EXTRACTION OF COLLAGENOLYTIC ENZYME FROM FISH VISCERA BY PHASE PARTITIONING (PEG/CITRATE) AND ITS POTENTIAL FOR INDUSTRIAL APPLICATION*

Internal viscera fish are potential sources of protein biomolecules of biopharmaceutical interest. However, this residue is frequently discarded inappropriately. The possibility to obtain byproducts of higher added value is a reality. Inside this view attention must be given to processes for the recovery and extraction of target molecules. However, the high cost of processing these residues is one of the obstacles to their reuse; techniques that facilitate their handling and make the process cheaper are desirable, such as extraction in a two-phase aqueous system. Thus, the aim of this study was to extract collagenolytic enzymes from common snook (Centropomus undecimalis) using a two-phase aqueous system (polyethylene glycol/citrate), according to the 24 factorial design, using as variables: molar mass of PEG (MPEG), PEG concentration (CPEG), citrate concentration (CCIT), pH, still, considering purification factor (FP), partition coefficient (K), and yield (Y). The collagenolytic activity of the crude extract was 102.41 U mg-1, after partitioning, was purified 3.91 times (MPEG: 8000; CPEG: 20.0%; CCIT: 20.0% and pH 6.0). Inhibition (U mg-1) was observed in benzamidine (22.51), TLCK (21.05), TPCK (21.29), and PMSF (23.05), signaling to be a serine-protease. The results showed the advantage of this semipurification technique as concerns to the low cost of extraction and purification, adding value to the fishing source material and allocating the residues from its processing to the industry.


INTRODUCTION
The need for reuse of waste is a practice that has been increasingly encouraged, especially organic solids, such as those from the fishing industry. During the processing of fish there is a lot of organic matter that is produced, among them: head, tail, fins, skin, scales and internal viscera (stomach, intestine, pyloric caecum, mesentery, among others), most of these materials are discarded. Special attention must be given to proteic macromolecules of that have high biotechnological interest. Recently, fishing waste has been the subject of research on their applications in human and animal health (Ideia et al., 2020;Oliveira et al., 2020).
From digestive viscera of aquatic organisms, it is possible to extract various proteases (pepsin, trypsin, chymotrypsin and collagenases and/or enzymes with similar catalytic functions) (Oliveira et al., 2017a;Gurumallesh et al., 2019;Silva et al., 2019). Enzymes with collagenolytic properties of aquatic organisms have already been recovered and partially purified from digestive residues of Neotropical fish species Oliveira et al., 2017b;Liu et al., 2019). The collagenolytic property (collagen cleavage) is quite required in industry, mainly biomedical and pharmaceutical, being applied in surgical procedures (Zhao et al., 2019), intracerebral therapies (Chen-Roetling et al., 2019), wound healing (Abood et al., 2018), extraction of collagen and production of bioactive peptides (Bhagwat and Dandge, 2018), fibrosis treatment (Villegas et al., 2018), and treatment of Dupuytren (Zhang et al., 2019), thus making these enzymes fall into the category of high commercial value proteases (Oliveira et al., 2017c).
Various methods of separating protein biomolecules for biopharmaceutical use have been used to obtain collagenolytic enzymes with a higher degree of purity, the main ones being chromatographic (ion exchange, molecular exclusion) Park et al., 2002). However, the high cost of obtaining it makes it necessary and desirable to constantly search for alternative sources and for extraction processes that are fast, efficient and low cost, such as liquid-liquid extraction, using the aqueous twophase system (ATPS) . ATPS is a form of partitioning where the system is formed by two distinct phases employing a polymer (usually, polyethylene glycol, PEG) and a salt (citrate, sulfate, phosphate, for example). In aquaculture, the extraction by ATPS has been successful in the recovery of microalgae products (Phong et al., 2018;Suarez Garcia et al., 2018), mainly ribulose-1,5-biphosphate carboxylase/oxygenase , proteases from hepatopancreas of pacific white shrimp (Senphan and Benjakul, 2014), and alkaline proteases from albacore tuna (Thunnus alalunga) liver  and spleen (Poonsin et al., 2017), and intestine from peacock bass (Cichla ocellaris) . Thus, this study aimed to partition a collagenolytic protein from digestive waste of common snook (Centropomus undecimalis), using two-phase aqueous system (PEG/citrate) for future biotechnological applications.

