Effects of the Addition of Non-Starter Lactic Acid Bacteria on Free Amino Acid Production During Cheese Ripening

Abstract

To accelerate cheese ripening, two non-starter lactic acid bacteria, Lactobacillus paracasei strain EG9 and Leuconostoc mesenteroides subsp. mesenteroides strain FM1, were isolated from ripened cheeses. The growth of EG9 at 10°C was not affected by 1.7% (w/v) sodium chloride, whereas the growth of Lb. paracasei subsp. paracasei JCM8130^T was suppressed. FM1 produced a higher amount of free amino acids (FAA) than Leu. mesenteroides subsp. mesenteroides ATCC8293^T in a skim milk culture. Experimental cheeses were produced to examine the effect of isolate additions on cheese ripening. The total FAA content of the cheese was significantly increased with EG9 addition during ripening. The parameters of a taste sensor were not affected by EG9 addition, but the pH and water content were significantly lower, and the texture showed greater hardening. It was suggested that EG9 is useful to accelerate FAA production in cheese without influencing taste sensor analysis.

Introduction

The production of fermented foods is one of the oldest food processing technologies known to man (Caplice and Fitzgerald, 1999). Cheese is the representative fermented food from milk produced by lactic acid bacteria (LAB) through ripening. Raw milk is a natural growth medium for microorganisms; the composition and quality of the microflora are variable (Ciprovica and Mikelsone, 2011). Compared with many other foods, relatively few microorganisms are able to grow in cheese because of its low redox, low pH, high salt, and low carbohydrate source environment (Swearingen et al., 2001).

Ripening cheese consists of a series of complex biochemical processes necessary for the development of cheese flavor and texture (Madkor et al., 2000). During cheese ripening, casein is degraded into peptides and amino acids by proteolytic enzymes. Even though ripening is an important process in cheese manufacturing, it is slow, and consequently expensive, and is not fully predictable or controllable (Fox et al., 1996). Accordingly, shortening the ripening time through a low-cost method will reduce cheese costs for both the producer and consumer (Raksakulthai et al., 2002). Many studies have focused on the acceleration of cheese ripening. These include studies on elevating the ripening temperature (Law and Haandrikman, 1997; Grazier et al., 1991; Grazier et al., 1993), high-pressure processing (Kolakowski et al., 1998), attenuated starter cultures (Law, 1999; Meijer et al., 1998), and addition of enzymes (Law and King, 1985; Laloy et al., 1998).

Among the many LAB, non-starter (NS) LAB dominate the cheese microbiota during ripening (Settanni and Moschetti, 2010). The sources of NSLAB likely include post-pasteurization contamination by airborne flora, the cheese-making equipment and ingredients, or survival post-pasteurization (Fitzsimons et al., 2001), and they are normally comprised of various mesophilic lactobacilli and pediococci species (Crow et al., 1995). NSLAB have been shown to accelerate the production of free amino acids (FAA) during ripening (Lynch et al., 1996) and to improve flavor development (Antonsson et al., 2003). They are also highly specialized to a specific hostile ecological niche, typically characterized by the presence of salt, low water, 4.9–5.3 pH values, low temperatures, and a deficiency of nutrients. Moreover, they have a key role in determining curd maturation and final cheese characteristics (Solieri et al., 2012). However, instrumental texture characteristics and taste characteristics have not been well investigated.

In this study, we attempted to obtain a novel NSLAB strain, produced a cheese with NSLAB as an adjunct starter, and evaluated the effect of NSLAB on FAA production, pH, and water content. Instrumental texture and taste characteristics were also investigated using Texture Profile Analysis (TPA) and a taste sensing system, respectively.

Materials and Methods

Strains and culture conditions    Lactobacillus paracasei subsp. paracasei JCM8130^T was purchased from Japan Collection of Microorganisms (RIKEN BRC, Tsukuba, Japan). Leuconostoc mesenteroides subsp. mesenteroides ATCC8293^T was obtained from the American Type Culture Collection (Manassas, VA, USA). Lactococcus lactis subsp. cremoris strain 01-1 and Lc. lactis subsp. cremoris strain 712 isolated from cheese starters were laboratory collections (Nomura et al., 1998). Bacteria were inoculated to fresh medium by a platinum loop and incubated at 30°C overnight to enhance their growth before use.

Lactobacilli MRS broth (DIFCO) and glucose yeast peptone (GYP) broth were used to cultivate LAB. GYP broth contained (L⁻¹) 10 g d-glucose, 10 g yeast extract, 5 g polypeptone, 2 g meat extract, 0.5 g Tween 80, 0.2 g magnesium sulfate heptahydrate, 0.01 g manganese(II) sulfate tetrahydrate, 0.01 g iron(II) sulfate heptahydrate, and 0.01 g sodium chloride (NaCl) at pH 6.8.

Isolation of NSLAB from cheeses    Nine ripened cheeses were used for NSLAB isolation; five cheeses were made at our laboratory and four were commercially sourced (Table 1). One gram of each cheese was homogenized using a T25 digital Ultra-Turrax homogenizer (IKA, Königswinter, Germany) with 50 mL of 2% trisodium citrate dihydrate solution maintained at 37°C. The curd solutions were adequately diluted with 0.85% NaCl and spread onto MRS and GYP agar plates. The plates were incubated at 15°C for 2 weeks. Colonies grown on agar plates were picked to 5 mL MRS broth and incubated at 30°C overnight. Isolates were stored at −80°C until use.

Table 1. Cheeses and media used for NSLAB isolation.

