Jellyfish sting is an important health problem for persons engaged in marine activities, and the symptoms of victims range from itch, pain, inflammation, and edema to cutaneous necrosis and even death (Cegolon et al., 2013). Venom from jellyfish Nemopilema nomurai is responsible for the symptoms of jellyfish envenomation. N. nomurai is widely distributed in the Yellow Sea and East China Sea every summer season (Sun et al., 2015a; Sun et al., 2015b). It has been regarded as a synonymy of Stomolophus meleagris in some research literatures (Liumin et al., 2011). In recent decades, blooms of N. nomurai have been of frequent occurrence and became more of a threat to humans in Chinese, Japanese, and Korean waters (Kawahara et al., 2006; Liumin et al., 2011; Xu et al., 2013; Yoon et al., 2014; Sun et al., 2015a; Sun et al., 2015b). This makes N. nomurai venom research more necessary and important.

Jellyfish N. nomurai venom is a complex mixture of peptides and proteins (Li et al., 2014) and has a variety of bioactivities such as hemolytic and antioxidant activities (Yu et al., 2005; Li et al., 2012; Li et al., 2013; Yue et al., 2021). According to the proteome and transcriptome analyses, metalloproteinases are main components in N. nomurai venom (Li et al., 2014). The crude venom extracted from nematocysts was identified to have significant metalloproteinase activity (Yue et al., 2017b) and played an important role in pro-inflammatory activity, edematogenic effects, lethality, and cytotoxicity (Lee et al., 2011; Yue et al., 2017b; Li et al., 2018; Yue et al., 2021). Through chromatography, some N. nomurai toxins were partially purified, such as hemolytic toxin SmTX (Li et al., 2013) and lethal toxin NnLF (Li et al., 2020). Protease in N. nomurai venom, isolated and identified by zymogram, was identified to contain many other protein components (Yue et al., 2017a). A purified protease with only one SDS-PAGE band was isolated from N. nomurai venom by HiPrep 26/60 Sephacryl S-200 column, which did not match any high similarity protease in LC-MS/MS analysis (Yue et al., 2021).

Several protein toxins were successfully purified from jellyfish venom by chromatography, such as cytotoxin CcTX-1 and antioxidant protein SmP90 (Rottini et al., 1995; Yang et al., 2003; Helmholz et al., 2008; Lassen et al., 2011; Lassen et al., 2012; Li et al., 2012; Horiike et al., 2015). But more toxins from jellyfish were partially purified (Bloom et al., 1998; Chung et al., 2001; Sanchez-Rodriguez et al., 2006; Li et al., 2011; Li et al., 2013; Li et al., 2018; Li et al., 2020). For example, Rastogi et al. (2017) identified a 95 kDa metalloproteinase in a partially purified product of Rhizostoma pulmo (barrel jellyfish) venom. But there are still no single metalloproteinase purified from the jellyfish venom, which limits the analysis of its molecular structure and action mechanism. In addition to the small amount of toxin and instability, synergistic effects of toxins may increase the difficulty of single metalloproteinase purification.

Synergistic effects in snake venom have been demonstrated to occur between different toxins with different patterns owing to the amounts of toxin purification, characterization, and pathological researches (Doley and Kini, 2009; Xiong and Huang, 2018). The synergetic pain activation mechanism of scorpion toxin BmP01 was deeply revealed based on its molecular structure researches (Yang et al., 2017). Metalloproteinases in snake venom showed different interactions to other subunits and components (Doley and Kini, 2009; Xiong and Huang, 2018). However, there is no report on the synergistic effects of jellyfish toxins, and the regulation mechanism of jellyfish venom metalloproteinase has not been revealed.

In the present study, the toxic protease components in N. nomurai tentacle tissue autolysis were purified by activity-guided chromatography. Protein components of the final purified products, which showed a synergistic effect of proteinase activity, were identified by LC-MS/MS analysis to explore the possible active components and regulation mechanism. This study may provide references for further research on the NnVMP and therapy of jellyfish stings.

