ARTICLE IN PRESS

International Dairy Journal 17 (2007) 937945 www.elsevier.com/locate/idairyj

Chemical analysis and sensory evaluation of Cheddar cheese produced with Lactobacillus acidophilus, Lb. casei, Lb. paracasei or Bidobacterium sp.
L. Onga, A. Henrikssonb, N.P. Shaha,
a

School of Molecular Sciences, Victoria University, Werribee. Campus, P.O. Box 14428, Melbourne, Victoria 8001, Australia b DSM Food Specialties Australia Pty Ltd., NSW, Australia Received 23 April 2006; accepted 30 November 2006

Abstract The sensory properties of probiotic Cheddar cheeses made using Lactobacillus acidophilus 4962, Lb. casei 279, Bidobacterium longum 1941, Lb. acidophilus LAFTIs L10, Lb. paracasei LAFTIs L26 or B. lactis LAFTIs B94 were assessed after ripening for 9 months at 4 1C. Probiotic cheeses except those with Lb. acidophilus 4962 were signicantly different (Po0.05) from the control without any probiotic organism. Acceptability of probiotic cheese with Lb. casei 279 was signicantly lower (Po0.05) than that of the control cheese with bitterness and sour-acid taste as the major defects. Concentration of acetic acid in probiotic cheeses was higher (Po0.05) than the control cheese. Vinegary scores did not inuence the acceptability of the cheeses (P40.05). Increased proteolysis in probiotic cheeses did not inuence the Cheddary attribute scores (P40.05). There were positive correlations (Po0.05) between the scores of bitterness and the level of water-soluble nitrogen. r 2007 Elsevier Ltd. All rights reserved.
Keywords: Cheddar cheese; Probiotic; Proteolysis; Sensory evaluation

1. Introduction A series of chemical and biochemical changes occur during cheese ripening including glycolysis, lipolysis and, most importantly proteolysis (Fox, Law, McSweeney, & Wallace, 1993). Proteolysis plays a critical role in determining the typical sensory characteristics and represents a signicant indicator of quality, as shown for Cheddar cheese (Fox & McSweeney, 1996). Proteolysis is caused by enzymes contained in milk (plasmin), rennet (pepsin and chymosin) and microbial enzymes released by starter cultures. The activities of these enzymes hydrolyze the caseins (as1-, as2-, b- and k-casein), which leads to the formation of large and intermediate size peptides. These peptides may be further hydrolyzed by proteolytic enzymes, originating from the microora [starter bacteria, non-starter lactic acid bacteria (NSLAB) and probiotic
Corresponding author. Tel.: +61 3 9919 8289; fax: +61 3 9919 8284.

E-mail address: nagendra.shah@vu.edu.au (N.P. Shah). 0958-6946/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.idairyj.2007.01.002

adjuncts] of the cheeses, into smaller peptides and free amino acids (FAAs), which are important for the development of Cheddar avour (Cliffe, Marks, & Mulholland, 1993; Lynch, Muir, Banks, McSweeney, & Fox, 1999; Thomas & Mills, 1981). Addition of lactobacilli in cheeses has been associated with an increased proteolysis and intensication of avour (Drake, Boylston, Spence, & Swanson, 1996; Lane & Fox 1996; Lynch et al., 1999; McSweeney et al., 1994). Proteolytic enzymes produced by certain probiotic adjuncts were also found to degrade bitter peptides (Koka & Weimer 2000). Broadbent et al. (2002) showed that lactococcal proteinases and rennet are responsible for the formation of bitter peptides from caseins in Cheddar cheese. Although bitter taste is considered a normal component of Cheddar avour, excessive bitterness may limit consumer acceptance of cheeses. Identication of an adjunct culture that produces a premium quality of Cheddar cheese will thus be useful to the industry.

ARTICLE IN PRESS
938 L. Ong et al. / International Dairy Journal 17 (2007) 937945

Glycolysis also inuences the avor of Cheddar cheese. It involves the conversion of lactose to lactic acid mainly by starter bacteria, reducing the pH of typical Cheddar to 5.15.4. During ripening, residual lactose (0.81.5%) is metabolized to predominantly L-lactate. Oxidation of lactate by starter bacteria, NSLAB or probiotic adjunct in general produces 1 mole of acetate and 1 mole of CO2 and consumes 1 mole of O2 per mole of lactate utilized (Fox et al., 1993). Acetate may also be produced by starter bacteria or probiotic adjunct such as Lactobacillus and Bidobacterium from lactose or citrate or amino acid and is usually present at fairly high concentrations in Cheddar cheeses (Fox & McSweeney, 1996). Acetate is considered to contribute to cheese avor, although high concentrations may cause off-avors. When Bidobacterium lactis was used in combination with Lactobacillus acidophilus strain Ki as the starter in Gouda cheese manufacture (Gomes, Malcata, & Klaver, 1995), there was a signicant effect on cheese avor after nine weeks of ripening, possibly due to the production of acetic acid by the Bidobacterium. Production of acetic acid in cheeses especially with the addition of probiotic adjunct thus requires careful examination. In our previous study (Ong, Henriksson, & Shah, 2006), six probiotic organisms (B. longum 1941, Lb. casei 279, Lb. acidophilus 4962, B. lactis LAFTIs B94, Lb. paracasei LAFTIs L26 and Lb. acidophilus LAFTIs L10) were used for the development of probiotic Cheddar cheese. These organisms were selected based on their acid and bile tolerance, adhesion to intestinal cell line, anticarcinogenic properties, oxygen sensitivity and ability to modify gut microora of human subjects (Crittenden et al., 2001; Lankaputhra & Shah, 1998; McIntosh, Royle, & Playne, 1999). These strains were shown to maintain their viability at high level of 47.0 log10 cfu g1 at the end of ripening period of 6 months at 4 1C. Addition of probiotic adjuncts also changed the proteolytic pattern and the organic acid proles of the cheeses. The objective of this study was to investigate the inuence of B. longum 1941, Lb. casei 279, Lb. acidophilus 4962, B. lactis LAFTIs B94, Lb. paracasei LAFTIs L26 and Lb. acidophilus LAFTIs L10 added individually and in combination, on sensory properties of cheddar cheeses after ripening for 9 months at 4 1C as affected by acetic acid production and proteolysis. 2. Materials and methods 2.1. Starter and probiotic organisms Freeze-dried cultures of Lactococcus lactis subsp. lactis and L. lactis subsp. cremoris obtained from DSM Food Specialties Pty. Ltd. (Moorebank, NSW, Australia) were used for cheese starter culture preparation. The organism were activated by culturing two times at 30 1C overnight in 12% (w/v) sterile reconstituted skim milk (RSM) containing 2% (w/v) glucose and 1.2% (w/v) yeast extract prior to use as bulk culture (2%, v/v).