Cheeses	Type	Manufacture	Starter	Isolation medium	Isolates
i	Gouda	Laboratory	Lc. lactis	MRS	AM1–10
				GYP	AG1–10
ii	Gouda	Laboratory	Lc. lactis	MRS	BM1–10
				GYP	BG1–10
iii	Gouda	Laboratory	Lc. lactis	MRS	–
				GYP	–
iv	Gouda	Laboratory	Lc. lactis	MRS	DM1–5
				GYP	DG1–5
v	Gouda	Laboratory	Lc. lactis	MRS	EM1–10
				GYP	EG1–10
vi	Gouda	Commercial (Japan)	Unknown	MRS	FM1–5
				GYP	FG1–5
vii	Gouda	Commercial (Dutch)	Unknown	MRS	GM1–5
				GYP	GG1–5
viii	Gruyere	Commercial (Switzerland)	Unknown	MRS	HM1–5
				GYP	HG1–5
ix	Parmigiano-Reggiano	Commercial (Italy)	Unknown	MRS	–
				GYP	–

Identification of isolates    Genotypic identification of the isolates was performed by 16S rRNA gene sequencing (Hagi et al., 2013). Fresh bacterial cells were suspended in 500 µL of sterile deionized water and then used as a polymerase chain reaction (PCR) template. For isolates in which a PCR product was not amplified, DNA preparation was carried out with a DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany). The 16S-derived universal primers 27F (5′-AGA GTT TGA TCC TGG CTC AG-3′) and 1525R (5′-AAG GAG GTG ATC CAG CC-3′) were used to amplify DNA. PCR amplification was performed with a total reaction mixture of 20 µL containing 1 µL of template, 0.8 µL of each primer, 2 µL of 10 × Ex Taq buffer, 1.6 µL of dNTP mixture, and 0.1 µL of TAKARA Ex Taq Polymerase Hot Start Version (Takara Bio, Otsu, Japan) by using the following program: denaturation at 95°C for 5 min, followed by 35 cycles of denaturation at 95°C for 30 s, annealing at 50°C for 30 s, and extension at 72°C for 90 s. PCR products were electrophoresed and recovered from the agarose gel with the QIAquick Gel Extraction Kit (Qiagen, Hilden, Germany). Purified fragments were sequenced with primers 27F and 800R (5′-CAT CGT TTA CGG CGT GGA C-3′) using a BigDye Terminator v3.1 Cycle Sequencing Kit and a 48-capillary 3730 DNA Analyzer (Applied Biosystems, Foster City, CA, USA). The sequences derived from the isolates were analyzed using the BLAST search tool (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Because the Lb. casei group (Lb. casei, Lb. paracasei, Lb. rhamnosus, and Lb. zeae) could not be differentiated by the abovementioned method, isolates classified as belonging to the Lb. casei group were tested with a species-specific PCR to distinguish members of the Lb. casei group (Desai et al., 2006).

Characterization of isolates    The physiological properties of NSLAB and their type strains were assessed using API 20 Strep (bioMerieux. Marcy l'Etoile, France) according to the manufacturer's instructions. Test preparations were incubated at 30°C, and readings were made after 4 and 24 h. The effects of low temperature and salt stress on the growth of the isolates were examined. Actively growing culture was inoculated at a rate of 1% (v/v) to MRS broth supplemented with 0, 1.7 or 4.0% (w/v) NaCl and incubated at 10, 15, or 30°C. Turbidity (OD620) was measured at 0, 6, 12, 24, 48, 72, 96, 120, 144, and 168 h cultivation with a Spectronics 21 Spectrophotometer (Bausch and Lomb, Rochester, NY, USA). The tests were performed in duplicate.

The FAAs produced by the isolates were measured (Nomura et al., 1998). The medium for the amino acid analysis was 10% reconstituted skim milk supplemented with 0.5% glucose and 0.5% yeast extract (DIFCO), since EG9 or FM1 seldom grew in skim milk. Actively growing culture was inoculated to the fresh medium at a rate of 1% (v/v) and incubated at 15°C for 3 days. The FAAs in the culture were analyzed using an amino acid analyzer (L-8900; Hitachi Ltd., Tokyo, Japan). In these experiments, type-strains were cultured under the same conditions as the control.

Experimental cheese manufacture    The isolates were cultured at 30°C for 48 h with 6.0 L of MRS broth. Cells were collected by centrifugation at 5,800 × g for 20 min at 4°C, washed once with 300 mL of 0.85% NaCl, suspended with the same solution, and stored at 4°C overnight. Conventional lactic fermentation starters were cultured at 30°C overnight in 300 mL of autoclaved skim milk.

Experimental cheeses were manufactured at our laboratory using standard Gouda-type cheese procedures (Nomura et al., 1998). Milk was obtained from Holstein cows at the Institute of Livestock and Grassland Science, NARO (Tsukuba, Ibaraki, Japan). Milk (30 kg) was pasteurized at 63°C for 30 min. A starter culture (300 mL) and NSLAB cell suspension (300 mL) were added to the milk and then incubated at 30°C. To elucidate the effect of adding the isolate, milk was inoculated with an excessive amount of NSLAB relative to the lactic fermentation starter. After 1 h of incubation, renneting was carried out at 30°C for 60 min using 0.75 g of rennet powder (Fromase 2200 TL Granulate; DSM Food Specialties, Heerlen, The Netherlands). The coagulated curd was cut into about 1–2 cm³ cubes with a curd knife. The cooking temperature was 36°C. After 1 h of cooking, the curd was stuffed into a plastic mold and pressed with a 1 kg weight for 4 h at 35°C. The curd was salted over a period of 16 h in an 8% NaCl solution at 10°C. Cheese curds were dried at 10°C for 1 week, then vacuum-packed and ripened for 180 days at 10°C.