As jellyfish stings pose a threat to humans in many marine activities in summer, especially swimming, the study of jellyfish sting mechanism is necessary for therapy researches. N. nomurai is a giant jellyfish widely distributed in the Yellow Sea and East China Sea, which blooms in summer in recent years (Sun et al., 2015a; Sun et al., 2015b). In N. nomurai venom, metalloproteinase is a major component (Li et al., 2014). NnVMPs showed azocasein hydrolysis activity and impacted the expression of many inflammatory factors (Li et al., 2019) and increased vascular permeability by directly degrading basement membrane components (Yue et al., 2021).

Metalloproteinases widely exist in venomous animals, including snakes, scorpions, spiders, jellyfish, etc. and play an important role in digestion and preying. Many different SVMPs are obtained by two or three steps of chromatography, such as CCSV-MPase (Cherifi et al., 2010), Atrase B (Sun and Bao, 2010), BmooMP α -II (de Queiroz et al., 2014), and CcD-II (Ami et al., 2017). Additionally, some metalloproteinases can be obtained by cDNA cloning and expression, such as rFIVa (Xu et al., 2006), Ahpfibrase (Zhang et al., 2010), and Jerdonitin (Zhu et al., 2010). Owing to the successful single protein purification, many SVMPs were studied in-depth in structures, functions, and mechanisms. SVMPs can be classified by their domain architecture into type P-I to P-III (Fox and Serrano, 2008). SVMPs are multifunction proteins, and they showed hemorrhagic, procoagulant, anticoagulant, and antiplatelet effects in envenomation cases (Fox and Serrano, 2005). Researches on the jellyfish venom purification are relatively difficult. Only a few toxins were highly purified from jellyfish venom, such as cytotoxin ClGp1, cytotoxin CcTX-1, neurotoxin CcNT, antioxidant protein SmP90 (Rottini et al., 1995; Yang et al., 2003; Helmholz et al., 2008; Lassen et al., 2011; Lassen et al., 2012; Li et al., 2012; Horiike et al., 2015), etc. Techniques most used in venom protein purification were activity-guided multidimensional chromatography, including size-exclude, ion-exchange, reversed-phase, and affinity chromatography. Lassen purified cytotoxin CcTX-1 and neurotoxin CcNT from Cyanea capillata in this program (Lassen et al., 2011; Lassen et al., 2012). Nevertheless, most jellyfish protein toxins could only be partially purified by multi-chromatography, such as a 95 kDa metalloproteinase from R. pulmo (barrel jellyfish) venom (Rastogi et al., 2017) and other hemolytic and lethal jellyfish toxins, such as SmTX from S. meleagris (Li et al., 2013) and NnLF from N. nomurai (Li et al., 2020). Due to the complex composition, it is not easy to get a single toxin component from jellyfish venom. In jellyfish venom, some different protein components have very similar physical properties such as molecule weight. In addition, toxins in jellyfish venom contained different subfamilies and may interact with other components. It was hard to efficiently isolate a single jellyfish toxin protein within two or three chromatography steps. In addition, the protein content in jellyfish nematocyst venom is too small an amount to do repeated chromatography, which makes it difficult to reveal the biological activity of a single toxin component. In this study, proteases from NnTNV were purified by ammonium sulfate precipitation, DEAE Sepharose Fast Flow, and Superdex 75 column chromatography successively. During the purification process, the proteinase activity of 80% saturation fraction and 0.2 M ion intensity fraction was increased from 314 U/mg to 1,204 U/mg and 1,065 U/mg, respectively. But none of the final fractions in Superdex chromatography showed significant proteinase activity to PBS (Figures 4C,D). However, when these fractions were mixed again by the same protein content, the mixed samples “A + B,” “C + D,” and “A + B + C + D” recovered the activity (Figure 5), which indicated that there was synergistic effect between the final fractions.

Synergistic effect is common in animal venom such as many snake toxins (Doley and Kini, 2009; Xiong and Huang, 2018). Currently, there is no universal method to study the synergistic effects of animal toxins. Laustsen proposed the Toxicity Score method to determine the presence of synergism in venom (Lauridsen et al., 2016). Venom synergistic effect is mainly revealed by the research of venom purification and molecular mechanisms. For example, SVMP may interact with phospholipase A₂ (Xiong et al., 2017). The enzyme activity of P-III metalloproteinase complex was regulated by its subunits (Doley and Kini, 2009). Although there has been no research in jellyfish metalloproteinase synergistic effect, unrevealed interaction of synergistic proteins may impact the bioactivity and purification studies.