The probiotic organisms, Lb. acidophilus 4962, Lb. casei 279 and B. longum 1941 were obtained from the Victoria University Culture Collection (Werribee, Victoria, Australia), while Lb. acidophilus LAFTIsL10, Lb. paracasei LAFTIsL26 and B. lactis LAFTIsB94 were obtained from DSM Food Specialties Pty. Ltd. B. longum 1941 (B. longum CSCC 1941) was originally obtained from the Commonwealth Scientic and Industrial Research Organization (CSIRO) (Highett, Victoria, Australia) while Lb. acidophilus 4962 and Lb. casei 279 were both originally obtained from the Australian Starter Culture Collection Center (ASCC, Werribee, VIC, Australia). All Lactobacillus strains were subcultured (1%, v/v) two times at 37 1C overnight in 12% (w/v) sterile RSM prior to use as a bulk culture (2%, v/v). Bidobacterium sp. were subcultured similarly using 1% inoculum in sterile RSM supplemented with 0.05% L-cysteine hydrochloride (Sigma-Aldrich, St. Louis, MO, USA). 2.2. Cheddar cheese manufacture Cheddar cheeses were made using 10 L pasteurized milk and 1.5% (v/v) inoculum of the mixed strain L. lactis subsp. lactis and L. lactis subsp. cremoris in a pair of custom made cheese vats. Probiotic bacteria were added as adjunct as shown in Table 1. Cheeses were manufactured according to the standard procedures of Kosikowski (1977) as described previously by Ong et al. (2006). The 9 batches were produced randomly in 5 days with the same batch of pasteurized milk and 2 replications were produced in the following weeks. The fresh cheese was removed from the mold, packed in oxygen barrier Cryovacs bags (Cryovacs Pty. Ltd., Fawkner, Victoria, Australia), heat-sealed with a Multivacs vacuum packaging equipment (Multivac Sepp Haggenmuller, Wolfertschwenden, Germany) and ripened at 4 1C for 9 months.
Table 1 Control and probiotic Cheddar cheeses with various combinations of probiotic bacteria Cheddar cheesea Batch Batch Batch Batch Batch Batch Batch Batch Batch 1 2 3 4 5 6 7 8 9 Probiotic adjuncts Concentration (v/v)b 1.2% 1.2% 1.2% 1.2% 1.2% 1.2% 0.4%/strain 0.4%/strain

None B. longum 1941 B. lactis LAFTIs B94 Lb. casei 279 Lb. paracasei LAFTIs L26 Lb. acidophilus 4962 Lb. acidophilus LAFTIs L10 BL 1941, LC 279, LA 4962 LAFTIs B94, LAFTIs L26, LAFTIs L10

a All Cheddar cheeses were produced with cheese starter culture (1.5%, v/v) of mix Lactococcus lactis subsp. lactis and L. lactis subsp. cremoris. b 1.2% v/v 12 mL of bulk culture of RSM added to every liter of pasteurized milk, 0.4% v/v 0.4 mL of bulk culture of RSM added to every liter of pasteurized milk.

ARTICLE IN PRESS
L. Ong et al. / International Dairy Journal 17 (2007) 937945 939

2.3. Sensory evaluation of cheeses Staff and students of Victoria University were recruited as the sensory panelists for a series of triangle tests (n 36), acceptance rating test (n 30) and evaluation for specic attributes (n 30). All panelists signed a Victoria University human subjects consent form. The panelists were familiar with basic sensory evaluation techniques and prior to sensory evaluation they participated in brieng sessions. Panelists had access to deionized water and unsalted soda crackers to help cleanse their palates. For sensory evaluation, cheese samples were removed from the refrigerator and cut into pieces (about 1.5 1.5 1.5 cm in size) and placed on white plates coded with random 3-digit numbers one hour prior to evaluation at room temperature (25 1C). Cheese cubes from the three replications of the same batch were mixed randomly so that all replications from the same batch were presented equal number of times. For the triangle test, panelists were asked to choose the odd cheese in a series of eight triangle tests conducted in two days (4 combinations each day). The probiotic cheeses were presented against the control cheese to nd out whether there were any signicant differences between the control and the probiotic cheeses. The cheeses were presented as AAB, ABA, BAA, BBA, BAB and ABB, where A was probiotic cheese and B was control cheese and each combination was presented an equal number of times. Signicant differences were determined using the method of Roessler et al. (1978). For acceptance rating test, panelists were instructed to evaluate their perception of the overall liking of the cheeses on a 10-point intensity scale (1 dislike extremely, 10 like extremely). Prior to tasting, panelists completed a questionnaire on frequency of cheese consumption (o1 once per wk, 23 times per wk, 45 times per wk, or 45 times per wk) and cheese preference (mild, medium or sharp matured cheese). Panelists evaluated all cheeses in two days (ve at each session), with a control cheese included as a reference in each session. The manner in which the treatment combinations were divided between the sessions and the order in which the cheeses were presented was randomized to minimize the carryover effects (Muir & Hunter, 1991, 1992). Panelists were instructed to cleanse their palates before proceeding to the next sample. Panelists evaluated specic attributes, which included Cheddary avour, bitterness, sour-acid, vinegary, creamy, hardness and crumbliness using a 10-point intensity scale. Panelists also evaluated all cheeses in two sessions (ve at each session including a control at each session), as described previously. A wide range of descriptive vocabulary (attributes) for Cheddar avour has been dened by Delahunty and Murray (1997) and Murray and Delahunty (2000). Drake et al. (1996) selected six specic attributes (bitter, sour-acid, oaky/nutty, creamy, rmness and crumbliness) for the sensory evaluation of Cheddar cheeses

