The drosomycin multigene family: three-disulfide variants from Drosophila takahashii possess antibacterial activity

Drosomycin (DRS) is a strictly antifungal peptide in Drosophila melanogaster, which contains four disulfide bridges (DBs) with three buried in molecular interior and one exposed on molecular surface to tie the amino- and carboxyl-termini of the molecule together (called wrapper disulfide bridge, WDB). Based on computational analysis of genomes of Drosophila species belonging to the Oriental lineage, we identified a new multigene family of DRS in Drosphila takahashii that includes a total of 11 DRS-encoding genes (termed DtDRS-1 to DtDRS-11) and a pseudogene. Phylogenetic tree and synteny analyses reveal orthologous relationship between DtDRSs and DRSs, indicating that orthologous genes of DRS-1, DRS-2, DRS-3 and DRS-6 have undergone duplication in D. takahashii and three amplifications (DtDRS-9 to DtDRS-11) of DRS-3 have lost WDB. Among the 11 genes, five are transcriptionally active in adult fruitflies. The ortholog of DRS (DtDRS-1) shows high structural and functional similarity to DRS while two WDB-deficient members display antibacterial activity accompanying complete loss or remarkable reduction of antifungal activity. To the best of our knowledge, this is the first report on the presence of three-disulfide antibacterial DRSs in a specific Drosophila species, suggesting a potential role of DB loss in neofunctionalization of a protein via structural adjustment.


DRS-6
were not expressed in all the developmental stages of Drosophila 14 . When challenged, the expression level of DRS, DRS-2 and DRS-3 were up-regulated but only DRS was strongly induced. DRS-1 and DRS-6 are not expressed even in the presence of microbes 13 . Homologous genes of DRS are also found in the melanogaster species group, all retained as a multigene family 15,16 .
In this work, we conducted a large scale of survey on all sequenced Drosophila genomes, from which we identified a complete set of DRS peptides in 14 species, including a new multigene family of DRSs (designated DtDRS) in Drosophila takahashii, a Southeast Asian species belonging to the takahashii subgroup. It is striking that two members with three DBs (DtDRS-11 and DtDRS-11d) possess antibacterial activity accompanying complete loss or remarkable decrease of antifungal function. A combination of sequence, structural and functional analyses suggests that a DB loss-mediated structural modification is likely implicated in the emergence of antibacterial function in the two WDB-deficient DRSs.

Results
Gene Expansion of the DRS Family in D. takahashii. By TBLASTN search of the whole-genome shotgun contigs (wgs) databases of Drosophila in GenBank ( March 18, 2015), we identified all DRS-type antifungal peptide genes in 14 species, which all belong to the Drosophila melanogaster species group 17 . Figure 2 shows their genomic position. As observed previously 13,16 , the genomic arrangement of the DRS family members is conserved in the Oriental lineage where three distinct clusters (C1 to C3) are separated by two long spacers of 18-38 kb, which is different from species from the ananassae and montium subgroups whose DRS gene clusters display a relatively scattered distribution pattern (Fig. 2). In view of basal position of these two subgroups within the melanogaster species group 18 , an ancient DRS could have undergone independent expansion between these basal species and the monophyletic Oriental lineage. In spite of overall conservation in the Oriental lineage, gene turnover frequently occurred in this family, which can be outlined as follows: (a) Seven paralogous genes are conserved among D. elegans, D. erecta, D. simulans and D. melanogaster, suggesting that they originated by early gene duplication in the common ancestor of the Oriental lineage; (b) D. yakuba and D. sechellia lost one member in C3, and one member in this cluster became a pseudogene in D. biamipes; (c) D. ficusphila contains the smallest gene number on account of loss in C2 and C3; (d) Both D. rhopaloa and D. takahashii have undergone gene duplication in C3, and in D. takahashii C2 has also expanded to four paralogous genes (Fig. 2).
