**1. Introduction**

Ideally, profound knowledge of biodiversity is the first step before any conservation action, sustainable managemen<sup>t</sup> or biological study is carried out in a given area [1,2]. The scientific community is not only aware that the world is facing a major extinction event [3], but also that many lineages are disappearing even before becoming known to science [4,5]. In the last decades, the use of molecular data has helped to improve our knowledge about biological diversity and the use of genetics as a tool for species recognition is now routine. Molecular data are more often used every day and are applied to the species identification [6–8] and species delimitation [9–12] of all kinds of living organisms, with applications even in food control quality [13].

In this context DNA barcoding analysis is the examination and comparison of short and stable fragments of DNA (DNA barcodes), usually mitochondrial, that represent genetic identifiers for a species [14], and have proven to be a reliable technique for taxonomy [15–17]. Nevertheless, researchers have to be cautious because despite being a power tool, barcoding analysis potentially can lead to misinterpretations if sequences used for comparison were generated from misidentified specimens [18]. Therefore, it is strongly recommended that molecular genetic tools are complemented

with morphological, bioacoustics, and ecological data [19]. In conclusion, molecular genetics open a path for more detailed taxonomic studies [20].

For biological works that are concerned with the central portion of South America, Paraguay is critical since the country is located in a confluence zone of di fferent ecoregions, such as Cerrado, Pantanal, Atlantic Forest, Chaco (Humid and Dry), and Southern Cone Mesopotamian Savanna, each of them having its own distinct origin and evolutionary history [21,22]. Additionally, Paraguay is key for works in northern Argentina, Uruguay, southwestern Brazil, and Bolivia. In spite of this biogeographical importance, Paraguay has been poorly explored, and in the current era of molecular genetics, the investigations that include genetic samples from Paraguay are extremely rare in herpetology. For instance, the natural history museum of Paraguay (Museo Nacional de Historia Natural del Paraguay) started its tissue collection for genetic analyses in this decade, whereas other neighbor countries began cryo tissue collections already some decades ago.

The herpetofauna from Paraguay, and specifically the squamate diversity, is still poorly known, evidenced by the fact that even in the last decade, and without the help of molecular tools, several new records for the country were made (e.g., *Ophiodes fragilis*, *Epictia vellardi*, *Chironius exoletus*, *Lygophis paucidens*, *Philodryas livida*, and *Micrurus silviae*) [23–28] and some species new to science were described (*Tropidurus lagunablanca* Carvalho, 2016; *Tropidurus tarara* Carvalho, 2016; *Tropidurus teyumirim* Carvalho, 2016; *Ophiodes luciae* Cacciali & Scott, 2015; *Phalotris normanscotti* Cabral & Cacciali, 2015) [29–31]. The incorporation of molecular genetics in taxonomy opened new pathways leading to a higher resolution in species delimitation, identifying several cryptic species. Some herpetologists in the region included genetic samples of Squamata from Paraguay (Gamble et al., 2012, Werneck et al., 2012, Morando et al., 2014, Recoder et al., 2014) although only very occasionally [32–35].

In 2015 we started a project of barcoding the reptile fauna (Squamata specifically) of Paraguay. It is important to mention here that the two genetic markers commonly used for barcoding analyses are the mtDNA genes 16S rRNA and Cytochrome Oxidase Subunit I (COI) [15,16,36–38] and both markers seem to work rather equally good for species identification. However, for the South American herpetofauna, 16S was more used than COI, and it is therefore better represented in GenBank for comparison. In addition, studies with 16S have shown not only good results in species recognition but also the systematic relationships among related species [37,39,40]. Thus, we decided to use sequences of the mtDNA gene 16S in our barcoding analysis.

During the project we gathered a lot of information about squamate diversity, and as a product of this work, some papers were published [41–45]. The use of DNA barcoding o ffers a starting point for recording the number of species that occur in a given region. Our results show how the use of DNA barcode data can augmen<sup>t</sup> and increase the accuracy of herpetological inventory surveys. Our barcoding study of the Paraguayan Squamata reveals the depth of taxonomic diversity in this country. Furthermore, our DNA barcode data represent the so far most comprehensive DNA barcode reference library for lizards and snakes of Paraguay. These reference data provide the scientific community with resources of numerous possibilities, ranging from species inventories, species identifications, taxonomic studies to wildlife tra fficking.