made with the addition of adjunct lactobacilli. In our study similar terms were used with some modications. Vinegary attribute was added to the attribute lists to determine the correlation between acetic acid concentrations obtained from instrumental analysis and human sensory perception of vinegary taste of the cheeses. Panelists received 3 main sessions of training prior to sensory evaluation. In the rst session panelists were trained for their ability to detect sour-acid taste, bitterness and vinegary taste of different concentration of lactic acid, caffeine and acetic acid in water. Detection threshold was determined using a series of duotrio tests. In the second session, panelists were trained to recognize the oaky/nutty avour of Cheddar cheese by using three different type of commercial cheese (mild, tasty and vintage cheeses). The attribute oaky/nutty, which corresponded to the avour characteristic of premium quality Cheddar cheese (Bodyfelt, Tobias, & Trout, 1988), was changed to Cheddary, which was dened as the general avours of Cheddar cheese (HulinBertaud, Kilcawley, Wilkinson, & Delahunty, 2000). Panelists were instructed to rate the Cheddary intensity using 110 scale (10 highest intensity, 1 lowest intensity) of three different types of commercial Cheddar cheeses (mild aged 3 months, tasty aged 9 months and vintage aged 412 months). In the third session, panelists were trained for their ability to detect sour-acid taste, bitterness and vinegary taste of different concentration of lactic acid, caffeine and acetic acid in a cream cheese. Detection threshold was determined using a series of duotrio tests. Panelists were also trained for their ability to rank products with different concentration of lactic acid, acetic acid and caffeine from lowest intensity to highest intensity in water and in cream cheese. Repeated testing was performed until panelists were competent to rank different intensity of lactic acid, acetic acid and caffeine both in water and in cream cheese. 2.4. Cheese composition Grated cheese samples were analyzed in duplicate for salt by the Volhard method 975.02 (AOAC, 1990), fat by the Babcock method 933.05 (AOAC, 1990), moisture by oven drying at 102 1C as method 926.08 (AOAC, 1990) and total protein by the Kjeldahl method 920.123 (AOAC, 1990) on a Kjeltec System II (with Digestion System 2000 and Distilling Unit 1002; Tecator, Hoganas, Sweden). The pH of cheese slurry prepared by blending 20 g of grated cheese with 12 mL of H2O (Australian Standard, 1989) was measured with a pH meter (Model 8417, Hanna Instruments Pty. Ltd., Singapore) after calibrating with fresh pH 4.0 and 7.0 standard buffers. 2.5. Acetic acid concentration The concentrations of acetic acid were determined using high performance liquid chromatography (HPLC) as per the method of Ong et al. (2006). The HPLC system

ARTICLE IN PRESS
940 L. Ong et al. / International Dairy Journal 17 (2007) 937945

consisted of a Varian 9012 solvent delivery system, a Varian 9100 auto-sampler, a Varian 9050 variable wavelength UV/Vis tunable absorbance detector and a 730 data module (Varian Inc., Palo Alto, CA, USA). An Aminex HPX-87H column (300 mm 7.8 mm, Bio-Rad Laboratories, Richmond, CA, USA) and a guard column with disposable cartridges H+ (Bio-Rad Laboratories) maintained at 65 1C were used for the analysis. Sulfuric acid (0.009 N), ltered through a 0.45 mm membrane lter (Millipore Corp., Bedford, MA, USA) was used as a mobile phase at a ow rate of 0.6 mL min1. The detection device was the ultravioletvisible detector set at 220 nm with running time of 30 min. 2.6. Assessment of proteolysis The water-soluble extracts (WSE) of the cheeses were prepared according to Kuchroo and Fox (1982). The nitrogen in the extract was estimated in duplicate by the Kjeldahl method. Non-protein nitrogen (TCA-SN) was estimated in 9 mL ltrate obtained after precipitation of the cheese homogenate with 12% TCA (Sigma-Aldrich). The extent of secondary proteolysis was assayed as the amount of nitrogen that was soluble in 5% PTA-SN (SigmaAldrich) extracts of 9 mL of the ltrate. The proteolytic patterns of Cheddar cheeses were also analyzed by assessing the percentage hydrolysis of as-casein and b-casein using sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) as per the method of Ong et al. (2006). The gel images were recorded using a Fuji lm intelligent dark box II with Fuji lm LAS-1000 Lite V1.3 software. The area and intensity of the bands were measured using Fuji lm image gauge V4.0 software (all from Fuji Photo Film Co. Ltd., Japan). Data from the analysis were expressed as the ratio of the area and the intensity of the band. The reduction in the intensity of bands during ripening with respect to the original intensity was expressed as percentage hydrolysis. 2.7. Statistical analysis (data treatment) Data analysis was carried out with Minitab statistical package (Minitab Inc., State College, PA, USA). One-way analysis of variance was used to nd out differences between means, with a signicant level at a 0.05. When signicant differences were found among treatments, means were compared using Tukeys test. The signicance in the differences of data obtained from the triangle test was determined using the expanded statistical tables of Roessler et al. (1978). Simple linear correlation analysis was used to determine a relationship between mean scores of sensory attributes and that of chemical analysis. 3. Results and discussions The results of the sensory analysis (n 36) based on a triangle test for differentiating between control and probiotic

Cheddar cheese are shown in Table 2. There was a signicant difference (Po0.05) between probiotic cheeses made using B. longum 1941, B. lactis LAFTIs B94, Lb. casei 279, Lb. paracasei LAFTIs L26, Lb. acidophilus LAFTIs L10 and the control cheese. Only the cheese made using Lb. acidophilus 4962 was similar to the control cheese (P40.05). Addition of Lactobacillus adjuncts has been reported previously to improve Cheddar cheese avour (Broome, Krause, & Hickey, 1990; McSweeney et al., 1994), but in other case they were responsible for avour defects (Lee et al., 1990). Acceptability of the control and probiotic cheeses in our study is shown in Table 3. There was no signicant difference (Po0.05) between acceptance scores of control cheeses presented at sensory sessions day 1 and day 2 (data not shown). Acceptance scores of other treatments presented at day 1 and day 2 could then be compared. The acceptance scores of cheeses with the addition of probiotic adjuncts were not signicantly different (P40.05) from the control except for cheeses

