Samples

A total of nine species of spiders representing six families according to current classification were collected from China and India (Table 1). Samples were preserved in 99.9% ethanol.

Table 1.

Specimens used in this study¹⁾

DNA extraction, amplification and sequencing of 12S and 18S rRNA gene fragments

Total genomic DNA was extracted from leg tissue following standard proteinase digestion and phenol-chloroform extraction procedures as described by Sambrook et al. (1989). A fragment of the mitochondral 12S rRNA gene was amplified by using primers 12St-L: 5′-GGTGGCATTTTATTTTATTAGAGG-3′ (Croom et al., 1991) and 12Sbi-H: 5′-AAGAGCGACGGGCGATGTGT-3′ (Simon et al., 1994), and a fragment of the nuclear 18S rRNA gene was also amplified by using primers 18S-ai: 5′-CCTGAGAAACGGCTACCACATC-3′ and 18S-b0.5: 5′-GTTTCAGCTTTGCAACCAT-3′ (Tautz et al., 1988). The polymerase chain reaction (PCR) was performed on PE 2400 (Perkin Elmer) thermocycler system in 30 μl volumes containing 1–3 μl of genomic DNA, 10 pM of each primer, each of four deoxynucleoside triphosphates at 250 μM, 1.5 mM MgCl₂, and 2 units of Taq Polymerase. PCR was carried out under the following conditions: hot-start at 95°C for 5 min; repeated 30 cycles; denaturation at 95°C for 40 s, primer annealing at 55°C for 50 s and primer extension at 72°C for 2 min; final extension at 72°C for 10 min. PCR products were separated on 2% agarose gel, and purified with DNA gel extraction kit (Vitagene). The purified DNA was sequenced using the BigDye^TM Terminator Cycle Sequencing Ready Reaction Kit on an ABI PRISM 310 Genetic Analyzer. An additional sequence of 12S rDNA was retrieved from GenBank (Table 1).

Cloning and sequencing of the 3′-end cDNA fragment of MaSp1 gene

Silk glands were dissected from Argiope amoena abdomens. The entire gland was removed and frozen immediately in liquid nitrogen. The gland was then crushed in a mortar and pestle under liquid nitrogen. All solutions, instruments and glassware were treated to inhibit RNAase activity. Total RNA was isolated by using Trizol®Reagent (Gibco), mRNA was isolated from total RNA with Oligotex™ mRNA Purification Kit (Qiagen). Single-strand cDNA was synthesized directly from mRNA using SMART™ cDNA Library Construction Kit (Clontech). Double-strand cDNA was synthesized by long-distance PCR using two adaptor primers offered by SMART™ cDNA Library Construction Kit (3′ PCR Primer: 5′-TTCTAGAGGCCGAGGCGGCC-3′; 5′ PCR Primer: 5′-AGCAGTGGTATCAACGCAGAGT-3).

Based on repetitive region's amino acid sequence (GGQG-GYGGL) of MaSp1 in spiders (Xu and Lewis, 1990), one primer (P_MaSp1: 5′-GGAGGACAAGGTGGATATGGCGGATTAGG-3′) was designed and synthesized. Double-strand cDNA of silk gland in Argiope amoena was used as a template for 3′ RACE. The 3′ end cDNA fragment of MaSp1 gene was amplified by the primer pair 3′ PCR Primer and P_MaSp1. A total of 30 μl PCR reaction mixture was composed of 3 μl of 1:100 dilution of double-strand cDNA of silk gland, 10 pM of each primer, each of four deoxynucleoside triphosphates at 250 μM, 1.5 mM MgCl₂, and 2 units of Taq Polymerase. Amplification was carried out on PE2400 (Perkin Elmer) thermocycler system under the following conditions: hot-start at 95°C for 5 min; repeated 30 cycles: denaturation at 95°C for 40 s, primer annealing at 50°C for 40 s and primer extension at 72°C for 2 min; final extension at 72°C for 10 min.

PCR products were separated on 2% agarose gel, purified with DNA gel extraction kit (Vitagene), and ligated into PMD18-T vector using T/A Cloning Kit (TaKaRa). The cloned DNA was transformed and replicated in Escherichia coli (JM109), plasmid DNA was purified using Plasmid Mini Kit (Watson). After identified by PCR using M13 and M13 Reserve as primers, positive clones were sequenced on LI-COR DNA sequencer (LI-COR) also using M13 and M13 Reserve as primers. A further eight sequences were retrieved from GenBank (Table 1).

Data analysis

The predicted C-terminal non-repetitive region's amino acid sequences of MaSp1 were obtained from the 3′-end cDNA fragment of MaSp1 gene by using DNAclub. The DNA sequences of 12S and 18S rRNA gene fragments, MaSp1 gene fragments and predicted C-terminal non-repetitive region's amino acid sequences of MaSp1 were aligned respectively with CLUSTAL × 1.8 (Thompson et al., 1997). The neighbor-joining (NJ) method was applied to infer relationships among taxa on the basis of a pairwise matrix of the distance from Kimura's two-parameter model, using MEGA 2.1 (Kumar et al., 2000). The maximum-parsimony (MP) analyses were conducted using heuristic search option of PAUP 4.0b (Swofford, 2000). And the Maximum-likelihood (ML) analyses were performed by use of DnaSP 3.5 and PHYLIP vers. 3.5c computer packages (Felsenstein, 1993).