Four experiments were conducted as follows using Lc. lactis subsp. cremoris 01-1 and 712 as a lactic fermentation starter and Lb. paracasei EG9 and Leu. mesenteroides FM1 as an adjunct starter. In experiment A, a control cheese was made with 01-1 and a trial cheese was made with 01-1 and EG9. In experiment B, a control cheese was made with 01-1 and a trial cheese was made with 01-1 and FM1. In experiment C, a control cheese was made with 712 and a trial cheese was made with 712 and EG9. In experiment D, a control cheese was made with 712 and a trial cheese was made with 712 and FM1. Cheese samples were gathered after 0, 7, 14, 30, 60, 90, 120, 150, and 180 days of ripening. The cheese was cut radially, the rind was removed, and it was minced finely and stored at −20°C until use.

Cheese analyses    LAB counts in experimental cheeses were measured by real-time PCR. One gram of a cheese sample was homogenized with 50 mL of 2% trisodium citrate solution at 37–40°C. DNA was extracted and purified from 100 µL of homogenate using the DNeasy Blood and Tissue Kit and used as templates for real-time PCR. The composition of the PCR reaction mixture was as follows: 2 µL of template, 5 µL of qPCR mix (Thunderbird qPCR Mix; Toyobo, Osaka, Japan), 0.3 µL of each primer, and diluted water was added to make 10 µL in total. Species-specific primers for each bacterial strain were used (Table 2). The PCR reaction condition was denaturation at 95°C for 1 min, followed by 39 cycles of denaturation at 95°C for 15 sec, annealing at 60°C for 30 sec, and extension at 65°C for 5 sec using the Bio-Rad CFX96 Real-Time System (Bio-Rad Laboratories, Hercules, CA, USA). For calibration, DNA prepared by a serial dilution series of EG9 cells was used as the standard template.

Table 2. Primers for real-time PCR.

Species	Primer	Sequence (5′-3′)	Target gene	Reference
Lc. lactis	68FCa33	GAT GAA GAT TGG TGC TTG CAC TTT GAA GAG	16S rRNA	Grattepanche et al., 2005.
	16SR308	CCT CTC AGG TCG GCT ATG TA
Lb. paracasei	PrI	CAG ACT GAA AGT CTG ACG G	16S–23S rRNA spacer	Walter et al., 2000.
	CasII	GCG ATG CGA ATT TCT TTT TC
Leu. mesenteroides	LcmesS	CCA GTT GTA ATG CGT TAT TAC C	23S rRNA	Elizaquivel et al., 2008.
	LcmesA	CAC AGC TTG TCC TTA TAG AAA A

FAAs of the cheeses were measured (Nomura et al., 1998). Water content was analyzed by oven-drying at 102°C (ISO/IDF, 2004). The pH was determined in a cheese slurry (1:4 cheese/water) using a pH meter.

The physical properties of the cheese were analyzed using TPA. Selected measurement parameters related to cheese texture were as follows: “hardness-1,” “hardness-2,” “cohesiveness,” “springiness,” and “adhesiveness” (Konstance and Holsinger, 1992; Messens et al., 2000). Cylindrical samples (12.7 mm diameter, 10.0 mm height) were taken from the cheese in three blocks, each for a triplicate analysis. Cheese samples were held in a water bath at 14°C until measurements were taken. Samples were compressed to 80% of their original height with a crosshead speed of 50 mm/min using the Model 5542 tensile tester (Instron Corp., Canton, MA, USA).

A taste sensor analysis was conducted using the SA-402B taste sensing system (Intelligent Sensor Technology, Atsugi, Japan). Taste was measured on the basis of the difference in membrane potential between a standard solution and a sample solution. The “relative value” as a first membrane potential change was the first taste, and the “change of membrane potential by adsorption” (CPA) value as a second change of re-measurement after washing was the after taste. The measurement parameters were “umami,” “saltiness,” “sourness,” “bitterness,” and “astringency” for the first taste, and “CPA-umami,” “CPA-bitterness,” and “CPA-astringency” for the after taste (Toko, 1996). Five grams of a minced cheese sample was homogenized with 20 mL of a 30-fold diluted standard solution for the SA-402B taste sensing system. After centrifugation at 10,000 × g for 20 min at 5°C, the fat was carefully removed and 0.25 mL of the supernatant was diluted to 50 mL with a 30-fold diluted standard solution as a sample for the taste sensor analysis. Following taste sensor measurements, sensor output values were converted into a “taste value” (Chikuni et al., 2010) using a data converting tool (Intelligent Sensor Technology). All analyses were performed in triplicate.

Statistical analysis    Data counts of LAB and FAA content were analyzed by a Student's t-test using a significance level of 0.05. TPA scores, pH, water content, and “taste value” of control and trial cheeses were analyzed by analysis of variance within each experiment. NSLAB were designated as the main fixed effect.

Results

Isolation of NSLAB    In an attempt to isolate NSLAB, nine ripened cheeses were used and applied to MRS and GYP agar plates. Colonies were observed on both plates for seven of the nine cheeses (Table 1), and no colonies formed from the remaining two cheeses. Colonies were picked up randomly from each plate and a total of 100 isolates were obtained.