Table 2 Differences between control and probiotic cheese as perceived by triangle test (n 36) after 9 months of ripening at 4 1C Cheddar cheesea No. of correct judgments 22 20 26 18 17 23 25 20 pb o0.05 o0.05 o0.05 o0.05 40.05 o0.05 o0.05 o0.05

Batch Batch Batch Batch Batch Batch Batch Batch

a

2 3 4 5 6 7 8 9

Batch codes used are as detailed in Table 1. All probiotic cheeses (Batch 2Batch 9) were presented against control cheese (Batch 1). b Probiotic cheeses are signicantly different with control cheese if Po0.05 (95% condence level). Signicance of triangle tests was determined using the expanded statistical tables of Roessler, Pangborn, Sidel, and Stone (1978). Table 3 Mean overall acceptance of control and probiotic Cheddar cheeses after 9 months of ripening at 4 1C (n 30) Cheddar cheesea Batch Batch Batch Batch Batch Batch Batch Batch Batch
a b

Meanb overall acceptance 6.3570.47a 6.3370.42a 4.9370.45ab 3.5970.38c 5.0370.38ab 5.6070.40ab 5.1070.35ab 4.3370.40bc 5.4070.29ab

1 2 3 4 5 6 7 8 9

Batch codes used are as detailed in Table 1. Means in a column followed by the same superscript letter are not signicantly different (P40.05). Results are expressed as mean of scores7standard error of means. 1 dislike extremely, 10 like extremely.

ARTICLE IN PRESS
L. Ong et al. / International Dairy Journal 17 (2007) 937945 941

made with the addition of Lb. casei (Batch 4) and mixture of ABC culture (Lb. acidophilus 4962, B. longum 1941, Lb. casei 279) (Batch 8), which received the lowest acceptance scores. The results thus show that Cheddar cheeses made with probiotic adjuncts B. longum 1941, B. lactis LAFTIs B94, Lb. paracasei LAFTIs L26 or Lb. acidophilus LAFTIs L10 received acceptance scores comparable with that of control cheese. Results from the questionnaire form completed by the panelists showed that the type of cheese preferred, sex, age and frequency of cheese consumption did not affect the acceptance scores (P40.05) (data not shown). About 80% of panelists who participated for the acceptance rating consumed cheese at 12 times per week or more, 70% were female, 73% were between the ages of 18 and 35 years and 53% preferred mild cheeses. Scores for specic attributes of the Cheddar cheeses are shown in Table 4. Out of the seven attributes assessed,

creamy, sour-acid, vinegary, bitterness and hardness of the cheeses were signicantly different among various cheeses (Po0.05). Cheeses with the addition of probiotic adjuncts received higher scores for bitterness, sour-acid taste and vinegary taste when compared with those without probiotic. Bitterness scores in cheese with Lb. casei 279 (Batch 4) and Lb. paracasei LAFTIs L26 (Batch 5) were signicantly higher (Po0.05) than those for the control cheese (Batch 1). Table 5 shows that there was a negative and signicant correlation between the bitterness scores and the acceptability of the cheeses (r 0.809, P 0.008). The resulting bitter off-avours have probably masked the liberated aroma component in the cheeses and may explain the reason for lower acceptance scores obtained. These results supported an earlier observation of El Soda, Madkor, and Tong (2000), who reported that addition of lactobacilli enhanced bitterness when incorporated into Cheddar cheese. In that study, the enhanced bitterness was

Table 4 Scores obtained from judging specic attributes of control and probiotic Cheddar cheeses after 9 months of ripening at 4 1C (n 30) Sensory attributes Cheddar cheesesa Batch 1 Flavor Cheddary Creamy Sour-acidic Vinegary Bitter Texture Hardness Crumbliness
a b

ANOVAb Batch 3 5.4070.33a 5.7370.37ab 4.9370.46ab 3.6770.53ab 5.5770.54ab Batch 4 5.1370.42a 6.0770.35ab 5.3370.53b 4.2370.52b 6.8370.50b Batch 5 5.5770.38a 6.5770.32b 4.0370.47ab 3.4070.51ab 6.3770.42bc Batch 6 5.4070.43a 6.0070.39ab 4.0370.49ab 3.6370.46ab 4.2770.53ab Batch 7 5.4070.44a 5.1070.31ab 4.8070.36ab 4.2070.53b 4.6770.38ab Batch 8 5.7270.45a 5.0070.41ab 4.3070.42ab 5.0070.41b 4.8370.51ab Batch 9 5.5770.32a 5.5370.31ab 3.8770.41ab 4.1770.41ab 4.5370.51ac F P

Batch 2 5.9370.35a 6.2070.36ab 3.7070.45ab 3.4770.50ab 3.7370.47a

5.9470.34a 5.2670.38a 3.2370.42ab 2.4470.39a 3.8470.63a

0.57 2.34 2.32 3.06 4.95

0.82 0.02 0.02 0.00 0.00

5.5270.49a 4.2770.47b 4.9770.41ab 4.4070.44bc 3.7770.41b 5.8070.46ac 5.6770.37abc 5..2770.53abc 5.5770.39abc 4.20 0.00 5.2470.60a 5.0370.58a 5.0370.51a 4.4370.58a 4.4770.56a 5.3070.53a 5.5770.44a 6.3370.53a 5.9770.62a 1.11 0.35

Batch codes used are as detailed in Table 1. One way analysis of variance of means. Means in a row followed by the same superscript letter are not signicantly different (P40.05). Results are expressed as mean of scores7standard error of means. (0 low intensity, 10 high intensity).

Table 5 Correlation between mean scores of sensory attributes and the mean scores of cheeses acceptability Sensory Attributes Correlation between mean scores of sensory attributes (r/P)a Acceptance Cheddary Creamy Sour-acidic Vinegary Bitter Hardness Crumbliness
a

Cheddary

Creamy

Sour-acidic

Vinegary

Bitter

Hardness

0.715 0.030 0.136 0.728 0.858 0.003 0.603 0.086 0.809 0.008 0.119 0.761 0.046 0.907

0.069 0.860 0.547 0.128 0.504 0.166 0.673 0.047 0.089 0.820 0.189 0.626

0.181 0.640 0.554 0.122 0.361 0.339 0.757 0.018 0.853 0.003

0.592 0.093 0.704 0.034 0.221 0.568 0.010 0.980

0.124 0.750 0.360 0.341 0.605 0.084

0.524 0.148 0.531 0.141

0.718 0.029

Correlation between mean scores of sensory attributes was expressed as r. There was signicant correlation between sensory attributes if Po0.05 (95% condence level).

