PCR and RAPD identification of L. plantarum strains isolated from ovine milk and cheese. Geographical distribution of strains

Abstract

Lactobacilli, and specifically Lactobacillus plantarum, are an important group of microorganisms in ovine cheeses, even though they are not ordinarily included in the starter cultures added. The present study effected counts of lactobacilli in Roncal Protected Designation of Origin (PDO) milk and cheese samples and isolated a total of 1026 strains. The strains were identified to species level by polymerase chain reaction (PCR) using L. plantarum-specific oligonucleotide primers, and the strains belonging to this species were then characterized by randomly amplified polymorphic DNA (RAPD). The percentage of L. plantarum present in the cheeses depended on the plant where the cheese was manufactured. Cluster analysis of the RAPD profiles obtained revealed seven main clusters. On comparing the strains, most of the strains present in the cheese were found not to have come from the raw milk.

Ewe's milk cheese, Polymerase chain reaction, Randomly amplified polymorphic DNA, Lactobacillus plantarum

1 Introduction

Roncal cheese is a ripened uncooked cheese made from raw ewe's milk in the Autonomous Community of Navarre in Spain and was the first cheese to be awarded an Appellation of Origin, or Protected Designation of Origin (PDO), in Spain. The dairies that manufacture this cheese collect ovine milk from the three main sheep-raising areas in Navarre.

The use of starter is optional, but ordinarily a freeze-dried industrial starter consisting of a culture of the mesophilic lactococci Lactococcus lactis subsp. lactis and L. lactis subsp. cremoris[1] is added.

The starter culture is of primary importance during actual cheesemaking, whereas the microorganisms that come from outside the starter, namely, the secondary microflora, play a major role over the course of ripening [2].

Mesophilic lactobacilli make up one of the most common groups of microorganisms in the secondary microflora present in cheeses. They are normally found in all types of cheese and are extremely important during ripening, when they reach high counts in such cheeses as Roncal, Fiore Sardo, Cheddar, Los Ibores, Comté, Dutch-type cheese, and Swiss cheese [3–12].

These microbes may enter from the source milk as well as adventitiously from the immediate surroundings during cheese manufacture [13–15].

Adventitious non-starter lactic acid bacteria (NSLAB) become the dominant bacterial population in the mature cheeses. It therefore appears vital to characterize and preserve the large numbers of mesophilic lactobacilli occurring as adventitious microflora, particularly in cheeses made from unpasteurized milk.

Recent research carried out on Roncal cheese shows that different species of Lactobacilli are present in high numbers (10⁸ colony-forming units (CFU) g⁻¹) even when not added to the cheese in the starter. Lactobacillus plantarum is one of the main species found. This species, one of the NSLAB, is a facultative heterofermenter and contributes to the final attributes of this cheese. L. plantarum is quantitatively the most important species of lactobacillus in many types of cheese, e.g., Cabrales, Mahón, and Los Ibores cheese [3,16,17].

Molecular methods have been employed in the identification and typing of different species of lactobacilli, including L. plantarum[18,19]. Molecular techniques offer a variety of advantages over other more conventional typing procedures [10,20,21], chief among them being their high discriminating power, their ability to type all strains, and their ability to produce data suitable for automatic statistical analysis [22].

The object of this study was to apply the polymerase chain reaction (PCR) and randomly amplified polymorphic DNA (RAPD) methods to identify and type L. plantarum strains in Roncal PDO milk and cheese to study the geographical distribution of the different strains according to the sample collection venue within the region of Navarre and to determine whether the strains present in the cheeses had their origin in the milk employed as raw material.

2 Materials and methods

2.1 Milk and cheese samples

Milk samples were collected at 22 sheep dairy farms located in three separate sheep-raising areas in Navarre, namely, the Atlantic coastal region, the Pyrenean region, and the northwestern region, in accordance with French standard NF V 04-150 [23]. Milk was collected on three separate occasions, each a week apart.

Six cheeses aged for 4 months were collected from four of the five cheesemaking plants operating under the Roncal PDO which used milk from the above-mentioned sheep-raising regions as raw material for cheesemaking. The milk used by the plants came from: for Dairy 1 (D1), the Atlantic coastal region; for Dairy 2 (D2), the Pyrenean region; for Dairy 3 (D3), the Atlantic coastal region, the northwestern region, and the Pyrenean region; and for Dairy 4 (D4), the northwestern region and the Pyrenean region.

