Infrniationcil Jouniul of' Food Sciencr.s rind Nutrifion (1997) 48. 169-1 76

Milk-coagulating enzymes of tuna fish waste as a rennet substitute

J.F.P. Tavares,' J.A.B. Baptista2 and M.F. Marcone3

'University of' Azores. Rua da Mae de Deus, 58, 9500 Pontu Delgada, Acores, Portugal, 2Department of Molecular and Medical Genetics, University of Toronto, Toronto, Ontario MSS IA8 and -'Department of Food Sciences, University of Guelph, Guelph. Ontario NlG 2 W l , Canada

Extraction and activation of the tuna zymogen is influenced by temperature, pH, salt concentration and freshness of the stomach tissue. The presence of 25% NaCl in the extraction process markedly enhanced the yield of tuna gastric enzyme. The milk- coagulating time for both tuna protease and rennet, at an incubation temperature of 32°C was dependent, at similar level, on the pH (5.5-6.4) of the milk as a substrate. Tuna protease was less sensitive to losses of activity than rennet at pH values above 6.4. Both enzymes became unstable beyond pH 7.0 and completely lost their activities at pH 8.0. The purpose of this study was to determine the best conditions for recovery of the fish enzyme and to compare its behavior to that of rennet.

Introduction

Many proteolytic enzymes induce coagulation of milk, and the importance of these proteases results not only from its ability to clot milk, but also from the relationship between milk clotting ability and the general proteolysis which the enzyme may produce.

Proteases, from various sources, differ in their catalytic and physical properties, and whether or not a particular enzyme would be suitable to use for a particular industrial appli- cation, depends on several factors. Some have the capacity to clot milk, but most of them are not suitable for cheese making because their hydrolytic action culminates in lower yields, loss of fat from the curd, and development of undesirable changes in texture and flavor during cheese aging (Delissiers et d., 1974). For example, the coagulant from the flowers of Cardon (Cyunuru curdunculus) is traditionally used in Portugal for making soft cheese (Serra Cheese) from sheep's milk. This plant protease increases its activity with decline in pH, and is superior to rennet in the production of this type

of cheese. However, its high proteolytic activity results in loss of yield and defective flavor and texture when used in the manufacture of Edam and Roquefort cheeses (Veira de SB et ul., 1970).

A large number of proteases from plant, animal, and microbial sources, have been inves- tigated as potential sources for milk clotting enzymes and possible rennet substitutes (Sardi- nas, 1972). The difficulties in finding a suitable natural coagulant from a single source have encouraged many researchers to try other routes, such as the modification of enzymatic properties, alteration of cheese-making proce- dures, and the use of enzyme mixtures from different sources.

The extraction of milk-clotting enzymes from fish stomach rnucosa for cheese manufacture would provide an inexpensive alternative to rennet substitutes, for domestic use or to export to cheese-producing nations, and could become a new food-related industry. In addition it would address a very important pollution and disposal

Correspondence to J.F.P. Tavares.

0963-7486/97/030169-08 0 1997 Journals Oxford Ltd

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y T

ufts

Uni

vers

ity o

n 10

/29/

14Fo

r pe

rson

al u

se o

nly.

170 J.EP Tavares et al.

problem, as a means of minimizing the waste associated with processed fish.

Having in mind the interest of the fishery by- products utilization, the purpose of this research was to determine the best conditions for the recovery of milk coagulating enzymes from tuna fish waste, and to compare their catalytic properties with those of commercial rennet.

Materials and methods

Reagents and fish material Spray dried non-fat milk was purchased from New Dundee Creamery, New Dundee, Ontario, Canada. Commercial rennet was provided by Dairy and Food Laboratories (Waukesha, WI). Bio-Rad protease substrate gel tablets were provided by the manufacturer. Unless otherwise specified, all reagents were obtained from E. Merck, Darmstad, Germany, and Sigma Chem- icals Company, St Louis, MO.

Atlantic tuna (Thunnus obesus) caught from fishery zones deep in the North Atlantic Ocean, in proximity of the Azores Islands was obtained from the University of Azores. Atlantic cod (Gadus morhua) caught 3 km off Atlantic Can- ada, was obtained from the Canada Institute of Fisheries Technology. The stomachs were removed from the fish after it was caught, frozen rapidly, and stored at -20°C.

