Role of casein on induction and enhancement of production of a bacterial milk clotting protease from an indigenously isolated Bacillus subtilis


R.K. Saxena, Department of Microbiology, University of Delhi South Campus, Benito Juarez Road, New Delhi 110021, India. E-mail:,


Aims:  To isolate and enhance the yield of a bacterial milk clotting protease (MCP) through process optimization and scale up.

Materials and Results: Bacillus subtilis was isolated as MCP producer with good milk clotting activity (MCA) per proteolytic activity (PA) index. The enzyme production was inducible with casein and enhanced with fructose and ammonium nitrate resulting in 571·43 U ml−1 of enzyme.

Conclusions:  Medium containing 4% fructose, 0·75% casein, 0·3% NH4NO3 and 10 mmol l–1 CaCl2, pH 6·0, inoculated with 4% (v/v) inoculum, incubated at 37°C, 200 rev min−1 for 72 h gave maximum production. A 6·67-fold increase in MCP yield with very high MCA per PA index was observed after final optimization indicating similarity to rennets.

Significance and Impact of the Study:  Mostly fungal MCPs have been reported. The MCA and MCA per PA index of this bacterium is comparable to that of many fungal reports and better than quite a few bacterial MCPs. Thus, this enzyme by B. subtilis has good probability of successful use in cheese production.


Rennet (E.C. 3·4·24·4) obtained from calf stomach is the single largest application of proteolytic enzymes in food processing for manufacture of cheese. The worldwide increase in cheese production, reduced supply and increasingly high prices of calf rennet have lead to numerous attempts to find suitable rennet substitutes (Cavalcanti et al. 2005). Several animal and plant proteases have been identified as possible rennet substitutes. However, rennet like enzymes from Rhizomucor, Mucor, Endothia parasitica, Aspergillus oryzae and Irpex lactis have received wide acceptability because of their high milk clotting (MCA) and low proteolytic activities (PA) (da Silveira et al. 2005). Owing to rapid growth and relatively inexpensive growth substrates, microbial milk clotting enzymes have become popular rennet substitutes. Mostly, investigations have been focused on fungal milk clotting proteases (MCP) with few sporadic reports on bacterial rennets (D’Souza and Pereira 1982; Cavalcanti et al. 2005). This is because of their high protease activity generating off flavours and bitter taste. Still, some bacterial MCP have been identified with the formation of Cheddar cheese having distinctive flavours (Preetha and Boopathy 1994). Among a dozen or so bacterial species, B. subtilis, B. mesentricus and B. cereus have been reported as potential sources of MCPs.

In the present study, screening for the isolation of bacterial MCP producer with high MCA and MCA per PA ratio was carried out. Subsequently, studies were carried out to enhance the yield of the MCP and the impact of various factors on the MCA per PA index was observed. Scale up in a 10L fermentor was carried out.

Materials and methods

Organism and culture condition

One hundred and eighty bacterial isolates obtained from soil were grown on nutrient agar slants (g l–1: peptone 5·0, beef extract 1·5, yeast extract 1·5, NaCl 5·0, pH 7·0 ± 0·2) at 37°C for 24 h and maintained at 10 ± 2°C in a B.O.D. incubator (Yorco Sales Pvt. Ltd, New Delhi, India) with regular subculturing.

Milk clotting protease producers were screened using casein agar plates as per to the procedure described in Arima et al. 1967. Isolates giving a clear zone of hydrolysis were grown in medium containing (w/v) glucose 3·0%, casein 0·5%, ammonium nitrate 0·2%, K2HPO4 0·1%, KH2PO4 0·05%, MgSO4.7H2O 0·05%, FeSO4.7H2O (0·001 mg 100 ml–1) and ZnSO4.7H2O (0·001 mg 100 ml–1). pH was adjusted to 6·0 using 1 mol l−1 NaOH/HCl before sterilization (10 psi and 20 min). At periodic intervals, samples were withdrawn and centrifuged at 8000 g for 15 min with supernatant used for assay of milk clotting activity (MCA) and proteolytic activity (PA).