2.2 Isolation of microorganisms and phenotypic characterization

Milk and cheese samples were plated on MRS agar (Difco, Sparks, USA) at pH 5.4 and incubated anaerobically (5% CO₂) at 31°C for 72 h to recover the mesophilic lactobacilli.

A representative number of colonies were collected randomly from each milk and cheese sample. The colonies isolated on the MRS agar were Gram-stained and examined for bacterial morphology and catalase activity.

2.3 PCR and RAPD identification

The PCR reaction specific to L. plantarum was carried out using the specific primers Lbpl1/Lbpl2 [24] and the semi-universal primers Lb1/Lb2 [25]. The RAPD reactions used primers OPA-3 [24] and P1 [26].

PCR and RAPD were performed in a volume of 50 µl containing 100 mM KCl, 20 mM Tris/HCl (pH 8), 0.1 mM ethylenediamine tetraacetic acid (EDTA), 1 mM dithiothreitol (DTT), 50% glycerol, 0.5% Tween-20, 0.5% Nonidet P-40, 1.5 mM MgCl₂, 0.25 mM of each deoxynucleoside triphosphate (dNTP), 1 unit Taq DNA polymerase (EcoTaq from Ecogen), and 0.1 µM primer (PCR) or 0.2 µM primer (RAPD). Colonies were picked with a sterile toothpick and transferred to PCR tubes.

PCR amplification was performed in a GeneAmp Thermal Cycler System 2.400 apparatus (Perkin-Elmer Applied Biosystems) programmed for 35 amplification cycles of 1 min at 94°C, 1 min at 54°C, and 1 min at 72°C, preceded by 5 min at 95°C and followed by a final extension of 5 min at 72°C.

The thermal cycle for the RAPD reaction using OPA-3 [24] was 45 cycles of 1 min at 94°C, 1 min at 36°C, and 1 min at 72°C, preceded by 5 min at 94°C and followed by a final extension of 7 min at 72°C. Using P1 [26] it was 50 cycles of 1 min at 94°C, 1 min at 42°C, and 1 min at 72°C, preceded by 6 min at 94°C and followed by a final extension of 7 min at 72°C.

Amplification products were separated on 1.5% agarose gel. A DNA molecular mass marker (Amplisize™ Molecular Ruler from Bio-Rad®) was used as standard. Gels were run in TEB electrophoresis buffer and stained in ethidium bromide solution (1 µg ml⁻¹). Type strains from the American Type Culture Collection (L. plantarum ATCC 8014) and the German Collection of Microorganisms and Cell Cultures (L. plantarum DSM 20174^T) were included in each reaction as positive controls.

The RAPD profiles were processed using the Gel Compar II® program, version 3.1 (Applied Maths, Kortrijk, Belgium).

2.4 Statistical analysis

Statistical analysis consisted of descriptive statistics, χ²-test, Student's t-test, and one-way analysis of variance and was performed using the SPSS program, version 10.1.

3 Results and discussion

3.1 Counts and PCR identification

3.1.1 Milk samples

Table 1 presents the Lactobacillus counts recorded by sheep dairy farm and region. Counts were between 10⁴ and 10⁶ CFU ml⁻¹ in the Atlantic coastal region and between 10⁴ and 10⁵ CFU ml⁻¹ in both the Pyrenean region and the northwestern region. These levels were similar to the counts reported earlier for milk used to make Roncal cheese [6,11] and other types of cheese, e.g., Idiazábal cheese [27].

1

Microbiological findings for the genus Lactobacillus (expressed as log CFU ml⁻¹ and log CFU g⁻¹) and percentage L. plantarum identified by the molecular PCR method in samples of ovine milk (n=3) and cheeses (n=6) ripened for 4 months from the Roncal PDO