ARTICLE IN PRESS
942 L. Ong et al. / International Dairy Journal 17 (2007) 937945 12.8670.65a 17.1272.63a 12.7871.33a 16.0570.25b 13.7970.22a 15.2371.16a 12.2472.29a 14.3073.18a 13.0771.03a Batch Batch Batch Batch Batch Batch Batch Batch Batch 1 2 3 4 5 6 7 8 9 39.0070.78a 40.6070.78a 43.3070.38a 39.8070.38a 39.0070.99a 43.3071.1a 40.6071.35a 39.2770.81a 39.3370.46a 1.7870.05a 1.8570.04a 1.9070.04a 1.7170.02a 1.7070.06a 1.9570.05a 1.9570.07a 1.7670.02a 1.7270.07a 31.4070.85a 33.1070.64a 30.4570.59a 33.2570.32a 32.5070.83a 31.0070.79a 30.6471.02a 31.0370.26a 31.8970.93a 27.4670.41a 25.5470.49a 28.4170.55a 25.8070.25a 26.8870.68a 25.5170.65a 26.7470.89a 27.7070.20a 28.0770.47a 5.4070.03a 5.2070.10a 5.2670.10a 5.2570.05a 5.3470.14 a 5.1570.13a 5.2270.17a 5.1470.03a 5.2170.04a 0.1570.01a 0.3070.05b 0.2670.01b 0.2570.01ab 0.2570.01ab 0.2070.01ab 0.1870.01ab 0.2370.01ab 0.2370.01ab 35.7871.04a 49.2172.08bc 49.4071.80bc 58.4772.25b 55.9371.95bc 52.9873.04bc 53.0573.10bc 47.1570.61c 45.3273.01ac 16.8370.50a 25.8270.98b 21.1871.80ab 23.7971.55b 19.4571.50ab 24.7171.89b 20.9871.42ab 22.6670.90ab 22.1770.95ab 4.4970.42a 9.2470.50b 7.1270.31bc 7.9170.55bc 6.8070.15ac 8.1170.32bc 6.6170.45ac 7.9970.57bc 8.1370.45bc 27.0773.31a 31.5571.96a 28.0871.47a 62.9471.20b 75.3874.20b 45.4574.30c 47.1771.22c 65.0873.00b 51.5071.77c Batch codes used are as detailed in Table 1. Results are expressed as mean7standard error of means; n 3 sets of data analyzed in duplicate. Means in column with like superscript letters do not differ (P40.05). c The concentration of soluble nitrogen is presented as percentage over total nitrogen. WSN water soluble nitrogen, TCA-SN trichloroacetic acid soluble nitrogen, PTA-SN phosphotungstic acid soluble nitrogen. d Percentage hydrolysis of casein.
b

Composition and proteolytic prolesb Cheddar cheesesa

3.2. Relationship between acetic acid concentration and sensory attributes of Cheddar cheeses Concentrations of acetic acid in cheese with Bidobacterium sp. (Batch 2 and 3), Lb. casei (Batch 4) and Lb. paracasei (Batch 5) were signicantly higher (Po0.05)

Moisture (%)

The results of the composition (percentage moisture, fat, protein, salt and pH), degree of proteolysis (percentage of WSN/TN, TCA-SN/TN, PTA-SN/TN, hydrolysis of as1-casein and b-casein) and acetic acid concentration of Cheddar cheeses are shown in Table 6. Simple linear correlation between the mean scores of the specic attributes and the chemical analysis is shown in Table 7. No signicant differences (P40.05) were observed among the experimental cheeses for the moisture, salt, fat and protein content, which conrmed our previous ndings that addition of probiotic microorganisms has no direct effect on cheese composition (Ong et al., 2006). The relationship between the main compositional variables (protein, fat, salt, moisture and pH) of the cheeses and their sensory characteristics were investigated (Table 7). There were no signicant correlations between the composition variables (protein, salt and moisture content) of the cheeses and sensory attributes scores. The fat content of the cheeses tended to show positive but not signicant correlation with the creamy scores of the cheeses (r 0.627, P 0.071). There were, however, negative and signicant correlation between the fat content and mean scores for hardness (r 0.731, P 0.025). The pH of probiotic cheeses was in general lower than that of control cheese, but this did not inuence the sour-acid scores of the cheeses. It is interesting to note, however, that vinegary scores were negatively and signicantly correlated to the pH of the cheeses (r 0.830, P 0.006) (Table 7).

Salt (%)

Fat (%)

Protein (%)

pH

Acetic acid (%)

WSNc (%)

3.1. Relationship between compositional variables of Cheddar cheeses and their sensory characteristics

Table 6 Inuence of probiotic microorganisms on composition and proteolytic proles of Cheddar cheeses after 9 months of ripening at 4 1C

TCA-SNc (%)

related to the complex peptidases system of the lactobacilli used. Table 5 also shows that the acceptance scores were inuenced by the sour-acidic scores (r 0.858, P 0.003). Higher levels of bitterness and development of more intense sour-acid avor in Cheddar cheeses containing adjunct cultures of lactobacilli have been reported by Lee et al. (1990) and Lynch et al. (1999). Lynch et al. (1999), however found that the bitterness scores were only higher at the early stage of ripening (up to 6 months); thereafter, the bitterness scores for the cheeses with adjunct lactobacilli were similar to those without adjunct lactobacilli. Both bitterness and the sour-acid taste of the cheeses needed to be controlled in order to produce probiotic cheeses with acceptable quality. Although there was a signicant correlation between Cheddary and acceptances scores, ANOVA showed that the Cheddary scores were not signicantly different (P40.05) among various cheeses (Table 4).

