Conodipine-P1-3, the First Phospholipases A2 Characterized from Injected Cone Snail Venom*

Cone snail venom is a wide source of active molecules that have potential pharmacological and biotechnological applications. Several proteins have been reported in the venom of cone snails. Here we describe the isolation and characterization of the sPLA2 Conodipines P1-3 from the injected venom of Conus purpurascens. We employed a combined proteotranscriptome approach to obtain the full sequences these Conodipines. The activity of Conodipine-P1 was assessed by a mass spectrometry-based method, which provides the first detailed Conodipine activity analysis. Graphical Abstract Highlights Three novel Conodipines P1-3 in the injected venom of Conus purpurascens. Conodipines P1-3 have consensus catalytic characteristics of sPLA2. We determined multiple modification sites in Conodipines P1-3. Evaluated the activity of Conohyal-P1 by a MS-based method. The phospholipase A2 (PLA2s) superfamily are ubiquitous small enzymes that catalyze the hydrolysis of phospholipids at the sn-2 ester bond. PLA2s in the venom of cone snails (conodipines, Cdpi) are composed of two chains termed as alpha and beta subunits. Conodipines are categorized within the group IX of PLA2s. Here we describe the purification and biochemical characterization of three conodipines (Cdpi-P1, -P2 and -P3) isolated from the injected venom of Conus purpurascens. Using proteomics methods, we determined the full sequences of all three conodipines. Conodipine-P1-3 have conserved consensus catalytic domain residues, including the Asp/His dyad. Additionally, these enzymes are expressed as a mixture of proline hydroxylated isoforms. The activities of the native Conodipine-Ps were evaluated by conventional colorimetric and by MS-based methods, which provide the first detailed cone snail venom conodipine activity monitored by mass spectrometry. Conodipines can have medicinal applications such inhibition of cancer proliferation, bacterial and viral infections among others.


In Brief
Cone snail venom is a wide source of active molecules that have potential pharmacological and biotechnological applications. Several proteins have been reported in the venom of cone snails. Here we describe the isolation and characterization of the sPLA 2 Conodipines P1-3 from the injected venom of Conus purpurascens. We employed a combined proteotranscriptome approach to obtain the full sequences these Conodipines. The activity of Conodipine-P1 was assessed by a mass spectrometry-based method, which provides the first detailed Conodipine activity analysis.

Graphical Abstract
The phospholipase A 2 (PLA 2 ) 1 superfamily (EC 3.1.1.4) comprises a broad range of small enzymes that catalyze the hydrolysis of phospholipids at the sn-2 ester bond to produce lysophospholipids and fatty acids (1,2). The fatty acid produced by the PLA 2 s catalysis can be potent mediators of inflammation and signal transduction and could act either as a second messenger or as a precursor of eicosanoids, (3). Additionally, the lysophospholipids exert a wide variety of biological effects on cells and tissues (4 -6). Fifteen groups (I to XV) and various subgroups of PLA 2 have been identified including five distinct types of enzymes: secreted PLA 2 s (sPLA 2 s), cytosolic PLA 2 s (cPLA 2 s), Ca 2ϩ -independent PLA 2 s (iPLA 2 s), platelet-activating factor acetyl hydrolases/oxidized lipid lipoprotein associated PLA 2 s (LpPLA 2 s), adipose PLA 2 s (AdPLA 2 s), and lysosomal PLA 2 s (LPPLA 2 s). Most of these enzymes have been extensively studied as they play a major role in the regulation of phospholipid turnover, membrane trafficking and fluidity, cell maturation, apoptosis, and the production of hormones such as prostaglandins (7)(8)(9)(10).
Venom-derived sPLA 2 s can be either monomeric or (noncovalent or covalent) homo-or hetero-dimers or larger oligomers. Despite this structural diversity, the mechanism of hydrolysis appears to be similar. Venom-derived sPLA 2 s are nonselective toward the fatty acid type at the sn-2 position of phospholipids. The catalytic cleft of the venom sPLA 2 s is surrounded by hydrophobic residues that in the presence of Ca 2ϩ bind the substrate. Four key residues are essential for catalysis sPLA 2 s: the His/Asp dyad, Tyr and Asp (5, 28 -30).
The sPLA 2 s in the venom of cone snails (conodipines) expand the reach of these enzymes. Conodipines (Cdpi) are composed of two chains (alpha and beta), and are classified within PLA 2 group IX. The first conodipine described was a partial sequence for conodipine-M (Cdpi-M), isolated from the dissected venom of C. magus (31). Recently, conodipines were reported from the transcriptome of C. consors (32), C. victorae (33), C. tribblei (34), C. lenavati (UniProt A0A0K8TTR8), C. monile (UniProt A0A161J284), C. ermineus (35) and C. geographus (36). Using sequences from tryptic peptides and the transcriptome, ten conodipines were found in the dissected venom of C. geographus, which differed in sequence and inter-cysteine loop spacing. Conodipines have a molec-ular mass near 14 kDa with their two subunits joined by one or more disulfide bonds. However, there is poor sequence homology between conodipines and other sPLA 2 s, despite having the conserved Asp/His catalytic dyad.