As the final fractions were multicomponent, the protein components of fractions C and D were identified by LC-MS/MS to reveal the possible mechanism of proteinase activity regulation in N. nomurai venom. However, few jellyfish toxins are well understood. A large number of protein toxins were matched in snakes, spiders, scorpions, bees, and other organisms (Table 1, Table 2) by searching in Tox-Prot. The unique peptide count of most matched proteins was only one, but some peptides matched many similar proteins in different organisms, such as peptide KENGRK which was matched to eight different zinc metalloproteinases. It made the peptide sequence more probable to represent the matched protein group. This means even the unique peptide count and cover percent were not high, these matched toxins from other organisms also provided very important references. Similar to the chromatography and SDS-PAGE results, toxins matched in fractions C and D showing shared and different components. These different toxins contained in fractions C and D might contain the regulator of proteinase activity, such as phospholipase A1, P-III metalloproteinase, P-I metalloproteinase, and metalloproteinase inhibitor.

In fraction C, the matched toxin proteins which related to proteinase activity were zinc metalloproteinase and serine protease. The peptide sequence matched to zinc metalloproteinase was KENGRK. It matched eight similar P-III metalloproteinases from different snakes. They can produce proteolytic activity and hemorrhage activity (Howes et al., 2003; Morine et al., 2008). This kind of metalloproteinase has metalloproteinase domain, disintegrin domain, and cysteine-rich domain (Gowda et al., 1994; Siigur et al., 2001; Siigur et al., 2004; Trummal et al., 2005; Takeda et al., 2007; Chen et al., 2008). Sequence KENGRK is located in the cysteine-rich domain. The peptide sequences matched to serine protease were CLGIR, TRTSWDEDIMLIR, and LPNKDR. CLGIR matched venom serine protease Bi-VSP from Bombus ignitus (bumblebee). TRTSWDEDIMLIR matched thrombin-like enzyme ancrod from Calloselasma rhodostoma (Malayan pit viper). LPNKDR matched thrombin-like enzyme cerastocytin from Cerastes cerastes (horned desert viper). These serine proteases have fibrinogenolytic activity (Marrakchi et al., 1995; Marrakchi et al., 1997; Dekhil et al., 2003; Choo et al., 2010; Choo et al., 2012).

In fraction D, the matched toxin proteins which related to proteinase activity were SVMP, SVMP inhibitor, and serine protease. Serine protease was matched by peptide sequence TRTSWDEDIMLIR which was also identified in fraction C. The sequence LENVEKEDGGPK was matched to SVMP from Crotalus molossus molossus (Northern black-tailed rattlesnake). This kind of P-I metalloproteinase impairs hemostasis in the envenomed animal (Gutierrez et al., 1995; Farsky et al., 2000; Patino et al., 2010; Bernardes et al., 2013). The sequence QASQKWGR was matched to SVMP inhibitor 02D01 from Echis ocellatus (ocellated saw-scaled viper). This protein may inhibit metalloproteinase activity in the venom gland through abundant pEKW and poly-His-poly-Gly peptides. The inhibition may be disengaged by dilution or physiochemical change (Wagstaff et al., 2008).

