Molecular evolution caught in action: gene duplication and evolution of molecular isoforms of prothrombin activators in Pseudonaja textilis (brown snake)

Abstract

Summary. Background: The evolution of structurally and functionally similar proteins with highly diverse physiological roles within a single organism is of great interest. Australian elapid snakes offer an excellent opportunity to study the molecular evolution of prothrombin activators. Venom from Pseudonaja textilis contains pseutarin C, a group C prothrombin activator. Its enzymatic subunit is structurally and functionally similar to mammalian factor (F) Xa, whereas its non-enzymatic subunit is similar to FVa. As vertebrates, the snakes also contain a system to activate prothrombin in their own blood during injury. These hemostatic factors are produced in the liver. Results: Here we describe the presence of two molecular forms of FX expressed in the liver of P. textilis. Both isoforms have molecular signatures and domain architecture of FX. However, one isoform shows ∼94% sequence identity with the snake FX from Tropidechis carinatus, whereas the other is much closer (90% identity) to the catalytic subunit of pseutarin C (PCCS). Real-time polymerase chain reaction reveals that the latter isoform is expressed ∼56 000 times lower in the liver of P. textilis. However, the isoforms are not expressed in the venom gland. Conclusion: A detailed analysis of deletions and insertions along with the sequence indicates that the second isoform is an intermediate caught in the evolution of venom prothrombin activator from the blood coagulation FX. Thus, this isoform represents a ‘molecular fossil’ and reveals the likely evolutionary path of recruitment of FX in the venom gland.

Introduction

Gene duplication and subsequent evolution are believed to result in the increase in rates of diversification and complexity of proteins [1,2], and thus play a critical role in the evolution of organisms by improving adaptability to a changing habitat or environmental conditions. Therefore, there is a tendency to have an increased number of genes or proteins in higher organisms. For example, it has been suggested that the ancestors of vertebrates had only ∼15 000 genes, while vertebrates have over 60 000 genes [3,4]. Duplicated genes and the derived proteins retain similar exon–intron organization and molecular scaffolding, respectively. The mammalian blood coagulation system is an excellent example of gene duplication and diversification [5,6]. Analysis of mammalian gene organization, protein structures and sequence identities suggest that most of the coagulation factors may have evolved by duplication, reduplication and functional diversification of only two different gene structures [7]. For example, a number of blood coagulation factors, including factor (F) X, FVII, FIX, protein C and protein Z, contain a γ-carboxylated glutamic acid (Gla) domain followed by two epidermal growth factor-like (EGF) domains and a serine protease (SP) domain. All of these proteins share similar gene organization as well as protein folding [8–10], and are therefore believed to have originated by the gene duplication and diversification of a common ancestral vitamin K-dependent SP gene. Although proteins with precise functional roles could be easily identified in their current status of evolutionary changes, intermediates in the evolution of the various blood coagulation factors have not yet been identified.

Our recent studies on prothrombin activators from Australian snake venom have shown their similarities to mammalian blood coagulation factors [11–13]. Group D prothrombin activators are structurally and functionally similar to blood coagulation FXa [11–15]. As vertebrates, these snakes also have a hemostatic system with a separate prothrombin activator system. The two systems perform completely divergent functions in the snake; the one in the venom is used as an offensive weapon, whereas the one in the plasma is used for the snakes’ own lifesaving hemostatic purposes. Recently, we determined the complete cDNA sequence of the blood coagulation FX from Tropidechis carinatus liver [16]. It is phylogenetically closer to its own venom prothrombin activator (vPA), trocarin D, compared with the FX of other mammalian and non-mammalian origin. Thus, we believe that the group D snake vPAs have evolved by the duplication of blood coagulation FX gene and subsequently was marked for tissue-specific expression in the venom gland to act as a weapon to attack the hemostatic system of prey [16].

