the description of several related species or subspecies, such as M. albuquerquensis patzcuarensis from Lake Pátzcuaro in the Central Plateau of Mexico. Methods Genetic results based on barcodes, morphology based on scanning electron and light microscopy images, and morphometric analyses were used to describe cryptic species within the M. albuquerquensis complex. Results The morphological analyses coincided partially with the genetic markers, suggesting the existence of at least two sibling species: M. albuquerquensis s. str. and M. patzcuarensis. A third species was genetically separated but was morphologically indistinguishable from the M. patzcuarensis group. Conclusions Hidden diversity has been a major problem in establishing real patterns of species distribution and genetic acquisition from megadiverse hotspots such as Mexico, where the Nearctic and the Neotropical regions of the Americas meet. Barcodes can help taxonomists to reveal and formally name these new species. Here, we describe two of three potential species highlighted by the use of barcodes: M. albuquerquensis s. str. in the northern semi-desert and M. patzcuarensis on the Central Plateau at more than 2000 m above sea level.


Introduction
Mastigodiaptomus is a freshwater calanoid genus with nine species in the Americas, four of them (M.montezumae, M. reidae, M. maya, and M. suarezmoralesi) are apparently endemic to Mexico, one is found in Guatemala (M.amatitlanensis) and one (M.purpureus) is from Cuba [1,2].In contrast, Mastigodiaptomus albuquerquensis (Herrick, 1895) is a common and widely distributed species, ranging from the southern USA to Central America [1].In Mexico, this species has been recorded from the Baja California Peninsula to the Yucatan Peninsula [1,3,4].
Mastigodiaptomus albuquerquensis s. l. is easily identified by a butterfly-like sclerotization on the basis of the right fifth leg of males.In the last 50 years, every specimen with this type of sclerotization has been assigned to this species.Barcoding has confirmed earlier suspicions of overlooked diversity [5].The morphological differences observed, such as notable changes in body size, were previously explained by the life history hypothesis, supporting the idea that morphological variations are the result of natural selection in response to particular ecological problems faced by species with wide distributional ranges [6].
Nevertheless, on the basis of subtle morphological characters, the population in Pa ´tzcuaro Lake was assigned to a different subspecies, M. albuquerquensis patzcuarensis [7].Previously Diaptomus lehmeri was described from Mexico City [8], although this species was later synonymized with M. albuquerquensis [9].
However, the genetic analysis of Copepoda demonstrated that under the name ''M.albuquerquensis'' is present a complex of at least two species in Mexico [5].Similarly, after one initial study, ecophysiological analyses, hybridization assays, and fine morphological details together with barcoding demonstrated the validity of another species, Leptodiaptomus garciai Osorio Tafall, 1942, which is micro-endemic to a saline crater lake located at 2365 masl [10].
The goal of the present work was to analyze in detail the morphology of the groups highlighted by the barcodes within Mastigodiaptomus albuquerquensis to ascertain the true status of this species complex and to define the taxonomic status of the surveyed populations.

Ethics statement
We collected from several freshwater ecosystems in Mexico.Zooplankton is not under any protection by Mexican laws; thus, no specific permits for this type of field studies are needed.

Sampling
Biological samples were obtained using a plankton net with a 50 mm mesh; the material was fixed and preserved in 100% ethanol.Samples were collected along a latitudinal gradient between 17 and 31uN, which is within the known distributional range of M. albuquerquensis.
The type locality of M. albuquerquensis (''water reservoir supplying Albuquerque, New Mexico'') [11] appears no longer to exist, but all the original samples were taken around the Rio Grande Basin (named the Rı ´o Bravo in Mexico, which constitutes the physical border between Mexico and the USA).After several surveys in the region, the closest place to the type locality for this species was ,200 km away and is part of the same biogeographic province; thus, we concluded that this material represents M. albuquerquensis s. str.Adults males and females taken here were considered as topotypes for the morphological and molecular analyses.The material examined and accession numbers of the sequences and coordinates are summarized in Table S1.