Here, we report the isolation and characterization of three conodipines, Cdpi-P1, -P2 and -P3 found the injected venom of C. purpurascens, the only fish-hunting cone snail species that inhabits the tropical Eastern Pacific region (Graphical Abstract Fig.). These are the first PLA 2 s reported from the injected venom of a cone snail. We have used combined proteomic-transcriptomic approaches, and have determined the full sequences of these Cdpi-Ps. Additionally, we assessed the activities Cdpi-P1 and injected venom using mass spectrometry-based methods and conventional PLA 2 assays.

EXPERIMENTAL PROCEDURES
Extraction of Injected Venom-Specimens of C. purpurascens (20 -50 mm) were collected from intertidal areas at the shores of the Republic of Ecuador and kept alive in aquaria for venom extraction. Extraction of injected venom of C. purpurascens was carried out according to the procedure of Hopkins et al. (37) with modifications (38).
Experimental Design and Statistical Rationale-The purpose of this study was to characterize PLA 2 s (conodipines) in the venom proteome of C. purpurascens. Therefore, we employed methods to maximize the number of conodipines identified, including strategies to identify parts of their sequences (Edman degradation, transcriptomic data mining; MALDI-MS of HPLC fractions, LC-MS/MS of HPLC fractions). The design and statistical rationale for each of the experiments conducted have been described in each subsection. The injected predatory venom was collected and screen from 27 individuals. The specimens with most conodipine content in their venom were selected for further analysis. Conodipine extraction was made based on produced biological quintuplicates. The eventual dataset includes 5 technical replicates of selected HPLC fractions, two technical replicates of reduced and alkylated but undigested venom, and two replicates of reduced, alkylated, and digested venom. The remaining samples are neither biological nor technical replicates but are subsets of venom components fractionated to maximize the potential of LC-MS/MS for conodipine identification.
Fractionation and Identification of Cdpi-Ps-Five to ten injected venom samples (ϳ80 l) were pooled and dissolved in 0.1% (v/v) TFA and separated by reversed phase HPLC (RP-HPLC) using a 4.6 mm ϫ 50 mm, 2.6-m-particle diameter, 100-Å-pore size Kinetex C8 column (Phenomenex, Torrance, CA) with a flow of 1 ml/min. The mobile phases were 0.1% (v/v) TFA (Solution A) and 0.1% (v/v) TFA in 60% (v/v) acetonitrile (Solution B). The venom was separated with an incremental linear gradient of 1% (v/v) B/min. All fractions were manually collected, lyophilized, and kept at Ϫ80°C before further use. The fractions corresponding to the conodipines were resuspended in 50 l of buffer; the conodipine concentrations were determined using an Epoch TM 2 microplate spectrophotometer (BioTeck ® , Winooski, VT) operating at ϭ 280 nm using the calculated extinction coefficient from ProtParam (39).
MALDI-TOF Mass Spectrometry of the Cdpi-Ps Fractions-Positive ion MALDI-TOF mass spectrometry was carried out on an Applied Biosystems Voyager-DE Pro spectrometer. Samples were dissolved in 0.1% TFA (v/v)/60% (v/v) acetonitrile, and applied on 4-hydroxy-3,5-dimethoxycinnamic acid (sinapinic acid) matrix. Mass spectra were obtained in the linear and reflector mode (M/⌬M resolution ϳ10000) using Calmix 1 and Calmix 2 (Applied Biosystems, Foster City, CA) as external calibration standards. An aliquot (1 l) of each RP-HPLC fraction was subjected for mass spectrometry.
Cdpi-P1 Partial Sequencing by Edman Degradation-An aliquot corresponding to fraction II was dried, redissolved in 0.1 M Tris-HCl (pH 6.2), 5 mM EDTA, 0.1% (v/v) sodium azide, and reduced with 10 mM DTT. Following incubation at 60°C for 30 min, the sample was alkylated at a final volume of 15 l with 20 mM iodoacetamide and 2 l of NH 4 OH (pH 10.5) at room temperature for 1 h in the dark. The reduced/alkylated subunits were separated by RP-HPLC and were dried for furthered N terminus sequence by Edman degradation on an Applied Biosystems Procise model 491A sequencer.
