Isolation of a Hop-Sensitive Variant of Lactobacillus lindneri and Identification of Genetic Markers for Beer Spoilage Ability of Lactic Acid Bacteria

Abstract

We have isolated a hop-sensitive variant of the beer spoilage bacterium Lactobacillus lindneri DSM 20692. The variant lost a plasmid carrying two contiguous open reading frames (ORF s) designated horB_L and horC_L that encode a putative regulator and multidrug transporter presumably belonging to the resistance-nodulation-cell division superfamily. The loss of hop resistance ability occurred with the loss of resistance to other drugs, including ethidium bromide, novobiocin, and cetyltrimethylammonium bromide. PCR and Southern blot analysis using 51 beer spoilage strains of various species of lactic acid bacteria (LAB) revealed that 49 strains possessed homologs of horB and horC. No false-positive results have been observed for nonspoilage LAB or frequently encountered brewery isolates. These features are superior to those of horA and ORF 5, previously reported genetic markers for determining the beer spoilage ability of LAB. It was further shown that the combined use of horB/horC and horA is able to detect all 51 beer spoilage strains examined in this study. Furthermore sequence comparison of horB and horC homologs identified in four different beer spoilage species indicates these homologs are 96.6 to 99.5% identical, which is not typical of distinct species. The wide and exclusive distribution of horB and horC homologs among beer spoilage LAB and their sequence identities suggest that the hop resistance ability of beer spoilage LAB has been acquired through horizontal gene transfer. These insights provide a foundation for applying trans-species genetic markers to differentiating beer spoilage LAB including previously unencountered species.

Beer has been recognized as a beverage with high microbiological stability. Among the beer spoilers, several species of lactic acid bacteria (LAB) are reported to be responsible for approximately 70% of spoilage incidents caused by microorganisms (2, 3). For this reason, species-specific identification methods based on PCR have been widely evaluated for potential applications to microbiological quality control (6, 19, 41, 42, 44). Although species-specific PCR tests are rapid and reasonably accurate, there are two problems for applying this approach to the quality control of breweries.

One problem is that the species-specific method is unable to distinguish intraspecies differences between beer spoilage strains and nonspoilage strains (9, 26, 28, 35). Hop compounds added to confer bitter flavor are reported to exert an antibacterial effect by acting as proton ionophores and dissipate transmembrane pH gradient, which prevents gram-positive bacteria, including most LAB, from growing in beer (24, 25, 27, 28, 40). Hop resistance ability has been known as the distinguishing character of beer spoilage strains of LAB and nonspoilage strains typically exhibit hop resistance ability considerably weaker than that of beer spoilage strains belonging to the same species (1, 9, 28, 34, 35). The presence of nonspoilage strains within a beer spoilage species inevitably leads to false-positive results as long as the brewers rely on the species-specific approaches.

The second and probably more important problem is that the species-specific approach is incapable of detecting unencountered species of spoilage bacteria that occasionally emerge in the brewing industry. Lactobacillus paracollinoides was recently proposed as a novel beer spoilage species (10, 32, 39). The 12 strains of L. paracollinoides, so far isolated from our brewery environments, exhibited very strong beer spoilage ability that was comparable with that of the most potent beer spoilage species, Lactobacillus brevis and Lactobacillus lindneri (32). Other novel beer spoilage species have also been recently described (7, 14). The presence of unencountered beer spoilage species poses a threat to brewers since one spoilage incident significantly damages the corporate brand. These two problems led us to explore genetic markers that are able to determine the beer spoilage ability of LAB independent of the species status of detected bacteria.