Pseutarin C, a group C prothrombin activator from Australian brown snake Pseudonaja textilis venom, is a multimeric protein complex similar to mammalian FXa–FVa complex [12,17,18]. The non-enzymatic subunit is similar to FVa, whereas the catalytic subunit (PCCS) is similar to FXa [12,14,15,17,18]. This prothrombin activator has probably evolved independently of group D prothrombin activators [18]. Because of the similarity between group D prothrombin activators and the catalytic subunit of group C prothrombin activators, it is interesting to understand their evolution. Therefore, we attempted to clone and sequence the blood coagulation FX from the liver tissues of P. textilis. Interestingly, we found two molecular isoforms of FX expressed in the liver of P. textilis. This is the first report of the existence of two isoforms of FX in the liver of any organism. Here we present the complete sequence of these two isoforms and their evolutionary relationship with the third isoform, PCCS, expressed exclusively in the venom gland. These studies will contribute significantly toward the understanding of the origin and evolution of snake vPA.

Materials and methods

Materials

Liver of P. textilis was purchased from the Venom Supplies Pty Ltd, Tanunda, Australia. The RNeasy RNA isolation kit, one-step Reverse transcription-polymerase chain reaction (RT-PCR) kit, Qiaquick gel extraction and PCR purification kit, and QiaPrep mini prep kit were purchased from Qiagen (Valencia, CA, USA). SMART RACE cDNA amplification kit and BD Advantage 2 PCR Enzyme system were bought from Clontech Laboratories Inc. (Palo Alto, CA, USA). Long-PCR enzyme mix was obtained from Fermentas (Hanover, MD, USA). pGEMT easy vector system was purchased from Promega Corp. Ltd (Madison, WI, USA). The ABI PRISM^® BigDye^® terminator cycle sequencing ready reaction kit was purchased from Perkin Elmer (Foster City, CA, USA). Oligonucleotides were custom synthesized from 1st BASE (Singapore). All other chemicals and reagents were of the purest grade available.

Methods

RNA isolation

Total RNA was isolated from the liver tissues of P. textilis using a rotor homogenizer and the RNeasy mini kit following the manufacturer's instruction. On column DNase digestion was carried out according to the manufacturer's instructions to get rid of any residual DNA contamination with the isolated RNA. RNA purity and quality was checked by agarose gel (0.8%) electrophoresis. Amount of RNA was quantified by checking its optical density.

RT-PCR

RT-PCR was carried out using Qiagen one-step RT-PCR kit in a final volume of 25 μL using 100 ng of total RNA. FX sequence of mammalian and non-mammalian sources including T. carinatus FX were aligned with that of trocarin D and PCCS sequence. Sense (Ps_FXF: 5′-ATTGAAAGGGAATGC-3′) and antisense (Ps_FXR: 5′-CTGTCTCCCTGGCATGCATC-3′) primers were designed based on the most conserved sequence.

The RT-PCR conditions used were: reverse transcription for 30 min at 50 °C followed by inactivation of reverse transcriptase and activation of Taq polymerase at 95 °C for 15 min. PCR was performed using the following conditions: 35 cycles of one step each at 94 °C for 1 min, 54 °C for 45 s, 72 °C for 1 min followed by a final extension at 72 °C for 10 min. PCR products were allowed to run on 1% agarose gel electrophoresis and were visualized by ethidium bromide staining. PCR products were purified using Qiagen PCR purification kit according to the manufacturer's instruction.

5′- and 3′-rapid amplification of cDNA ends (RACE)

5′- and 3′-RACE libraries were created and amplifications were carried out with Clontech's SMART RACE cDNA amplification kit and Advantage 2 PCR Enzyme system. Universal primer mix (UPM: 5′-CTAATACGACTCACTATAGGGCA AGCAGTGGTATCAACGCAGAGT-3′; 5′-CTAATACGACTCACTATAGGGC-3′) from the SMART RACE kit and gene-specific oligonucleotides 5′GSP1 (5′-TCAGCAGTGGGAAGGCAGGCAGGAACC-3′) and 3′GSP1 (5′-GGTTCCTGCCTGCCTTCCCACTGCTGA-3′) were used initial amplification. These gene-specific primers were designed based on the sequences obtained by RT-PCR. Then, the primary PCR products were reamplified using a second set of gene-specific primers [designed a few hundred bp upstream and downstream of 5′Ps_GSP1 and 3′Ps_GSP1 (5′GSP2: 5′-TGCAGAAGTGCCAACAGTTACCATTG-3′ and 3′GSP2: 5′-CCTTCCCACTGCTGATTTTGCCAACC-3′), respectively] and nested UPM (5′-AAGCAGTGGTATCAACGCAGAGT-3′) from the Clontech kit.