RNA Extraction, RNA-Sequencing, Transcriptome Assembly and Conodipines-P1-6 Sequencing-The sequences of the precursors of Cdpi-P1-6 were determined from the transcriptome of the dissected venom of C. purpurascens as described by Mö ller et al., 2017 (40). Briefly, a single adult specimen of C. purpurascens was dissected, the venom duct removed and placed in TRIZOL (Invitrogen, Carlsbad, CA). The RNA was extracted from venom duct and prepared utilizing Illumina Poly(A)-Truseq preparation protocol. Samples were sequenced on Illumina Next-Seq 500 sequencing platform to produce 21 million forward and reverse reads. The paired-end data was then assembled the de novo assembler Trinity (41) to produce a final FASTA dataset containing 96,000 independent assembled contig sequences. The N terminus sequence of the Cdpi-P1 obtained by Edman degradation was used to identify the protein from the BLAST (Basic Local Alignment Search Tool) database generated from the transcriptome (42). The signal region and predictions of the cleavage sites were determined using SignalP 4.0 (43), NetChop 3.1 (44) and ProP 1.0 server (45). Sequence alignments were performed by Clustal Omega (46).
Cdpi-P1-3 Proteomic Analysis-Top-down and bottom up proteomics strategies were used to determine the sequences and modifications of Cdpi-P1-3. For the bottom up analysis, the RP-HPLC fractions were reduced and alkylated with 10 mM DTT and 20 mM iodoacetamide. Centrifugal filters were used to eliminate excess alkylating agent. Proteins were resuspended in 25 mM ammonium bicarbonate (pH 7.8) containing 1 mM of CaCl 2 . Then, proteins were digested with a LysC/trypsin mix (Promega, Madison, WI) at an enzyme/protein ratio of 1:50 (w/w) at 37°C for 18 h. After the digestion, each sample was resuspended in 0.1% (v/v) FA, and the protein digests were analyzed by an Orbitrap Fusion™ Lumos™ Tribrid™ Mass Spectrometer (Thermo Fisher, San Jose, CA) using a 75 ϫ 25 mm, 2 m, C18 column. Additionally, top-down proteomics was performed on the intact protein fractions, and the reduced/alkylated subunits using a Thermo Scientific 75 ϫ 15 mm, 3 m, C8 column. The analyses were performed using electron transfer dissociation (ETD). The parameters used for the MS acquisitions were: survey scan ϭ 1 (scan), mass range ϭ 750 -4,000 (m/z), maximum injection time ϭ 50 ms, resolution ϭ 240,000. The MS/MS acquisition parameters were: scans ϭ 1 scans, cycle time ϭ 3 s, maximum injection time ϭ 35 ms, isolation with ϭ 0.7 Da, ETD activation time ϭ 50 ms and dynamic exclusion was set to exclude selected precursor ions for 60 s after 10 fragmentation selections in 30 s. Spectra were processed and analyzed with Proteome Discoverer™ Software (Version 2.0) using the SEQUEST algorithm (47,48) using the Conus (taxon ID 6490) database found in UniprotKB. This database was created using the 2018 releases of the SwissProt, and TrEMBL databases from UniProt and adding the sequences of the precursors of Cdpi-P1-3 from the transcriptome resulting in 9677 sequences. For Proteome Discoverer, the search parameter settings were 10 ppm precursor mass tolerance, 0.2 Da product mass tolerance, unspecific enzyme with 2 missed cleavages and 0.1 FDR. Cysteine carbamidomethylation was selected as a fixed modification, and Pro hydroxylation, Glu carboxylation, N terminus cyclization (for Glu and Gln), and deamidation amidation were selected as variable modifications. The search parameters used for the peptide mapping analysis using the software Thermo BioPharma™ Finder™ were: 10 -20 signal/noise threshold and 4 ppm mass tolerance for peak detection, 5 ppm with a minimum confidence of 0.80 and high protease specificity for trypsin for peptide identification and mass accuracy. sPLA 2 Assay for the Activity of Cdpi-P1-3-PLA 2 activity was determined using sPLA 2 assay kit (Cayman Chemical Company, Ann Arbor, MI) following the manufacturer's protocol. To achieve reproducible results, the amount of sPLA 2 s added to the well should cause an absorbance increase between 0.01 and 0.1/min. The samples were diluted to reach the enzymatic activity at this level. The assay buffer was 25 mM Tris-HCl, pH 7.5, containing 10 mM CaCl 2 , 100 mM KCl, and 0.3 mM Triton X-100. Assay buffer was used as a blank, and Apis mellifera venom PLA 2 (1 g/ml) was used as a positive control. We used 0.25 l of injected venom from C. purpurascens diluted to a final volume of 100 l and ϳ1 g/ml of Cdpi-P1. The optical density was measured every minute at ϭ 405 nm using Thermomax microplate reader (Molecular devices, San Jose, CA). All tests were carried out by triplicate or more and mean values were calculated. Enzymatic activity was expressed as the increase in absorbance per minute. Specific activity was expressed as micromoles of fatty acid released per minute per milligram of protein.