PTA-SNc (%)

as1-CNd (%)

b-CNd (%)

ARTICLE IN PRESS
L. Ong et al. / International Dairy Journal 17 (2007) 937945 Table 7 Correlation between results of chemical analysis and mean scores of sensory attributes Sensory attributes Correlation between results of chemical analysis and mean scores of sensory attributes (r/P)a Moisture Flavor Cheddary Creamy Sour-acidic Vinegary Bitter Texture Hardness Crumbliness Acceptance Salt Fat Protein pH Acetic acid WSNb TCA-SNb PTA-SNb as-CNb b-CNb 943

0.391 0.298 0.111 0.776 0.100 0.798 0.086 0.825 0.124 0.751 0.340 0.374 0.060 0.878 0.119 0.761

0.116 0.776 0.185 0.633 0.389 0.301 0.055 0.888 0.469 0.202 0.533 0.140 0.155 0.690 0.355 0.348

0.030 0.939 0.627 0.071 0.120 0.759 0.137 0.724 0.295 0.441 0.731 0.025 0.529 0.143 0.072 0.854

0.160 0.682 0.475 0.196 0.335 0.378 0.055 0.888 0.044 0.910 0.221 0.568 0.418 0.263 0.094 0.809

0.328 0.388 0.247 0.521 0.139 0.721 0.830 0.006 0.241 0.531 0.453 0.221 0.611 0.081 0.189 0.627

0.138 0.722 0.446 0.229 0.388 0.302 0.271 0.481 0.240 0.534 0.594 0.091 0.192 0.621 0.136 0.728

0.807 0.009 0.425 0.254 0.515 0.156 0.376 0.318 0.755 0.046 0.329 0.388 0.385 0.306 0.564 0.114

0.335 0.379 0.142 0.716 0.121 0.756 0.579 0.102 0.110 0.777 0.035 0.930 0.128 0.742 0.100 0.798

0.296 0.439 0.757 0.018 0.218 0.574 0.618 0.076 0.033 0.932 0.078 0.843 0.201 0.605 0.128 0.742

0.414 0.268 0.116 0.766 0.368 0.342 0.369 0.328 0.627 0.071 0.291 0.448 0.022 0.955 0.648 0.059

0.014 0.972 0.530 0.142 0.003 0.995 0.124 0.750 0.003 0.994 0.425 0.255 0.293 0.444 0.012 0.976

a Correlation between mean scores of sensory attributes and chemical analysis was expressed as r. There is signicant correlation between sensory attributes scores and results of chemical analysis if Po0.05 (95% condence level). b WSN water soluble nitrogen, TCA-SN trichloroacetic acid soluble nitrogen, PTA-SN phosphotungstic acid soluble nitrogen, CN casein.

as compared with the control cheese (Table 6). Lb. casei is known to produce acetic acid (Shihata & Shah 2000). On the other hand, Bidobacterium sp. produced acetic acid and lactic acid from lactose via a fructose-6-phosphate shunt pathway (Dinakar & Mistry, 1994). The fermentation pathway results in 3 mole of acetic acid and 2 mole of lactic acid per 2 mole of glucose, therefore generating a theoretical molar ratio (acetic:lactic) of 3:2 (Scardovi & Trovatelli 1965). The level of acetic acid in the Cheddar cheeses was assessed during sensory evaluation as the level of vinegary taste as shown in Table 4. During the course of training, panelists average threshold of acetic acid was determined using a series of duotrio tests. It was found that their acetic acid threshold in water was less than 0.006% (v/v). The vinegary taste, however, was more difcult to be detected in the presence of other components such as fat, protein, acid and salt. When different concentrations of acetic acid were mixed into a product such as cream cheese, the detection threshold increased more than ten fold to 0.060.100% (v/w). The detection thresholds both in water and in cream cheese, however, were much lower than the concentration of acetic acid in experimental Cheddar cheeses (Table 6). Production and accumulation of acetic acid in the probiotic cheeses were reected in the sensory scores. Although all cheeses with adjuncts probiotic organisms received higher vinegary attribute scores (Table 4), there was no signicant correlation between vinegary and acceptability scores (Table 5). Similarly, there was no signicant correlation between the mean scores of vinegary

attribute and acetic acid concentration of the cheeses (r 0.271, P 0.481; Table 7). This indicated that, although panelists were able to detect the presence of acetic acid in the probiotic Cheddar cheeses, they were unable to detect the differences in the concentration of acetic acid between different batches of cheeses. Results also indicated that the vinegary avor of the cheeses in this study was still within the acceptable range. 3.3. Relationship between proteolytic pattern of Cheddar cheeses and their sensory characteristics Higher levels of proteolysis were detected in probiotic cheeses after 9 months of ripening at 4 1C as indicated by the release of more water-, TCA- and PTA-soluble peptides in the probiotic cheeses (Table 6). The soluble nitrogenous compounds provided by casein proteolysis contribute directly to cheese avour and texture (Fox et al., 1993). These nitrogen fractions generally increased as cheeses aged, corresponding to the continued breakdown of casein and large peptides into small peptides and amino acids by the action of starter culture enzymes and residual rennet (Lau, Barbano, & Rasmussen, 1991). The levels of proteolysis was particularly high in cheeses with the addition of Lb. casei 279 (Batch 4) and Lb. paracasei LAFTIs L26 (Batch 5), because both strains have considerably higher proteolytic activity (Ong et al., 2006). At the end of the 9 months ripening period at 4 1C, percentage hydrolysis of as1-CN in all probiotic cheeses except B. longum 1941 (Batch 3) cheese was signicantly higher (Po0.05) than in control cheeses.

ARTICLE IN PRESS
944 L. Ong et al. / International Dairy Journal 17 (2007) 937945