Materials
Azocoll substrate, Tris, polyethylene glycol (PEG) of 400, 3350 and 8000 molar mass, dehydrate tri-sodium citrate and monohydrate citric acid salts used were obtained from Sigma-Aldrich®, Saint Louis, USA. HCl-Merck, Darmstadt, Germany. All other reagents were used of analytical grade.

Extraction of collagenolytic enzyme
Digestive viscera of common snook (Brazilian Genetic Heritage n°A25441A) were collected from fish sellers in the public market of São José, Recife, Pernambuco, Brazil. The viscera were transported under refrigeration in ice bags to the Bioactive Technology Laboratory, LABTECBIO, Federal Rural University of Pernambuco, UFRPE. The enzymatic extraction was done by the description by Oliveira et al. (2019). Primarily, 30 g of intestinal viscera were macerated in 50 mM Tris-HCl extraction buffer, homogenized (IKA RW 20D S32, Guangzhou, China) for 5 min., at 10,000-12,000 rpm (4°C), and centrifuged at 12,000 x g (Sorvall® Superspeed RC-6 Centrifuge, Thermo Fisher Scientific, Waltham, MA, USA) for 30 min. at 4°C. The resulting supernatant was preheated to 55ºC for 30 min., was centrifuged again. The resulting supernatant was called crude enzyme extract, used in later analyzes, as shown in Figure 1. The common snook (Centropomus undecimalis) crude extract was subjected to ATPS PEG/citrate technique to extract and pre purify of collagenolytic protease ( Figure 1).

Protein determination
The protein concentrations of all samples were determined by the method of bicinchoninic acid (BCA) according to Smith et al. (1985).

Extraction of aqueous two-phase systems (ATPS)
For enzyme extraction, the ATPS runs were formed in different concentrations according to the 2 4 -factorial design, described in Table 1. The solution of sodium citrate (30% w w -1 ) was prepared by mixing solutions of dihydrate tri-sodium citrate and monohydrate citric acid to obtain pH values (6.0, 7.0 and 8.0) and different volumes of these solutions were transferred to 15 mL-graduated tubes containing different PEG concentrations with different molar masses (400, 3350 or 8000 g mol -1 ). This mixture was added the crude extract accounting for 20% (w w -1 ) of total mass and after water addition up to a final mass of 5 g, the suspensions were then homogenized in a vortex for 1 min., and the two phases were separated by settling for 1 h. The influence of independents variables poly(ethylene glycol) molecular mass (M PEG ), PEG concentration (C PEG ), sodium citrate concentration (C CIT ), and pH, was verify under the response variables protease collagenolytic partition (K), yield (Y) and purification factor (PF) according to a 2 4 -factorial complete design, composed of 16 runs; all independent variables were evaluated at two levels (+1 and -1) (Table 1), with three replicates at the central point (0) 3/8 used to estimate the experimental error as described by (Barros-Neto et al., 2003). After the experiments, the statistical analysis of the data was performed using the Statistica 8.0 program (Statsoft Inc., 2008).

Determination of ATPS parameters
The protease collagenolytic partition coefficient (K) was determined as the ratio of collagenolytic activity (U mL -1 ) in the PEG-rich top phase (CAt) to that in the salt-rich bottom phase (CAb): K = CA T /CA B .
The yeld (Y) was determined as the ratio of the total collagenolytic activity in the top phase to that in the crude extract and expressed as a percentage: Y= (CA t x V t / CA ce x V ce ) x 100; where, V t and V ce are the volumes of the top phase and the crude extract, respectively.
The purification factor (PF) was defined as the ratio of the specific activity in the top phase to its initial value in the crude extract before partition: PF = (CA t /Pt)/(Ca ce /Pt ce ); where Pt (protein in phase rich PEG) and Pt ce (protein in crude extract) are the total protein concentrations (mg mL -1 ) and CA t (collagenolytic in phase rich PEG) and CA ce (collagenolytic in crude extract) are collagenolytic activities (U mL -1 ).