Agar Plate Activity Assay-PLA 2 activity was evaluated by indirect radial digestion in an agar plate according to the methods described by Gutierrez et al. (49). A solution of 25 ml of 1% (w/v) agarose dissolved in phosphate-buffered saline (PBS) pH 7.2 was melted in a boiling water bath and cooled to 50°C. One ml of human whole blood (Bioreclamation IVT, Westbury, NY) was washed three times with (PBS) pH 7.2 using centrifugations at 1900 ϫ g. Three hundred microliters of washed erythrocytes were combined with 300 l of 1:3 egg yolk solution in PBS and 250 l of 0.01 M CaCl 2 . The solution was added to the agar solution, stirred, and immediately poured into a 9 ϫ 9-cm plastic Petri dish. Cylindrical holes of ϳ2 mm in diameter were punched into the gel and emptied by suction. Samples were prepared in duplicates as follows: positive control containing 20 g/ml of A. mellifera venom PLA 2 , 0.5 l of C. purpurascens injected venom, and the Cdpi-P1 fraction (20 g/ml) in a final volume of 10 l in 0.2 mM sodium acetate pH 6.0, 0.15 mM NaCl. The holes were filled with a volume of 8 l from each sample. Empty holes (negative controls) were filled with 8 l of buffer. The plate was covered and incubated for 20 h at 37°C.
Hemolytic Activity-The hemolytic effects of C. purpurascens injected venom on red blood cells (RBC) were evaluated using human blood (50). One ml of whole human blood (Bioreclamation IVT, Westbury, NY) was diluted in 10 ml PBS, pH 7.4, and then centrifuged at 1900 ϫ g for 15 min at 4°C. The RBCs were washed three consecutive times. The pellet rich in erythrocytes was resuspended in PBS to a final volume of 10 ml. A hundred microliters of the purified RBCs were combined with 2, 5, 10, 15, and 20 l of injected venom. Final volume was adjusted to 300 l with PBS. Positive control (maximum hemolysis) consisted of 200 l of PBS in 1% (v/v) Triton X-100 and 100 l RBC, whereas the negative control contained 200 l of PBS and 100 l RBC. The samples were incubated for 3 h at 37°C in an incubator and then centrifuged at 1900 ϫ g for 15 min at 4°C. The absorbance of the supernatant was determined at ϭ 540 nm. All samples were tested by triplicate. Percentage of hemolysis was calculated as follows: Direct Assessment of the Activity by Mass Spectrometry-PLA 2 activity was determined by mass spectrometry. For that purpose, A. mellifera venom PLA 2 was used as a positive control in a final concentration of 10 g/ml. Venom activity was determined using a volume of 3 l of injected venom from C. purpurascens and the Cdpi-P1 fraction in a final concentration of ϳ10 g/ml. A stock solution of 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) (Avanti Polar Lipids, Inc, Alabaster, AL) was prepared at 1 mg/ml diluted with methanol. Reaction was started by adding DOPC to a final concentration of 100 g/ml. The final assay volume was 30 l adjusted with the 50 mM sodium acetate (pH 7.4), 10 mM CaCl 2 , 100 mM KCl.
Calcium-dependent activity was determined using the conditions described above, using 25 mM ammonium acetate (pH 7.4) as buffer, 100 mM KCl, 2 mM EDTA. Control assays were A. mellifera venom PLA 2 , injected venom and the RP-HPLC fraction II without DOPC.
Samples were incubated for 3 h while gently agitated at 37°C. Direct flow injection analysis (FIA) was performed on an Agilent 6530 Accurate Mass Q-TOF, calibrated with the mass range set to 3200 m/z and an extended dynamic range of 2 GHz. Analysis was performed in positive mode with the following settings: capillary voltage, 4000 V; nozzle voltage, 0 V; fragmentor voltage, 170 V; nebulizer pressure, 20 psi; drying gas, 11 L/min at 350°C, and sheath gas, 12 L/min at 400°C. The mass range was set from 100 to 1500 m/z, with a 1.0 spectra/s acquisition speed for the MS and MS/MS modes (51). The mobile phase used was 50% (v/v) HPLC H 2 O/50% (v/v) acetonitrile/0.1% (v/v) FA. The flow was 0.5 ml/min. The injection volume of sample was 10 l.
Additional MS of venom sPLA 2 s activities were carried out using a Thermo Scientific Velos Pro TM with a HESI probe. The analysis was performed by direct injection of 2 l of sample and ran isocratically over 10 min in 90% (v/v) methanol/10% (v/v) acetonitrile in positive mode. The autosampler was set at 37°C.

Fractionation of Cdpi-P1-3 and Characterization-
The RP-HPLC profile of the injected venom from C. purpurascens is shown in Fig. 1. The peaks eluted between 39 and 42 min (peaks I to VII) were collected and saved for further analysis.
