Initial autolysis is necessary for full activation of P. excavatus proteases
As we could see, the proteolytic activity towards casein increased and peaked through a 15-day autolysis. The presence of sodium azide could inhibit the bacterial growth. The autolytic process could therefore trigger the release of proteases from the earthworm's tissues and exert a subsequent degradation of keratin and lipids, thus reducing the mixture's viscosity. The increase of proteolytic activity over time suggested that activation of these enzymes had occurred, probably through self-proteolysis that cleaved parts of the zymogens.
Purification protocol for P. excavatus proteases
The acetone precipitation of the P. excavatus lysate was more effective than ammonium sulfate (AS) precipitation because of more impurity removal, higher proteolytic activity recovery (data not shown) and less time consuming since subsequent dialysis is not necessary.
The presence of the 33 kDa protein in faction FIII-3 could support the hypothesis of zymogen degradation mentioned earlier since it could be the zymogen of the 31 kDa peptide, and its presence in the SEC fraction probably resulted from the incomplete initial autolysis. Similar results were reported in the study on L. rubellus earthworm (Cho et al. 2004) in which a 44 residues were cleaved off from a 283-residue-zymogen to release the fully active proteolytic enzyme. This hypothesis, however, requires further validation through sequencing of our FIII-3a and FIII-3b proteins.
Generally, the two-step chromatography of AEX and HIC was sufficient for the purification of FIII-1, FIII-2, and FII, since their SEC and SDS-PAGE profiles represented pure proteins. Fraction FI actually contained two isozymes with close MW and pI, which could not be further separated. For FIII-3 and FIV, SEC was necessary to achieve the highest purity.
P. excavatus proteases possess dual pH optima
The dual pH optima has not been reported for the proteases from L. rubellus and E. fetida. However, this characteristic was found for the intestinal serine protease of red flour beetle (Tribolium castaneum), whose optimal pH was determined to be at 4 and 8.5 (Oppert et al. 2003). (Choi et al. (1989)) studied the proteases from the parasitic protozoa Toxoplasma gondii and discovered that they catalyzed most effectively at pH 6 and 8.5. Likewise, the dual pH optima characteristic has been observed in various hydrolases such as β-glucuronidase from human seminal plasma (Gupta and Singh 1983), Staphylococcus sp. xylanase (Gupta et al. 2000), reptile lysozyme (Thammasirirak et al. 2006) and Rhizopus lipase (Upadhyay et al. 1989).
(Gupta et al. (2000)) hypothesized that the xylanase in their study might contain two distinct active sites that could perform catalysis at two distinct pH levels of 7.5 and 9.2, respectively, although no such enzyme has been reported before. In another study, the aspartate protease Plasmepsin I from Plasmodium falciparum was characterized, revealing the existence of two states of this protease as monomer and aggregated oligomer (Xiao et al. 2007). These two co-existing states resulted in the dual pH optima of the enzyme as determined experimentally. Since all protease fractions from P. excavatus in our study were completely inhibited by PMSF (Table 1), they would not harbor any active sites rather than the typical catalytic triad of serine proteases. Additionally, no aggregation was observed by SEC chromatography performed at pH 8.5. We therefore hypothesize that the P. excavatus proteases existed in both monomeric and aggregated oligomeric form in our assays; and the aggregation might be triggered at strong alkaline pH, for instance pH 11.
The inhibitory effect of PMSF towards all fractions revealed that P. excavatus proteases are serine proteases, since PMSF is a specific irreversible inhibitor for this group of proteases (James 1978). FIII-3 was to some degree inhibited by TPCK, which is specific for chymotrysin-like protease (Table 1). However, it was not able to hydrolyze the chymotrypsin-like specific BTpNA substrate. Therefore, it was not possible to classify this protease. Two fractions FI and FII were also unambiguous since they had no activity towards both BApNA and BTpNA. In contrast, FIV was more likely a trypsin-like protease due to its specific hydrolysis of BApNA and specific inhibition by SBTI. Two fractions FIII-1 and FIII-2 displayed much lower hydrolytic effect on BApNA in comparison to FIV but were not inhibited by SBTI, thus it is still questionable if they were actually trypsin-like proteases.
On the other hand, the sequence alignment study revealed a considerable similarity between FIII-1 and FIII-2 fragments with the trypsin-like lumbrokinase fragments from L. rubellus and E. fetida. However, the sequence homology obtained in our study was expected to be higher because of the close evolutionary relationship between these earthworm species. Cho et al. reported extremely high conservation of the N-terminal 20-22 residues between L. rubellus protease fractions (Cho et al. 2004). The sequence alignment within P. excavatus fractions was not conducted due to insufficient information from the MS/MS data. Therefore mass spectrometric analysis for these proteases should be further elaborated to obtain their full sequences.
The isozymes expressed strong hydrolytic activity towards both fibrinogen and fibrin
(Park et al. (2007)) discovered a protease from Flammulina velutipes that showed both fibrinolytic and fibrinogenolytic activity. This enzyme could perform hydrolyses without the presence of any activators, while human plasminogen is an inactive precursor and strictly requires tPA or urokinase for its conversion into fibrinolytic plasmin. In our experiment, all fractions except for FII displayed remarkable fibrinolysis, which was two to three times stronger than human plasmin (data not shown). They rapidly degraded the fibrinogen monomer as well. Therefore, the P. excavatus proteases would have a different catalytic mechanism towards these two substrates than human plasminogen. Moreover, each fraction experimentally displayed a distinct catalytic rate, thus probably having different kinetic parameters such as K
, and K
Applicability of P. excavatus proteases
Pure proteins are generally less stable in water due to the absence of natural intracellular buffering and ionic conditions. Therefore, good storage conditions are required to maintain their biological activity. Nakajima et al. found that the protease fractions from Japanese L. rubellus could maintain approximately 80% of their activity after five years in 100 mM Tris-HCl buffer at pH 8 (Nakajima et al. 2000). Interestingly, our results revealed that water is more appropriate than phosphate buffer as a storage medium for P. excavatus proteases. In addition, the presence of plasminogen activators was declared to be unnecessary for the enzymes that could be able to hydrolyze both fibrinogen and fibrin (Park et al. 2007). These properties are favourable for convenient and cost-effective formulation of these enzymes. (Cho et al. (2004)) reported that all six protease fractions from L. rubellus had similar caseinolytic activity, only one of which exhibited remarkable fibrinolysis. This fraction was the first earthworm protease to be investigated for thrombosis therapy in Korea. Therefore, the proteases from P. excavatus characterised in the present study seem promising candidates for that purpose in Vietnam. Fraction FIII-2 is the most interesting fraction because of its strong fibrinolysis activity and high stability over long-term storage.