Currently two genetic markers have been reported to determine the beer spoilage ability of LAB and designated trans-species genetic markers due to their discriminatory ability that transcends species status (21, 31, 33). The first trans-species genetic marker, horA, was found on 15.1-kb plasmid pRH45 identified in strong beer spoilage L. brevis strain ABBC45 (22, 23, 36). It was subsequently shown that HorA acts as an ATP-dependent multidrug transporter and confers hop resistance on LAB (20, 22, 43). Interestingly, PCR analysis based on the nucleotide sequence of horA demonstrated that horA is a genetic marker that transcends species status in differentiating the beer spoilage ability of a wide variety of lactobacilli (21). Nonetheless, L. brevis ABBC45^C, a variant that lost pRH45, still exhibited residual beer spoilage ability, indicating the presence of a horA-independent hop resistance mechanism (35). It was further shown that a small number of L. brevis and L. paracollinoides strains are able to grow in beer despite the absence of horA homologs, indicating multiple genetic markers are required for the concept of trans-species genetic markers to be practically applied in breweries (38).

The second trans-species genetic marker, ORF 5, was found in the excised DNA region of 23.4-kb plasmid pRH45II identified in L. brevis ABBC45^C (33). The partial loss of pRH45II, coupled with the loss of pRH45, resulted in complete loss of the beer spoilage ability of L. brevis ABBC45 (33). The excised DNA region of pRH45II contained the ORF 1 to 7 region that was implicated in horA-independent hop resistance mechanism (30, 33). Biochemical data indicate that a proton motive force-dependent multidrug transporter mediates horA-independent hop resistance (35). On the basis of 11 to 12 transmenbrane domains of the deduced protein, ORF 5 was tentatively selected as a candidate for the putative proton motive force-dependent multidrug transporter and demonstrated to differentiate the beer spoilage ability of L. brevis and L. paracollinoides strains (31, 33). But this genetic marker has never been extensively evaluated for the differentiation of the beer spoilage ability of other species of LAB and the combination of the trans-species genetic markers has not thus far been proposed for practical applications to microbiological control of breweries.

L. lindneri has been reported to be the second most frequent beer spoilage species of LAB (2, 3). So far nonspoilage strains of L. lindneri have not been found (17, 29) and some brewing microbiologists argue that, unlike other beer spoilage species of LAB, L. lindneri is an innate beer spoiler. In this study, we isolated a hop-sensitive variant of L. lindneri DSM 20692 and carried out genetic characterization of this strain, leading to new insights into the character of the ORF 1 to 7 region. The results of this work may allow the species-independent detection of hop-resistant LAB of importance to the brewing industry, including those that have not yet been characterized.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Strains of LAB were grown anaerobically at 25°C in MRS broth (Merck, Darmstadt, Germany, pH adjusted to 5.5 with HCl). Other strains isolated from brewery environments were cultured anaerobically at 25°C in TGC medium (Nissui Pharmaceutical, Tokyo, Japan). Anaerobic conditions were generated by AnaeroPack (Mitshubishi Gas Chemicals, Tokyo, Japan). Cells were stored in MRS broth containing 20% glycerol at −80°C.

Hop compounds.

Isomerized hop compounds were obtained from Simon H. Steiner Hopfen GmbH (Mainburg, Germany) as a concentrated hop extract. The iso-α-acid contents were determined, using high-performance liquid chromatography (8). The concentration of hop compounds in the medium was expressed as the total iso-α-acid contents.

Isolation of hop-sensitive variant.

We attempted to isolate hop-sensitive variants of L. lindneri DSM 20690^T and DSM 20692 by repeatedly subculturing the strains by inoculating 10⁵ cells in 10 ml fresh MRS broth (pH 5.5) every 3 to 4 days. After the 10th and 15th subcultures at 30°C, single colonies were examined for the presence or absence of hop resistance. An isolate that failed to grow in MRS broth (pH 5.5) containing 100 μM hop compounds was defined as a hop-sensitive variant.

DNA extraction and plasmid profile analysis.

The DNA was extracted, as described previously (21). The extracted DNA was subjected to electrophoresis on 0.5% agarose gels (Seakem Gold agarose, BMA, Rockland, ME) in TAE buffer (0.04 M Tris-acetate, 0.01 M EDTA [pH 8.0]) and stained with ethidium bromide. Lambda phage DNA digested with HindIII was used as a molecular weight marker.

Hop adaptation of L. lindneri.