D. takahashii contains 11 homologous genes of DRS (termed DtDRS-1 -DtDRS-11) and one pseudogene with a premature stop codon in the signal peptide-encoding region (named PseudoDtDRS) (Fig. 3). Unlike DRSs from other Drosophila species, the DtDRS multigene family contains three members (DtDRS-9 to DtDRS-11) without WDB. The loss attributes to mutations of codons encoding two cysteines (Cys 1 and Cys 8 , TGT) into a codon of Phe (TTT) and a stop codon (TGA). From the genome of D. lutescens, a sibling species of D. takahashii, we also amplified a WDB-deficient DRS (designated DlDRS, Supplementary Figs S1 and S2), suggesting that the history of these unique DRS molecules could trace back to the common ancestor of the takahashii subgroup. DtDRSs and DRSs both share 15 identical sites, including six cysteines, three glycines (Gly 5 , Gly 9 , and Gly 31 ), two acidic residues (Asp 1 and Glu 42 ), two tryptophans (Trp 14 and Trp 40 ), one serine (Ser 4 ), and one histidine (His 32 ) (numbered according to DRS) (Fig. 3). Two of them (Asp 1 and Trp 14 ) have been identified as functionally important residues of DRS involved in the interaction with fungi 3,7,8 (Fig. 1).  constitute a monophyletic clade in the phylogenetic tree (Fig. 4A), it is reasonable to infer that the loss of WDB in D. takahashii occurred only once during evolution. Figure 4B shows the synteny relationship between DtDRS  Hydrophobic or aromatic residues are shadowed in green, hydrophilic in cyan, acidic in red, and cysteines in yellow. Secondary structure elements (cylinder: α -helix; arrow: β -strand) and disulfide bridge connectivities are extracted from the structural coordinates of DRS with WDB represented by a dotted line. "x" in pseudoDtDRS indicates position of one nucleotide deletion resulting in a premature stop codon. and DRS genes, giving strong support for the same conclusion based on the phylogenetic analysis ( Fig. 4A), in particular, for the branches with low bootstrap values (e.g. < 50%).

Molecular Characterization of DtDRS Genes.
To isolate cDNA clones encoding DtDRS-1 to DtDRS-11, we designed a series of degenerate forward primers (Table S1) in combination with 3AP to perform RT-PCR 19 . Our RT-PCR experiments confirmed that four DtDRS genes were transcriptionally active in D. takahashii adults without experimental stimulus, including DtDRS-1, DtDRS-2, DtDRS-6, and DtDRS-11 (Fig. 5A), which correspond to their orthologs in D. melanogaster (DRS, DRS-5, DRS-4, and DRS-3). All these D. melanogaster genes are also transcriptionally active in adults in the absence of experimental infection 14 . DtDRS-4 is a gene whose transcription depends on microbial stimulus, indicating its inducible feature. This appears to be different from its othologous gene -DRS-2. In adult D. melanogaster, DRS-2 is transcribed in a constitutive manner (Fig. 5A). Using degenerate DtDRS-3/5-F and DtDRS-7/8-F primers, we failed to amplify PCR products for four genes (DtDRS-3, − 5, − 7 and − 8) from the first-strand cDNA templates prepared from both non-challenged and challenged adult fruitflies. Using degenerate DtDRS-2/4/9-10-F, we obtained PCR products from the challenged or non-challenged cDNA template, but all clones sequenced carry inserts encoding DtDRS-2 or DtDRS-4 without DtDRS-9 and DtDRS-10. Among these untranscribed genes, DtDRS-3/DtDRS-5 are orthologous to DRS-1 and DtDRS-7/DtDRS8 to DRS-6 ( Fig. 4B), and interestingly these two orthologous genes in D. melanogaster are also transcriptionally inactive in adults and other developmental stages and even after challenge, suggesting an overall conserved transcriptional pattern between the two multigene families. However, considering only adult fruitflies analyzed for DtDRS genes, it is likely that these untranscribed genes are functional in other developmental stages or in response to specific microbial infections given that they are conserved over more than 10 million years. Sequence analysis of the isolated cDNA clones revealed