From the 16S rRNA gene sequence analysis, 88 isolates could be identified (Table 3): 28 isolates of Lb. paracasei, 6 isolates of Leu. mesenteroides subsp. mesenteroides, 3 isolates of Leu. mesenteroides, 11 isolates of Lc. lactis subsp. lactis, 31 isolates of Lc. lactis, 8 isolates of Kocuria koreensis, and 1 isolate of Brevibacterium sp. Of those identified isolates, Lactobacillus sp. and Leuconostoc sp. were presumed as potential NSLAB. It was difficult to determine whether the isolated Lc. lactis was a NSLAB because it is a dominant microflora of cheese starters. Lb. paracasei was regarded as an NSLAB participating in cheese ripening. Leuconostoc sp. is a heterofermentative LAB. Two strains, Lb. paracasei EG9 and Leu. mesenteroides subsp. mesenteroides FM1, were selected for further investigations.

Table 3. Identification of the isolates by BLAST and species-specific PCR.

Isolate	Species	Accession No.	Identity
EM1–EM3, EM6–EM10, EG1–EG10, HM1–HM5, HG1–HG5	Lactobacillus paracasei*	KX851554.1	100%
FM1–FM3, FM5, FG1, FG2	Leuconostoc mesenteroides subsp. mesenteroides	CP014610.1	99–100%
FG3	Leuconostoc mesenteroides	GQ253515.1	97%
FG4	Leuconostoc mesenteroides	HF562842.1	99%
FM4	Leuconostoc mesenteroides	KU301265.1	99%
AM4, AM10, BM9	Lactococcus lactis subsp. lactis	AP012281.1	99–100%
EM4	Lactococcus lactis subsp. lactis	CP002365.1	99%
BM7, BM10, BG4–BG6, BG8	Lactococcus lactis subsp. lactis	CP009054.1	99–100%
BM6	Lactococcus lactis subsp. lactis	CP010050.1	100%
AM3, DM5	Lactococcus lactis	EU074844.1	98–99%
EM5	Lactococcus lactis	JX267125.1	100%
AM1–2, AM5–AM9, AG1, AG3–AG7, AG10, BM1–BM3, BM8, BG1–BG3, BG7, BG9–BG10, GM2, GG2	Lactococcus lactis	KU324909.1	99–100%
BM4, BM5	Lactococcus lactis	KX156195.1	99–100%
DM1, DM2, DM4, DG1, DG2	Kocuria koreensis	JQ659396.1	99%
DM3, DG4, DG5	Kocuria koreensis	LT223585.1	98–99%
DG3	Brevibacterium sp.	KR055018.1	99%
AG2, AG8, AG9, FG5, GM1, GM3–GM5, GG1, GG3–GG5	Uncultivable	—	—

*  Species of these isolates were confirmed by species-specific PCR to distinguish members of the Lb. casei group.

Characteristics of NSLAB    The physiological properties of NSLAB and their type strains were investigated (Table 4). As with JCM8130^T, EG9 did not show β-galactosidase activity and did not utilize lactose by 24 h. However, acidification from lactose was observed after 48 h. FM 1 assimilated only lactose by 24 h, and no reaction was observed with other carbohydrates. After 48 h, EG9, FM1 and ATCC8293^T assimilated more carbohydrates.

Table 4. Physiological properties of NSLAB isolates and their type strains.

	Lb. paracasei		Leuc. mesenteriodes
Property	EG9	JCM8130T	FM1	ATCC8293T
Acetoin from pyruvate	+	+	+	+
Hippurate hydrolysis	+	+	−	−
Aesculin hydrolysis	−	+	−	+
Pyrrolidonyl arylamidase	+	+	−	−
α-Galactosidase	−	−	−	+
β-Glucuronidase	−	−	−	−
β-Galactosidase	−	−	+	+
Alkaline phosphatase	−	+	−	−
Leucine aminopeptidase	+	+	+	+
Arginine dihydrolase	−	−	−	−
Acid from*
d-ribose	− (−)	− (−)	− (−)	− (−)
l-arabinose	− (+)	− (−)	− (+)	+ (+)
d-mannitol	+ (+)	− (−)	− (−)	+ (+)
d-sorbitol	− (−)	− (−)	− (−)	− (−)
lactose	− (+)	− (−)	+ (+)	+ (+)
d-trehalose	+ (+)	+ (+)	− (+)	+ (+)
Inulin	− (−)	− (−)	− (−)	− (−)
d-raffinose	− (+)	− (−)	− (+)	+ (+)
starch	− (−)	− (−)	− (−)	− (−)
glycogen	− (−)	− (−)	− (−)	− (−)

*  In parentheses, the readings after incubation for 48 h.

The effect of NaCl on the growth of NSLAB and their type strains was investigated at 10, 15, and 30°C. At 10°C, the growth of EG9 was faster than that of JCM8130^T and turbidity reached 1.0 at 96-h cultivation, while that of JCM8130^T was around 0.5 (Fig. 1). The growth of EG9 was not affected by 1.7% NaCl, but it was retarded at 4.0%. The growth of JCM8130^T was suppressed at both 1.7% and 4.0% NaCl at 10°C. The growth of both EG9 and JCM8130^T was suppressed by 1.7% and 4.0% NaCl at 15°C (Fig. 1) and 30°C (data not shown). From these results, it was suggested that at 10°C, EG9 showed higher NaCl tolerance compared to JCM8130^T. On the other hand, the growth rates of FM1 and ATCC8293^T were suppressed at 1.7% and 4.0% NaCl at 10°C. The low temperature decreased their final turbidity. It was considered that the growth characteristics of FM1 under salt stress were not different from those of ATCC8293^T.