Preliminary MALDI-TOF mass spectrometry of the purified RP-HPLC fractions peaks designed as II and VI yielded an average molecular mass of 13,546 Da and 13,371 Da, respectively (data not shown). Additionally, the intact protein masses of the chromatographic fractions were determined by direct injection in the Orbitrap Fusion™ Lumos™ Tribrid™ Mass Spectrometer. The spectra deconvolution showed isotopically unresolved masses of 13,535.80 Da for the fraction peak II ( Fig. 2A) and 13,370.46 Da for the fraction peak VI (Fig. 2B), which is within the range of the reported conodipines (31). The other chromatographic peaks (III to VI) contained different hydroxylation states of p21a, a conotoxin with a 9.3 kDa molecular mass (38).
The two fractions containing the putative conodipines were reduced and alkylated yielding two peaks with molecular masses of 8595.11 Da and 5634.82 Da for fraction II ( Fig. 2A) and 8591.83 Da and 5490.67 Da for fraction VI (Fig. 2B), which confirmed that these are heterodimeric proteins, which is characteristic of the previously reported conodipines. Notably, fraction II was more concentrated and of higher purity than fraction VI. Fraction VI major component was the nonhydroxylated state of p21a. For that reason, the conodipine in fraction II (Cdpi-P1) was selected for testing in the activity assayed.
Cdpi-P1 Partial Sequencing by Edman Degradation and Transcriptome Assembling-The reduced and alkylated subunits corresponding to the fraction peak II were separated by RP-HPLC and subjected to N terminus sequencing by Edman degradation. We were unable to sequence the ␤-subunit, presumably because it had a blocked N terminus (pyroglutamate) (Fig. 3); however, we obtained the first 34 residues N terminus of the ␣-subunit, yielding a KESCTKHSNGCSTPLR-LPCQEYFROACDIHDNCY sequence (O ϭ hydroxyproline). A preliminary BLAST search with the obtained N terminus sequence yield identities from 67% to 79% with the N terminus sequence of the conodipines reported from C. geographus (36), and an identity of 53% with the alpha chain of Cdpi-M from C. magus, the only conodipine isolated so far from Conus dissected venom (31). Sequence homology along with the molecular mass and the heterodimeric composition were strong indications of conodipines present in the injected venom of C. purpurascens. The partial N terminus sequence of the conodipine-P1 obtained by Edman degradation was used to search in the transcriptome of the venom duct of C. purpurascens. We found a set of six types of conodipines coded by the transcript: Cdpi-P1 to -P6 (Fig. 3) (GenBank accession MK493027-MK493032). The six conodipine precursor sequences have 178 residues, and the sequence found by Edman degradation of the first 34 amino acids was the same for all Cdpi-Ps. The predicted protein precursor sequence starts with a 24-residue signal peptide followed by the sequence that includes the ␣ and ␤-subunits. Five different alphas (␣ 1 -␣ 5 ) and four different betas (␤ 1 -␤ 4 ) were found in the transcriptome. The combina-  tion of subunits ␣ 1 ␤ 1 was assigned as Cdpi-P1, whereas the combination of ␣ 1 ␤ 2 , ␣ 2 ␤ 2 , ␣ 3 ␤ 3 , ␣ 4 ␤ 4, and ␣ 5 ␤ 2 were Cdpi-P2, -P3, P4, P5, and -P6 respectively (Fig. 3). The signal region and the cleavage sites for all six transcriptome-derived sequences were determined and their calculated molecular masses are shown in Table I. A BLAST search with the whole Cdpi-P1, -P2 and -P3 transcriptome sequences shows similarities with other conodipines and sPLA 2 s. The higher sequence identities ranged between 58 and 61% with the transcriptome-based conodipines from C. lenavati (Cdpi-L) (UniProtKB A0A0K8TUY3_ 9CAEN), C. tribblei (Cdpi-T) (UniProtKB A0A0C9RYL5_ 9CAEN), and C. geographus (Cdpi-G5, -G4 and -G3) (UniProtKB W4VS60_CONGE, W4VSC8_CONGE, and W4VSJ3_CONGE) (Fig. 4).
Cdpi-Ps Proteomics Sequence Analysis-RP-HPLC fractions I to VII were subjected to LysC/trypsin digestion and then analyzed by LC-MS/MS. The digested peptides sequences were obtained by peptide mapping analysis using the modified Conus database information. Only sequences with a confidence score of 80% or more were used. The MS/MS data of the peak II fraction provided 81% coverage for ␣ 1 -subunit and 100% coverage for ␤ 1 -subunit. This combination corresponded to Cdpi-P1 (␣ 1 ␤ 1 ) (Fig. 5A). It should be noted that there are distinctive residues that differentiate ␣ 1 from ␣ 2 and ␤ 1 from ␤ 2 . Raw MS/MS spectra that distinguished Cdpi-P1 are shown on supplemental Fig. S1 and S2. These results agreed with the deconvoluted masses obtained for the native and reduced alkylated form of fraction II.