## **2. Materials and Methods**

## *2.1. Study Area*

Paraguay is located in the center of South America (Figure 1) between parallels 18◦18- and 27◦30- S, and the meridians 54◦19- and 62◦38- W; with a total surface of 406,752 km2. The country is divided by the Paraguay River into two portions: the Occidental Region (commonly known as "Chaco") with an area of 246,925 km<sup>2</sup> (60.7% of the country), and the Eastern Region (or Oriental Region) with a surface of 159,827 km<sup>2</sup> (39.9% of the country).

**Figure 1.** Administrative divisions of Paraguay, showing the two regions known as "Occidental Region" or Chaco (light brown) and "Oriental Region" (light green), divided by the Paraguay River (highlighted in blue in the map).

The topography of the Chaco region is a flat savanna with few small isolate hills in the center/north and east. Bad drainage creates vast flooded areas especially south and east, north-west has some dunes formation and is extremely dry. The oriental region is undulated with hills and the highest point in the center is well irrigated by tributaries of the Paraguayan and Parana rivers basins.

The climatic conditions vary in a northwestern–southeastern gradient, being more humid and cooler in the southeast. The mean annual temperature in the whole country is about 23 ◦C, being 24.5 ◦C in the western region and 22.5 ◦C in the eastern region. It is important to note that there are two seasons, the wet season, in which it rains frequently, coincide with the warm period from October to April and the dry season from May to September, where rain is less frequent and coincides with the coldest period.

There is a big difference with respect to the variation in temperature, given that the mean maximum is 25 ◦C, but the absolute maximum temperature could reach around 50 ◦C, especially in the northwestern portion of the country in January or February. The coldest month is July, and the absolute minimum temperature can be −6 ◦C in the south. Nevertheless, in Paraguay the "true" winter usually does not last longer than 16 days each year. Thus, Paraguay has a warm/hot climate during most parts of the year.

## *2.2. Data Collection*

Even though there are some blanks in the areas sampled, the coverage of collecting sites in this study is rather vast (Figure 2). Nevertheless, there are two ecoregions in the Occidental Region (Cerrado Chaqueño and Médanos del Chaco), from where we have no samples. The methods used in the field were the traditional techniques for herpetology: Active searching at di fferent times of the day and night, examining potential shelters (e.g., barks, logs, caves, mud, leaf litter, etc.) [46] (Figure 3). Fast moving lizards (e.g., *Ameiva* and *Teius*) were collected using compressed air rifles [47]. Additionally, some habitats, such as ant nests and swamps, were dug looking for hypogeal organisms [48], and floating vegetation was sampled using a trawl net (Figure 3). In total, 147 days of fieldwork were accounted for this project, and about 400 specimens collected.

It is important to highlight that the exotic lizard *Hemidactylus mabouia* is currently widely distributed in the country and now is part of the Paraguayan herpetofauna, and thus also included in the study. These genetic data may help in future studies about the colonization of the species, which in Paraguay has been recorded in the Concepción, San Pedro, Central, Alto Paraná, and Itap úa departments [49].

Reptiles that were captured alive were euthanized with a pericardial injection of a solution of embutramide, mebezonium iodide, and tetracaine hydrochloride (T-61 ®, Intervet International GmbH, Unterschleissheim, Germany) or Sodium Thiopental (Tiopental Sódico ®, Biosano, Chile). The Secretaría del Ambiente from Paraguay (Currently "Ministerio del Ambiente y Desarrollo Sustentable") authorized the collecting of specimens though permits SEAM [Secretaría del Ambiente] N◦ 004/11 and 009/2014. Exportation permits for tissues and specimens were also issued by the same authority through the permits SEAM N◦ 002/14, 016/2016, and 084/2016.

**Figure 2.** Ecoregions of Paraguay, showing the collecting sites (brown dots) for this project.

**Figure 3.** Sampling methods during fieldwork included diverse techniques to search in different environments.

After euthanasia, tissue samples were taken either from the muscle of the thigh, tongue, finger clips, tail (when regenerated), or liver. Tissues were preserved in vials containing 98% non-denatured ethanol, and stored at −20 ◦C as soon as possible.

Hemipenes of Squamata were everted after euthanasia, with an injection of 70% ethanol after manually everting the organs. All specimens were fixed with a solution of 36% formalin and 96% ethanol in the proportion of 5:1000 (e.g., 5 mL formalin in 1 L ethanol), injected in the body cavity, thighs, and thickest part of the tail. Following fixation, the specimens were maintained in 70% ethanol.