Fig. 1.

Bacterial growth curves at 10°C and 15°C at various NaCl concentrations. (a) JCM8130^T, (b) EG9, (c) ATCC8293^T, and (d) FM1. ◆; NaCl 0%, ■; NaCl 1.7%, ▴; NaCl 4.0%. Points are means of duplicate experiments.

The FAA contents produced by EG9 were almost the same as those of JCM8130^T, whereas FM1 produced a much higher amount of FAAs than ATCC8293^T after 72-h incubation (Fig. 2). Therefore, it is possible that FM1 has a higher proteolytic ability.

Fig. 2.

Comparison of free amino acids (FAA) content in the culture of type strain and NSLAB. Bacteria were cultured at 15°C for 72 h. a) Strain JCM8130^T/EG9, b) Strain ATCC8293^T/FM1. Open bar; type strain, solid bar; NSLAB. Values are means of triplicate determinations.

Analysis of experimental cheese

Bacterial amounts    The EG9 population in experiments A and C was measured at approximately 1.0 × 10⁷ population number g⁻¹ at 0 days of ripening and gradually increased until 180 days (Fig. 3a, 3c). On the other hand, the FM1 population in experiments B and D gradually decreased during ripening (Fig. 3b, 3d). Although FM1 cells inoculated into milk reached 3.0 × 10⁸ colony forming units (cfu) mL⁻¹, FM1 was 1.0 × 10⁵ population number g⁻¹ at 0 days of ripening in experimental cheese D. The cause for the low FM1 population number was unclear, although FM1 might be inhibited by competition with strain 712.

Fig. 3.

Changes in bacterial amounts during cheese ripening. a) experiment A, △; 01-1 in control, ▴; 01-1 in trial, ■, EG9 in trial. b) experiment B, △; 01-1 in control, ▴; 01-1 in trial, ◆, FM1 in trial. c) experiment C, ○; 712 in control, ●; 712 in trial, ■, EG9 in trial. d) experiment D, ○; 712 in control, ●; 712 in trial, ◆, FM1 in trial.

Proteolysis    The total FAA content of the 180-day ripened cheeses was significantly increased by EG9 addition, especially in experiment A (Fig. 4). “Umami” components such as l-Glu and l-Asp were abundantly included. After 30-day ripening, the amino acid content of the EG9 addition cheese was significantly higher than that of the EG9 non-addition cheese (Table 5). It appeared that the amino acid level of the EG9 addition cheese after 30-day ripening was higher than that of the EG9 non-addition cheese after 180 days of ripening. In experiments B and D, there was no difference between the total FAA contents of control cheese and FM1 addition cheese (Fig. 4).

Fig. 4.

Total FAA content of cheeses after 180-day ripening. Open bar; control cheeses, solid bar; trial cheeses. Values are means of triplicate determinations. * Significant difference between control and trial (P < 0.05).

Table 5. FAA content in experiment A cheeses.

	Control				Trial
	0-day	30-day	90-day	180-day	0-day	30-day	90-day	180-day
	(mg g−1)				(mg g−1)
Asp	0.48	0.78	1.23	2.01	1.22	2.63*	4.08*	6.15*
Thr	0.08	0.20	0.52	1.01	0.14	0.49	2.12*	4.40*
Ser	0.26	0.41	0.75	1.79	0.39	1.10*	3.90*	8.47*
Glu	1.29	2.71	5.92	11.21	2.71	9.95*	21.72*	43.05*
Pro	2.60	3.48	9.58	15.18	2.91	5.31*	7.56*	11.41*
Gly	0.19	0.47	0.64	1.38	0.31	1.18*	3.51*	6.68*
Ala	0.80	0.99	1.17	1.97	1.99	2.96*	5.52*	9.16*
Val	0.77	1.06	1.09	2.10	0.78	2.46*	6.31*	11.52*
Met	0.35	0.56	1.18	2.94	0.42	0.82	2.94*	6.37*
Ile	0.90	0.67	1.03	1.23	0.85	1.12*	1.40*	0.00
Leu	1.16	1.63	2.50	6.10	1.70	5.77*	13.42*	26.59*
Tyr	1.28	1.10	1.25	1.83	1.38	1.96*	3.18*	5.52*
Phe	1.28	1.59	1.61	3.13	1.35	3.07*	6.09*	9.83*
γ-ABA	0.27	0.35	0.09	1.29	0.30	0.43*	0.46	1.56
His	0.64	0.58	0.84	0.98	0.72	1.03*	1.43*	2.05*
Lys	1.18	1.18	1.27	2.09	1.25	2.58	5.14*	8.92*
Trp	1.18	1.96	4.05	6.37	3.25	2.93*	5.75	3.71*
Arg	1.50	1.16	1.31	2.53	1.52	3.47	5.62*	7.37*

γ-ABA; γ-aminobutyric acid.

Data are means of triplicate determinations.

*  Significant difference between control and trial in the same ripening days (P < 0.05).

Physical properties    The values of pH, water content, and TPA parameters for 180-day ripened cheeses were measured (Table 6). In experiment A, the values of pH and water content of the EG9 addition cheese were significantly lower than in the control. In experiment C, the pH value of the EG9 cheese was significantly lower than that of the control. The “hardness-1” and “hardness-2” parameters of the EG9 cheese in experiment C were also significantly higher than those of the control. On the other hand, in experiments B and D made with strain FM1, there were no differences between the control and trial for all measured parameters.