some polymorphic sequences for the transcribed DtDRSs ( Fig. 5B; Supplementary Fig. S2). Peptide Identification. There was no precedent for a DRS with three DBs in Drosophila reported so far. To study the potential function of these unusual peptides, we chose DtDRS-11 and DtDRS-11d, a cloned polymorphic cDNA sequence of DtDRS-11, as representatives for chemical synthesis. They both differ by four residues (L3K, M13A, T35S and E43M) (Fig. 6A). Oxidized DtDRS-11 and DtDRS-11d were produced via in vitro folding from their reduced peptides, with retention time of 18.5 and 22.5 min, respectively, on a C 18 column (Fig. 6B). Their experimental molecular weights (MWs) were 4800.16 and 4859.8 Da, as determined by MALDI-TOF (Fig. 6C), matching their calculated MWs (Fig. 6A). To study the potential structural and functional effect of WDB in DRS-3, we also chemically synthesized and oxidized its WDB-deficient variant (termed DRS-3-WDB) (Fig. 6). In addition, using a prokaryotic system, we prepared recombinant DtDRS-1, the ortholog of DRS, for comparison with the WDB-deficient DRSs at structural and functional levels. The reason we chose recombinant expression of DtDRS-1 was because there was difficulty in the chemical synthesis of this peptide with four disulfide bridges. From the chemical nature, peptides derived from recombinant or chemical synthesis are the same so long as they are characterized by standard biochemical techniques, such as RP-HPLC, MALDI-TOF and circular dichroism (CD), as described in this work. Recombinant DtDRS-1 was eluted at 22.4 min of retention time and an experimental molecular mass of 4904 Da, well matching its theoretic molecular mass of 4902 Da (Fig. 6). The eluted peptides were further purified by RP-HPLC to ensure their purity > 95%. Functional Divergence between Three-and Four-Disulfide DRSs. To assess potential antimicrobial function of DtDRS-1, DtDRS-11, DtDRS-11d and DRS-3-WDB, we firstly assayed their effect on a series of filamentous fungi and the yeast Candida albicans. As a result, we found that DtDRS-1 had highly similar antifungal spectrum and potency to DRS, both inhibiting the growth of Aspergillus fumigatus (strain CEA17 other than YJ-407), A. nidulans (strains A28 and RCho15), A. niger, Geotrichum candidum, and Neurospora crassa with lethal concentrations (C L ) ranging from 0.1-2.6 μ M (Table 1). Like DRS, DtDRS-1 is also a strictly antifungal peptide without activity on the bacteria tested here. The most remarkable discovery here is that DtDRS-11d has lost its antifungal function but evolved activity on two Gram-positive bacteria Bacillus megaterium and Micrococcus luteus (Fig. 7A) with a C L of 0.98-1.08 μ M (Table 1). Similarly, DtDRS-11 is also an antibacterial peptide but with some activity against N. crassa (Table 1). Different from these two naturally-occurring WDB-deficient DRSs, the engineered DRS-3-WDB exhibited antifungal activity on two species (G. candidum and N. crassa) with a C L of 3.72-4.89 μ M but no activity on the bacteria used here (Table 1).
Because many antibacterial peptides kill their targets via a membrane disruption mechanism 20 , we examined a possible impact of DtDRS-11 and DtDRS-11d on membrane permeability of B. megaterium cells via propidium iodide (PI), a fluorescent nucleic acid-binding dye. The results showed that these two peptides at 5× C L caused an immediate fluorescence increase upon exposure of the peptides even though the effect is much milder than that observed with the positive control meucin-18 21 , indicating that bacterial membrane integrity was affected. On the contrary, no fluorescence increase was observed after B. megaterium cells were exposed to vancomycin at 10× C L (Fig. 7B). To evaluate the stability of the WDB-deficient peptide DtDRS-11, we assayed its antimicrobial activity in water, insect saline or insect haemolymph. In these three environments, DtDRS-11 displayed similar activity (Fig. 8), revealing its resistance on insect blood proteases.