Cloning and sequencing

All PCR products were subjected to 1% agarose gel electrophoresis followed by ethidium bromide staining and purified either by PCR purification kit or by gel-extraction kit. Subsequently, the PCR products were ligated into pGEMT easy vector (Promega). The ligated vectors were transfected into DH5α competent cells using heat shock. Ampicillin (100 mg mL⁻¹) was used for selection against antibiotics. Blue and white colony screening was performed on LB agar plates using 80 mg mL⁻¹ X-gal and 0.5 mL L⁻¹ of 100 mm isoprophyl-β-D-thiogalacto-pyranoside (IPTG).

Plasmids were extracted using Qiagen miniprep kit and checked by EcoRI digestion to determine the size of inserts. Each selected clone was sequenced four to six times in both directions. All sequencing reactions were carried out using BigDye terminator v3.1 kit with ABI PRISM 3100 automated DNA sequencer following manufacturer's instructions.

Sequence analysis

Sequence analysis was carried out using BLASTX program at the National Centre for Biotechnology Information website (http://www.ncbi.nlm.nih.gov). Sequence alignments were carried out using the GeneDoc/DNAMAN program. Prediction of signal peptide and N-glycosylation site was carried out using PSORT and NetGlyc prediction site, respectively, at the Expasy website (http://tw.expasy.org/).

Real-time quantitative expression analysis

Real-time PCR was used to quantify the relative expression of two isoforms of FX in P. textilis liver. Expression of the vPA PCCS in the liver as well as PFX1 and PFX2 in the venom gland were also measured. β-actin was used as an internal control, as described earlier [19]. Four sets of primers (forward and reverse) were designed specific to PFX1 and PFX2, PCCS and β-actin, respectively, from the cDNA sequences (GenBank accession nos DQ017705, DQ017706, AY260939, and AY734452, respectively) for real-time PCR: (i) FAM-FX1: 5′-GAAGGTGACCAAGTTCATGCTCTCCTACAAATTTCTCTGGCTTAGTTCT-3′ and RevFX1: 5′-AACAACTCTGATATCAGGGCTTGG-3′; (ii) FAM-FX2: 5′-GAAGGTCGGAGTCAACGGATTTGTTCCTCCTCAAAAAGCCTATAAGT-3′ and RevFX2: 5′-CACATTTTCAGAGAACTGGATAGGG-3′; (iii) FAM-PCCS: 5′-GAAGGTGACCAAGTTCATGCTCTCCCAAAAAAAGCCAGGAATTC-3′ and RevPCCS: 5′-CACATTTTCAGAGAACTGGATAGGGGTCTTC-3′ and (iv) JOE Act: 5′GAAGGTCGGAGTCAACGGATTAACTGGGATGACATGGAGAAGATTTGGC-3′ and Rev_act: 5′-CTTGGGGTTCAGGGGAGCTTCTGTC-3′. Each primer was tested for specificity. The 5′ ends (italicized part) of the FAM primers are complementary with FAM UniPrimer^TM while the 5′ end of JOE primer (italicized part) is complementary with JOE UniPrimer^TM. Real-time PCR was set up using FAM UniPrimer, JOE UniPrimer, specific primer mix (including FAM-FX1, RevFX1 with JOE-Actin, RevAct, FAM-FX2, RevFX2 with JOE-Actin, RevAct or FAM-PCCS, RevPCCS with JOE-Actin, RevAct), reaction mix S-plus, dNTPs and Platinum^® Taq polymerase and cDNA template corresponding to 200 ng RNA μL. The reactions were performed in an ABI PRISM 7000 system under the following program: 96 °C for 4 min; 20 cycles of 96 °C for 15 s, 60 °C for 5 s, 72 °C for 4 s; 24 cycles of 96 °C for 15 s, 55 °C for 20 s, 72 °C for 40 s.