Collagenolytic assay
Collagenolytic activity was performed according to Oliveira et al. (2020). First, in microtube (2 mL), 5 mg of the azocoll, 500 μL of 50 mM Tris-HCl (pH 7.5) and 500 μL of enzymatic sample were added, incubated at 55°C for 30 min., under constant agitation. Then, the reaction was stopped (200 μL of 40% trichloroacetic acid, TCA) for 10 min., to then be centrifuged (10,000 x g for 10 min at 4°C). The reading was done at 520 nm. All tests were performed in triplicate.

Inhibitor sensitivity
The evaluation of sensitivity to inhibitors was carried out according to Oliveira et al. (2019). Briefly: in Eppendorf (2 mL), in the proportion of 1:1, the partitioned sample was subjected to Benzamidine, ethylenediaminetetraacetic acid (EDTA), Phenylmethylsulfonyl fluoride (PMSF), Tosyl-L-lysine chloromethyl ketone (TLCK), Tosyl phenylalanone chloromethylKylethyl ketone (TPCK), for a period of 30 min., to then determine the collagenolytic activity (U mg -1 ). Data are expressed as mean ± SD. These data were statistically analyzed by ANOVA, followed by a post hoc (Tukey) test, when indicated. Differences between groups were accepted as significant at the 95% confidence level (p <0.05).

RESULTS
The two phases formation of intestine extracts was not observed in runs 1, 3, 9 and 11 after the addition of the crude extract in the system ( Table 2). The results obtained of partition coefficient (K) indicating that the collagenolytic protease has higher affinity for the upper phase of the system (rich in PEG), since K >1 values were obtained. It was observed that only in tests 5 and 7 the enzyme partitioned for lower phase of the system.
The effect of independent variables and their interactions evaluated of factorial design showed that K was positively influenced by the PEG molar mass (M PEG ) ( Table 3).
In the runs formed by the polymer with molar mass 8.000 g mol -1 higher values of partition coefficient were obtained (13.17 and 14.78), indicating once again the affinity of the collagenolytic protease for the upper phase of system (rich in PEG). The interaction between the M PEG and pH (1X4) had a significant negative effect under the K for p <0.05.
This interaction presented the antagonistic effect since a lower pH value combined with the higher polymer molar mass (8000 g mol -1 ) favored the enzymatic partition for the polymeric phase. For the activity Yield (Y) values above 100% were observed in the upper phase of the system in the extraction of the collagenolytic protease from C. undecimalis (Table 2). As well as, in the statistical analysis of K, it was possible to observe a strong significative positive influence of M PEG under Y in the upper phase (Y PEG ) and the negative influence of the interaction between M PEG and pH (1x4) for p <0.05 (Table 3).
The highest purification factors (PF) 3.91 was obtained on the run 6, formed with PEG molar mass 8000 (g mol -1 ) in 20.0% (w w -1 ) concentration, 20% (w w -1 ) of citrate concentration at pH 6.0 (Figure 2). Just like for K and Y activity, the independent variable M PEG also had the strong significative positive effect on PF, already, the polymer concentration had a significative negative effect for this variable response. The interaction between variables the M PEG , C PEG and pH (1x2x4) was positively significant (p <0.05) for obtaining for this variable (Table 3).