The selected isolate was identified at MIDILABS Inc., (Newark, DE, USA) using 16s RNA analysis.

Analytical procedures

Assay for MCA

The activity determination was based on the time required to clot reconstituted instant milk powder (100 g l−1) modified by the addition of 0·01 mol l–1 CaCl2. An amount of the crude enzyme extract (0·5 ml) was added to a test tube containing 5 ml of skim milk preincubated at 37°C for about 10 min. The end point was decided as being the time required for the first appearance of a ‘grainy’ texture (Arima et al. 1967). Results were expressed as MCA units.

The amount of enzyme required to clot milk in 1 min is defined as containing 400 MCA units (Tubesha and Al-Delaimy 2003).

Assay for PA

The modified Anson’s method as described in Arima et al. 1967 was used.

One unit of PA is defined as the amount of enzyme which yields the colour equivalent to 1 μmol of tyrosine per min.

Production of MCP

Process optimization by one variable at a time approach was followed to determine the factors affecting production of MCP. Two percent inoculum (OD620 of 0·6–0·8) raised in nutrient broth at 37°C was used unless specified otherwise.

Induction of the MCP

Initially, the nature of enzyme secretions was investigated by replacing casein with different nitrogen substrates (Table 1).

Table 1.   Production of milk clotting protease by Bacillus subtilis in basal medium supplemented with different sources
Growth mediumProteolytic activity (U ml–1)Clotting time (min) MCA (U ml–1)MCA per PA
  1. MCA, milk clotting activity; PA, proteolytic activity; CSL, corn steep liquor.

Basal medium + 0·5% yeast extract0·28 
Basal medium + 0·5% tryptone0·19 
Basal medium + 0·5% peptone0·39 
Basal medium + 0·5% CSL0·30 
Basal medium + 0·5% casein hydrolysate0·378 : 0349·69147·30
Basal medium + 0·50% soybean meal0·22 
Basal medium + 0·50% beef extract0·25 
Basal medium + 0·50% gelatin0·63 
Basal medium + 0·25% casein0·325 : 4270·10219·06
Basal medium + 0·50% casein0·284 : 5980·20219·06
Basal medium + 0·75% casein0·193 : 33112·56592·42
Basal medium + 1·0% casein0·393 : 27115·83297·0
Basal medium + 1·25% casein0·303 : 23118·45394·83
Basal medium + 1·50% casein0·373 : 20120·31325·16
Basal medium + 1·75% casein0·223 : 32113·60516·36
Basal medium + 2·0% casein0·254 : 5581·36325·44

Process optimization in batch fermentations

To determine the most favorable source of carbon and nitrogen for enzyme production, various simple and complex carbon (2% w/v) sources (Fig. 1) and inorganic and complex nitrogen (0·2% w/v) sources (Fig. 2) were supplemented individually in the production medium. The concentrations of the selected carbon (1–5% w/v) and nitrogen (0·1–0·5% w/v) sources were optimized. Different metal chlorides (CaCl2, MgCl2, BaCl2.2H2O, MnCl2, KCl, NaCl, FeCl3, ZnCl2 and AlCl3 at concentration of 10 mmol l–1) were assessed for effect on production.

Figure 1.

 Effect of carbon sources on production of milk clotting protease.

Figure 2.

 Effect of nitrogen sources on production of milk clotting protease.

pH (4–8), temperature (25–50°C), inoculum (1–8% v/v) and agitation rate (100–300 rev min−1 at intervals of 50) were evaluated for obtaining optimal MCP yield up to 96 with sampling at 12-h intervals unless otherwise specified.

Scale up in fermentor

Scale up was carried out in 10 l bioreactor (Bioflo IV, New Brunswick Scientific Inc. Co., USA) with 7·5 l optimized medium (pH 6·0). Fermentation was initiated with an air flow of 10 l min−1, 200 rev min−1 and 37°C using 4% (v/v) inoculum. Samples were withdrawn periodically over 70 h to determine pH, CFU and MCA.