L. lindneri DSM 20692 and DSM 20692^NB were repeatedly subcultured in 10 ml MRS broth (pH 5.5) containing hop compounds at progressively higher concentrations in increments of 50 μM. The subcultures were performed until the growth no longer occurred within 4 days. The maximally hop-adapted strain is defined as a strain that was adapted to grow at the maximum concentration of hop compounds. DSM 20692^NB, the hop-sensitive variant, did not develop significant hop resistance ability and failed to grow in MRS broth (pH 5.5) containing 100 μM hop compounds even after hop adaptation. Therefore a strain that grew with 50 μM hop compounds was used as a maximally hop-adapted strain for DSM 20692^NB.

Assay for drug resistance.

Cells of hop-adapted L. lindneri DSM 20692 and DSM 20692^NB were washed with and resuspended in sterile deionized water at a concentration of 10⁸ cells/ml. Five microliters of the cell suspension was spotted on MRS agar (pH 5.5), containing a range of concentrations of the following drugs: isomerized hop compounds (0 to 500 μM), ethidium bromide (0 to 20 μg/ml, Sigma, St. Louis, MO), daunomycin (0 to 40 μg/ml, Wako, Osaka, Japan), nisin (0 to 0.8 μg/ml, Sigma), rhodamine 6G (0 to 20 μg/ml, Sigma), cetyltrimethylammonium bromide (CTAB) (0 to 4 μg/ml, Wako), novobiocin (0 to 16 μg/ml, Sigma), chloramphenicol (0 to 16 μg/ml, Wako), and penicillin G potassium (0 to 0.4 μg/ml, Wako). After 4 days of incubation at 25°C, drug resistance was determined from the extent of growth on each agar plate. The maximum concentration that allowed growth was defined as drug resistance in this study.

Southern blot analysis.

To obtain probes for Southern blot analysis, PCR was conducted with the primer sets designed on the basis of the nucleotide sequences of horA and each ORF found in the ORF 1 to 7 region using the DNA extracted from L. brevis ABBC45 as templates. The primers are listed in Table 1. Taq DNA polymerase and reaction buffer were supplied as a kit (TaKaRa EX Taq, Takara Bio, Shiga, Japan). The PCR mixtures were placed in a thermal cycler (GeneAmp PCR System 9700, Applied Biosystems, Foster City, CA). The PCR conditions for amplification of horA and each ORF of the ORF 3 to 7 region were described previously (21, 31). The cycling profile for the ORF 1 and ORF 2 amplifications consisted of an initial heating at 94°C for 2.5 min, followed by amplification of 30 cycles for denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 1 min. The final extension was set at 72°C for 3 min.

TABLE 1.

Primer sequences to generate specific probes for Southern blot analysis^a

ORF	Specific primers	Nucleotide sequence	Homologous protein (accession no.)
horA	LbHC-1	5′-ATCCGGCGGTGGCAAATCA-3′	L. brevis HorA (BAA21552.1)
	LbHC-2	5′-AATCGCCAATCGTTGGCG-3′
1 (horB)	AORF1F	5′-ATGTTGTATTTGACTTGTATCAGG-3′	No homologyb
	AORF1R	5′-AACAGGGATTTAATAATGTATCGG-3′
2 (horC)	AORF2F	5′-GTCAACGAAGACAAAGGAGCTCTC-3′	No homology
	AORF2R	5′-GGGCGAACCGTGAACAAATAG-3′
3	AORF3F	5′-TAACCGAACACTGCGAACTGCTGA-3′	Leuconostoc lactis putative transposase (CAA58055)
	AORF3R	5′-GATTAGTTTATCATCTCAAGAACG-3′
4	AORF4F	5′-AACGATTGTTGTTCCTGC-3′	L. plantarum glycosyltransferase (NP_785047)
	AORF4R	5′-ATCGTTTGCACCGTGTAG-3′
5	AORF5F	5′-CATTGGCCTTGGCTTACT-3′	Oenococcus oeni hypothetical protein (ZP_00069477)
	AORF5R	5′-AGGTTTTGAACGGTAATC-3′
6	AORF6F	5′-CCAATCACTACCCTTTAC-3′	L. plantarum teichoic acid glycosylation protein (NP_786130)
	AORF6R	5′-AATTAGTTTGCTGAAGCC-3′
7	AORF7F	5′-CTGCGTCGACCACGTTCATCAC-3′	Lactococcus lactis nicking enzyme TraA (NP_047290)
	AORF7R	5′-GAGTTTTAGTAATATTAGTGCTGG-3′