Table 6. The value of pH, water content, and TPA parameters for 180-day ripened cheeses.

				TPA parameters
Experiment	LAB	pH	Water (%)	Hardness-1 (kg)	Hardness-2 (kg)	Cohesiveness	Springiness (mm)	Adhesiveness
A	01–1	4.89±0.04	41.49±1.15	30.21±6.94	23.64±4.60	0.068±0.043	1.19±0.40	−0.0039±0.0039
	01-1/EG9	4.76±0.04*	37.79±1.02*	49.84±8.02	30.57±2.61	0.0052±0.063	1.28±0.73	−0.0092±0.0096
B	01-1	4.92±0.07	39.79±0.80	28.99±2.35	21.79±1.44	−0.015±0.16	0.97±0.24	−0.0099±0.015
	01-1/FM1	4.91±0.65	37.94±0.89	29.65±5.40	23.70±5.12	0.031±0.094	1.18±0.34	−0.0065±0.0082
C	712	4.86±0.02	42.30±1.00	15.86±1.52	11.37±2.05	0.036±0.10	1.60±0.82	−0.0036±0.0051
	712/EG9	4.62±0.46*	41.31±0.83	33.77±2.53*	21.17±1.65*	0.018±0.087	1.40±0.26	−0.0089±0.0098
D	712	4.98±0.04	39.38±0.19	25.55±3.65	17.90±2.23	0.081±0.012	1.43±0.12	−0.0013±0.0004
	712/FM1	4.98±0.04	40.34±0.98	19.79±2.29	13.58±1.32	0.077±0.013	1.46±0.23	−0.0014±0.0009

Data are means and standard deviation (±SD) from three trials each with triplicate assays.

*  Significant difference between control and trial within each experiment (P < 0.05).

Taste sensor analysis    None of the measurement parameters were significantly different between the control and test cheeses, although the parameter values of saltiness and sourness in the trial cheese of experiments A and C were relatively higher than in the control (Table 7). No difference in taste value was observed in FM1 addition cheese.

Table 7. Taste sensing analysis value of experimental cheeses after 180-day ripening.

		First-taste					After-taste
Experiment	Starter	Umami	Salty	Sour	Bitter	Astringent	CPA-umami	CPA-bitter	CPA-astringent
A	01-1	0.33±0.052	−0.13±0.40	−2.17±0.22	−1.63±0.058	1.61±0.32	0.033±0.13	−0.17±0.027	−0.05±0.010
	01-1/EG9	0.22±0.041	0.71±0.66	−0.98±0.43	−1.89±0.30	1.62±0.38	0.014±0.11	−0.25±0.025	−0.08±0.026
B	01-1	0.28±0.16	0.87±0.75	−1.82±0.43	−1.95±0.27	1.79±0.47	−0.11±0.14	−0.35±0.022	−0.18±0.028
	01-1/FM1	0.28±0.21	1.15±0.84	−1.71±0.41	−2.01±0.36	1.800.59	−0.20±0.11	−0.42±0.035	−0.27±0.024
C	712	0.36±0.16	1.30±0.85	−1.75±0.38	−2.29±0.37	1.76±0.57	−0.58±0.19	−0.49±0.034	−0.35±0.028
	712/EG9	0.31±0.072	1.62±0.97	−0.27±0.52	−2.61±0.41	1.53±0.57	−0.28±0.14	−0.51±0.033	−0.42±0.031
D	712	0.36±0.21	1.49±0.95	−1.59±0.42	−2.31±0.37	1.72±0.60	−0.43±0.093	−0.54±0.029	−0.49±0.032
	712/FM1	0.36±0.22	1.54±1.06	−1.71±0.56	−2.41±0.37	1.67±0.59	−0.31±0.12	−0.58±0.022	−0.55±0.042

Data are means and standard deviation (±SD) from three trials each with triplicate assays.

No difference in statistical analysis within each experiment (P < 0.05).

Discussion

For cheese making, LAB was added to pasteurized milk as a lactic fermentation starter. The starter bacteria form the dominant species of the cheese in the early ripening period, but NSLAB gradually increase during ripening (Settanni and Moschetti, 2010). It has been suggested that NSLAB are related to cheese ripening (Lynch et al., 1996; Solieri et al., 2012). However, the effects on TPA and a taste sensor analysis were not sufficiently investigated. In this study, we attempted to isolate NSLAB from ripened cheeses and evaluate the effects on cheese ripening, including TPA parameters and taste values. As a result, two strains, Lb. paracasei EG9 and Leu. mesenteroides subsp. mesenteroides FM1, were obtained, and the effects of temperature and salt on their growth were investigated. Strains EG9 and FM1 were capable of growing in 1.7% NaCl at 10°C, similar to the conditions of cheese ripening. In addition, at 10°C, EG9 showed better growth and a higher salt-tolerance than the type-strain JCM8130^T. These properties might be suitable for NSLAB growth in cheese.