DISCUSSION
The common snook has been identified as a promising species for supplying enzymes, such as trypsin, chymotrypsin (Concha-Frias et al., 2016), leucine aminopeptidase, carboxypeptidase A, lipase, α-amylase, acid phosphatases, alkaline phosphatase (Jimenez-Martinez et al., 2012), and collagenases . The richness in proteases of the sea bass intestinal viscera has aroused interest for investigations that may lead to the reuse of its residues and at the same time provide alternative sources of biomolecules. Silva et al. (2019) extracted protease from common snook intestinal viscera using only maceration and centrifugation techniques. Here, the aqueous two-phase system (ATPS) was chosen due to the efficiency in the extraction of collagenolytic proteases described for waste from fishery resources, mainly in the efficient extraction of proteases (Kuepethkaew et al., 2017;Phong et al., 2018;Poonsin et al., 2017;Sripokar et al., 2017;Suarez Garcia et al., 2018;Oliveira et al., 2020).
ATPS was a fast method for extracting proteases of digestive viscera from common snook, corroborating with what was described by Porto et al. (2007) and Amaral et al. (2020) when they submitted proteases to the same partitioning conditions, observing the non-formation of phases when displacement of salt and polymer concentrations were below the binodal curve for PEG 400 g mol -1 . The PEG phase potentialized the enzymatic activity, meeting that described by Oliveira et al. (2020) in which the elimination of inhibitors from the enzymatic extract of peacock bass is reported, favoring the increase in collagenolytic activity (pH 8.0; 12.5% PEG 8000 g mol -1 ; 10.0% phosphate). Kuepethkaew et al. (2017) reported partitioning using ATPS (25% PEG 1000 g mol -1 ; 15% MgSO 4 ; pH 5.0) as an efficient recovery technique with good yield and high degree of purification; while Ketnawa et al. (2013) reported an increase in yield (365.53%) when performing a second ATPS cycle (15% PEG 2000 g mol -1 ; 15% sodium citrate; pH 8.5) in digestive viscera extracts of giant catfish (Pangasianodon gigas) for partitioning alkaline proteases, indicating that ATPS can be expanded to industrial production.
The statistical analysis of results showed that the independent variable M PEG had a strong influence for all the variables evaluated in the present study (K, Y PEG and PF). M PEG had a positive influence indicating that the higher molar mass of the polymer (8000 g mol -1 ) favored the enzymatic partition for the upper phase of the ATPS, improved Y activity, and increased the purification factor.
Generally, with the increase in molar mass and concentration of polymer, occurs the K values decreasing. This phenomenon is known as the volume exclusion effect, since, the available free volume is reduced for the protein in the PEG-rich phase, leading to the partition of the biomolecules to the lower phase of the system (Iqbal et al., 2016;Nadar et al., 2017).
However, as increase in the polymer molar mass occurs simultaneous increases in your hydrophobicity, due to increase in hydrophobic area to hydrophilic group ratio (Leong et al., 2015) which results in an increase in the number of hydrophobic interactions formed between the protein and the polymer molecules. Collagenolytic proteases have hydrophobic regions that favor their interaction with PEG and their partitioning to the upper phase (PEG-phase) of the system. Others authors report this behavior when studying the partitioning of these enzymes in ATPS formed by PEG-salt (Rosso et al., 2012;Oliveira et al., 2020).
After the partitioning step, the samples were submitted to sensitivity to inhibitors. The extracts of the common snook (C. undecimalis) showed a marked reduction in activity when subjected to classical inhibitors of serine proteases, mainly trypsin and chymotrypsin, presenting a significant difference in all treatments (p <0.05). Similar inhibition results by Benzamidine, TLCK, and PMSF were found for the pre-purified collagenolytic enzyme from smooth weakfish (Cynoscion leiarchus) (Oliveira et al., 2017a). Enzymes belonging to the class of metalloproteases activities are significantly reduced after exposure to EDTA. In a parallel, the correlation of the effects of the inhibitors on the enzymatic action can indicate the class of the target enzyme.
According to the results described herein, it is likely that the purified enzyme in question is of the serine protease class, corroborating with those described by Teruel and Simpson (1995), Roy et al. (1996), Kim et al. (2002)

CONCLUSIONS
In this study, collagenolytic protease was successfully extracted and semi purified from internal viscera from common snook using the biphasic partition method. Because it is a fast and simple technique when compared to the most improved chromatographic methods, it was possible to prove its feasibility for large-scale application in fish farming residues. All intestinal waste was reused, suggesting that this practice can be included in management plans to reduce environmental impact. The salt used (citrate) is not considered a polluting element and the polymer used can also be reused, which would reduce processing costs. The partitioned enzyme was recovered with a high degree of purification similar to the collagenolytic serine-proteases extracted using chromatographic techniques. Thus, the use of fish residues as a promising raw material in the supply of collagenolytic proteases for biopharmaceutical applications is suggestive, mainly to compose formulations of healing ointments.