Morphological observations
The specimen analysis was performed following the current standards for the taxonomic study of diaptomid copepods [12].Structures of each specimen considered to have taxonomic value were dissected, and appendages were mounted in glycerin.Taxonomically important structures were illustrated with the aid of a camera lucida.Some specimens were prepared for SEM (scanning electron microscopy) for the observation of microcharacters.Abbreviations in the descriptive section are as follows: P1-P4, first to fourth swimming legs; Exp, exopod; Enp, endopod; Bsp, basis; s, seta(e); ae, aesthetasc; sp, spine; sps, spiniform process; Fu, furcal rami.
To complement the morphological analysis, the type specimens of M. albuquerquensis patzcuarensis were examined (slides 04057, 04058, 04059, and 04060) as well as specimens of M. albuquerquensis from Arizona, USA (slides 02965, 02966, 02967, and 02968); all of them deposited in the Staatliches Museum fu ¨r Naturkunde, Karlsruhe, Germany.
Finally, meristic magnitudes transformed as the square root of a + a ¯of adult females and males from the different populations were examined with a principal component analysis (PCA) performed with the Multi Variate Statistical Package (MVSP 3.21).
In females, 20 morphological features were considered in the PCA: total body length, prosomal length, first prosomite length, A1 length (posterior or anterior to the furcal rami), urosomal length without Fu, urosomal length with Fu, Fu length, length (L)to-width (W) ratio of the rostral spines, genital double-somite L, genital double-somite W, L/W ratio of genital double-somite, point of insertion of the left spine on the genital double-somite (%), point of insertion of the right spine on the genital double-somite (%), L/W ratio of the left spine on the genital double-somite, L/W ratio of the right spine on the genital double-somite, distance between spines on the right prosomite wings, distance between spines on the left prosomite wings, L ratio of ExpP5/EnpP5, L ratio of EnpP5/distal setae of EnpP5, and the number of eggs carried in ovisacs.A total of 94 adult females from Pa ´tzcuaro Lake, Cuitzeo, Hueco Tanks, Rancho Grande, Laguna Bustillos, Arizona, La Cruz, El Salvador, La Goleta dam, Ignacio Ramı ´rez dam, and km 55 pond were analyzed in the PCA.
For the males, 12 morphological features were considered in the PCA: total body length, prosomal length, first prosomite length, urosomal length without Fu, urosomal length with Fu, Fu length, L/W ratio of the rostral spines, L/W ratio of the 20 th segment of A1, L ratio of the 21 st /20 th segments of A1, L ratio of spines on the 10 th /11 th segments of A1, L ratio of 20 th /21 st segments of A1, L/W ratio of the rostral spines, and L/W ratio of the right spine on the first urosomite.A total of 69 adult males from Pa ´tzcuaro Lake, Cuitzeo, Laguna La Cruz, El Salvador, Hueco Tanks, Rancho Grande, Laguna Bustillos, Arizona, La Goleta dam, Ignacio Ramı ´rez dam, and km 55 pond were analyzed in the PCA.

Molecular markers
A total of 53 specimens were sequenced.Genomic DNA was extracted using a membrane-based approach employing AcroPrep 96, 1 ml filter plates with 3.0 mm glass fiber media over a 0.2 mm BioInert membrane (PALL) [13], or a modified HotShot technique [14].A 658 bp segment of COI was amplified using LCOI490 and HCO2198 primers [15] or the new Zplank primers [16].The 12.5 ml PCR reaction mixes included 6.25 ml of 10% trehalose, 2 ml of ultrapure water, 1.25 ml of 10X PCR buffer, 0.625 of MgCl 2 (50 mM), 0.125 ml of each primer (0.01 mM), 0.0625 ml of each dNTP (0.05 mM), 0.625 ml of Taq polymerase (New England Biolabs or Invitrogen), and 2.0 ml of DNA template.PCR products were visualized on pre-cast agarose gels (E-Gelsß, Invitrogen), and the most intense products were selected for sequencing.Sequence analysis was carried out at the Canadian Centre for DNA Barcoding using standard protocols [17].
Products were labeled with a BigDyeß Terminator v. 3.1 Cycle Sequencing Kit (Applied Biosystems, Inc.) and sequenced bidirectionally on an ABI 3730 capillary sequencer.

Sequence alignment and analysis
Bidirectional sequences were edited using CodonCode v. 3.0.1 (CodonCode Corporation, Dedham, Massachusetts) and uploaded to the Barcode of Life Database (BOLD; http://www.barcodinglife.org).Sequence data, trace files, and primer details for all specimens are available within the BOLD project files Mastigodiaptomus of Mexico (MALB), Zooplankton of Mexico and Central America (ZPLMC), Zooplankton II (ZPII), and Zooplankton III (ZMIII).Accession numbers for BOLD and GenBank are listed in Table S1.
We used the tools provided by BOLD to obtain the Id tree: Sequence divergences were calculated using the Kimura twoparameter (K2P) distance model [18].Neighbor-joining (NJ) trees of K2P distances were created to provide a graphic representation of the patterning of divergence between species (Figure S1).Additionally, to compare these results in Mega 5 [19], sequences with at least 604 bp (except two from Rancho Grande to avoid a singleton, see Figure S1) were aligned with CLUSTAL W using default parameters, and phylogenetic relationships of all haplotypes were then inferred using maximum likelihood analysis.A discrete Gamma distribution was used to model evolutionary rate differences among the sites (five categories (+G, parameter = 0.7043)).All positions with less than 0% site coverage were eliminated.In all cases, a sequence of Mastigodiaptomus cf.nesus was used as the outgroup.

