DNA Barcodes of Tetragnathid Spiders (Araneae: Tetragnathidae) in Malaysia

Although diverse groups of spiders are present in tropical countries, very little attention has been given on various types of differences that exist within these groups, particularly for the groups of tropical spiders in Malaysia. The morphological and genitalia characteristics are traditionally used to identify species variants. However, in this study, a molecular approach was utilized to produce a more precise and accurate result in an effort to identify, delineate and verify the species. Mitochondria-encoded cytochrome oxidase I (COI) and nuclear-encoded 18S rRNA (18S) genes) were adopted to establish DNA bacodes for 17 species of tetragnathid spiders (Araneae, Tetragnathidae) in Malaysia. Generally, the molecular data of tetragnathid spiders was consistent with their classification based upon morphological characteristics, though species boundary of Opadometa grata and Leucauge decorate could not be resolved by 18S gene.


INTRODUCTION
T he long-jawed, orb-weaver, spider family Tetragnathidae, contains 47 genera (World Spider Catalog, 2020) with at least 967 species throughout the world. It comprises three subfamilies, namely the Leucauginae Caporiacco 1955, Metainae and Tetragnathinae Menge 1866. The genus Tetragnatha Latreille 1804, is the genus type of this family (Alvarez-Padilla and . Members from this family are diverse in morphological and behavioral characteristics. Many of these characteristics are homoplasius to Araneidae and Nephilidae (Griswold et al., 1998;Alvarez-Padilla and Hormiga, 2011). However, the morphological features within the family vary significantly, with some displaying distinctive characteristics that do not fit into the typical tetragnathid morphology. The body length of tetragnathids range between 2-23 mm. The habitats used by tetragnathid spiders in tropical and subtropical ecosystems are highly O n l i n e

F i r s t A r t i c l e
plate and fertilization ducts, where only the copulatory ducts are found to be present (Alvarez-Padilla et al., 2009). In Malaysia, the genus Tetragnatha is the only genus that has been documented within the Tetergnathinae subfamily. The genetics of tetragnathid species from different parts of the world have been well-studied (i.e. Levi, 1980;Hormiga et al., 1995;Pan et al., 2004;Blackledge et al., 2009;Alvarez-Padilla et al., 2009;Dimitrov and Hormiga, 2011). These previous studies adopted combination sets of different markers to identify the relationships between the studied species, but the determination of the most suitable and reliable markers for taxonomic identification is important to reduce the cost and time (i.e. Fang et al., 2000;Astrin et al., 2006). Nonetheless, the genetic data on Malaysian tetragnathid species is still unknown, with no molecular studies being reported to date.
The objective of this study was to determine the sequence diversity of two gene fragments (mitochondriaencoded cytochrome oxidase I (COI) and the nuclearencoded 18S rRNA (18S)) in 17 Malaysian tetragnathid species, and thereby attempting to establish DNA bacodes for members of the subfamilies Leucauginae (Leucauge, Opadometa, Mesida and Tylorida) and Tetragnathinae (Tetragnatha). The COI and 18S sequences were analyzed independently. The results were then crossexamined with the morphological characteristics of the regional tetragnathid species to categorize the specimens' subfamily. The genetic information derived from the results of this molecular analysis would be highly useful for future species identification and taxonomic verification of Malaysian tetragnathids.