Results

cDNA sequence of FX of P. textilis

Blood coagulation FX in the plasma is primarily synthesized in the liver. To obtain the full-length FX cDNA sequence, RT-PCR followed by RACE was performed using the total RNA extracted from the liver tissues of P. textilis. RT-PCR using the sense and antisense primers (Ps_FXF and Ps_FXR) yielded a product of ∼1.2 kb size. The PCR products were purified, cloned and sequenced. Interestingly, sequencing data indicated the presence of two distinct types of clones. To complete their sequences, both 5′- and 3′ RACE libraries were amplified using gene-specific RACE primers (5′Ps_GSP1 and 3′Ps_GSP1) and UPM (see Materials and methods for details). The resultant PCR products were reamplified using nested gene-specific primers (5′Ps_GSP2 and 3′Ps_GSP2) and the nested UPM to reduce the non-specific background amplification. After reamplification, clear bands of ∼600 bp and ∼800 bp were obtained in 5′- and 3′-RACE, respectively. Resultant PCR products were purified, cloned into pGEMT vector and sequenced. The cDNA and the deduced amino acid sequences, as expected, showed 100% identity in the overlapping regions obtained from the sequence of the RT-PCR product. Assembled sequencing data for RT-PCR and RACE confirmed the existence of two types of FX cDNA in P. textilis liver; we named the isoforms as PFX1 (1452 bp) and PFX2 (1392 bp). Both PFX1 and PFX2 cDNA sequences have been submitted to GenBank (DQ017705 and DQ017706, respectively).

PFX1 and PFX2 share high percentage of identity in cDNA sequence (∼85%) as well as at protein level (∼81%). The deduced amino acid sequence also shows ∼40–50% identity with other known mammalian and piscine FX, as well as with the vPAs with identical domain architecture (Gla – EGF-I – EGF-II – SP). Interestingly, cDNA sequence of PFX1 shows much higher identity (∼94%) to the FX of another Australian elapid, T. carinatus, compared with PFX2 (∼79%). On the contrary, PFX2 cDNA sequence shows comparatively higher identity to its homologs in the venom (∼90% and ∼81% identity with PCCS and trocarin D, respectively) compared with PFX1 (∼76% identity with both PCCS and trocarin D). Thus there are two distinct molecular forms of FX present in the liver of P. textilis, one similar to the blood coagulation FX and the other similar to its vPA. As FX has the same domain architecture and sequence similarity with blood coagulation FIX and VII [8–10], we also compared the level of identity of PFX1 and PFX2 with them. Both P. textilis FX isoforms show only ∼30–35% identity with both FIX and FVII. Therefore, it is quite likely that PFX1 and PFX2 are not snake FIX or FVII.

Mammalian FX has a pre–pro leader sequence [20,21]. The leader sequences in both PFX1 and PFX2 are identical and consist of 40 amino acid residues each. The predicted processing site of the signal peptide occurs between the residues Ala20 and Glu21 (SignalP 3.0: http://www.cbs.dtu.dk/services/SignalP/), which is the same as in T. carinatus FX. The propeptide region and a short internal tripeptide are excised during proteolytic maturation to yield a two-chain mature zymogen. The propeptide region plays a critical role in the γ-carboxylation of Glu residues in the Gla domain [22]. The motifs for enzyme recognition for proteolytic cleavage at both the propeptide and the internal peptide are conserved in PFX1 and PFX2 cDNA sequence and are therefore expected to be similarly processed.