Sample	Region	Location/plant	Lactobacillus mean±S.E.M.	%L. plantarum
Milk	Atlantic coastal region	1	5.41±0.38	31
		2	5.86±0.20	7
		3	6.03±0.06	38
		4	5.47±0.33	35
		5	4.15±0.38	75
		6	4.21±0.09	43
		7	4.40±0.32	29
		8	5.13±0.09	24
		9	5.00±0.23	20
		10	4.96±0.24	21
		11	6.39±0.06	41
		12	5.78±0.44	20
		13	6.14±0.08	14
		14	4.51±0.21	29
		15	4.82±0.37	62
		16	5.97±0.31	34
	total Atlantic region		5.26±0.10	33
	Pyrenean region	17	4.14±0.40	36
		18	4.71±0.20	nd
		19	3.68±0.24	20
	total Pyrenean region		4.18±0.19	19
	northwestern region	20	4.50±0.15	22
		21	3.81±0.24	25
		22	4.06±0.24	nd
	total northwestern region		4.13±0.14	16
	total all regions		4.96±0.09	29
Cheese		Dairy 1	7.65±0.13	4
		Dairy 3	7.39±0.08	9
		Dairy 4	7.34±0.09	6
		Dairy 5	7.11±0.15	21
		total	7.37±0.06	10

nd, not detected.

Significant differences were recorded between the farms (data not shown). By region the differences were also significant (P<0.05) among the farms, the values for the Atlantic coastal region always being higher than the values for the other two regions, which were similar to each other. This could be ascribable to the milder climate in the Atlantic coastal region than in the other two regions or perhaps to characteristics derived from the region's geographical features.

After the counts, PCR was applied to 306 Gram-positive, catalase-negative, bacillus or coccobacillus strains isolated from the milk samples (Fig. 1) to identify the strains that belonged to the species L. plantarum (Table 1).

1

PCR specific to L. plantarum. Lanes 5 and 10, molecular mass markers (2000 bp, 1500 bp, 1000 bp, 700 bp, 500 bp, 400 bp, 300 bp, 200 bp, 100 bp, 50 bp); lanes 1–3, 6–9, and 11, L. plantarum strains amplified using genus-specific primers (Lb1/Lb2) and species-specific primers (Lbpl1/Lbpl2); lanes 1–3, strains from milk samples; lanes 6–9 and 11, strains from cheese samples; lane 4, non-L. plantarum lactobacillus; lane 12, L. plantarum DSM 20174^T; lane 13, L. plantarum ATCC 8014; lane 14, negative control.

The χ²-test results (data not shown) did not reveal any relationship between the percentage of L. plantarum found and the origin of the milk samples by dairy or sheep-raising region. This finding disagreed with the observations published by Corroler and co-workers [28] in a similar study on the population of ‘wild’ lactococci in raw milk from the region of the Camembert PDO. They reported a geographical component in the distribution of the subspecies L. lactis subsp. lactis and subsp. cremoris, with a prevalence of cremoris (70%) in the Le Bessin region and a predominance of lactis (80%) in the Bocage region.

3.1.2 Cheese samples

Table 1 also gives the levels observed by cheesemaking plant. Counts were similar to those previously reported for Roncal cheese [6,11] and in such other cheeses as Cheddar, Herrgard, and Pecorino Sardo [29–31].

The values were indicative of uniform Lactobacillus counts among the different Roncal PDO cheesemaking plants considered in this study. This same finding has been reported earlier for both the Roncal PDO and the Idiazábal PDO [6].

On the whole, mean lactobacillus counts were 10⁴–10⁶ CFU ml⁻¹ in the milk and 10⁷ CFU g⁻¹ in the cheeses. The increase of one or more log units attests to the important role played by these microorganisms during ripening of this type of cheese, even though they were not included in the starter culture added. This same finding of increased numbers and significant involvement by these microorganisms in cheese ripening and, more specifically, in proteolysis has been reported by different workers in a variety of cheeses [7,8,32].

PCR identification of L. plantarum (Fig. 1) was carried out for the 720 Gram-positive, catalase-negative, bacillus and coccobacillus strains isolated from the Roncal cheeses in the same manner as for the strains isolated from the milk samples. The percentage share of this species by cheesemaking plant appears at the bottom of Table 1 and agreed with the findings reported for such other types of cheese as Cheddar cheese [33].

The χ²-test results (data not shown) revealed a linkage between cheese origin venue, i.e., cheesemaking plant, and the percentage of L. plantarum observed. In other words, even though the counts for the genus Lactobacillus were rather uniform (at around 10⁷ CFU g⁻¹) among the different cheesemaking plants, there appeared to be differences with respect to the percentage share of L. plantarum present at the individual plants.

The percentage share of L. plantarum in the counts was higher in the milk samples than in the cheese samples after 4 months of ripening.