Structural Basis of Functional Divergence.
To understand the structural basis of antibacterial activity in both DtDRS-11d and DtDRS-11, we compared their CD spectra with those of the three antifungal DRSs, including DRS, DtDRS-1 and DRS-3-WDB (Fig. 9). It is known that DRS adopts a rigid and compact structure with a high content of α -helix (25%) and β -sheet (29.5%) 22 . The CD spectra of DRS were identified by maxima at 188 nm and minima at 207 nm, indicative of the presence of a CSα β structure 23 . In addition to these two typical signals, it had one negative band arround 217-218 nm (Fig. 9), previously seldom observed in members from the same structural superfamily, such as scorpion Na + channel toxins 23,24 . The negative band at this position is usually ascribed to β -sheet and its presence thus reveals a high content of β -sheet residues in DRS, as mentioned above. The CD spectrum of DtDRS-1 was nearly the same with that of DRS (Fig. 9A), in accordance with their functional similarity (Table 1). In comparison with DRS, the three WDB-deficient peptides displayed clearly visible modifications in their CD spectra: (a) In DtDRS-11d and DRS-3-WDB, the negative band at 217-218 nm disappeared       whereas DtDRS-11 remained but the intensity slightly decreased as compared to DRS (Fig. 9B-D), indicating that these three WDB-deficient peptides had a lower content of β -sheet than DRS; (b) The CD spectra of DtDRS-11d and DtDRS-11 both crossed the baseline once at 192 nm, blue-shifted 3 nm relative to DRS (195 nm) (Fig. 9), and their negative minima were also blue-shifted from 206 nm of DRS to 203 nm of DtDRS-11d and 204 nm of DtDRS-11 ( Fig. 9). No such shift was observed in DRS-3-WDB. In the two antibacterial variants, the shifted minima next to 202 nm, a signal for random coli, suggesting their structures were more flexible than DRS and DRS-3-WDB.
The CD spectra were analyzed by CDSSTR to estimate percentages of peptide secondary structure element contents 25 (Table 2). For all calculations, the NRMSD values 26 ranged from 0.015 to 0.025 (Table 2), suggesting a good correlation between them. The results showed that the thee WDB-deficient peptides had similar α -helical contents (18-21%) to DRS (20%) ( Table 2), indicating that the loss or deletion of WDB led to no remarkable impact on the α -helical formation. However, such modification resulted in a significant reduction in the β -sheet content from 27% of DRS to 22% of DRS-3-WDB and 16% of DtDRS-11 and DtDRS-11d, in line with the disappearance or the intensity decrease of the band at 217-218 nm in their CD spectra ( Fig. 9; Table 2). According to the unordered contents, we can rank the structural rigidity of these peptides as follows: DRS = DtDRS-1 > DRS-3-WDB > DtDRS-11 > DtDRS-11d, suggesting that more structural flexibility derived from the evolutionary loss of WDB in DtDRS-11 and DtDRS-11d might be a direct cause of the emergence of antibacterial activity from a four-disulfide DRS scaffold. This analysis also provides a reasonable structural explanation for the lack of antibacterial activity in DRS-3-WDB.