During injury, FX is activated to FXa by TF-VIIa [23,24] or IXa–VIIIa [23,25] complexes through a proteolytic cleavage at Arg194-Ile195 (human FX numbering) and removal of the activation peptide. The activation sites in PFX1 and PFX2 cDNA are conserved. However, activation peptides of PFX1 and PFX2 show interesting differences. PFX1 has a longer activation peptide (57 residues), which is nearly identical (single-residue difference) to that of T. carinatus FX [16]. In contrast, PFX2 has a comparatively shorter activation peptide (27 residues) [18] which is almost identical (same single-residue difference) to the PCCS.

Like all other mammalian FX, as well as group D prothrombin activators, both PFX1 and PFX2 deduced sequence have 24 conserved cysteine residues. Based on homology with mammalian and T. carinatus FX, all 12 disulphide bonds are expected to be conserved in both the PFX1 and PFX2 of P. textilis.

One crucial post-translational modification for vitamin K-dependent blood clotting factors is γ-carboxylation [26–28]. The recognition site for γ-carboxylase is located in the propeptide region of FX. Two critical residues (Phe16 and Ala10) for the recognition of carboxylase [29] are conserved in both FX isoforms. There are 12 Glu residues in the Gla domain of both types of FX of P. textilis; all are at identical positions compared with their venom counterpart PCCS [12,18], FX of T. carinatus [16] and group D prothrombin activators trocarin D and hopsarin D [11,13]. We have shown that all except Glu38 residues are carboxylated in trocarin D [11] and hopsarin [13]. Based on the homology (Fig. 1), we propose that only the first 11 Glu residues are carboxylated in both isoforms of P. textilis FX. Unlike mammalian FX [30–32], group D prothrombin activators (such as trocarin D) are not hydroxylated at Asp63 [11]. Although this residue is conserved in both PFX1 and PFX2, similar to T. carinatus FX, β-hydroxylation at this position is yet to be determined.

Figure 1.  Alignment of amino acid sequence of PFX1 and PFX2 with Tropidechis carinatus factor (F) X and group C and group D venom prothrombin activators (vPA). Identical amino acid residues are shaded in black. Residues showing >70% and >50% identity are shaded in dark gray and light gray, respectively. Regions showing maximum variation are boxed. The mature protein starts with Ala (+1) at the amino terminus of the light chain of the mature zymogen. Boundaries of each domain are marked by arrows. The putative signal peptide and the propeptide are shown by double line (from −40 to −21) and dotted line (from −20 to −1) above the sequence. Putative γ-carboxylated Glu residues, O- and N-glycosylation sites are marked by rectangle, circle and triangle, respectively.

We have shown that unlike mammalian FX [33,34], the vPAs, such as trocarin D from T. carinatus and hopsarin D from Hoplocephalus stephensi, are glycosylated in their light and heavy chains [11,13]. These prothrombin activators have O-glycosylation at Ser52 position of the light chain and N-glycosylation on the Asn254 residue of the heavy chain. Similarly, N- and O-glycosylation are also found in PCCS from P. textilis [12,18]. These sites are conserved in T. carinatus FX, PFX1 and PFX2 protein sequences. In addition, PFX1 has two potential N-glycosylation sites within the activation peptide, at identical positions compared with T. carinatus FX, whereas PFX2 has only one. However, the existence and the functional role of glycosylation at these positions in snake plasma FX are yet to be determined experimentally.

Expression analysis of PFX1 and PFX2

We studied the expression level of the two isoforms of FX compared with a housekeeping gene β-actin in the liver of P. textilis using universal primers. Preamplified cDNA was used as template. The difference in expression levels was calculated based on 2^−ΔΔCt method [35]. The expression of PFX1 is 55 878 times higher than that of PFX2 (Fig. 2). However, PFX1 and PFX2 transcripts in the venom gland and PCCS transcript in the liver were found below the detection limit of the highly sensitive real-time PCR detection system used (Fig. 2). So, our results confirm that PFX1 and PFX2 are exclusively expressed in the liver, and PCCS in the venom gland. Expression of PCCS was ∼80 times higher in the venom gland compared with PFX1 expression in the liver (Fig. 2).