All experiments were run in triplicate, sets repeated with similar results and the mean values presented. The values, when analysed by Student’s t-test, were within the 95% CL.


Isolation of the organism

Preliminary screening indicated 17 bacterial MCP producers of which an isolate subsequently identified as Bacillus subtilis, showing highest MCA per PA ratio (288·48 in 72 h after incubation) was selected for further investigations.

Enzyme induction

Here, casein acted as an inducer of enzyme synthesis as its replacement or absence yielded variable proteolytic (caseinolytic) activity with loss in MCP production (Table 1). Using 0·75% w/v casein, 112·56 U ml−1 MCP was obtained with best MCA per PA ratio. Higher concentrations yielded higher MCA values with decline in MCA per PA ratio.

Process optimization

Negligible enzyme production (12·0 U ml−1) was observed in absence of carbon source (Fig. 1). A 9·3-fold increase in MCA was observed with glucose. Evaluation of various carbon sources for production resulted in fructose yielding 212·76 U ml−1 of enzyme followed by dextran (194·83 U ml−1) and sorbitol (165·43 U ml−1). Optimization of fructose concentration further increased production yielding 278·06 U ml−1 at 4% w/v with a decline thereafter. Production of MCP in presence of a single nitrogen source (casein) was considerably less compared (Fig. 2) to the yield obtained through a synergistic effect shown by (a) casein and ammonium nitrate (276·06 U ml−1) and (b) casein and peptone (271·89 U ml−1), while diammonium hydrogen phosphate, ammonium chloride and ammonium sulphate completely inhibited production. Maximum MCA units of 300 U ml–1 was obtained with 0·3% (w/v) ammonium nitrate. Metal ions like Ca2+, Mg2+, Fe2+ and Na+ significantly affected enzyme yield. Highest units were obtained with Ca2+ (357·3 U ml−1) followed by Mg2+(328·0 U ml−1), Fe2+ (320·0 U ml−1) and Na+ (312·17 U ml−1).

Maximum production was obtained at initial pH 6·0. The production fell drastically at pH 5·0 with no growth and production observed below this pH. Production was observed in the range of 25–50°C with maximum production achieved at 37°C. Higher temperature decreased production with no growth beyond 50°C.

Varying inoculum size rather than the inoculum age led to a significant impact on the production of MCP (Fig. 3). Highest production was obtained at an inoculum size of 4% v/v inoculum (490·20 U ml−1) at an OD620 of 0·6–0·8. Agitation had a significant impact on production increasing the yield to 571·43 U ml−1 at 200 rev min−1. Increasing the agitation rate further gave no yield enhancement.

Figure 3.

 Effect of inoculum on production of milk clotting protease (% v/v) bsl00041 1%; bsl00043 2%, bsl00000 3%; bsl00001 4%; bsl00066 5%; bsl000726%.

Scale up in fermentor

The production was scaled up under controlled conditions in a 10L fermentor. pH of the medium was maintained at 6·0. Optimum MCP production was obtained at 37 ± 1°C, aeration of 9·0 l min−1 and agitation rate of 250 rev min−1. Highest MCA value was obtained in 42 h with a productivity of 14·024 U ml–1 h–1 along with an MCA per PA value of more than 1·0 (Fig. 4).

Figure 4.

 Milk clotting protease production profile of in a 10 L fermentor. bsl00041 MCA (U ml−1) bsl00066 pH bsl00001 CFU (10 11 ml−1).


Microbial rennets or MCPs possess distinct and worth while advantages not available with animal rennet in terms of cost, relatively unlimited and stable supplies, etc. (Sardinas 1972). Till date, fungal sp. especially Mucor, Rhizopus, Endothia, Rhizomucor and Aspergillus have been exploited for MCPs. Some Bacillus sp. notably B. subtilis, B. cereus, B. licheniformis, etc. (Srinivasan et al. 1964; D’Souza and Pereira 1982) and others like Myxococcus (Poza et al. 2003) and Nocardiopsis (Cavalcanti et al. 2005) have also been reported as producers. The lack of research on bacterial MCP may be due to their higher PA leading to formation of bitter peptides (Channe and Shewale 1998). This investigation aimed at isolating a bacterium with high MCA and MCA per PA index.