Nomenclatural acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article.This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN.The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix ''http://zoobank.org/''.The LSID for this publication is: urn:lsid:zoobank.org:pub:FF036BBC-2CB6-4C1D-932F-9C8FB8377C82.The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS.

Molecular analyses
DNA barcoding, barcode index numbers (BIN), and 2% of MOTU provided by BOLD suggest the presence of five possible species based on the COI haplotypes (Figure S1, Fig. 1).The results of the K2P analyses in BOLD and the maximum likelihood method were identical.The latter analysis found only one tree (Fig. 1) on which bootstrap percentages are shown.Average values are shown in Table 1.
The first group belongs to the species M. albuquerquensis, found in the northern semi-desert region, and is recognized as the strict form.This group forms a consistent cluster, and all localities for this species are the closest ones to the no longer extant type locality.The second main species is apparently restricted to the Central Plateau in localities close to Lago de Pa ´tzcuaro.This taxon, described as M. albuquerquensis patzcuarensis Kiefer, 1938 [7], is in fact a full species, Mastigodiaptomus patzcuarensis, and also forms a consistent cluster.
A third group suggested by the genetic analyses seems to be sympatric with M. patzcuarensis in systems such as Cuitzeo, and which appears in several other water bodies in the Central Plateau, is here regarded as M. cf.albuquerquensis.Two other groups remaining as possible different species are the specimens from Rancho Grande to Zacatecas, with an average divergence of 4.4% from the main cluster, and two specimens from Papasquiaro B, with an average divergence of 4.6%.As these sequences are shorter than the others and present low-to middle-quality trace files, they are here regarded as members of M. albuquerquensis s. str. on the basis of their morphology and geographic location in the northern semi-desert region.

PCA analysis
From the biplot, a gradient was observed from axis 1 to axis 2, representing the meristic features in the specimens.Axis 1+2 explained 82.6% of the variability in females (Fig. 2A) and 91.8% in males (Fig. 2B).Two defined groups were found in each case: one group included the surveyed specimens from Pa ´tzcuaro, Cuitzeo, La Cruz, and Goleta.The second group included the available specimens from Hueco Tanks, Arizona, Rancho Grande (Zacatecas), Ignacio Ramı ´rez, El Salvador, km 55 pond, and Bustillos (Fig. 2A, B).

Morphological analyses
The morphology of A1, the mouthparts, and thoracic legs in females and males, as well as the fifth leg from the males, are similar along all latitudinal gradients studied.Morphological similarities are described in this section; the detailed morphological differences between the genetically different materials will be described later in this paper.
Swimming legs.P1 with three-segmented Exp and twosegmented End; P2-P4 with three-segmented Exps and Enps.The armature formula of P1-P4 observed in the material analyzed here and shown in Table 2 has been described and illustrated for other Mastigodiaptomus species [1,20].Only two females from Zacatecas (12.5% of the surveyed population) showed two spines on Enp3P4 (on the left side).
From the PCA analyses, the variability is explained mainly by features related to the body length and secondarily by the number of eggs carried by females in axis 1, whereas the length of the A1 related with the Fu is more important in axis 2 (Fig. 2A).
In males, magnitudes related to the body length are important to explaining the variability in axis 1, whereas the L/W ratio of rostral spines is important in axis 2 (Fig. 2B).
In contrast to the genetic analyses, the PCA suggested the presence of only two sibling species with disjunct distributions in Mexico (Fig. 2), both coincident with two of the three main groups according to the barcodes, namely, M. albuquerquensis s. str.and M. patzcuarensis.Females.1.47 to 1.87 mm of total body length, including Fu (Fig. 5A).Short spines on rostrum, 1.5-3.6 times longer than wide, with rounded tip (Fig. 5C, D).Antennule 25-segmented, short, ending at distal margin of anal somite (Fig. 5A); females from Zacatecas with one or two antennular segments beyond Fu (25-75 mm); in females from Ignacio Ramı ´rez, the A1 is from the same level as the Fu to 50 mm beyond it; tiny hair-like setae on ventral surface of each prosomal somite (Fig. 5E, F).