Sample collection and identification
Random sampling was employed during the sample collection. Tetragnathid spiders were individually caught by hand following visual detection. The sampling was performed throughout the day and night, and was conducted at various types of diverse habitats available throughout Malaysia. The collected specimens were immediately preserved in individual jars containing 70% (v/v) ethanol, and subsequently stored in a freezer at -20°C.
For species identification, the preserved specimens were first observed under a 50× dissecting microscope (Amscope, USA and identified to species based on the morphology descriptions of the Leucauge and Opadometa (Yoshida, 2009;, Mesida (Jager and Praxaysombath, 2011), Tylorida (Tanikawa, 2004;Jager and Praxaysombath, 2009;Kulkarni, 2014) and Tetragnatha (Okuma, 1987(Okuma, , 1988 subfamilies, and further supported by other documented morphological references (Barrion and Litsinger, 1995;Murphy and Murphy, 2000;Song et al., 2002;. In this study, a total of 17 tetragnathid species were collected from 10 different locations in Malaysia (Table I). The Leucauginae subfamily is represented by the genera Leucauge (five out of six species), Opadometa (two out of three species), Mesida (two out of two species) and Tylorida (three out of four species). The Tetragnathinae subfamily is represented by the genus Tetragnatha (five out of 16 species). The genus Larinioides cornutus and Gasteracantha cancriformis were used as the basis and main reference for the COI gene, while the genus Gasteracantha kuhlii and Cyclosa conica were used as the basis and main reference for the 18S gene. Both groups of main references used were obtained from the GenBank and were used as outgroups (Table I).

DNA extraction, polymerase chain reaction and sequencing
For DNA extraction, two or more legs (depending on the spider's size) were removed from the preserved specimen, rinsed several times with distilled water and then transferred into a 1.5-mL microcentrifuge tube containing lysis buffer and homogenized in liquid nitrogen for at least 5 min. Then, Proteinase K was added and the homogenized sample was incubated overnight at 65 o C in a water bath before heating to 95 o C for 10 min to deactivate the proteinase K. Finally, the genomic DNA was extracted from the resultant sample using the Qiagen DNeasy Tissue Kit according to the manufacturer's recommended instructions.

Amplification of the mitochondrial COI and nuclear 18S gene fragments
For 18S gene fragment, the full length 18S region was amplified with the universal primer set 18Sa (3'-ATTAAAGTTGTTGCGGTTA-5') and 9r (3'-GATCCTTCCGCAGGTTCACCTAC-5') (Alvarez-Padilla, 2008). The total reaction volume used was 50 µL, consisting of 10 µL of dH 2 O, 5 µL of each primer (0.5 mM), 25 µL of the master mix (Lucigen, USA) and 5 µL (ca. 250 ng) of the DNA template. 'Touchdown' condition for the PCR reaction was set at 94 o C for 2 min, followed by 13 cycles of 94 o C for 20 s, 60 o C for 35 s and 65 o C for 30 s and then 21 cycles at 94 o C for 15 s, 48 o C for 35 s and 65 o C for 30 s and then a final 72 o C for 3 min (Dzulhelmi, 2016).
For COI, the PCR amplification was performed in the same manner except using the universal forward primer LCOI1490 ( 3 ' -G G T C A A C A A AT C ATA A A G ATAT T G G -5 ' ) and reverse primer HCOI2198 (3'-TAAACTTCAGGGTGACCAAAAAATCA-5') (Alvarez-Padilla, 2008), and the thermal cycling was performed at 94 o C for 2 min, followed by 34 cycles of O n l i n e

F i r s t A r t i c l e
94 o C for 20 s, 50 o C for 35 s and 65 o C for 30 s, and then a final 72 o C for 3 min (Dzulhelmi, 2016). Each PCR-amplified sample was resolved by agarose gel electrophoresis mixed with ethidium bromide to enable viewing of the amplified nuclear of the 18S gene under ultra-violet illumination. Finally, the products were then sent to Genomics Bioscience and Technology Co. Ltd., for purification and sequencing.

Multiple sequence alignment and sequence analyses
The electophoragram of forward and reverse sequences was checked manually and assembled using BioEdit Sequence Alignment Editor Version 7.0.5 (Hall, 2005). Multiple alignments of the cleaned sequences were performed using Clustal X version 1.81 (Thompson et al., 1997). The lengths of the COI and 18S gene fragment DNA sequences were 625 bp and 930 bp, respectively. The COI and 18S rRNA sequences obtained from this study were deposited into the National Center for Biotechnology Information (NCBI) GenBank DNA sequence database (Table I).