The FAA production was found to be accelerated by the addition of EG9 without affecting taste values, but the addition of FM1 did not influence cheese ripening. Both cell envelope-associated proteinases and intracellular peptidases play a role in protein degradation during cheese ripening. Since intracellular peptidases are released by autolysis at cell death, cheese ripening is strongly affected by the quantity of bacteria, regardless of life and death. Generally, within three months of manufacture, the starter bacteria that are unable to survive in the selective conditions of ripening cheese (low water, pH, and temperature, coupled with a high salt environment) decline (Fitzsimons et al., 2001). Thus, the quantity of bacteria was measured with a PCR-based method in this study. The amount of EG9 was maintained until the later ripening period (Fig. 3a, 3c). The total FAA contents of EG9 cheeses were significantly higher than in the control (Fig. 4). The total FAA level of the control cheese of experiments A and C was increased to 65.1 and 137.3 mg g⁻¹ after 180-day ripening, respectively. This is at the same level as common semi-hard cheeses (Kabelová et al., 2009). Therefore, it was suggested that the liberation of FAAs was enhanced and cheese ripening was accelerated by EG9 addition. Aminopeptidase liberates N-terminal amino acid from peptides. The aminopeptidase of Lc. lactis, known as PepA, is known to be related to cheese ripening (Laan et al., 1998). The high aminopeptidase activity might also be found in Lb. paracasei EG9. The whole genome sequence of EG9 was analyzed and it was revealed that there were some known gene groups related to proteolysis (Data not shown). The mechanism of high level FAA production is expected to be clarified by expression analysis of these genes.

FAA production was enhanced by EG9 addition without affecting the taste value. Although taste sensor analysis is useful for detecting differences between samples for parameters detectable by sensors, organoleptic evaluation cannot be completely replaced. It is necessary to evaluate the effect of EG9 addition on the sensory properties of cheese by further investigations. The pH level and water content were lowered by EG9 addition, and the texture showed enhanced hardening. Cheese hardness is affected by the following factors: water content, low pH, and compression degree of the cheese curd (Juan et al., 2007). There was a concern that the hardness of trial cheeses in experiment C was not related to water content but to a lower pH. Moreover, the relationship with compression degree was not clear because it was not analyzed. In the trial cheese in experiment A, a lower pH would directly influence hardness. However, the water content of trial cheeses in experiment B was slightly lower than that of common Gouda cheese, which is 38–48%. TPA parameters and pH values did not differ from the control. The reason for this could not be determined by this study. There were no significant differences in the physical properties between control and trial cheeses in experiment D due to the lack of differences in pH and water content. It is necessary to improve the pH decrease and hardening of cheese with EG9 addition. In this study, NSLAB was inoculated at the same level as a lactic fermentation starter to clarify the effect of NSLAB additions. It could be improved by adjusting the NSLAB inoculum rate.

Conclusions

Lb. paracasei EG9 was isolated as a NSLAB from old ripened cheese. This study revealed that the addition of EG9 to the cheese-making process made it possible to accelerate FAA production without affecting the taste value.

Acknowledgments    The authors would like to thank Enago (www.enago.jp) for the English language review.