Lactobacillus casei/Lactobacillus paracasei are generally thought to be able to thrive in cheese more readily than L. plantarum; nevertheless, if the latter species is present in the milk, it usually stays on in the cheese until advanced stages of ripening [17].

3.2 RAPD results

Primer OPA-3 (Fig. 2a,b) yielded several different patterns of bands. Pattern A was the most common among the strains from the milk samples (present in 78%) and pattern D the most common in the strains from the cheese samples (present in 39%).

2

RAPD performed on L. plantarum using primer OPA-3. Lane 8, molecular mass markers (2000 bp, 1500 bp, 1000 bp, 700 bp, 500 bp, 400 bp, 300 bp, 200 bp, 100 bp, 50 bp); lane 4, L. plantarum DSM 20174^T; lane 12, L. plantarum ATCC 8014. a: Milk samples: Lanes 1–3, 5–7, 9, 11, 13, and 14, primary band pattern A; lanes 10 and 12, band pattern B; lane 15, band pattern C. b: Cheese samples: Lanes 1–3, 6, 11, 13, and 15, band pattern D; lanes 5, 7, 9, 12, and 14, band pattern B; lane 10, band pattern C.

Primer P1 (Fig. 3a,b) also yielded a number of band patterns, the most common being pattern F present in 55% of the strains from the milk samples and in 20% of the strains from the cheese samples. This pattern matched the pattern for L. plantarum strain DSM 20174^T from the German collection.

3

RAPD performed on L. plantarum using primer P1. Lane 8, molecular mass markers (2000 bp, 1500 bp, 1000 bp, 700 bp, 500 bp, 400 bp, 300 bp, 200 bp, 100 bp, 50 bp); lane 5, L. plantarum DSM 20174^T; lane 12, L. plantarum ATCC 8014. a: Lanes 1 and 6, band pattern G; lanes 2, 5, 7, 11, and 15, band pattern F; lanes 9 and 13, band pattern H; lane 10, band pattern I; lanes 4, 12, and 14, band pattern J; lane 3, band pattern K (no bands). b: Cheese samples: Lanes 1, 6, 7, 10, and 12, band pattern J; lanes 3, 5, and 11, band pattern F; lanes 4 and 15, band pattern H; lanes 9 and 13, band pattern L; lane 14, band pattern K (no bands).

The Gel Compar® program was used to obtain a similarity dendrogram (Fig. 4) to compare the strains in the milk samples from the dairy farms and the strains in the cheese samples from the cheesemaking plants to which the dairies supply milk. The purpose was to try to ascertain whether the strains present in the cheeses came from the milk supplied by the different dairies.

4

Simplified dendrogram for the RAPD patterns of the L. plantarum strains in the milk and cheese obtained using the unweighted pair group method with arithmetic mean (UPGMA).

The RAPD results yielded seven primary clusters (Ia–VI) at a similarity level of 50–60%, along with a further 18 less important clusters (Table 2). Comparison showed that most of the strains present in the cheeses did not come from the source milk. Thus, cluster Ia comprised strains from the Atlantic coastal region and from Dairy 5, which did not obtain milk from that region. In contrast, cheesemaking Dairies 1 and 4, which did use milk from that region, contributed practically no strains to the cluster. The same was true for clusters Ib and III.

2

Number of isolates per cluster in the cluster analysis of all isolates performed using the unweighted pair group method with arithmetic mean (UPGMA)

RAPD group	Similarity (%)	Sheep-raising region			Dairy
		Atlantic	Pyrenean	Northwestern	Dairy 1	Dairy 2	Dairy 3	Dairy 4
Ia	64	16	1				3	14
Ib	66	30	4					3
II	63	1	1				1	5
III	58	7	1			1		3
IV	57		2			2		4
V	55	10	1	3	4			2
VI	51					6		5
1		1
2		1
3				1
4	52	1			1	3
5	81		1		1
6					1
7	61	2
8		1
9		1
10						1
11							1
12							1
13	60				1			2
14						1
15							1
16						1
17					1
18		1

These findings seem to suggest that the strains colonizing the cheeses mostly do not come from the source milk used as raw material for making the cheeses but rather come from other sources subsequent to milking, such as transportation systems, the cheesemaking plant environment, and the cheesemaking equipment. Another possibility is of course that these strains are in fact present in the milk at levels so low as to be undetectable and that they subsequently do extremely well in the cheeses made from the milk.