Metagenomics strategies are successfully employed to isolate and identify enzymes with novel biocatalytic activities, or secondary metabolites from uncultivable components of microbial communities from various environmental samples [18]. A previous study has constructed a plasmid library containing approximately 50,000 clones from metagenomic DNA isolated from marine water samples [16]. The marine water samples were collected from the South China Sea (21°28′N, 109°07′E). The pH, temperature, and salinity of the collected seawater sample were 8.2, 15 °C, and 32‰, respectively. Plasmid DNA from randomly picked clones is digested with EcoRI. Different insert DNAs 1–15 kb long, with an average of approximately 4.0 kb, are obtained. This result confirms that the marine metagenomic library contained DNA molecules from uncharacterized genomes. The marine metagenome of naturally occurring microrganisms is also demonstrated to contain an immense pool of genes. Most of these genes are not represented in pure and enriched cultures established under certain selective conditions [19]. In the present study, a metagenomic library was screened using a sequence-based screening strategy. Consequently, an interesting recombinant clone, named pGXAG59, was identified. pGXAG59 possibly contained a serpin gene, and had its insert further characterized. The insert DNA had a length of 1128 bp, and had no good match with known genes at the DNA level in the database. However, insert DNA was most similar with some serpins at the amino acid level. Based on the database comparison, the cloned gene on pGXAG59 was considered as novel, and was named Spi1C. The gene had a long open reading frame of 642 bp, with a mol% G + C content of 49.92, encoding 214 amino acids. The predicted relative molecular weight (Mr) was approximately 24 kDa, and the isoelectric point was 4.33. These values were consistent with the observed Mr of 20–40 kDa of most serpin proteins [20]. Given that the unstable index was 24.90, Spi1C was considered as a relatively stable serpin. The deduced amino acid sequence of Spi1C was searched in the National Center for Biotechnology Information (NCBI) and Expert Protein Analysis System (ExPaSy) databases. The results showed that Spi1C protein shared some moderate similarities with serpins from Salinibacter ruber M8 (39% identical and 60% similar) and Spirosoma linguale DSM 74 (43% identical and 57% similar). Similarities with three other possible partial proteinase inhibitor I4 serpins were also found. The proteinase inhibitor I4 serpins were from Dyadobacter fermentans DSM 18053 (GenBank accession No. YP_003088034; 39% identical and 60% similar), from Arthrospira maxima CS-328 (GenBank accession No. ZP_03274562; 38% identical and 59% similar), and from Cyanothece sp. PCC 7822 (GenBank accession No. YP_003888593; 37% identical and 58% similar). A similarity with the hypothetical protein BACCELL_01031 (GenBank accession No. ZP_03676704) from Bacteroides cellulosilyticus DSM 14838 (36% identical and 60% similar) was also observed.

Multiple alignments of the deduced amino acids of Spi1C with the most homologous serpin proteins (NCBI database) are shown in Figure 1. The alignments revealed that the target enzymes had lower overall amino acid homologies with other serpins. The percentages of identity with the serpins from bacteria were slightly higher than with those from eukaryotic homologues. The amino acid sequence comparison revealed that the deduced Spi1C peptide shared conserved active site residues with other bacterial serpin members. An amino acid N-terminal extension presented only in the bacterial serpins (Figure 1). Therefore, Spi1C may be classified as a new proteinase inhibitor I4 serpin [21,22].

A phylogenetic tree based on the neighbor-joining method located the Spi1C protein in a distinct clade from the serpins found in other microorganisms (Figure 2). Such a placement suggested a relatively high level of divergence from bacterial serpins. The phylogenetic analysis also revealed that the Spi1C protein is not distinctively grouped with the eukaryotic homologues. This finding reflected the considerable dissimilarities between eukaryotes and prokaryotes. Sequence comparisons between the Spi1C protein and other serpins demonstrated differences clustered at the 5′ end of the coding sequence [23].

To investigate the biochemical properties of Spi1C, the gene was subcloned in a frame with a 6-histidine tag sequence into the expression vector pETBlue-2. The clone was then expressed in E. coli BL21(DE3)pLysS. The initial analysis of the crude cell lysate showed that the bacteria containing the recombinant plasmid pETBlue-2-Spi1C produced a substantial amount of the expected recombinant protein. In contrast, this protein was not detectable in the culture of the bacteria containing the parent vector pETBlue-2. The cell extracts of Spi1C were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). An increased expression of the ~28 kDa protein was observed in the cell extracts of recombinant Spi1C compared with the control. The molecular weights of the proteins were similar with those of recombinant Spi1C. Hence, Spi1C was considered as intracellularly expressed without any modification. The recombinant Spi1C protein was then purified with nickel-nitrilotriacetic acid (Ni-NTA) metal-affinity chromatography (Figure 3). The purified proteins produced a single band on the SDS-PAGE gels. Their molecular weights concurred with those deduced from the amino acid sequences of the recombinant protein.