Discussion
Gene Duplication and Positive Selection. Gene duplication followed by positive selection represents a major event in the evolution of immune genes, presumably due to the need to cope with rapidly diversifying pathogens. However, three classical statistic models, including M2a and M8 implemented in PAML 27 , and mechanistic-empirical model (MEC) implemented in Selecton, which takes into account the physicochemical properties of amino acids 28 , all detected no positive selection signals in the DtDRS multigene family (data not shown), in agreement with several previous studies on the evolution of DRS in other Drosophila species 15,16 . It is known that the absence of positive selection is common to Drosophila antimicrobial peptide (AMP) gene families 29 , which might be related to two factors: (a) non-coevolving saprophytic organisms Drosophila meet, and (b) multiple AMP genes induced by infection, both leading to selection for speed and efficiency of expression of AMPs towards infection rather than amino acid modification via accelerated evolution 16 . From a functional viewpoint, the absence of adaptive amino acid substitutions in these two multigene families (DRS and DtDRS) is also likely due to the constraint of their potential house-keeping functions beyond immunity in development, diapause, fertility and lifespan 30,31 . Also, the power limitation of statistical approaches is another reason of detecting no positive selection because our experimental data have clearly demonstrated that cysteine mutations-associated functional diversification had occurred between three-and four-disulfide-bridged members of the DRS family.

Contribution of Gene Duplication to D. takahashii. Several lines of evidences suggest that although
D. takahashii has more DRS genes by duplication, its DRS-based antifungal immunity could be similar to D. melanogaster: (a) Firstly, DRS is an important component of antifungal defense in D. melanogaster 32 . The ortholog of DRS in D. takahashii (DtDRS-1) possesses nearly the same potency against filamentous fungi (Table 1); (b) Secondly, DtDRSs and DRSs exhibit a similar transcriptional pattern, both having five transcriptionally active orthologs in adult fruitflies; (c) Thirdly, genes derived from the C2 cluster all are transcriptionally inactive in our study. However, some members of the DtDRS family conferring antibacterial immunity are not still reported in D. melanogaster. It remains an open question whether these species-specific duplicates contributes to other biological processes, as mentioned above 31 . Given that gene duplicates tend to have divergent expression patterns 33 , a detailed comparison of these differentials between the DRS and DtDRS families will help understand the biological and evolutionary significance of gene duplication in D. takahashii.

DB Loss and Functional Neofunctionalization.
As mentioned in Introduction, WDB is one highly exposed disulfide bridge related to peptide function. Deleting the WDB of the scorpion Na + channel toxin BmKM1 dramatically reduced its potency due to destruction of a local functional region stabilized by this WDB 34 . Evidence in favor of functional importance of WDB in the DRS family members include: (a) All the four-disulfide DRS homologs characterized so far (e.g. DRS-2 and DtDRS-1) exhibit strictly antifungal activity with a rigid structure 14 ; (b) Functional exertion of DRS depends on a rigid scaffold stabilized by the WDB to sustain its scattered functional sites onto the molecular surface (Fig. 1). This is consistent with the absence of  Table 2. Comparison of secondary structure element contents (%) of DtDRS-11d, DtDRS-11 and DRS-3-WDB with DRS. Note: The secondary structure element contents were estimated from the CD data by CDSSTR. NRMSD (normalized root-mean-square deviation) was used to compare how well the best calculated structure correlates with the experimental data 26 .
Scientific RepoRts | 6:32175 | DOI: 10.1038/srep32175 antifungal activity in DtDRS-11d and the weak antifungal activity in DtDRS-11 even if they both possess nearly identical functional amino acids to DRS (Figs 1 and 2). The increase in the unordered content accompanying the decrease in the β -sheet content in both DtDRS-11d and DtDRS-11 could attribute to the N-terminal rigid structure destroyed due to the WDB loss and thus a flexible N-terminus renders the functional Asp1 in a position unsuitable for interaction with fungi 7,8 (see Fig. 1). On the contrary, a conformationally flexible structural region is functionally important in peptide's binding to bacterial membrane. For example, the long N-terminal loop is a key functional region of insect defensins in bacterial killing 35 and a series of mutational experiments have shown that a well-defined CSα β -type defensin structure is not an advantage in its antibacterial function 36 . This reasonably explains why a structurally more rigid four-disulfide DRS lacks antibacterial activity while a structurally looser three-disulfide DRS possesses such activity. Apart from the cysteine loss leading to structural and functional changes described here, reduction of DBs has also been found to unmask potent antimicrobial activity of human β -defensin-1 37 . In addition, recent studies demonstrated that DBs in several cysteine-rich antibacterial peptides (e.g. human β -defensin-3, the designed NvBH, and porcine PG-1) are dispensable for their function [38][39][40] . Taken together, all these observations support a role of the WDB loss in developing antibacterial activity from a rigid scaffold. The membrane-disruptive activity of DtDRS-11d and DtDRS-11 (Fig. 7B) suggests their ability in forming an amphiphilic architecture in a membrane environment via structural flexibility 21 .