These SVMP-like proteins in fractions C and D might be the important factors in regulating azocasein hydrolysis activity. Zinc metalloproteinases matched in fraction C were P-III metalloproteinase which contains metalloproteinase domain, disintegrin domain, and cysteine-rich domain (Gowda et al., 1994; Siigur et al., 2001; Siigur et al., 2004; Trummal et al., 2005; Takeda et al., 2007; Chen et al., 2008). Some of this kind of P-III metalloproteinase were confirmed that they can constitute complexes with C-type lectin homodimers light chains, such as RVV-X (Gowda et al., 1994; Takeda et al., 2007) and CA-1 (Yamada et al., 1996). RVV-X is a P-III metalloproteinase complex isolated from Daboia siamensis (Eastern Russel’s viper) which can activate coagulation factor X by cleavage of Arg-|-Xaa bonds (Gowda et al., 1994; Takeda et al., 2007). The heavy chain (Protein ID, Q7LZ61) was matched in fraction C. In RVV-X complex, the heavy chain is the catalytic subunit of activating coagulation factor X, and the two light chains are regulatory subunits of binding the Gla domain of factor X (Takeya et al., 1992). Similar to RVV-X, this protein group can selectively hydrolyze factor X to Xa, but some of them were also confirmed having azocasein hydrolysis activity such as the protein zinc metalloproteinase–disintegrin-like bothropasin from Bothrops jararaca (Mandelbaum et al., 1982). However, the zinc metalloproteinase–disintegrin-like atrase B isolated from Naja atra (Chinese cobra) did not show hydrolysis activity to fibrin, azocasein, and BAEE (Cherifi et al., 2010). Therefore, it was speculated that a zinc metalloproteinase which can hydrolyze azocasein was contained in fraction C, and its activity could be activated by binding with regulatory subunits contained in fraction D. Differences from the C-type lectin subunits of RVV-X, the regulatory subunits in fraction D, might affect the active center of the zinc metalloproteinase, similar to allosteric regulation. According to Table 2, fraction D might contain similar proteins to P-I SVMP and SVMP inhibitor. This indicated that there might have been an interaction between them that could inhibit the metalloproteinase activity. The inhibition of SVMP inhibitor may be disengaged by dilution or physiochemical change (Wagstaff et al., 2008).

To sum up the regulation mechanism, inactivated fractions C and D would recover the metalloproteinase activity by binding regulator or disaggregating inhibitor. Fraction C contained a zinc metalloproteinase which could hydrolysis azocasein under the activation of regulatory subunits in fraction D (Figure 6A). In another explanation, the interacted metalloproteinase and inhibitor in fraction D would disengage to activate metalloproteinase activity when fractions C and D were mixed (Figure 6B).

The spectra in N. nomurai genome database were also searched. The whole-genome database was uploaded to NCBI by Hak-Min Kim's team (Kim et al., 2019). LC-MS/MS analysis did not match any toxins in this database. BLAST of SVMP, which was matched in fractions C and D, also had no significant similarity found. Although the nucleotide sequences of N. nomurai toxins were not obtained, some important peptide sequences were transformed into nucleotide sequence. The nucleotide sequence of 21 amino acids shown in Table 4 was obtained from the two N. nomurai mitochondrion proteins listed in Table 3. Through the comparison of the protein sequence and nucleotide sequence of ATP synthase F0 subunit 8 and NADH dehydrogenase subunit 4 L in N. nomurai, two interesting differences were found. In N. nomurai, the isoleucine (Ile, I) codon AUA would translate to methionine (Met, M), and the termination codon UGA would translate to tryptophan (Try, W). This should be paid special attention in further expression works. Based on this nucleotide sequence table, three peptides, which were matched to zinc metalloproteinase, SVMP, and SVMP inhibitor, were transformed into nucleotide sequence as shown in Figure 7. These sequences may provide references for further studies interested in NnVMP.

It has been a problem for a long time that jellyfish venom purification has been limited by the complex protein components. As shown in each purification step, different proteins may have similar physical properties and are hard to be completely isolated by chromatography. Low protein content also limit sample reloading of chromatography. As a new revealed property of N. nomurai toxins, the synergistic effect of proteinase activity may impact further activity-guided chromatography studies. Although jellyfish venom purification is difficult, it might reveal more venom properties in the purification process and provide references for further studies.

In conclusion, the toxic protease components from NnTNV were isolated by ammonium sulfate precipitation, DEAE Sepharose Fast Flow, and Superdex 75 column chromatography successively. After the three purification steps, the main proteinase activity was lost, but it could be recovered by mixing again. This is the first time that the synergistic effect of jellyfish proteinase activity has been revealed. Through LC-MS/MS analysis, it has been shown that the proteinase activity might have been contributed by metalloproteinases and regulated by metalloproteinase subunit or metalloproteinase inhibitor. Three important peptide sequences were transformed into nucleotide sequences to provide more information on N. nomurai metalloproteinases. The results could help further research in jellyfish toxins purification and expression. The synergistic effect might be a new entry point in purifying N. nomurai metalloproteinases and regulators.