We found isoforms of the conodipines formed by Pro hydroxylation (Hyp) on Pro14, Pro18, and Pro25 of the ␣ 1subunit (Fig. 5A). Hyp was also found on Pro25 with the Edman degradation analysis and in the native and reduced alkylated deconvoluted masses were a ϩ16 Da was observed ( Fig. 2A). In the case of the Cdpi-P1 ␤1-subunit, we found evidence the presence of a pyroglutamate on the N terminus along with hydroxylation at Pro7 (Fig. 5A and supplemental Fig. S2).
Reduced and alkylated samples of fraction peak II and VI were subjected to top-down proteomic analysis via ETD fragmentation. Fig. 6 shows the fragmentation maps of the two subunits ␣ 1 and ␤ 1 for fraction II and ␤ 2 for fraction VI. Each subunit showed a characteristic charge distribution, which was a key for subunit identification. Data on the matched fragments are included in the supplementary information (supplemental Fig. S5).
Altogether, intact protein deconvolution, and top-down and bottom-up approaches showed that fraction II contained Cdpi-P1 and fraction VI contained Cdpi-P2 and Cdpi-P3 with the latter at a lower concentration. Given the purity and abundance of fraction II, Cdpi-P1 was selected for use in subsequent activity assays.
Agar Egg Yolk Plate Assay and Hemolytic Activity of Cdpi-P1 and Injected Venom-The activity of Cdpi-P1 and injected venom were determined by the formation of halos in agarose-erythrocyte egg yolk gels as a consequence of the enzymatic hydrolysis of lecithin to lysolecithin, and latter lysis of red blood cell membranes (Fig. 7A). All three, the positive control, A. mellifera PLA 2 , C. purpurascens-injected venom and Cpdi-P1 formed halos indicating the presence of PLA 2 activity.
A dose-response curve showed the direct hemolytic activity of C. purpurascens venom is shown on Fig. 7B. These results confirmed that injected venom from C. purpurascens can induce hemolysis in human erythrocytes. Two microliters of injected venom can induce 60% of hemolysis while a plateau between 85 and 95% is reached above 10 l of venom. sPLA 2 Assay for the Activity of Cdpi-P1 and Injected Venom-We determined PLA 2 activities using a colorimetric assay based on the hydrolysis of a 1,2-dithio analog of diheptanoyl phosphatidylcholine that serves as a substrate for most sPLA 2 s. C. purpurascens injected venom and Cdpi-P1 were tested using the A. mellifera PLA 2 as the positive control (Fig.   7C). We determined that C. purpurascens injected venom have an active PLA 2, as the specific activity for Cdpi-P1 was 449.7 Ϯ 7.9 mol/min/mg (n ϭ 5), compared with A. mellifera PLA 2 , which showed an 87.9 Ϯ 2.1 /min/mg (n ϭ 5) specific activity.
Direct Assessment of the Activity of Cdpi-Ps by Mass Spectrometry-The direct measurement of venom PLA 2 activity by LC-ESI-MS was carried out by monitoring by the formation of the lysophospholipid and fatty acid (products of the hydrolysis). We followed the formation of the hydrolysis products of DOPC after treatment with Cdpi-P1 (Fig. 8). The peak corresponding to DOPC showed a molecular mass of 786.59 Da. After Cdpi-P1 activity, a lysophospholipid with a mass of 522.35 Da was formed along with oleic acid with a mass of 282.28 Da. The same results were observed when A. mellifera PLA 2 was used as the positive control and with the C. purpurascens-injected venom, where the DOPC peak disappeared after the treatment and corresponding lisophospholipid was formed (Fig. 9). Other control experiments were performed using the venom from the snakes, Agkistrodon piscivorous piscivourous (APP) and Agkistrodon piscivorous locustoma (APL). Both snake venoms have been reported to have sPLA 2 s (52). Additionally, we tested the activity of the injected venoms from the fishhunting cone snails, C. ermineus and C. striatus. All the above, except the C. striatus venom, formed the lysophospholipid (522 Da), indicating PLA 2 activity (supplemental Fig. S6). MS analysis indicated the presence of molecular masses in the range of sPLA 2 in all venoms except in the venom of C. striatus (data not shown).