12S rDNA data set

The aligned sequences of 12S rDNA fragments for nine species of spiders consisted of a total of 269 sites, including 191 variable and 102 parsimony informative sites. The fragments have a high average A/T content of about 75.5%; this result is consistent with the statement that arthropod mitochondrial genomes in general tend to be highly A+T biased (Crozier and Croizer, 1993). The majority of transition values were lower than transversion ones. Sequence differences (transition + transversion) among these spiders ranged from 0.6% to 33.9%, with an average of 24.7%. It is noteworthy that average sequence difference (23.0%) between the genus Nephila and other tetragnathid genera was somewhat higher than that between Nephila and the araneid genera (21.4%) (Table 2).

Table 2.

Percent difference of base substitution (upper triangle) and values of transitions/transver-sions (lower triangle) for 12S rRNA gene fragment among 9 species of spiders

NJ, MP and ML trees of the 12S rDNA fragments with Ornithoctonus huwena as outgroup indicate that the Araneidae, Nephila, Linyphiidae and Tetragnathidae clustered together, while the theridiid genus (Achaearanea) formed another single clade (Fig. 1). In NJ and ML trees, two species of Nephila clustered with the araneid clade, the linyphiid genus (Hylyphantes) and two tetragnathid genera (Tetragnatha and Leucauge) formed another monophyletic clade; but in the MP tree, the linyphiid genus clustered with the group which consists of the araneid genera (Gasteracantha, Argiope) and Nephila.

Fig. 1.

1Neighbor-joining (NJ), maximum parsimony (MP) and maximum-likelihood (ML) trees based on 12S rDNA sequence data of nine species of spiders. Bar equals 0.05 (NJ) and 10 (ML) units of Kimura's two-parameter distance. Numbers above branches in these trees are Bootstrap values at least 50% of the 1,000 bootstrap replications.

18S rDNA data set

The aligned sequences of 18S rDNA fragments for eight species of spiders consisted of 738 total sites, including 190 variable and 58 parsimony informative sites. In contrast to the 12S rDNA fragments, the 18S fragments have a lower average A/T content about 49.6%, and the majority of transition values were higher than transversion ones (Table 3).

Table 3.

Percent difference of base substitution (upper triangle) and values of transitions/transversions (lower triangle) for 18S rRNA gene fragment among 8 species of spiders

NJ, MP and ML trees of the 18S rDNA fragments with Heptathela hangzhouensis and Ornithoctonus huwena as outgroups generated very similar tree topologies one another (Fig. 2). The theridiid genus (Achaearanea) clustered with the group which consists of Nephila and the araneid genera (Gasteracantha, Argiope), while the tetragnathid genus (Tetragnatha) and the linyphiid genus (Hylyphantes) formed another monophyletic clade.

Fig. 2.

Neighbor-joining (NJ), maximum parsimony (MP) and maximum-likelihood (ML) trees based on 18S rDNA sequence data of eight species of spiders. Bar equals 0.02 (NJ) and 10 (ML) units of Kimura's two-parameter distance. Numbers above branches in these trees are Bootstrap values at least 50% of the 1,000 bootstrap replications.

MaSp1 gene data set

The sequences of the cDNA fragment (GenBank Accession no.  AY263390) of MaSp1 gene can be divided into two regions: (1) a repetitive region that codes for an alternating alanine-rich and glycine-rich domain and (2) a non-repetitive coding region. The predicted C-terminal non-repetitive region of MaSp1 consisted of predicted 102 amino acids, its sequence was rather homologous to published amino acid sequences of other eight species of spiders, for example, the amino acid sequence of Argiope amoena had a homology of 72.4% to that of Nephila clavipes (Fig. 3).

Fig. 3.

Allignment of the C-terminal non-repetitive regions predicted Amino acid sequences of MaSp1 of Argiope amoena and other eight species of spiders. The mark (#) indicates where amino acid sequence has not revealed, the short line (–) indicates where an amino acid is deleted.

The aligned 3′-end non-repetitive region's cDNA sequences of MaSp1 gene for nine species of spiders consisted of 310 total sites, including 215 variable and 149 parsimony informative sites. The MaSp1 gene fragments have a lower average A/T content about 54.6%, and the majority of transition values were higher than transversion ones. The average sequence difference (18.1%) between the genus Nephila and other tetragnathid genera was markedly higher than that between Nephila and the araneid genera (13.2%) (Table 4).

Table 4.