Phylogenetic tree reconstruction
The phylogenetic trees were constructed based on maximum parsimony (MP) and maximum likelihood (ML) analyses for the 18S and CO1 gene fragments, respectively, using (PAUP) software version 4.0 (Swofford, 2003). For the ML and MP analysis, the best model was suggested by Akaike Information Criterion (AIC), which was performed using the PhyML 3.0 software (Guindon et al., 2010). The phylogenetic trees were reconstructed using HKY85 and GTR model for COI and 18S, respectively. Pairwise genetic distance and neigbour joining distances of both the COI and 18S genes were calculated using the Kimura-two-parameter model. A full heuristic search was used for the MP analysis. The tree reliability was then estimated by bootstrapping the phylogenetic tree data with 1000 replications of data sets for MP, and 100 replications for ML. In addition, Bayesian Inference (BI) analysis, using four chains from the Markov Chain Monte Carlo (MCMC) generations until it reached less than 0.01 was also performed for both the 18S and COI datasets using the MrBayes version 3.1.2 software (Ronquist and Huelsenbeck, 2003).

RESULTS
The phylogenetic analyses produced almost similar topologies for both the COI (BI and ML analyses) and 18S (BI and MP analyses) gene fragments, but displayed different bootstrap support values. Both genes revealed two main clusters, corresponding to the Leucauginae and Tetragnathinae subfamilies. The Leucauginae cluster further formed two separate subclusters and one subcluster consisted of members from the genera  (Table II).
The distance matrix calculated using the COI gene fragment sequences showed a wider range of values for each studied species, compared to 18S gene fragments. The existence of a higher number of informative sites in the COI genes could have been responsible for the wider range of genetic distances garnered compared to the more O n l i n e