Because calcium plays an important role in the lipolytic activity of different venom sPLA 2 s (1, 31), we evaluated the activity of different venoms in absence of Ca 2ϩ using a buffer with no calcium that contained EDTA (36). Fig. 10 shows that no lysophospholipid was formed in the absence of calcium for the A. mellifera PLA 2 and the C. purpurascens-injected venom, indicating that in absence of calcium the PLA 2 activity was abolished. The same results were obtained with C. ermineus, APP, and APL venom (data not shown). DISCUSION Here we describe the isolation and characterization of three conodipines (Cdpi-P1, -P2, and -P3) found in the injected venom of C. purpurascens. These are the first conodipines described in injected venom, the actual concoction deliver to FIG. 5. Determination of Cdpi-P1, -P2 and -P3 sequence by bottom up proteomics. Sequence coverage map was achieved after the trypsin/LysC digestion of the fraction II (A) and fraction VI (B). Peptide mapping analysis was performed by mass spectrometry. Gray bars indicate the identified digested peptides. Fragments sequences matches were determined using the conodipines sequences from the transcriptome. A, Fraction II sequence coverage corresponded to subunits ␣ 1 and ␤ 1 , which match Cdpi-1. B, Fraction VI coverage corresponded to ␣ 1 , ␣ 2 , and ␤ 2 -subunits, which subunits combination are a 1 ␤ 2 and ␣ 2 ␤ 2 and match to Cdpi-P2 and Cdpi-P3 respectively. Covered fragments are in gray, Hyp in blue, carbamidomethyl cysteine in green and pyroglutamate in yellow. Cleavage residues for Trypsin/LysC are in magenta. Bold residues correspond to then Edman degradation obtained sequence. Some of the key MS/MS spectra are showing on the supplemental Figs. S1 to S4. For more details about the spectra and masses of the corresponding fragments, refer to the supplementary material. the prey. Other Cdpis have been found in the dissected venom of C. magus and of C. geographus (31,36); however, because no conodipines had been found in the injected venom, it was suggested that these enzymes were not involved in the envenomation process (32). The isolation of Cdpi-Ps clearly show that these enzymes are part of the envenomation strategy of certain cone snail species. Conodipine sequences have been obtained from the transcriptome of several cone snail species, which might be an indication of their prevalence in cone snail venom and even a key component for envenomation (34,(53)(54)(55).
Conodipines are heterodimers, however, their alpha and beta subunits are encoded by a single precursor with a signal peptide sequence and a peptide linker between the two subunits. Like other venom PLA 2 s, conodipines have high intra and interspecies sequence variability as shown in the phylogenetic tree of known conodipines (supplemental Fig. S7). Different isoforms were reported within the same species (54). In our case, six conodipine isoforms were found in the transcriptome, with higher sequence variability on the beta subunit. All Cdpi-Ps precursors have the same signal and remarkable similarity in their subunits linker. The catalytic His-Asp dyad is conserved as in the other conodipines precursors reported.
The complete sequences of Cdpi-P1, -P2, and -P3 from C. purpurascens-injected venom were obtained by a combination of proteomics (MS/MS) and transcriptomic (RNAseq) data analysis. The mature Cdpi-Ps contains nine cysteines on the ␣-subunit and three on the ␤-subunit. Using homology modeling (supplemental Fig. S8), we surmised that the third cys- teine (Cys19) in the alpha subunit is the one involved in intermolecular disulfide bonding. Our proteomic analysis shows that these conodipines are hydroxylated at Pro in both the alpha and beta subunits and several hydroxylationrelated isoforms were observed. This differential and selective hydroxylation has been reported in conotoxins and it is found in the venom of C. purpurascens (37,38,56,57). The importance of Hyp modification in conopeptides is still unclear, but it is possible they could contribute to the formation of hydrogen bonds and facilitate the binding strength and selectivity toward a target (58). For the Cdpi-Ps, there is the possibility that the Hyp might be involved in enzyme dimerization and stabilization. In addition to Hyp modifications, we found that all Cdpi-Ps contain N-terminal pyro-Glu/Gln residues. This moiety is a major modification in secretory proteins, and it is thought to assist in receptor-ligand interactions and in the stabilization of the N terminus against degradation (59).
Venom sPLA 2 s have been reported as either as monomeric or multimeric complexes with at least one subunit catalytically active. The mechanism of action of multimeric PLA 2 s is more complex when compared with their monomeric counterparts. One of the first examples of a heterodimeric PLA 2 is the snake ␤-bungarotoxin (60); heterodimeric sPLA 2 s have been isolated from scorpion venom such as Imperatoxin (61). Other examples include Heteromtoxin (62) and Hemilipin (63), which have a large PLA 2 active subunit, and a small subunit linked by a disulfide bond. The function of the small subunit is unknown, although they have been categorized as members of the Kunitz protease inhibitor superfamily (64). Likewise, the catalytic residues of the conodipines are conserved and located on the alpha subunit, which is larger than the beta. However, there is a high sequence variability on the beta subunit, and they have no sequence relationship with the Kunitz protease inhibitor superfamily.