Serpin inhibitors commonly contained two types of inhibition-activity and non-inhibition-activity members. The activity-serpin inhibitors are effective against the serine and cysteine protease families (also including the tissue protease and Cys-Asp families) [1]. The activity-serpin inhibitors (specially the Kunitz and Kazal family members) use reactive site loop structure conformational change and consequent kinetic trapping of an enzyme intermediate to effect inhibition [24]. To examine the inhibitory specificity of the sample to serine proteases, the purified Spi1C protein from a marine metagenome was assayed for inhibitory activity against three different proteases of elastase, trypsin, and α-chymotrypsin (Figure 4). The purified Spi1C protein inhibited trypsin and α-chymotrypsin, but had no effect on elastase. Among the serine proteases, Spi1C protein had inhibitory activity against trypsin, with an inhibitory constant K_i of 1.52 × 10⁻⁸ M. However, Spi1C had a comparatively weaker inhibition toward α-chymotrypsin, with an inhibitory constant of 1.79 × 10⁻⁸ M. Yang et al. have found a novel serine protease inhibitor named bicolin from the venom of Vespa bicolor Fabricius. The K_i of the bicolin toward trypsin was 5.5 × 10⁻⁷ M and no inhibition was detected to elastase and α-chymotrypsin, respectivley [21]. Lu et al. have also reported that a novel serine protease inhibitor (Bungaruskuni) from Bungarus fasciatus venom had a K_i of around 6.1 × 10⁻⁶ M to chymotrypsin [22]. Using different concentrations of Spi1C, the V_max values of trypsin and α-chymotrypsin remained consistent (Figure 5). These data indicated that Spi1C was a competitive inhibitor. The specific inhibitor activity of Spi1C to trypsin and α-chymotrypsin were 6940 and 3640 U/mg, respectively. Using trypsin as the chromogenic substrate, the optimum pH and temperature of Spi1C against trypsin activity were 7.0–8.0 and 25 °C, respectively (Figure 6). In 2009, Torres-Castillo et al. have found that the serpin OsTI 2 from the seeds of Opuntia streptacantha only inhibits the trypsin-like proteinases present in P. truncatus, P. americana, Acheta sp., and Gryllus sp. [20].

Serpins are easily identified by conserved amino acid motifs because of their high degree of structural conservation. The relationship between the function and structure of the Spi1C protein is interesting. Many kinds of serpins have been separated and characterized from living organisms using culture-dependent techniques [1]. In the current study, a novel serpin gene was identified from a marine metagenome library. However, some serpins may not have inhibitory activity. In the hinge region in the RCL conserved domain, the residues are large and polar (e.g., Lys, Asp, Glu, and Ser); in inhibitory serpins, this region is occupied by small residues (e.g., Ala, Thr, and Ser). Laskowski and Kato [23] have pointed out that P1 site residues of RSL conserved domain in trypsin inhibitors are positively charged residues (Arg or Lys). In the chymotrypsin inhibitor, the residue are usually large and hydrophobic (Leu, Phe, or Tyr). The P1-reactive-site residue of the venom basic protease inhibitor IX has been identified as Asn17, whereas the same position in 2 RCLs of Spi1C is possible Gly and Phe, respectively (Figure 1). Recently, the P1 site residues His and Asn for α-chymotrypsin binding have also been reported [10]. The specificity of Kunitz/bovine pancreatic trypsin inhibitors toward serine proteases is closely associated with P1 amino acids. However, aside from surrounding the reactive site, the residues present in the weak contact loop are also important in different interactions with various serine proteases [23,25]. Thr and Leu were often at the P1 and P1′ positions of RCL, and the protein contained most of the conserved residues [26,27]. Apart from the strange RCL sequence, Spi1C had the inhibitory characteristics of trypsin and α-chymotrypsin. The functional characterization of Spi1C may provide new insights into the relationship among the sequence, structure, and activity of known serpins [27]. Elucidating the catalytic mechanism via studies of the three-dimensional structures of the Spi1C protein using X-ray diffraction methods is an interesting research direction.