Isolation of tuna fish gastric proteases The stomachs received frozen in dry ice, were partially thawed, split, cleaned, and briefly rinsed three times in tap water. The inner mucosa linings were peeled away from the outer muscular layer, chopped in small pieces using a sharp knife and homogenized using a Polytron homogenizer in equal weight of dis- tilled water (DW), containing 25% of total weight (tissue plus DW) of NaCl. The homoge- nate was stored at -20°C and on the following day was transformed into a slurry using a combination meat grinder/Waring blender, and centrifuged in a Beckman L8-70 M ultracen- trifuge unit, at 35 000 X g during 60 min at 4°C. The pellet was twice more extracted, as descri- bed above. The clear supernatants, called ‘crude pepsinogen’, were combined, concentrated approximately 10-fold, using a Buchi rota- evaporator at room temperature or freeze dried using a Labconco lyophilizer, and stored at -20°C. This gastric extract was used for enzyme

comparative studies and Colby cheese making. The study of the temperature influence on the

yield of tuna gastric protease extraction, was carried out at 4°C and room temperature. Tuna protease was extracted with 25% NaCl solution (w/v), using different holding times, ranging from 0 to 3 h, prior to the enzyme activation at pH 5.0.

Milk clotting assay For many years, the standard substrate for measurement of milk-clotting activity has been skim milk powder reconstituted in 10 mM CaCl, (Berridge, 1952). Visual turbidity in buffered k-casein is used as an index of milk- clotting activity.

The tuna gastric protease activity was assayed by the measurement of coagulation time and milk curd firmness in fresh spray-dried non-fat milk (97% TS) dissolved in a 10 mM CaC1, solution to a concentration of 12% (w/v) milk solids (Berridge substrate), and equilibrated overnight to ensure the milk pro- teins solubilization. Sodium azide (0.02% w / ~ ) was added to prevent microbial growth. The reconstituted milk was warmed, for 1 h prior to the enzyme extract addition, and kept at the coagulation temperature of 32°C.

The influence of tuna protease extract on the rate of milk aggregation (viscosity), was tested in concentrations of 2, 3, 4, and 5 mYlOO ml of milk, using the Nameter Viscometer (Nameter Co., Edison, NJ - IBM compatible computer via analog-to-digital converter), at 32°C (Figure 8). The comparison of milk clotting by the extrac- ted enzymes with known standard rennet was also performed using the apparatus described by Sommer & Matsen (1935). Then, one ml of diluted enzyme was added to 25 ml portions of Berridge substrate in wide mouthed bottles, and the clotting time measured, in bottle revolu- tions, at 32°C. The first appearance of visible flakes on the moving glass surface was taken as the end point. I

The pH-induced physico-chemical changes of casein in milk, and their effect on the structure formation have recently been exten- sively studied (Snoeren et al., 1984; Van Hooydonk et al., 1984; Heertie et al., 1985; Roefs et al., 1985). Drastic changes in the casein micellar system can occur by acid- ification of milk, especially in the pH range 5.0-6.0. For the standardization of the milk

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y T

ufts

Uni

vers

ity o

n 10

/29/

14Fo

r pe

rson

al u

se o

nly.

renneting process, the pH values around 6.3 (mainly soft cheeses), and around 6.6 (mainly hard cheeses) are of interest. The curd acidity at drainage determines not only the basic structure of the cheese by controlling the calcium phos- phate level, but also the amount of rennet retained in the curd. The last aspect is important during the cheese ripening process.

Activation of zymogens The action of the tuna gastric zymogens was achieved by modified Brewer's methodology. Activation was accomplished by adding 0.1 NHCI to pH 4.0, holding at 20-25°C for 60min and re-adjusting de pH to 5.0 with 0. I N NaOH. Completeness of zymogens acti- vation was checked by determination of milk clotting activity (MCU), using the Sommer & Mat sen apparatus. According to these research- ers, the MCU/ml was defined by the following equation:

MCU/ml = 100 X T,IT, X C,/C,

where T, is coagulation time of standard (revolutions), T, is coagulation of unknown (revolutions), C, is concentration of standard, and C, is concentration of unknown.