Figure 2.  Real-time polymerase chain reaction (PCR) studies of tissue-specific expression of PFX1, PFX2 and pseutarin C catalytic subunit (PCCS). Amplification curves (delta Rn) of PFX1, PFX2 and PCCS in liver and venom gland tissues of Pseudonaja textilis are shown. (A) PFX1-PCCS in the liver; (B) PFX2-PCCS in the liver; and (C) PCCS, PFX1 and PFX2 in the venom gland. β-actin was used as internal control in each reaction. No template control (NTC) includes FAM background and JOE. Delta Rn represents the normalized reporter signal minus the baseline signal established in the first two cycles of PCR. Each curve was plotted from the average of triplicates. Threshold was set at 0.2 (denoted by dotted line). Quantitative analyses of gene expression of PFX1 and PFX2 in the liver (D) and those of PFX1 and PCCS in the venom glands (E) are shown. Each Ct value represents the mean and standard deviation of triplicate reactions. Fold of difference in expression level was calculated based on the 2^−ΔΔCt method [35]. Threshold was set at 0.2. In each reaction, the FAM signal corresponds to amplification of target DNA (PFX1 or PFX2 or PCCS) and the JOE signal corresponds to amplification of β-actin.

Discussion

Gene duplication and recruitment of new genes are not static or one-time events but recurrent phenomena, giving the organism required advantages for adaptation in response to any environmental changes [36,37]. As the amount of sequence information is relatively small, it is difficult to obtain a clear picture of an evolutionary path because of lack of intermediates. The evolutionary intermediates help us define putative stages in the evolutionary process. Here we describe the sequence of PFX2, an intermediate in the evolution of snake vPA. This gene/protein appears to be trapped during evolution and acts as a molecular fossil.

Prothrombin activators in the snake venoms are classified into four groups depending on the cofactor requirements for optimal activity [38]. Among them group C and group D prothrombin activators are serine proteinases and are structurally and functionally similar to the blood coagulation factors. Group D prothrombin activators are similar to mammalian FXa [11,13,39–42]. As vertebrates, snakes contain a blood coagulation system and hence FX in their plasma. We recently showed that the blood coagulation FX from T. carinatus expressed in the liver is similar to its own venom counterpart, trocarin D [16]. Based on phylogenetic analysis, we proposed that the group D prothrombin activators have evolved by the duplication and diversification of the FX gene [16]. Thus, some Australian elapids are unique compared with any other organism in possessing two parallel prothrombin activator systems. These two systems play important roles in distinctly different functions. One system exists in the plasma and acts as a hemostatic factor, whereas the other is present in the venom and acts as a toxin. Their tissue-specific expression and regulation are also different.

Pseudonaja textilis venom contains group C prothrombin activator [12]. It has catalytic and non-catalytic subunits [12]. PCCS shows ∼77% identity with trocarin D and mammalian FXa [18]. As group C and group D prothrombin activators evolved independently, we studied FX expressed in the liver of P. textilis. Interestingly, we found two isoforms of FX: PFX1 and PFX2. PFX1 is similar to snake plasma FX: (i) it shows higher sequence identity with T. carinatus FX compared with PCCS and group D prothrombin activators; (ii) the activation peptide is 57 residues long and nearly identical to plasma FX of T. carinatus (Fig. 3A); and (iii) it does not have a 13-residue insert in the SP domain that is found in PCCS [12,18] and group D prothrombin activators [11,13] (Fig. 3B). On the other hand, PFX2 is closer to PCCS, the enzymatic subunit of vPA, pseutarin C: (i) it has higher sequence identity with PCCS compared with T. carinatus FX; (ii) the activation peptide is much shorter with only 27 residues, and is almost identical to PCCS (Fig. 3A); and (iii) similar to venom counterparts, it has an insert in the same position of the serine proteinases domain (Fig. 3B). However, this insert is shorter compared with those of PCCS and group D prothrombin activators. Phylogenetic analysis of FX-like proteins (Fig. 4) shows that in group C and group D prothrombin activators such proteins are present, separate in two clades. The phylogenetic trees constructed using partial sequence analysis of genes that encode for 16S rRNA and cytochrome b [43] also separates snakes into identical clades. Further, the sequence of the inserts in the heavy chain of trocarin D is relatively different to that of PCCS (Fig. 3B), which indicates that group C and group D prothrombin activators most likely have evolved independently from the ancestral FX molecule. T. carinatus FX and PFX1 group together in phylogenetic analysis, which supports our assumption that PFX1 is the functional FX in P. textilis. In contrast, PFX2 groups together with PCCS.