The protein expression vector used was pETBlue-2. E. coli NovaBlue (Novagen) strain was used as the screening host for the pETBlue-2 bearing inserts. The bacterial strain for the expression of the recombinant protein was E. coli BL21(DE3)pLysS (Novagen). All E. coli bacterial cultures were grown at 37 °C on Luria–Bertani (LB) agar plates or media. Where appropriate, the media were modified with 100 μg ampicillin/mL, 50 μg carbenicillin/mL, 15 μg tetracycline/mL, or 34 μg chloramphenicol/mL.

All DNA manipulations, including cloning and subcloning, transformation of E. coli cells, and polymerase chain reaction (PCR), were performed according to standard techniques [28] or following the instructions of the manufacturer, unless indicated otherwise. Protein preparation and analysis, including protein extraction from E. coli, protein quantification, and SDS-PAGE, were performed as described by standard protocols [29].

DNA sequence analysis was performed using the BigDye Terminator Cycle sequencing kit on an ABI Prism 3700 DNA analyzer (Applied Biosystems, USA). Protein translation was carried out using the web-based translation tool at the ExPaSy homepage [30]. Sequence predictions were retrieved from the protein and nucleotide databases at NCBI Entrez page [31]. Sequence similarity searches were performed with the BLAST 2.0 program. The amino acid sequence alignment of Spi1C with homologous proteins was done using the AlignX program, a component of the Vector NT1 suite (InforMax, North Bethesda, MD, USA) using the blosum62mt2 scoring matrix. Based on the comparison of the deduced amino acids, a putative gene (Spi1C) encoding a novel serpin was identified and further characterized. A phylogenetic tree was constructed using the neighbor-joining method with MEGA 4.0 software [32]. Boot-strapping values were used to estimate the reliability of the phylogenetic reconstructions (1000 replicates).

The Spi1C nucleotide sequence was amplified from the plasmid pGXAG59 isolated from a serpin-producing clone. PCR was carried out in a total volume of 50 μL containing 2.5 mM MgCl₂, 10 mM Tris-HCl (pH 8.4), 50 mM KCl, 0.2 mM deoxynucleotide triphosphate, 0.4 μM each primer, 1.0 unit Vent DNA polymerase (NEB, USA), and 10 ng of plasmid template. The restriction enzyme sites (underlined) for BamHI and HindIII were designed in the forward primer/reverse primer (5′-TTAGGATCCGATGTTCCTTATGAACGCC-3′/5′-ATAAGCTTCTCCGGCTGCATCACTTTC-3′). The PCR cycling consisted of denaturation steps (96 °C for 2 min), 30 cycles at 94 °C for 40 s, 58.5 °C for 30 s, and 72 °C for 2 min, as well as a final extension step at 72 °C for 10 min. After amplification, the PCR product mixture was digested by BamHI as well as HindIII, and directly ligated into the pETBlue-2 (Novagen) expression vector cleaved with the same enzymes. The resulting recombinant plasmids were transferred into NovaBlue (Novagen) competent cells and placed on LB selection plates. After overnight incubation at 37 °C, positive white colonies were picked for the isolation of the recombinant expression plasmid. The recombinant expression plasmid was then introduced into E. coli BL21(DE3)pLysS to express the target protein.

The transformed bacterial cells were cultured in an LB medium containing 50 μg carbenicillin/mL and 100 μg chloramphenicol/mL at 37 °C. Protein expression was induced by the addition of 1 mM isopropyl-β-d-thiogalactopyranoside when the optical density at 600 nm reached 0.6. After incubation for an additional 6 h, the cells were harvested, washed twice with phosphate-buffered saline (pH 7.6), and lysed by sonication in 10 mL of 20 mM Tris-HCl (pH 8.0). The lysate was centrifuged twice at 30,000× g for 20 min at 4 °C. About 1 mL of the supernatant was diluted with 5 mL of column buffer (20 mM Tris-HCl, pH 8.0, 10 mM imidazole, 300 mM NaCl). The diluted supernatant was then applied onto an equilibrated Ni-NTA column containing 1 mL of Ni-NTA-agarose (Novagen). After washing with 5 mL of column buffer containing 10 mM imidazole, the recombinant protein was eluted with column buffer containing 250 mM imidazole and the target fraction (~600 μL) was collected. Immediately after elution from the column, 1,4-dithiothreitol (Promega) was added to the protein solution to a final concentration of 5 mM. The protein concentration was determined using a Bio-Rad protein assay kit, with bovine serum albumin as the standard. The protein purified with the Ni-NTA column was used for enzyme activity assays.

The Spi1C nucleotide sequence was deposited in GenBank under accession number JF815525.