Sensory evaluation of cheese Colby cheese samples, manufactured with ren- net and tuna protease, according to Kosikowski methodology (1978) and, aged for one month at 12"C, were evaluated by 100 consumers. Of the

100 judges, 60 were female and 40 male with a large proportion between the age of 18 and 25 (79%) and 26-45 years of age (21%). Ninety per cent of judges consumed mild cheddar cheese types. Four replications (rep) were evaluated by groups of 25 consumers. Cheese samples were presented at room temperature in covered 30ml sample cups which had been coded with 3-digit and randomized among judges. They were instructed to use water and unsalted crackers to clear their palates between samples. Preference for flavour, texture, and overall preferences were evaluated using the 9-pt hedonic scale (9 corresponded to very good). In addition, acceptability was assessed as definitely accepted and definitely rejected. Data were collected on the computerized sensory data collection system. Hedonic data were analysed by ANOVA and Tukeys Honestly Different Test, using SAS (version 1 .O).

Reproducibility The data shown in Figures 1-8 are means of triplicate determinations.

Results and discussion

Influence of NaCl concentration, tissue freshness and temperature on the yield of tuna gastric protease extraction The yield of the extraction, increased with increasing amounts of sodium chloride from 15% to 30% (w/v). Salt concentration of 15%

Figure 1. Effect of NaCl concentration on the yield of tuna gastric protease during extraction process. Assay conditions were 12% milk solids, 10 mM CaClz and 32°C of renneting temperature.

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y T

ufts

Uni

vers

ity o

n 10

/29/

14Fo

r pe

rson

al u

se o

nly.

172 ./.El? Tavares et al.

Figure 2. Influence of tissue freshness on the yield of tuna gastric protease. Assay condition were as described in the legend to Figure 1.

slightly accelerated activation, whereas of 25% and 30% markedly enhanced the yield of tuna gastric proteases. The enzyme activation step was carried out by acidification to pH 5.0, just prior to its use. Figure 1 illustrates the effect of sodium chloride concentration on the yield of tuna gastric protease extractions.

The tuna gastric proteases, extracted from fresh and frozen stomachs, were tested at the same concentrations. Figure 2 shows that fresh- ness had a positive influence on milk-clotting activity, and also demonstrates a lag phase of 96 h, probably due to tissue autolysis, followed by a rapid increasing of enzyme activity

(approximately double of the frozen tissue) from 144 h after activation.

The results in Figure 3 illustrate the superior effect of low temperature during extraction, for all holding times. The tuna protease was partially inactivated at room temperature, and for the same temperature, the level of the activity was reduced with increased holding time.

Influence of p H on the protease activation Tuna gastric enzyme activity was compared with cod gastric enzyme and standard rennet, using different activation pHs, ranging from 4.0 to 7.0, and keeping the same activation holding

Figure 3. Influence of temperature on the yield of tuna gastric protease extraction. Assay conditions were as described in the legend to Figure 1 .

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y T

ufts

Uni

vers

ity o

n 10

/29/

14Fo

r pe

rson

al u

se o

nly.

Milk-coagulating enzymes 173

Figure 4. Effect of pH on the yield of tuna gastric protease extraction. Assay conditions were as described in the legend to Figure I

time. The assays were carried out in Berridge substrate at 32°C of renneting temperature. Figure 4 shows no significant differences between tuna, cod gastric enzymes and standard rennet, in the range of pH 4.0-6.0, during 60min. For pH 7.0, milk coagulation was not observed, using tuna and cod gastric enzymes.

Effect r f p H and renneting temperature on milk coagulation time The effects of milk pH on the milk-coagulation by tuna gastric enzymes, standard calf rennet, and mixture of tuna and bovine gastric enzymes (90: 10) were compared to those of Fox ( 1969)

(Figure 6). The milk-clotting time, for tuna protease and calf rennet was pH-dependent at the same level on pH values 5.5-6.3, but the first one, was less pH-dependent for higher pHs than was that of rennet and bovine pepsin enzymes. The tuna gastric enzyme was capable of coagulating milk at a fairly rapid rate up to pH 6.8. When tuna gastric enzyme was mixed with bovine pepsin, there was a large increase in clotting time with an increase in milk pH. For bovine pepsin, this pattern was similar at the higher pH values. The milk pH was adjusted with 0.1 MNaOH or 0.1 MHCl, and all the assays were carried out at 32°C of renneting

Figure 5. Effect of CaClz and milk fat on the milk coagulation. Assay conditions were as described in the legend to Figure I .

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y T

ufts

Uni

vers

ity o

n 10

/29/

14Fo

r pe

rson

al u

se o

nly.

174 J.El? Tavares et al.

30

25 - d

20 E

E F p 15

0 10 i

5 5

- Tuna Gastric Enz. - Standard Rennet Enz.