Figure 3.  Structural comparison of FX and its homologs in snakes and the evolution of vPAs from FX. (A) Alignment of the nucleotide sequence of the activation peptide of PFX1 and PFX2 with the blood and vPAs. PFX1 contains a longer activation peptide nearly identical to T. carinatus FX, whereas PFX2 possesses a shorter activation peptide nearly identical to the vPAs, such as PCCS and trocarin D. (B) Alignment of the nucleotide sequence of the heavy-chain segment showing the progressive insertion during the evolution of vPA from the blood coagulation factor. (C) Schematic diagram showing the probable evolutionary path in the recruitment of FX protein as toxin in the venom.

Figure 4.  Phylogenetic relationship between FX and FX-like vPAs. A phylogenetic tree was constructed from the distance matrix using ClustalW algorithm. FX molecules are shown in open boxes and FX-like vPA are shown in shaded boxes. Nodes representing the postulated gene duplications are marked by solid arrowhead. The tree was plotted using the software dnaman. Nucleotide sequences were obtained from the following sources: human FX (AF503510); bovine FX (X00673); murine FX (BC003877); rat FX (NM017143); rabbit FX (AF003200); zebrafish FX (AF519546); T. carinatus FX (AY651849); PFX1 (DQ017705); PFX2 (DQ017706); trocarin D (DQ017707); pseutarin C catalytic subunit (AY260939); FX-like vPA from Oxyuranus scutellatus (AY940204), Oxyuranus microlepidotus (AY940205), Pseudechis porphyriacus (AY940207), Hoplocephalus stephensii (AY940208) and Notechis scutatus (AY940206).

Based on Ohno's model [2,3] and the ‘neofunctionalization’ hypothesis [44,45], we suggest a possible pathway of evolution of PCCS from the blood coagulation FX of P. textilis. The insertions, deletions and mutations designate the evolutionary path of a gene/protein. In the case of FX-like proteins, deletion in the activation peptide segment and insertions in the heavy chain help to track the evolutionary path. These events seem to have occurred in several steps and our data suggest that PFX2 is an intermediate form in the evolution of PCCS (Fig. 3). Gene duplication of the ancestral PFX1 gene (aPFX1) probably gave rise to the ancestral PFX2 (aPFX2), which evolved to the present PFX1 isoform (Fig. 3C). This aPFX2 accumulated an insertion of 27 bp in its heavy chain and a deletion of 87 bp in its activation peptide. Subsequently, there was a second gene duplication event. The expression of one of the duplicated aPFX2 was ‘turned off’ (or significantly lowered) in the liver. The event controlling the expression may have occurred either before or after the second duplication. The second of the duplicated aPFX2 accumulated more insertion in its heavy chain, probably in at least two steps, and subsequently its expression was ‘turned on’ in the venom gland to evolve as the current form of PCCS (Fig. 3C). Quantitative real-time expression analysis (Fig. 2) shows that the expression of PFX2 in the liver is ∼56 000 times lower than functionally important PFX1. Assuming that mRNA levels reflect the protein levels, the concentration of PFX2 would be too low (2.5 pM PFX2 compared with 136 nm PFX1, calculated based on mammalian plasma FX concentration) to have any significant physiological role in hemostasis. Thus, PFX2 in the liver of P. textilis is an evolutionary by-product and possibly a redundant gene and hence one may consider it as an intermediate in the evolution of PCCS from the FX gene.

Acknowledgements

This work was supported by the Academic Research Grants from the National University of Singapore.