Descriptions of the two species
The dorsal process of the fourth pediger is variable in shape within populations: approximately 33% of the surveyed specimens exhibit a high, frontally concave projection; 17% of females bear a low, convex projection, and the rest have a very low projection (Fig. 5F).
Distal margin of the fifth pediger delicate; central margin rounded (Fig. 5G).Right wing not projected with one dorsal and one ventral spine.Distance between the two spines is from 45 to 90 mm (Fig. 5E).Left wing is strongly projected, bearing spines separated by 32 to 75 mm (Fig. 5F).
The length ratio of Exp/Enp of P5 is 1.9 to 3.2 (average = 2.3), Enp bi-segmented, with long setae apically (Fig. 5I).Two spines on Exp2; Exp3 with spines on medial margin (Fig. 5H).Ornamentation of lateral margin of Exp3 variable: without spines (in the female from Arizona) or with one (in females from Zacatecas as in Fig. 5H) or two spines (in females from Chihuahua).
Double genital somite 1.0-15 (average = 1.2) times longer than wide; egg sacs carrying from 12 to 40 eggs.Left spine inserted dorsally on the proximal region of the somite (approximately 35% of total double genital segment length).Fu with hair-like setae on Table 2. Setation formula for major armament of swimming legs (P1-P4) in the females and males in all the species described here.the medial margin and spine-like setae on the lateral margin (Fig. 5J).
First urosomal segment with one strong, curved spine on right side, averaging 4.6 times longer than wide (Fig. 6G); left side bearing one tiny, hair-like seta only observable laterally.A dorsal, thin spine present on right wing in males from Arizona, Laguna Bustillos, and Hueco Tanks (Fig. 6I); absent in males from Zacatecas (Fig. 6H).Left wing with one strong, ventral spine plus one weak, hair-like seta dorsally (Fig. 6H).
Syn: Diaptomus albuquerquensis patzcuarensis Kiefer, 1938.Material examined.15 adult females and ten adult males, ethanol preserved from Pa ´tzcuaro, Michoaca ´n; 15 adult females and 11 adult males from Cuitzeo, Michoaca ´n; nine adult females and five adult males from La Cruz, Guanajuato; two adult females and six adult males from La Goleta, Estado de Mexico; three adult females and two adult males of the Staatliches Museum fu ¨r Naturkunde Karlsruhe, from Pa ´tzcuaro, Michoaca ´n, Me ´xico (slides labeled as ''Mastigodiaptomus albuquerquensis patzcuarensis, Typus'' by Kiefer number 4057 -one dissected female, 4058two adult females, 4059 -one dissected male, and 4060 -one adult male).
Females.Total body length is 0.9-1.3mm (average = 1.14 mm), including Fu (Fig. 7A).In the type material, it was possible to observe a group of spine-like setae on Enp1 in the antennae (Fig. 7C) and the three toothed pectens on the outer margin of mandibular Enp2 (Fig. 7D).
Rostrum with spines 3.6 times longer than wide (Fig. 7E).A1 long: one to three antennular segments (25 to 175 mm) are beyond the distal margin of Fu (Fig. 7A).Cuticle surface with hair-like setae on ventral and distal margins from second to fifth prosomal somites in females from Pa ´tzcuaro (Fig. 7F) but nude in females from La Cruz, Cuitzeo, and Goleta.Dorsal process of fourth prosomal somite always convex anteriorly, pitted (Fig. 7G).
Exp1 of P5 is 1.7-2.1 times longer than the bi-segmented Enp (average = 1.8) (Fig. 7J).The Enp is 4-7 times longer than the very short apical setae; Exp3 with spines on medial margin and a reduced group of spines (Fig. 7I) or no spines on the lateral margin (Fig. 7K).Genital double somite with parallel scars ventrally (Fig. 7L, M), 1.2 times longer than wide, carrying 2-4 eggs in the sac (Fig. 7N).Left spine inserted dorsal or ventrally in a high position on the genital double somite (approximately 40% of total length segment).Hair-like setae along the lateral and medial margins of the furcal rami; these setae of similar length in both margins (Fig. 7L).
Right wing of fifth pediger with two spines: one well developed ventral spine and one tiny dorsal spine (Fig. 8G).Left wing with similar ornamentation (Fig. 8H).First abdominal segment with one strong or short spine on the right side (Fig. 8I, G).Left side with one tiny seta (Fig. 8H).Right, dorsal projection of fourth abdominal segment not strongly bulbose.Medial margin of Fu pilose (Fig. 8I).