References

•  Antonsson,  M.,  Molin,  G., and  Ardö,  Y. (2003). Lactobacillus strains isolated from Danbo cheese and their function as adjunct cultures in a cheese model system. Int. J. Food Microbiol., 85, 159-169.
•  Caplice,  E. and  Fitzgerald,  G. F. (1999). Food fermentation: role of microorganisms in food production and preservation. Int. J. Food Microbiol., 50, 131-149.
•  Chikuni,  K.,  Oe,  M.,  Sasaki,  K.,  Shibata,  M.,  Nakajima,  I.,  Ojima,  K., and  Muroya,  S. (2010). Effects of muscle type on beef taste-traits assessed by an electric sensing system. Anim. Sci. J., 81, 600-605.
•  Ciprovica,  I. and  Mikelsone,  A. (2011). The influence of ripening temperature on diversity of non-starter lactic acid bacteria in semi-hard cheeses. Rom. Biotechnol. Lett., 16, 155-162.
•  Crow,  V. L.,  Coolbear,  T.,  Gopal,  P. K.,  Martley,  F. G.,  McKay,  L. L., and  Riepe,  H. (1995). The role of autolysis of lactic acid bacteria in the ripening of cheese. Int. Dairy J., 5, 855-875.
•  Desai,  A. R.,  Shah,  N. P., and  Powell,  I. B. (2006). Discrimination of dairy industry isolates of the Lactobacillus casei group. J. Dairy Sci., 89, 3345-3351.
•  Elizaquivel,  P.,  Chenoll,  E., and  Aznar,  R. (2008). A TaqMan-based real-time PCR assay for the specific detection and quantification of Leuconostoc mesenteroides in meat products. FEMS Microbiol. Lett., 278, 62-71.
•  Fitzsimons,  N. A.,  Cogan,  T. M.,  Condon,  S., and  Beresford,  T. (2001). Spatial and temporal distribution of non-starter lactic acid bacteria in Cheddar cheese. J. Appl. Microbiol., 90, 600-608.
•  Fox,  P. F.,  Wallace,  J. M.,  Morgan,  S.,  Lynch,  C. M.,  Niland,  E. J., and  Tobin,  J. (1996). Acceleration of cheese ripening. Antonie van Leeuwenhoek. 70, 271-297.
•  Grattepanche,  F.,  Lacroix,  C.,  Audet,  P., and  Lapointe,  G. (2005). Quantification by real-time PCR of Lactococcus lactis subsp. cremoris in milk fermented by a mixed culture. Appl. Microbiol. Biotech., 66, 414-421.
•  Grazier,  C. L.,  Bodyfelt,  F. W.,  McDaniel,  M. R., and  Torres,  J. A. (1991). Temperature effects on the development of Cheddar cheese flavor and aroma. J. Dairy Sci., 74, 3656-3663.
•  Grazier,  C. L.,  Simpson,  R.,  Roncagliolo,  S.,  Bodyfelt,  F. W., and  Torres,  J. A. (1993). Modelling of time-temperature effects on bacteria populations during cooling of Cheddar cheese blocks. J. Food Process Eng., 16, 173-190.
•  Hagi,  T.,  Sasaki,  K.,  Aso,  H., and  Nomura,  M. (2013). Adhesive properties of predominant bacteria in raw cow's milk to bovine mammary gland epithelial cells. Folia Microbiol., 58, 515-522.
• ISO/IDF. (2004). Cheese and processed cheese. Determination of the total solids content. Reference method. ISO 5534:2004 (IDF 4:2004), Brussels, Belgium: International Dairy Federation.
•  Juan,  B.,  Trujillo,  A. J.,  Guamis,  V.,  Buffa,  M., and  Ferragut,  V. (2007). Rheological, textual and sensory characteristics of high-pressure treated semi-hard ewes' milk cheese. Int. Dairy J., 17, 248-254.
•  Kabelová,  I.,  Dvořáková,  M.,  Čížková,  H.,  Dostálek,  P., and  Melzoch,  K. (2009). Determination of free amino acids in cheeses from the Czech market. Czech J. Food Sci., 27, 143-150.
•  Kolakowski,  P.,  Reps,  A., and  Babuchowski,  A. (1998). Characteristics of pressurized ripened cheeses. Pol. J. Food Nutr. Sci., 48, 473-484.
•  Konstance,  R. P. and  Holsinger,  V. H. (1992). Developments of rheological test methods for cheese. Food Tech., 1, 105-109.
•  Laan,  H.,  Tan,  S. E.,  Bruinenberg,  P.,  Limsowtin,  G., and  Broome,  M. (1998). Aminopeptidase activities of starter and non-starter lactic acid bacteria under simulated Cheddar cheese ripening conditions. Int. Dairy J., 8, 267-274.
•  Laloy,  E.,  Vuillemard,  J. C., and  Simard,  R. E. (1998). Characterization or liposomes and their effect on Cheddar cheese properties during ripening. Lait. 78, 401-412.
•  Law,  B. A. (1999). Cheese ripening and cheese flavour technology. In  Law  B. A (Ed.), Technology of cheesemaking. Sheffield, UK: Sheffield Academic Press. pp. 163-192
•  Law,  B. A. and  Haandrikman,  A.. (1997). Proteolytic enzymes of lactic acid bacteria. Int. Dairy J., 7, 1-11.
•  Law,  B. A. and  King,  J. C. (1985). The use of liposomes for the addition of enzymes to cheese. J. Dairy Res., 52, 183-188.
•  Lynch,  C. M.,  McSweeney,  P. L. H.,  Fox,  P. F.,  Cogan,  T. M., and  Drinan,  F. D. (1996). Manufacture of Cheddar Cheese with and without adjunct Lactobacilli under controlled microbiological conditions. Int. Dairy J., 6, 851-867.
•  Madkor,  S. A.,  Tong,  P. S., and  EI Soda,  M. (2000). Ripening of Cheddar cheese with added attenuated adjunct cultures of lactobacilli. J. Dairy Sci., 83, 1684.
•  Meijer,  W.,  Dobldaar,  C., and  Hugenholtz,  J. (1998). Thermoinducible lysis of Lactococcus lactis subsp. cremoris SKII: Implications for cheese ripening. Int. Dairy J., 8, 275-280.
•  Messens,  W.,  Van de Walle,  D.,  Arevalo,  J.,  Dewettinck,  K., and  Huyghebaert,  A. (2000). Rheological properties of high-pressure-treated Gouda cheese. Int. Dairy J., 10, 35-367.
•  Nomura,  M.,  Kimoto,  H.,  Someya,  Y.,  Furukawa,  S., and  Suzuki,  I. (1998). Production of γ-Aminobutyric acid by cheese starters during cheese ripening. J. Dairy Sci., 81, 1486-1491.
•  Raksakulthai,  R.,  Rosenberg,  M., and  Haard,  N. E. (2002). Accelerated Cheddar cheese ripening with an aminopeptidase fraction from squid hepatopancreas. J. Food Sci., 67, 923-929.
•  Settanni,  L. and  Moschetti,  G. (2010). Non-starter lactic acid bacteria used to improve cheese quality and provide health benefits. Food Microbiol., 27, 691-697.
•  Solieri,  L.,  Bianchi,  A., and  Giudici,  P. (2012). Inventory of non-starter lactic acid bacteria from ripened Parmigiano Reggiano cheese as assessed by a culture dependent multiphasic approach. Syst. Appl. Microbiol., 35, 270-277.
•  Swearingen,  P. A.,  O'Sullivan,  D. J., and  Warthesen,  J. J. (2001). Isolation, characterization, and influence of native, nonstarter lactic acid bacteria on Cheddar cheese quality. J. Dairy Sci., 84, 50-59.
•  Toko,  K. (1996). Taste sensor with global selectivity. Mater. Sci. Eng. C., 4, 69-82.
•  Walter,  J.,  Tannock,  G. W.,  Tilsala-Timisjarvi,  A.,  Rodtong,  S.,  Loach,  D. M.,  Munro,  K., and  Alatossava,  T. (2000). Detection and identification of gastrointestinal Lactobacillus species using denaturing gradient gel electrophoresis and species-specific PCR primers. Appl. Environ. Microbiol., 66, 297-303.