-d- Tuna Gastric Enz. + Bovine Pepsin Enz.

-. Bovine Pepsin Enz. (Fox)

~~?~~~~~~$~~~~ Milk pH

Figure 6. Influence of milk pH on milk clotting activity by different proteolytic enzymes. Assay conditions were as described in the legend to Figure I .

temperature. Enzyme concentrations were adjusted in order to have equal milk clotting time at pH 6.3.

Although Norris & Mathies (1953) affirmed that the observed apparent optimal temperature, for proteolytic enzyme activity, was near 42"C, the results shown in this study clarify that the best temperatures, using tuna gastric enzyme and standard calf rennet, for milk-clotting protease activation, occurred around 30-38°C (Figure 7). The experiments were carried out using reconstituted Berridge substrate and the enzyme activation at pH 5.0, using the same holding time.

Effect of milk calcium concentration on milk clotting activity Addition of calcium chloride to milk, prior to renneting, not only increases the calcium con- centration, but also decreases the pH of the milk. Both factors accelerate the renneting process. The pH effect is small, but significant ( 1 mMCaC1, decreases the pH by about 0.03 units).

In these experiments using tuna gastric enzyme a strong effect of CaCI2 (Berridge substrate) on the rate of milk coagulation was found. Calcium ion probably not only speeds the rate of aggregation and thereby the spatial

Figure 7. Influence of milk incubation temperature on the clotting activity, using tuna gastric protease and standard calf rennet. Assay conditions were 12% milk solids, 10 mM CaCI2, incubation temperature 15-38°C.

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y T

ufts

Uni

vers

ity o

n 10

/29/

14Fo

r pe

rson

al u

se o

nly.

Milk-coanulatinn enzymes 175

80

70

60

- 50 ?

3 40

3 30

E

.- Ir

5 20 n

10

0

- 2ml of tuna gastric enzyme extract I lOOml of milk - 3ml of tuna gastric enzyme extract I lOOml of milk - 4ml of tuna gastric enzyme extract I 100ml of milk - 5ml of tuna gastric enzyme extract I 1 OOml of milk

I

Figure 8. Effect of tuna gastric protease concentration on the rate of milk aggregation (viscosity). Assay conditions were as described in the legend to Figure I .

distribution of the structural elements, but also the intra- and inter-particle bond strength. Figure 5 illustrates the relative effect of CaCI2 and milk fat on the milk coagulation.

Evaluation qf Colby cheese: sensory analysis Consumer acceptability of Colby cheese made with rennet and tuna protease was determined. ANOVA showed that treatment effect, using different enzymes, was significant for flavour, texture, and overall preference. These sig- nificant effects indicated that there was a high degree of variability within the treatments. Flavour scores for rep 2 were lower than for the other replications (Table 1). Tukeys Honestly Different Test did not reveal any significant difference between replication means for over- all preference, indicating that the difference

here probably resulted from a difference in variability around the mean.

Mean scores (M) and standard deviation (SD) for treatment effects are shown in Table 2. Cheese made with tuna protease was less preferred for flavour, texture and overall prefer- ence than that made by rennet. SD obtained are large indicating a spread in the ratings. Exam-

Table 2. Mean scores* for hedonic ratings of Colby cheeses made with rennet and tuna protease

Overall Flavour Texture preference

Treat. M SD M SD M SD

Rennet 5.15 2.16 5.90 1.94 5.60 1.97 Tuna 4.27 2.37 4.52 2.31 4.18 2.46

Table 1. Mean scores* for replications of Colhy cheeses made by rennet and tuna protease

*n = 100. M = mean scores. SD = standard deviation.

Overall Flavour Texture prejerence

Rep. M SD M SD M SD

1 4.67 2.26 5.31 2.13 5.08 2.27 2 4.37 1.87 4.80 2.02 4.56 1.94 3 5.31 2.35 5.44 2.16 5.42 2.22 4 4.56 2.35 5.23 2.34 4.61 2.43

*n = 75. M = mean scores. SD = standard deviation.

Table 3. Percent of judgements in low preference and high preference categories

Flavour Texture Overall preference preference preference

Low High Low High Low High (%) (%) (%) (%) (%) (%)

Rennet 51 15 23 25 37 18 Tuna 64 07 64 13 67 12

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y T

ufts

Uni

vers

ity o

n 10

/29/

14Fo

r pe

rson

al u

se o

nly.