Percent difference of base substitution (upper triangle) and values of transitions/transversions (lower triangle) for 3′-end non-repetitive region's cDNA sequence of MaSp1 gene among 9 species of spiders

NJ and MP trees based on the 3′-end non-repetitive region's cDNA sequences of MaSp1 gene or the predicted amino acid sequences of C-terminal non-repetitive region of MaSp1 with Dolomedes tenebrosus as an outgroup are also in agreement with the traditional view that genus Nephila should be placed in the family Araneidae rather than Tetragnathidae (Fig. 4). Three species of the genus Nephila is sister to other two araneid genera Araneus and Argiope. And then the clade is sister to genus Tetragnatha of the family Tetragnathidae.

Fig. 4.

The neighbour-joining (NJ) and Maximum parsimony (MP) trees resulting from analysis of the 3′-end non-repetitive regions cDNA sequences of MaSp1 gene and C-terminal non-repetitive regions amino acid sequences of MaSp1 of nine species of spiders. Bootstrap values at least 50% of the 1,000 bootstrap replications are shown above branches in these trees.

The phylogenetic position of the genus Nephila

Although the genus Nephila possesses morphological characters that are believed to be synapomorphies of the family Tetragnathidae (Roewer, 1942; Brignoli, 1983; Song et al., 2001), it also has many characters that are not shared by other tetragnathid genera. These characters include the labium longer than wide, transverse grooves on the book lung covers, and the metatarsi and tarsi together longer than the patellae and tibiae (Levi and Eickstedt, 1989). On the other hand, it also bears many characters in common with the family Araneidae, such as a distinct epigynum, characteristic integral structure of male palpal organ and striking sexual size dimorphism (female giantism). Hence, it seems to be difficult to judge the phylogenetic position of the genus Nephila by means of morphological methods alone.

In the present study, the NJ, MP and ML trees based on the sequences of 12S rRNA or 18S rRNA gene fragments, and the NJ and MP trees based on the 3′-end non-repetitive region's cDNA sequences of MaSp1 gene or the C-terminal non-repetitive region's amino acid sequences of MaSp1, revealed that the phylogenetic position of the genus Nephila is closer to the araneid genera than to tetragnathid. It is notable that the NJ, MP and ML trees of the 12S and 18S rDNA fragments corroborated the statements that Araneidae and Tetragnathidae do not form a sister group in the superfamily Araneoidea (Coddington and Levi, 1991; Hausdorf, 1999). Although a linyphiid genus or a theriddiid genus included in the analyses intervened between Araneidae and Tetragnathidae in those reconstructed phylogenetic trees (Figs. 1 and 2), the genus Nephila was always connected directly with the araneid genera. Our results showed that Nephila should be transferred back to the family Araneidae, or the genus and its subfamily might be recognized as a separate clade in the superfamily Araneoidea and should be removed from either Araneidae or Tetragnathidae. In the later case, the subfamily Nephilinae might be elevated to family level after being redefined and redelimited.

The use of the C-terminal non-repetitive region of MaSp1 as a molecular marker

As the web frame and lifelines of spiders, dragline silk plays an important role in spider's life. Dragline silk is a two-protein fiber and, MaSp1 (major ampullate spidroin-1) and MaSp2 (major ampullate spidroin-2) genes which are active in the major ampullate gland are thought to be responsible for the dragline silk. MaSp1 gene encodes a kind of protein molecule that contains the repetitive region and C-terminal non-repetitive region (Xu and Lewis, 1990; Van Beek et al., 2002). The repetitive region exhibits a pattern of alternating alanine-rich, crystal-forming blocks that impart the silk's unmatched strength in the natural world and glycine-rich amorphous blocks implicated in providing elasticity in the silk filament. In contrast to the repetitive region, the C-terminal non-repetitive region is hydrophilic, thus it proved to be necessary to maintain the soluble or liquid crystalline state of silk molecules along the gland duct before being secreted from silk gland (Anthoula et al., 2002). The alignment of predicted amino acid sequences of C-terminal non-repetitive region of MaSp1 of nine species of spiders shows this region has considerably conserved amino acid sequences (Fig. 3).

The results from the NJ and MP trees based on the 3′-end non-repetitive region's cDNA sequences of MaSp1 gene or the C-terminal non-repetitive region's amino acid sequences (Fig. 4) agree with the phylogenetic relationships among these spiders in traditional classifications. They are also similar to the results based on sequences of 12S and 18S rRNA gene fragments (Figs. 1 and 2) on the phylogenetic placement of Nephila. At least, it indicates that both the 3′-end non-repetitive region's cDNA sequences of MaSp1 gene and the C-terminal non-repetitive region's amino acid sequence of MaSp1 have a certain value in phylogenetic analysis of spiders. In addition, the MaSp1 gene has no intron, and consists of a single exon alone (Beckwitt et al, 1998; Hayashi and Lewis, 2000). Therefore, amplification of the 3′-end partial DNA of MaSp1 gene directly from genomic DNA of spider is feasible. That is to say, not via complicated reverse transcription from mRNA, the 3′-end non-repetitive region's DNA sequences of MaSp1 gene and the C-terminal non-repetitive region's amino acid sequence of MaSp1 can be obtained conveniently and economically.