The increased proteolysis in probiotic cheeses showed that the cheeses ripened faster than controls. The results were expected to inuence the Cheddary scores of the cheeses, but Table 4 shows that the Cheddary scores were not signicantly different between batches. The scores of commercial cheeses used during the training session included 3.1170.54, 5.6070.41 and 8.9070.62 for mild, tasty and vintage cheeses, respectively. The average scores of Cheddary attributes as shown in Table 4 indicated that the experimental cheeses were comparable with the commercial tasty cheese. The ndings also show that the probiotic adjuncts used in this study did not increase the level of perceived maturity. The control cheese, despite having lower WSN, TCA-SN and PTA-SN had a Cheddary score comparable to the probiotic cheeses (Table 4). Previously, excessive proteolysis was associated with the increase in bitterness (Lee et al., 1990). Table 7 shows that there was a positive and signicant correlation between the scores of bitterness and the level of WSN (r 0.755, P 0.046). According to Lemieux and Simard (1992), bitterness develops when small to medium-size hydrophobic peptides produced by the coagulant and microbial enzymes accumulate to the levels that exceed desirable taste thresholds. Residue 193209 of the C-terminal region of b-casein and residue 19 of the N-terminal region of as1casein have been associated with bitterness in cheese (Lemieux & Simard, 1992). Reverse phaseHPLC chromatograms of the Cheddar cheeses in our study indicated the presence of b-casein (f 193209) and as1-casein (f 19) (data not shown). Broadbent et al. (2002) performed regression analysis of bitter avour scores from the trained sensory panel and concentration of b-casein (f 193209) and as1casein (f 19) and found that these peptides had positive correlation with bitterness. Hydrolysis of these peptides is associated with decreased bitterness in Cheddar cheese, but it appeared that proteolytic enzymes produced by the probiotic adjunct might not be able to convert the bitter peptides to nonbitter peptides. The correlation between the proteolytic enzymes produced by the probiotic bacteria and the hydrolysis of bitter peptides in Cheddar cheese is currently being investigated. A level of salt in moisture (S/M)44.5% is necessary to prevent the development of bitterness in cheese (Mistry & Kasperson, 1998). These ratios reect the amount of free water available for microbial growth. A lower level of S/M and high water activities allow excessive bacterial growth, promote excessive proteolysis and lipolysis and in turn lead to defective body, texture (open, soft, greasy), avor (unclean, bitter) and consumer unacceptability (Beresford & Williams, 2004). The moisture content of the experimental cheeses was slightly higher than average moisture content of commercial cheeses. The higher moisture content resulted in the lowering of the percentage of S/M. Although there was no signicant difference (P40.05) between the salt content of the cheeses, some cheeses such as those with Lb. casei 279

(Batch 4) and Lb. paracasei LAFTIs L26 (Batch 5), had only about 4.3% S/M. The percentage of S/M would probably be one of the factors that contribute to the increase in bitterness. The most important contributor to the continuous casein matrix of a Cheddar cheese is as1-CN (Lawrence, Creamer, & Gilles, 1987), and hydrolysis of this casein is believed to be responsible for the softening of Cheddar cheese texture (Grappin, Rank, & Olson, 1985). Cheese with the addition of B. longum 1941 (Batch 2), Lb. casei 279 (Batch 4) and Lb. paracasei LAFTIs L26 (Batch 5) received signicantly lower hardness scores than the control (Table 4). The scores however did not inuence the acceptability of the cheeses (Table 5) and no correlation was observed between the degree of proteolysis (percentage of WSN/TN, TCASN/TN, PTA-SN/TN) and panelists scores for hardness (Table 7). 4. Conclusion Sensory panelists perceived all probiotic cheeses except that with Lb. acidophilus 4962 to be signicantly different than the control cheese without probiotic. Acceptability of probiotic cheese with Lb. casei 279 was signicantly lower than that of the control cheese with bitterness and souracid taste as the major defects. Concentration of acetic acid in cheeses with Bidobacterium sp., Lb. casei 279 and Lb. paracasei LAFTIs L26 was signicantly higher than the control cheese. Vinegary scores, however, did not signicantly inuence the acceptability of the cheeses and panelists were not able to detect the various concentrations of acetic acid between the different batches of cheese. Concentrations of soluble nitrogen and hydrolysis of casein were higher in probiotic cheeses, but increased proteolysis did not signicantly inuence the Cheddary attribute scores of the cheeses. There was, however, a positive and signicant correlation between the scores of bitterness and the level of water-soluble nitrogen. Although there was no signicant correlation between the composition variables (protein, salt and moisture content) and sensory attributes scores, cheeses with the lowest percentage of S/M received the highest bitterness scores. Some strains used in the study have potential for use in the production of probiotic Cheddar cheese with acceptable quality. Acknowledgments This study was made possible through a grant from Australian Research Council (ARC) under a Strategic Partnership with Industry for Research and Training scheme in collaboration with DSM Food Specialties Pty. Ltd. (NSW, Australia). References
AoAC. (1990). AoAC Ofcial methods of analysis (15th ed., Vol. 2). Arlington, VA, USA: Association of Ofcial Analytical Chemists.