In addition to the absence of WDB, one might argue that the target's alteration in the two naturally occurring WDB-deficient variants is also likely associated with other amino acid site mutations. To answer this question, we compared amino acid sequences between the four-disulfide DRSs (antifungal) and the WDB-deficient homologues (antibacterial) (Fig. 3) and found that these homologues only contain two group-specific residues at sites 12 (Met) and 37 (Phe) (numbered according to DtDRS-11) whose location respectively corresponds to the N-terminal loop (n-loop) preceding the α -helix and the γ -core linking two β -strands of DRS (Fig. 1), two regions previously identified as key antibacterial elements of insect defensins 35,36,41 . These two sites are occupied by hydrophobic side-chains and are situated on one side of the molecule, especially in the functional region of the structurally similar antibacterial insect defensins, providing a structural basis for its antibacterial function. Therefore, if we consider that the rigid structure destruction by the loss of WDB is a prerequisite for the target's alteration in DtDRS-11 and DtDRS-11d, the two group-specific residues could play a secondary role in further increasing the molecular flexibility following the loss of WDB. This is further strengthened by the structural and functional data of DRS-3-WDB which lacks the two specific residues (Fig. 6). Artificial deletion of WDB in DRS-3 leading to no target's transfer suggests that the evolutionary emergence of antibacterial function in an ancestral four-disulfide DRS scaffold is a gradual process, in which the WDB loss and mutations in key regions are involved.
It is long accepted that DBs have been added to proteins during evolution to enhance their stability for a fluctuating cellular environment 42,43 . DB reshuffling is also found in the evolution of an ape placental ribonuclease 44 . However, the loss of DBs in protein evolution is rarely reported. Herein we show that evolutionary loss of DBs might represent a new mechanism for functional diversification of antifungal peptides. For the DtDRS multigene family, the WDB loss can be considered as an evolutionary advantage for neofunctionalization of duplicated copies in a specific lineage through increasing structural flexibility to alter the target of a member.

Materials and Methods
cDNA Cloning. Microbial challenge was performed by pricking of D. takahasii adults with a thin needle previously dipped into a concentrated microbial culture of Micrococcus luteus (Gram-positive bacterium) and Neurospora crassa (filamentous fungus). Total RNA was prepared from either non-challenged or challenged D. takahasii adults with Total RNA Isolation Reagent and its reverse transcription to the first-strand cDNA was performed by the EasyScript First-Strand cDNA Synthesis Kit primed by a universal oligo(dT)-containing adaptor primer (dT3AP) 19 . Reverse transcription PCR (RT-PCR) was carried out by a forward primer designed based on the genomic DNA sequence of a predicted DtDRS gene (Table S1) combined with the universal reverse primer 3AP 3 . PCR products were ligated into pGM-T vector and resultant recombinant plasmids were transformed into E. coli DH5α . Recombinant plasmids were sequenced with T7 and SP6 primers.

Preparation of Peptides.