^aPrimer sequences were deisgned based on the nucleotide sequences of horA and the ORF 1 to 7 region identified in L. brevis ABBC45^C (21, 33).

^bORF 1 exhibited identities with some of the AcrR family regulators but the level of identity was slightly less than the threshold value (30%) set for the previous study (33).

Labeling of the probes and Southern blot analysis were performed in accordance with the manufacturer's protocol (DIG High Prime Labeling and Detection Starter Kit I, Roche Diagnostics, Mannheim, Germany). The whole genomic DNA and the EcoRI digests were separated on 0.5% agarose gels as described earlier and subjected to Southern blot analysis at a hybridization temperature of 42°C and a washing temperature of 65°C in 0.5× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) with 0.1% sodium dodecyl sulfate (SDS). Southern blot analysis was also performed under less stringent conditions (hybridization temperature: 35°C, washing temperature: 42°C in 0.5× SSC with 0.1% SDS).

Determination for presence of genetic markers.

A total of 130 strains of LAB and other bacteria were examined for the presence of the homologs of ORF 1 and ORF 2, hereafter renamed horB and horC, respectively. To avoid confusion, horB and horC homologs identified in L. lindneri DSM 20692 are designated horB_L and horC_L, respectively, in this study. The previously reported genetic markers, ORF 5 and horA (21, 33), were also investigated in an identical fashion. The nucleic acids were extracted from bacterial strains as described previously (21). The primer sets listed in Table 1 were used to amplify horB, horC and ORF 5 homologs. PCR was performed in the same manner as described in the probe preparations for Southern blot analysis. The specific primers and the procedures for the amplification of horA were previously described by Sami et al. (21). The PCR products were separated on a 2% agarose gel by electrophoresis.

For positive strains, the Southern blot analysis was conducted with the probes specific to horB, horC, ORF 5, and horA to determine whether the PCR products are the homologs of the genetic markers. The experimental conditions for Southern blot analysis were described earlier.

Evaluation of beer spoilage ability.

To prepare beers with standard microbiological stability of the Japanese pilsner type, degassed commercial lager beers containing 55 μM hop compounds were adjusted to pH 4.2 with 5 N NaOH. Ten ml of beers was dispensed in 15-ml sterile polypropylene tubes and inoculated with approximately 3 × 10³ cells/ml of LAB strains and environmental isolates. The inoculated beers were incubated anaerobically at 25°C and examined regularly for visible growth for up to 60 days.

DNA sequencing of horB_L and horC_L identified in L. lindneri DSM 20692.

The DNA region containing horB_L and horC_L was amplified by PCR using the forward primer (5′-ATCCCTCTCTTCGATTGATGGTTG-3′) and the reverse primer (5′-TACCTCAAGAACATTCTCTGCCTC-3′). These primers were designed based on the flanking DNA regions of horB and horC found in L. brevis ABBC45^C (33). The cycling profile consisted of an initial denaturation at 94°C for 2.5 min, followed by amplification of 30 cycles for denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 4 min. The final extension was set at 72°C for 8 min. The PCR product was sequenced by primer walking method, using RISA-384 (Shimadzu, Kyoto, Japan).

Analysis of sequenced DNA region.

The ORF analysis and G+C content analysis of each ORF were performed, using ORF Finder (National Center for Biotechnology Information, Bethesda, MD), on the basis of the nucleotide sequence. Identification of homologous proteins was based on BLASTP analysis, the program of which was provided by the National Center for Biotechnology Information. Hydropathy analysis was conducted using Genetyx Ver.6 (Genetyx Corporation, Tokyo, Japan). Nucleotide sequence comparisons of the horB and horC regions were carried out, using DNASIS Pro software package (Hitachi Software Engineering, Tokyo, Japan).