176 J . E P Tavures et ul.

Table 4. Percent definitely accepted and definitely rejected assessments

Definitelv accepted Definite1.v rejected Treatment (%I (%I

Rennet 23 Tuna 15

16 49

ination of the frequency for scores obtained for each cheese showed a very high proportion in the dislike range for all cheese (Table 3). This was particularly evident for flavour, suggesting the cheeses had not sufficiently ripened when the test was conducted at 1 month of age.

Examination of the acceptedhejected data (Table 4) supports the evidence that the cheese was not liked. Only 23% would definitely accept the rennet, and reduced to 15% for the tuna protease. These data were supported by the comments of judges, and it became difficult to get people to participate as negative feed-back occurred. The cheese made with the tuna enzyme developed beads of moisture on the surface when it was cut indicating a poor ability to hold water.

Conclusions

The yield of tuna gastric enzyme during extraction process increased with increasing

~ ~~

amounts of NaCl from 15% to 25% (w/v). The tissue freshness had a positive effect on the yield of the enzyme extraction. At low tem- perature (4°C) the yield of the enzyme extrac- tion was always higher than at room tem- perature, for all holding times (0-3 h), using 25% (w/v) salt (NaCl) concentration. The tuna gastric enzyme activity (milk coagulation time) was very similar to that of standard rennet, for enzyme activation pH values between 4.0-6.0. The milk pH influence on the milk coagulation by tuna gastric protease had a pH dependence profile similar to that of calf rennet for pH values, ranging from 5.5-6.3, but the first one was less sensitive to losses of activity above pH 6.4.

Addition of CaCl, to the milk prior to renneting affected the rate of the micellar aggregation. There was an influence of milk renneting temperature on the tuna gastric milk clotting activity. The cheese samples were submitted to a taste panel. None of the cheeses were of acceptable commercial quality, prob- ably because they were not sufficiently ripened at time of testing. The cheese samples were highly variable between replications. It appears that the use of tuna protease, as rennet sub- stitute, is technologically feasible, but further studies to characterize the efficacy of this enzyme in cheese processing are necessary to better evaluate their industrial potential.

References Berridge NJ (1952): An improved method of observing the

clotting of milk containing rennin. J. Dairy Res. 19,

Berridge NJ (1955): Purification and assay of rennin. Methods Enzymol. 2, 69-77.

Brewer P, Helbig N & Haard N F (1984): Can. Insf. Fuod Sci. Technol. J. 17, 38-43.

Delissier JP, Mercier JC & Ribadeau Dumas B (1974): Reproduction nutrition development. Annal. Biol. Ani- male Biochim. Biophys. 14, 343.

Fox PF (1969): Milk-clotting and proteolytic activities of rennet, and of bovine pepsin and porcine pepsin. Proc. Rex 36, 427.

Heertie I. Visser J & Smits P (1985): Structure formation of acid milk gels. Food Microstruct. 4, 267-277.

Kosikowski F (1978): Cheese and Fermenfed Milk Foods, 2nd edn, pp. 256-258. New York: F.V. Kosikowski.

Norris EIR & Mathies JC (1953): Preparation, properties and crystallization of tuna pepsin. J. Biol. Chem. 204,

328-329.

673-680. Roefs SPFM, Walstra P, Dalgleish DG & Harme DS

(1985): Preliminary note on the change in casein micelles caused by acidification. Nerh. Milk Dairy J. 39: 119-122.

Sardinas JL (1972): Microbial rennets. Adv. Appl. Micro- bid. 15, 39-73.

Sommer HH & Matsen H (1935): The relation of mastitis to rennet coagulability and curd strength of milk. J. Dairy Sci 18, 741.

Snoeren THM, Klol HJ, Van Hooydonk ACM & Danman AJ (1984): The voluminosity of casein micelles. Milchw. 39, 461-463.

Van Hooydonk ACM, Olieman C & Hagedoorn HG ( 1984): Kinetics of the chymosin-catalysed proteolysis of K-casein in milk. Nefh. Milk Dainj J. 38,

Vieira de S i FC & Barbosa M (1970): Paper presented at 207-222.

l8rh International Dairy Congress, Sydney.

Int J

Foo

d Sc

i Nut

r D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y T

ufts

Uni

vers

ity o

n 10

/29/

14Fo

r pe

rson

al u

se o

nly.