Among all sPLA 2 isoforms, the amino acid residues involved in the Ca 2ϩ binding and catalysis are highly conserved. The His/Asp dyad and a Tyr are also highly conserved in the active site of these enzymes (65,66). Conodipines have the conserved His/Asp dyad that corresponds to the residues His-30 and Asp-31. Additionally, Asp-28 has been suggested as an active residue because it may possibly form hydrogen bonds with His-30 (9,31). A study with species of the phylum Cnidarian, Mollusca, and Turnicata found a conserved catalytic site with the sequence X-Cys-Asp-X-His-Asp-X-Cys-Tyr-X-Cys (67), which matches with the domain observed in all Cdpi-Ps and other Cdpis. The precise role of these residues and the other conserved amino acids has not been determined. Although the His/Asp dyad has been a hallmark of these enzymes, other group of PLA 2 enzymes with a His/Lys dyad has been described in snake venom (68 -74). No His/Lys dyad conodipines have been reported so far. In addition to the calcium binding dyad, other venom PLA 2 s have an extra calcium binding motif of Trp/Tyr-Cys-Gly-X-Gly. It is notable that this motif is not present in any conodipines, which also underwrites conodipines as unique group of sPLA 2 .
We assessed the activity of Cdpi-P1-3 by colorimetry and an agarose-erythrocyte egg yolk gel assay. We found enzymatic activity for the venom fraction containing Cdpi-P1 and for the C. purpurascens injected venom. A direct assessment of Cdpi-P1 activity was performed by MS indicating that is a fast and reliable method for PLA 2 s activity studies. We evaluated the DOPC lipolysis in presence of Cdpi-P1 and the injected venom of C. purpurascens, C. ermineus and Agkistrodon piscivorous. We found that the activity of Cdpi-Ps is Ca 2ϩ -dependent. The use of MS to follow the enzymatic activity of PLA 2 s has been reported, as venom toxins have been tested in cultured neurons and in cardiolipin (75,76). It has been shown that MS can be used as a screening method with high sensitivity and specificity. These MS-based techniques not only allow the characterization of enzymes, but also enable studies on the lipid composition of membranes. The use of MS to determine the activity of venom enzymes by MS is promising for a fast and reliable method for enzyme kinetics. We recently reported that other venom enzyme's activities, such as hyaluronidases, could be follow by MS (40).
Venom sPLA 2 s could be potent hemolytic agents (77-81); i.e. C. loroisii venom produced significant alterations in the hematological parameters, including the reduction of the red blood cells and induced edema (82). Our results show that the injected venom of C. purpurascens exerts a notable hemolytic effect. This may be because of Cdpi-P isoforms and/or also because of other active enzymes as proteases present in the venom (83,84). The damage of the red cell membrane by the action of venom molecular components could be a direct or a multistep process, where not only does the sPLA 2 s participate in the hemolysis process, but also other enzymes.
Conodipines, similar to other peptides and proteins present in venom, may be naturally tailored by cone snails to efficiently adapt to a variety of different preys (85). We not only found different conodipines in C. purpurascens, but we also found that some individuals of this species kept in our lab did not express measurable quantities of conodipines and no PLA 2 activity was detected in their injected venom. It would be important to determine the factors that conditioned the selective expression of these venom enzymes by some specimens. Furthermore, that we found PLA 2 activity in the venom of C. ermineus, but not in C. striatus, which demonstrate that not all fish-hunting cone snails express these enzymes. Additionally, it should be noted that conodipines are not only restricted to fish-hunting species, as conodipines have been found on the closely related worm-hunting species, C. tribblei and C. lenavati. This demonstrates that conodipines are found in cone snails of their prey preference.
It is unclear how conodipines have evolved and adapted within this genus. The intraspecific and interspecies variability and adaptations shown so far indicates a biochemical conundrum that requires further study. However, it has been recently proposed that group IX PLA 2 s, belonging to the phylum Mollusca, Cnidarian and Turnicata, share a common evolutionary origin with group XIV PLA 2 s of bacteria and fungi. Furthermore, there are indications that group IX PLA 2 s go through species-specific duplication and that evolutionary pressures in different lineages promoted ancestral molecule modifications (67).
Venom PLA 2 s may have medical and therapeutic applications, such as prevention of prion peptide-induced cell death (84), viral inhibition (85,86), hepatoxicity protection (86), inhibition of cancer cell proliferation (88), neuroinflammatory modulation (87), and mediation of antigen delivery (88). Given the large intraspecific and interspecies structural variability shown by venom PLA 2 s along with their associated activity, it seems that their use can be optimized for specific applications and the discovery and testing new PLA 2 s will be important for a variety of uses in research, medicine and industry.
Disclaimer-Certain commercial equipment, instruments, or materials are identified in this paper in order to specify the experimental procedure adequately. Such identification is not intended to imply recommendation or endorsement by the National Institute of Standards and Technology, nor is it intended to imply that the materials or equipment identified are necessarily the best available for the purpose.

DATA AVAILABILITY
This article contains supplemental Fig. S1-S8, supplemental Table S1 and other data files. Raw and processed bottom up proteomics data has been deposited to MassIVE (https:// massive.ucsd.edu/ProteoSAFe/static/massive.jsp) project ID MSV000083408. Exported results can also be found in supplemental excel files Pur_4L_II_TD_Results and Pur_4L_VI_ TD_Results.