Linear DtDRS-11 and DRS-3-WDB were chemically synthesized by ChinaPeptides Co., Ltd. (Shanghai, China) and DtDRS-11d by SBS Genetech Co., Ltd (Beijing, China). A dimethyl sulfoxide (DMSO)-based method, previously employed for the synthesis of the three disulfide-bridged Tityus kappa toxin and some iberiotoxin analogs 45 , was used to prepare oxidized products of DtDRS-11d in an alkaline environment with some modifications. In brief, crude synthetic peptides were dissolved in 100 μ l of 10% DMSO/ H 2 O solution (v/v) with a peptide concentration of 2 mM. Following 30 min of incubation at room temperature, 900 μ l of 0.1 M Tris-HCl buffer (pH 8.5) was added to give a final peptide concentration of 0.2 mM. The mixture was incubated at 25 °C for 48 h. Peptides were purified to homogeneity by reversed-phase high-pressure liquid chromatography (RP-HPLC) with a C 18 column (Agilent Zorbax 300SB, 4.6 mm × 150 mm, 5 μ m). Elution was carried out with a linear gradient from 0 to 60% acetonitrile in 0.05% (v/v) TFA(v/v) within 40 min at a flow rate of 1 ml/min. For DtDRS-11 and DRS-3-WDB oxidative refolding, peptide samples were dissolved in 0.1 MTris-HCl buffer (pH 8.5) to a final concentration of 0.5 mM and incubated at 25 °C for 48 h. Peptides were purified by RP-HPLC.
For recombinant preparation of DtDRS-1, we chose a mutation strategy to make its expression vector from that of DRS 46 given only one amino acid difference (S29V) between them (Fig. 3). Firstly, we designed two back-to-back primers (DRS-S29V-F and DRS-S29V-R) (Supplementary Table S1) to construct the recombinant plasmid pGEX-6P-1-DtDRS-1 by using pGEX-6P-1-DRS as template for inverse PCR 46 . Methods for the expression and purification of DtDRS-1 have been described in our previous paper that reported the work of the first prokaryotic production of DRS 46 . In brief, the pGEX-6P-1-DtDRS-1 plasmid was transformed into E. coli BL21(DE3)pLysS host cells and the expression of a fusion protein product (glutathione-S-transferase (GST)-DtDRS-1) was induced by 0.1 mM IPTG at an OD 600 of 0.6. E. coli cells were harvested after induction for 4 hr at 37 °C. Fusion proteins were acquired from the supernatant of E. coli cell lysate after sonication, followed by affinity chromatography with glutathione-Sepharose 4B beads (GE Healthcare, USA). After washing by PBS buffer (140 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , pH 7.3), fusion proteins were on-column digested with enterokinase (Sinobio Biotech Co. Ltd, Shanghai, China) at 4 °C overnight. Finally, RP-HPLC was applied to separate DtDRS-1 from GST in the same condition as described above.
Purity and molecular masses of all peptides were determined by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) on a Kratos PC Axima CFR plus (Shimadzu Co. LTD, Kyoto, Japan). Circular Dichroism Spectroscopy. Circular dichroism (CD) spectra of all peptides described here were recorded on Chirascan ™ -plus circular dichroism spectrometer (Applied Photophysics Ltd, United Kingdom) at room temperature from 185 to 260 nm with a quartz cell of 1.0 mm thickness. Spectra were measured at a peptide concentration of about 0.10-0.15 mg/ml in water. Data were collected at 1 nm intervals with a scan rate of 60 nm/min. Secondary structure elements of peptides were estimated in DICHROWEB, an online server for protein secondary structure analyses from CD spectroscopic data (http://dichroweb.cryst.bbk.ac.uk/html/home.shtml). The mothed used was CDSSTR that implements the variable selection method by performing all possible calulations using a fixed number of proteins from the reference set 6 optimised for 185-240 nm 25 . This method probably produces the most accurate analysis results 47 .
Antimicrobial Assays. Antimicrobial activity of peptides was assessed by the inhibition zone assay 3,48 .
Membrane permeability assay was performed according to the method previously reported 48 . Sources of microbial strains used here are listed in Supplementary Table S2.