An isolate with good MCA and MCA per PA values, identified as Bacillus subtilis was selected for further studies. Initial studies showed that the enzyme was inducible with casein. Supplementation of amino acids or peptides or their generation because of some basal PA can act as inducer for MCP production in some bacteria (Porto et al. 1996). Substrate activation for enzyme induction has been reported for milk clotting enzyme production by Absidia ramosa (Sannabhadti and Srinivasan 1977). MCP production from B. subtilis has been noted in skim milk medium (Srinivasan et al. 1964), while Proflo and casein have been reported for Mucor (da Silveira et al. 2005). As casein is the prime constituent of skim milk powder, its role in induction of enzyme synthesis is evident in these investigations (Thakur et al. 1990). Noticeable increase in production with skim milk powder has been reported for Rhizopus and Mucor respectively (Thakur et al. 1990; Preetha and Boopathy 1994).

Carbon sources stimulated rennet production, such as glucose in M. meihei (da Silveira et al. 2005). Our results show that fructose resulted in maximum stimulation as in Aspergillus (Shata 2005) whereas for A. niger, starch was a better carbon source (Channe and Shewale 1998). Catabolite repression by fructose at concentrations more than 4% w/v is similar to other organisms using different carbohydrates at higher concentrations (Jorai and Buxton 1994). Supplementation of casein with ammonium nitrate or peptone gave significant yields. Thakur et al. (1990) reported ammonium nitrate for M. meihei while soybean flour and meat peptone have been reported for Aspergillus niger and Nocardiopsis (Srinivasan and Dhar 1990, Cavalcanti et al. 2005).

Metal ions are known to affect MCP production. Enhancement in presence of Mg2+ and Fe2+ has been reported for Amylomyces rouxii (Yu and Chou 2005). Skim milk powder and CaCl2, a metal ion essential for milk clotting, exhibited a synergistic effect in enhancing the production of MCP from A. niger (Channe and Shewale 1998) while having no effect from Penicillium citrinum (Abdel-Fattah et al. 1972).

Optimal MCP production at pH 6·0 has been reported from Fusarium subglutinans (Ghareib et al. 2001) and Mucor (da Silveira et al. 2005). However, D’Souza and Pereira (1982), Cavalcanti et al. (2005) and Poza et al. (2003) observed production by B. licheniformis, Nocardiopsis and M. xanthus at pH 7·0, 7·2 and 7·6 respectively, while M. mucedo (Mashaly et al. 1981) gave maximum production at pH 9·0. 30°C has been reported to be optimal for A. rouxii and Mucor respectively (Tubesha and Al-Delaimy 2003,Yu and Chou 2005) and 35°C for production from Mucor sp. (Thakur et al. 1990). But production of MCP from B. subtilis fell considerably at this temperature.

Optimization finally yielded a 6·67-fold increase in the production of MCP (88·88 U ml−1 to 571·43 U ml−1) along with an increase in the MCA per PA index. The scale up studies on a 10L fermentor clearly establishes its scalability for large scale production. The MCA yield has been found to be comparable to A. niger (Channe and Shewale 1998), Rhizomucor (Preetha and Boopathy 1994), Mucor (Tubesha and Al-Delaimy 2003), A. rouxii (Yu and Chou 2005) and much higher than milk clotting rennets from M. xanthus (Poza et al. 2003) and Nocardiopsis (Cavalcanti et al. 2005).

On summarizing, it can be concluded that MCP from B. subtilis can be considered as a suitable alternative to the conventional rennet as well as to the currently available fungal rennets though it would require additional efforts at further enhancing the yield.


Authors acknowledge with sincere thanks the help of Lata, Jasmine, Saurabh and Rekha for critically evaluating the manuscript.