Discussion
The detailed analyses of the morphology described above (and supplemented in Tables 3 and 4) show slight but important differences between the two species, in agreement with the barcoding.
However, the barcoding results suggest the presence of one more species from Central Mexico, herein named M. cf.albuquerquensis, with disjunctive populations from the other two.Although the specimens were analyzed in detail by their morphological attributes, it was not possible to ascertain whether they belong to a different species.It is possible these specimens correspond to another species described as Mastigodiaptomus lehmeri Pearse, 1904 [8], and later synonymized with M. albuquerquensis [21].Unfortunately, type material for M. lehmeri is nonexistent, and the type locality is quite vague, stated simply as Mexico City.All systems with M. cf.albuquerquensis in this study are from localities less than 150 km from Mexico City, except for the specimens from Cuitzeo.To disentangle this latter species, it is necessary to prepare hybridization assays among the three species similar to those previously prepared to uncover Leptodiaptomus garciai, an endemic from a saline crater lake [10].If this result indicates a different biological species, it would be the first Mastigodiaptomus with no morphological features allowing identification to species level.This analysis will be part of a future work due to the need to work with live specimens collected as close as possible to Mexico City.
On the other hand, the micro-structural conformation of anatomic features related to reproduction (i.e., antennules and fifth legs) in female and male freshwater copepods has been informative for species recognition elsewhere [20,22], and these morphological differences are consistent with reproductive isolation [10] or genetic differentiation in the COI gene among species [23].Both genetic analyses and morphology, have been employed to successfully discriminate sibling species of marine and freshwater copepods [24,25], cladocerans [23], rotifers [26], polychaetes [27], and others.These descriptions based on morphology and molecular data, are part of the integrative taxonomy [28], a new approach where different types of evidence converge to support the species discovery that currently is revitalizing this discipline [29].

Morphological features
Following the description of Mastigodiaptomus albuquerquensis [11], the hyaline membrane inserted on the caudal side of right basipodite of the male fifth leg (which is butterfly-like) was widely used to identify to this species regionally [1].However, the genetic analysis in the present survey has shown that at least two different species share the same morphological structure in the fifth leg in males, in the genital field in females, and in all the mouthparts and the antennular ornamentation in both sexes.Therefore, additional features must be considered to distinguish species; for instance, the analysis of tegument features has been suggested to distinguish Mastigodiaptomus species [30].According to this idea, and following the information obtained with the gen COI and PCA analyses, the combination of the following morphological characteristics is present in M. albuquerquensis s. str.: females with hair-like setae on the lateral margin of the Fu shorter than the setae on the medial margin; prosomites with very short hair-like setae on the ventral region; short rostral spines; A1 ending before the distal-most margin of Fu, central margin of fifth prosomite rounded, delicate; dorsal process of fourth pediger variable in shape within the populations; fifth leg endopod bearing a long distal seta; and ovisac carrying a large number of eggs .In addition to the tegumental features present in females, males of M. albuquerquensis have a smooth hyaline membrane on the 20 th right antennular segment, and the spine of the 10 th segment is short, not reaching the distal margin of the bearing segment.
In contrast, females of M. patzcuarensis can be distinguished from those of M. albuquerquensis s. str.by the following combination of characters: the hair-like setae on lateral and medial Fu margins with the same length; prosomites with long, hair-like setae along the distal margin; long rostral spines; A1 ending after the distalmost margin of Fu, central margin of fifth prosomite angled; shape of the dorsal process of the fourth pediger always the same within populations (high with concave anterior margin); fifth leg endopod bearing a short distal seta; and ovisac carrying a small number of eggs (2)(3)(4).Males of M. albuquerquensis have a thin, serrated hyaline membrane on the 20 th right antennular segment, and the spine tip of the 10 th segment is beyond the distal margin of the bearing segment.
For freshwater copepods, the differences in body length have been considered to be due to the limitation of nutrients [31], trophic state [32], or predation [32,33].However, in populations inhabiting tropical latitudes, this variation does not exceed dozens of microns within the same population (across seasonal changes) or between differentially distributed populations.When M. albuquerquensis is compared with M. patzcuarensis, the body length varies by at least 600 mm; a corporal difference of this magnitude can be used as an additional feature to distinguish between these species.

Figure 1 .
Figure 1.Id tree inferred by using the maximum likelihood method as described in the Methods section.doi:10.1371/journal.pone.0085019.g001

Table 4 .
Morphometric characters in the surveyed males.