RESULTS

Isolation and biochemical characterization of a hop-sensitive variant.

We attempted to isolate hop-sensitive variants from L. lindneri DSM 20690^T and DSM 20692. As a result, we were unable to isolate a hop-sensitive variant from DSM 20690^T after the 10th and 15th passages at 30°C in MRS broth (pH 5.5). In contrast, a hop-sensitive variant was successfully isolated from L. lindneri DSM 20692 by subculturing this strain after the 10th passage and designated DSM 20692^NB. As shown in Table 2, DSM 20692^NB exhibited considerably diminished hop resistance ability compared with the wild-type strain DSM 20692. The hop resistance ability of this variant was no longer inducible even after repeated subcultures with 50 μM hop compounds. The maximally hop-adapted DSM 20692^NB was also unable to grow in beer even after 90 days of inoculation into beer, while the wild-type strain grew within 10 days. This result indicates that DSM 20692^NB completely lost beer spoilage ability.

TABLE 2.

Resistance of DSM 20692 and DSM 20692^NB to various drugs

Drug	Resistance (μg/ml)a
	DSM 20692	DSM 20692NB
Hop compounds	400	50
Ethidium bromide	5	1.25
Novobiocin	4	1
Cetyltrimethylammonium bromide	1	0.5
Daunomycin	20	20
Rhodamine 6G	10	10
Chloramphenicol	2	2
Nisin	0.2	0.2
Penicillin G potassium	0.1	0.1

^aDrug resistance was determined based on confirmation of cell growth on MRS agar (pH5.5) containing the drugs. Drug resistance is expressed in μg/ml except for hop compounds, which are expressed as total iso-α-acid concentrations in μM.

The resistance of L. lindneri DSM 20692 and DSM 20692^NB to various drugs was also investigated. As the results showed, resistance to ethidium bromide, novobiocin, and CTAB was found to be reduced, concomitant with the decrease in hop resistance ability of DSM 20692 (Table 2). In contrast, resistance to the other drugs remained unchanged.

Plasmid profile comparison and Southern blot analysis.

The plasmid profiles of L. lindneri DSM 20692 and DSM 20692^NB were compared. As a result, two plasmids appear to be missing (Fig. 1a). The Southern blot analysis was performed with probes specific to each ORF identified in the ORF 1 to 7 region, a DNA region previously implicated in the hop resistance ability of L. brevis ABBC45^C (33). Each probe specific to ORFs 1, 2 and 3 hybridized with the larger plasmid found in DSM 20692, indicating this plasmid contains homologs of ORFs 1, 2 and 3. In contrast, probes specific to ORFs 1, 2 and 3 failed to hybridize with the whole genomic DNA or EcoRI-digested DNA isolated from the hop-sensitive variant DSM 20692^NB, indicating the DNA region containing these three ORFs was lost in the variant. Figure 1b shows the result using the probe specific to ORF 2.

FIG. 1.

Southern blot analysis with probes specific to ORFs 1, 2, and 3. (a) Genomic DNA extracted from L. lindneri DSM 20692 (lane 1) and L. lindneri DSM 20692^NB (lane 2) was subjected to agarose gel electrophoresis. EcoRI-digested genomic DNA obtained from DSM 20692 (lane 3) and DSM 20692^NB (lane 4) was also separated on agarose gels. Lane M contains molecular size standards (λ DNA digested with HindIII). (b) Southern blot analysis was performed using probes specific to ORFs 1, 2, and 3. This figure shows an example using the ORF 2-specific probe. The order of the lanes is identical to that in a. The unnicked form of the plasmid and the nicked (open circular) form of the plasmid are indicated by arrows.