ARTICLE IN PRESS
L. Ong et al. / International Dairy Journal 17 (2007) 937945 Australian Standard. (1989). Methods of chemical and physical testing in the dairy industrydetermination of pH. Australian Standard 2300.1.6. Beresford, T., & Williams, A. (2004). The microbiology of cheese ripening. In P. F. Fox, P. L. H. McSweeney, T. M. Cogan, T. P. Guinee (Eds.), Cheese: Chemistry, physics and microbiology (3rd ed., Vol. 1) (pp. 287317). London, UK: Elsevier Academic Press. Bodyfelt, F. W., Tobias, J., & Trout, G. M. (1988). Sensory evaluation of cheese. In F. N. Bodyfelt (Ed.), The sensory evaluation of dairy products (pp. 300376). New York, USA: Van Nostrand Reinhold. Broadbent, J. R., Barnes, M., Brennand, C., Strickland, M., Houck, K., Johnson, M. E., et al. (2002). Contribution of Lactococcus lactis cell envelope proteinase specicity to peptide accumulation and bitterness in reduced-fat Cheddar cheese. Applied and Environmental Microbiology, 68, 17781785. Broome, M. C., Krause, D. A., & Hickey, M. W. (1990). The use of nonstarter lactobacilli in Cheddar cheese manufacture. The Australian Journal of Dairy Technology, 45, 6773. Cliffe, A. J., Marks, J. D., & Mulholland, F. (1993). Isolation and characterization of non-volatile avors from cheese: Peptide prole of avor fractions from Cheddar cheese, determined by reverse-phase high performance liquid chromatography. International Dairy Journal, 3, 379387. Crittenden, R. G., Morries, L. F., Harvey, M. L., Tran, L. T., Mitchell, H. L., & Playne, M. J. (2001). Selection of Bidobacterium strain to complement resistant starch in symbiotic yoghurt. Journal of Applied Microbiology, 90, 268278. Delahunty, C. M., & Murray, J. M. (1997). Organoleptic evaluation of cheese. In T. M. Cogan, P. F. Fox, R. P. Ross (Eds.), Teagasc, Dublin proceedings of the 5th cheese symposium (pp. 9097). Fermoy, Ireland: National Dairy Products Research Centre. Dinakar, P., & Mistry, V. V. (1994). Growth of Bidobacterium bidum in Cheddar cheese. Journal of Dairy Science, 77, 28542864. Drake, M. A., Boylston, T. D., Spence, K. D., & Swanson, B. G. (1996). Chemical and sensory effects of a Lactobacillus adjunct in Cheddar cheese. Food Research International, 29, 381387. El Soda, M., Madkor, S. A., & Tong, P. S. (2000). Adjunct cultures: Recent developments and potential signicance to the cheese industry. Journal of Dairy Science, 83, 609619. Fox, P. F., Law, L., McSweeney, P. L. H., & Wallace, J. (1993). Biochemistry of cheese ripening. In P. F. Fox (Ed.), Cheese: Chemistry, physics and microbiology (2nd ed., Vol. 1) (pp. 343367). London, UK: Chapman & Hall. Fox, P. F., & McSweeney, P. L. H. (1996). Proteolysis in cheese during ripening. Food Review International, 12, 457509. Gomes, A. M. P., Malcata, F. X., & Klaver, F. A. M. (1995). Incorporation and survival of Bidobacterium sp. strain Bo and Lactobacillus acidophilus strain Ki in a cheese product. Netherlands Milk and Dairy Journal, 49, 7195. Grappin, R., Rank, T. C., & Olson, N. F. (1985). Primary proteolysis of cheese proteins during ripening. A review. Journal of Dairy Science, 68, 531540. Hulin-Bertaud, S., Kilcawley, K. N., Wilkinson, M. G., & Delahunty, C. M. (2000). Sensory and compositional relationships between commercial Cheddar-avored enzyme-modied cheeses and natural Cheddar. Journal of Food Science, 65, 10761082. Koka, R., & Weimer, B. C. (2000). Investigation of the ability of puried protease from Pseudomonas uorescens RO98 to de-bitter cheese. International Dairy Journal, 10, 7579. 945 Kosikowski, F. (1977). Cheese and Fermented Milk Foods. Ann Arbor, MI, USA: Edward Brothers Inc. (pp. 228260). Kuchroo, C. N., & Fox, P. F. (1982). Soluble nitrogen in Cheddar cheese: Comparison of extraction procedures. Milchwissenschaft, 37, 331335. Lane, C. M., & Fox, P. F. (1996). Contribution of starter and adjunct lactobacilli to proteolysis in Cheddar cheese during ripening. International Dairy Journal, 6, 715728. Lankaputhra, W. E. V., & Shah, N. P. (1998). Adherence of probiotic bacteria to human colonic cells. Bioscience and Microora, 17, 105113. Lau, K. Y., Barbano, D. M., & Rasmussen, R. R. (1991). Inuence of pasteurization of milk on protein breakdown in Cheddar cheese during aging. Journal of Dairy Science, 74, 727740. Lawrence, R. C., Creamer, L. K., & Gilles, J. (1987). Texture development during cheese ripening. Journal of Dairy Science, 70, 17481760. Lee, B. H., Laleye, L. C., Simard, R. E., Holley, R. A., Emmons, D. B., & Giroux, R. N. (1990). Inuence of homofermentative Lactobacilli on physicochemical and sensory properties of Cheddar cheese. Journal of Food Science, 55, 386390. Lemieux, L., & Simard, R. E. (1992). Bitter avor in dairy products. II. A review of bitter peptides from caseins: Their formation, isolation and identication, structure masking and inhibition. Lait, 72, 335382. Lynch, C. M., Muir, D. D., Banks, J. M., McSweeney, P. L. H., & Fox, P. F. (1999). Inuence of adjunct cultures of Lactobacillus paracasei ssp. paracasei or Lactobacillus plantarum on Cheddar cheese ripening. Journal of Dairy Science, 82, 16181628. McSweeney, P. L. H., Walsh, E. M., Fox, P. F., Cogan, T. M., Drinan, F. D., & Castelo-Gonzalez, M. (1994). A procedure for the manufacture of Cheddar cheese under controlled bacteriological conditions and the effect of adjunct lactobacilli on cheese quality. Ireland Journal of Agriculture and Food Reserve, 33, 186192. McIntosh, G. H., Royle, P. J., & Playne, M. J. (1999). A probiotic strain of L. acidophilus reduces DMH-induced large intestinal tumours in male Sprague-Dawley rats. Nutrition and Cancer, 35, 153159. Mistry, V. V., & Kasperson, K. M. (1998). Inuence of salt on the quality of reduced fat Cheddar cheese. Journal of Dairy Science, 81, 12141221. Muir, D. D., & Hunter, E. A. (1991, 1992). Sensory evaluation of Cheddar cheese: Order of testing and carryover effects. Food Quality and Preference, 3, 141145. Murray, J. M., & Delahunty, C. M. (2000). Selection of standard reference terms in a Cheddar-type cheese avour language. Journal of Sensory Studies, 15, 179199. Ong, L., Henriksson, A., & Shah, N. P. (2006). Development of probiotic Cheddar cheese containing Lb. acidophilus, Lb. paracasei, Lb. casei and Bidobacterium spp. and the inuence of these bacteria on proteolytic patterns and production of organic acid. International Dairy Journal, 16, 446456. Roessler, E. B., Pangborn, R. M., Sidel, J. L., & Stone, H. (1978). Expanded statistical tables for estimating signicance in pairedpreference, paireddifference, duotrio and triangle tests. Journal of Food Science, 43, 940944. Scardovi, V., & Trovatelli, L. D. (1965). The fructose-6-phosphate shunt as a peculiar pattern of hexose degradation in the genus Bidobacterium. Analytical Microbiology Enzymology, 15, 1927. Shihata, A., & Shah, N. P. (2000). Proteolytic proles of yoghurt and probiotic bacteria. International Dairy Journal, 10, 401408. Thomas, T. D., & Mills, O. E. (1981). Proteolytic enzymes of starter bacteria. Netherlands Milk and Dairy Journal, 35, 255273.