## *2.3. Molecular Protocols*

We used two different methods of DNA extraction. For sets containing few samples (usually eight or fewer), we used the DNeasy® Blood & Tissue Kit of Qiagen® (Hilden, Germany), whereas for sets of 96 samples we used the fiberglass plate [50]. Both methods are detailed below. The DNA was isolated from tissues whenever possible, or taken from preserved specimens that had been stored for a considerable time in 70% ethanol at room temperature in some cases.

For the DNeasy® method, we used tissue fragments of ~2 mm2. When buffers formed precipitates, they were warmed up at 56 ◦C before use. All reagents for this protocol are included in the kit. Tissues were digested adding 180 μL (all values are for individual samples) of ATL Buffer and 20 μL of proteinase K. Samples in that mix were incubated in a rocking platform at 56 ◦C for 4 to 12 h until the tissue was completely lysed.

Following digestion, 200 μL of AL lysis Buffer + 200 μL of ethanol (98%) were added. This mix was centrifuged (8000 rpm) in DNeasy® Mini spin columns, discarding all the flow-through. Then, 500 μL of AW1 washing Buffer was added and centrifuged (8000 rpm) discarding the flow-through. Finally, 500 μL of AW2 washing Buffer was added and centrifuged (14,000 rpm) discarding the flow-through. The final elution was made with 200 μL of AE Buffer, after an incubation of one minute, followed by centrifugation (8000 rpm).

For the fiberglass extraction method, we used tissue fragments of ~1 mm2. Specifications of reagents used in this protocol, are detailed in Table S1. Initially, the samples were washed with 50 μL (values per sample) of a solution of 1× Tris-Ethylenediaminetetraacetic acid (TE) Buffer to remove the remaining ethanol, for ~15 h. Following, the samples were digested with 50 μL of a solution of Vertebrate lysis Buffer and proteinase K (10:1), and incubated in a rocking platform at 56 ◦C for 12–24 h.

Once the samples were digested, the DNA extraction was made adding 100 μL of Binding Buffer and centrifuging at 2800 rpm. These products were transferred to a Pall® (Cortland, NY, USA) AcroPrep® filter plate, where the plate was vacuumed for 2 min. Then it was added 180 μL of Washing Buffer 1 and vacuumed again for 2 min. Posteriorly, it was added 750 μL of the Washing Buffer 2 and vacuumed for 2. Then TE Buffer was used to elute the DNA, adding 50 μL and incubating it for 2 min at 56 ◦C.

We amplified fragments of the mtDNA 16S gene with forward and reverse reactions using the following primers respectively: F: L2510 (5--CGCCTGTTTAACAAAAACAT-3-) and R: H3056 (5--CGGTCTGAACTCAGATCACGT-3-) [6]. The master mix cocktail used for amplification was of 1 μL of the DNA template, 2.5 μL of Y Buffer, 4 μL of dNTPs, 0.5 μL of TaqPol, 1 μL of MgCl2, 1 μL of the forward and reverse primers, and 14 μL of distilled water, reaching a final volume of 25 μL. Amplification reactions were performed in an Eppendorf Mastercycler® pro (Hamburg, Germany) thermocycler using the following PCR conditions: initial denaturation 2 min (94 ◦C)—[denaturation 35 s (94 ◦C)—hybridization 35 s (48.5 ◦C)—elongation 60 s (72 ◦C)] × 40—final elongation 10 min (72 ◦C) [51]. Sequencing was performed using a BigDye® Terminator (ThermoFisher Scientific®, Waltham, MA, USA) with the following cycling conditions: 1 min at 95 ◦C, 30 × [10 s at 95 ◦C, 10 s at 50 ◦C, 2 min at 60 ◦C], with 10 μL of reaction volume.

Additionally to our own samples, we included non-Paraguayan data from sequences downloaded from GenBank, selecting preferably those sequences associated with museum vouchers, to avoid common problems of misidentifications in that repository [52–54]. In most cases, we downloaded only sequences from species represented in our samples, except for *Bothrops* and *Amphisbaena*. In these two cases, we downloaded samples from all the species present in Paraguay, because of the difficulty of these taxa for morphological identification. In the case of *Micrurus,* there is only one sequence of a species present in Paraguay available in GenBank (JQ627286). A particular case was the only available sample of *Vanzosaura rubricauda* (AF420716) uploaded in the framework of a lizards' phylogeny [55]. That specimen (MRT 05059) from Vacaria, Estado de Bahia, Brazil, actually is *V. multiscutata* [35]. Nevertheless, it was included in the analysis to evaluate the clustering with the genetic sample from Paraguay.

Codes of sequences downloaded from GenBank, plus accession numbers of sequences generated in this work, are available in the Table S2.