DISCUSSION
Two distinct and independent sequence alignments of the COI and 18S gene fragments produced two almost identical tree structures with only minor differences. It can be hypothesized that the lineage history of the mitochondrial and nuclear DNA could have caused the slight differences in the tree topology. For the family Tetragnathidae, the two subfamilies and three sub-groups were clearly identified, delineated and resolved within the both the COI and 18S phylogenetic trees (Fig. 1). The Leucauge-Opadometa and Mesida-Tylorida were both found clustered within the Leucauginae-group, while the Tetragnatha were clustered in the Tetragnathinae-group. Some of these groupings were strongly supported by high bootstrap values of greater than 70% (Hillis and Bull, 1993).
These phylogenetic trees, based on the BI and MP analyses of the 18S gene and the BI and ML analyses of the COI gene, were highly comparable and corroborated with the documented internal relationships hypothesis of the Tetragnathidae family group (Alvarez-Padilla et al., 2009;Alvarez-Padilla and Hormiga, 2011). When taken together, the reconstructed phylogenetic trees in this study strongly support the separate groupings of Leucauge-Opadometa and Mesida-Tylorida within the subfamily Leucauginae are coherent with their morphological characteristics. Both these findings suggest that the COI and 18S genes are reliable genetic markers which can be used to identify, delineate and verify the natural groupings within the Tetragnathidae subfamilies (Alvarez-Padilla et al., 2009), as well as for other spider families, which are hypothetically highly similar and strongly related based on their morphological criteria (Astrin et al., 2006). Comparative analysis of both the COI and the 18S gene fragment sequences of Leucauge and Opadometa were located within the same clade, suggesting they may represent a sister relationship. However, this hypothesis required further verification by adopting multiple gene sequence analyses with larger sample size. Nevertheless, Alvarez-Padilla and  showed that Opadometa was closely related to Leucauge based on the numerous similarities observed in their morphology, behaviour and DNA sequences. Table III.-Genetic pairwise distance (%) of the 18S gene sequences among the 17 tetragnathid species [Leucauge (5 species), Mesida (2 species), Opadometa (2 species), Tetragnatha (5 species) and Tylorida (3 species)] plus the two outgroups (Larinioides cornutus and Gasteracantha cancriformis). The distances were calculated by using the Kimura-two-model (Kimura, 1980). The distances were calculated by using the Kimura-two-model (Kimura, 1980  To date, the reclassification of Opadometa species is still ongoing, but has never been successfully revised due to the difficulty of procuring the scarce male specimen that is necessary for precise species identification. However, the widely accepted view is that the general morphology of male Opadometa highly resembles the smaller-sized male of Leucauge species (Alvarez-Padilla and . As such, Opadometa species is grouped under Leucauge species (Yoshida, 2009). Similarly, L. argentina formed a separate subclade in the phylogenetic tree (Fig. 1). Although L. argentina was categorized as a sister group to the Leucauge-Opadometa family group in this study, it was the only small-sized species (< 6 mm) found in the present study.
Within the Mesida-Tylorida group, members share many similar morphological characteristics (Tanikawa, 2001;Alvarez-Padilla and Hormiga, 2011) and so these two genera can only be distinguished by minor differences in their outer appearance (Tanikawa, 2001(Tanikawa, , 2004Kulkarni, 2014). Tylorida striata was recognized as a sister-clade to M. argentiopunctata and the Mesida species (Tanikawa, 2001) based on a cladistic analysis of their morphological characteristics. Based on a previous study conducted by comparing their morphological characteristics, behaviour and DNA sequence, Tylorida was identified as a sister group to Orsinome species (Alvarez-Padilla and . However, since the specimens used in this study failed to include the Orsinome species from Malaysia, this observation cannot be fully verified as yet. Meanwhile, the genus Tetragnatha is the only genus from the subfamily Tetragnathinae that has a readily available record in Malaysia. The genus Tetragnatha was documented as a sister clade in the grouping, as inferred from their genetics, morphology and behavior patterns, which were all highly similar to the Glenognatha and Pachygnatha (Alvarez-Padilla and Hormiga, 2011), albeit both genera are yet to be documented in Malaysia (Dzulhelmi et al., 2014a, b;Norma-Rashid and Li, 2009). In spite of their similarities, the Tetragnathinae-group was found to differ significantly from the Leucauginae-group in terms of their morphological characteristics (Alvarez-Padilla et al., 2009). Nevertheless, the species boundaries among Tetragnatha spp. in Malaysia could not be fully resolved. Tetragnatha pinicola from the present study was identical to some of the unpublished COI sequences of T. ceylonica from Malaysia.
A detailed morphological analysis of these T. ceylonica samples would be beneficial in confirming their species identity. There is also a possibility of the shared DNA barcodes among closely related members, hence a comparative genetic analysis between T. ceylonica and T. pinicola by adopting various genetic markers would be needed to unveil the differences or to identy if they are a species complex. The species status of T. pinicola was also complicated by the distinct COI genetic distances (>16%) between Malaysian strain and those from the GenBank (i.e., China, Germany, Italy, Spain Switzerland). This discrepancy also deserves additional research efforts to resolve its species status.
The tree-based taxa clustering in the present study indicated that the molecular evidence provided by the COI gene supports the morphological hypothesis. This DNA identification method has, therefore, been found to be scientifically reliable and useful for spider taxonomic studies. An important point to note is that the current findings indicate that the COI gene has adequate variable regions and is more informative than the 18S gene in resolving intra-and inter-specific relationships among the Tetragnathid species. For example, a single marker from the COI gene is already sufficient for studying the genetic relationships in other spider species (i.e. Garb et al., 2004;Tanikawa et al., 2006;Vink et al., 2009;Smith et al., 2012;Muslimin et al., 2015). Nonetheless, some groups of spiders achieve better results from using other genetic markers (i.e. Croom et al., 1991;Fang et al., 2000;Astrin et al., 2006), or a combination of several genetic markers (i.e. Benjamin et al., 2008;Alvarez-Padilla et al., 2009;Su et al., 2011;Franzini et al., 2013).

CONCLUSION
This study utilized COI and 18S gene sequences to the DNA barcodes of 17 studied tetragnathid spider species in Malaysia. However, their genetic diversity maybe underestimated because the low sample size used in the present study may not sufficient to disclose their intraspecific variation. Therefore, it is highly recommended that a wider range of the Malaysian tetragnathid species with larger sample size should be used in future studies to further compare and contrast the compatibility of these two gene sequences. Testing on other genetic markers is also warranted to resolve the species boundaries among some of the complex taxa.