Southern blot analysis was also conducted using probes specific to ORFs 4, 5, 6, and 7. As the results showed, no hybridization was observed for L. lindneri DSM 20692 and 20692^NB. Figure 2 shows the result with the probe specific to ORF 5. These results suggest the ORF 4 to 7 region is deficient in L. lindneri DSM 20692. The Southern blot analysis was also conducted with the probe specific to ORF 5 under less stringent hybridization and washing conditions but the presence of ORF 5 was not detectable in L. lindneri DSM 20692, suggesting this strain lacks an ORF 5 homolog (data not shown).

FIG. 2.

Southern blot analysis with probes specific to ORFs 4, 5, 6, and 7. (a) Genomic DNA, extracted from L. lindneri DSM 20692 (lane 1) and L. lindneri DSM 20692^NB (lane 2), was subjected to agarose gel electrophoresis. EcoRI-digested genomic DNA obtained from DSM 20692 (lane 3) and DSM 20692^NB (lane 4) was also separated on agarose gel. EcoRI-digested genomic DNA obtained from L. brevis ABBC45^C was used as a positive control (lane 5). Lane M contains molecular size standards (λ DNA digested with HindIII). (b) Southern blot analysis was performed using probes specific to ORFs 4, 5, 6, and 7. This figure shows an example using the ORF 5-specific probe. The order of the lanes is identical to that in a.

Correlation between the presence of genetic markers and beer spoilage ability of LAB.

ORF 1 and ORF 2, hereafter renamed horB and horC, respectively, were evaluated as genetic markers for differentiating the beer spoilage ability of LAB and other bacteria, and compared with the previously reported trans-species genetic markers ORF 5 and horA (21, 33). A total of 130 strains were examined using PCR with primers specific to each genetic marker and the amplified products were subjected to Southern blot analysis with probes specific to the respective genetic markers. As the results showed, 49 strains out of the 51 beer spoilage LAB strains were found to possess homologs of horB and horC (Table 3). No false-positive reactions were observed for nonspoilage LAB or frequently encountered brewery isolates examined in this study.

TABLE 3.

Correlation between beer spoilage ability and the presence of genetic markers^a

Strain no.b	Species	horB	horC	ORF5	horA	Beer spoilage ability		Strain no.	Species	horB	horC	ORF5	horA	Beer spoilage ability
ABBC3	L. brevis	+	+	+	+	+
ABBC34	L. brevis	+	+	+	+	+
ABBC37	L. brevis	+	+	+	+	+
ABBC42	L. brevis	+	+	+	+	+
ABBC43	L. brevis	+	+	+	+	+
ABBC44	L. brevis	−	−	+	+	+
ABBC45	L. brevis	+	+	+	+	+
ABBC45C	L. brevis	+	+	+	−	+
ABBC46	L. brevis	+	+	+	+	+
ABBC56	L. brevis	+	+	+	+	+
ABBC64	L. brevis	+	+	+	+	+
ABBC69	L. brevis	+	+	+	+	+
ABBC70	L. brevis	+	+	+	+	+
ABBC76	L. brevis	+	+	+	+	+
ABBC77	L. brevis	+	+	+	+	+
ABBC78	L. brevis	+	+	+	+	+
ABBC79	L. brevis	+	+	+	+	+
ABBC84	L. brevis	+	+	+	+	+
ABBC85	L. brevis	+	+	+	+	+
ABBC86	L. brevis	+	+	+	+	+
ABBC99	L. brevis	+	+	+	+	+
ABBC100	L. brevis	+	+	+	+	+
ABBC104	L. brevis	+	+	+	+	+
ABBC400	L. brevis	+	+	+	+	+
ABBC402	L. brevis	+	+	+	+	+
ABBC403	L. brevis	+	+	+	+	+
ABBC404	L. brevis	+	+	+	+	+
ABBC405	L. brevis	+	+	+	+	+
ABBC407	L. brevis	+	+	+	+	+
ABBC408	L. brevis	+	+	+	+	+
ABBC4	L. brevis	−	−	−	+	−
ABBC12	L. brevis	−	−	−	−	−
ABBC36	L. brevis	−	−	−	+	−
ABBC65	L. brevis	−	−	−	+	−
ABBC67	L. brevis	−	−	−	+	−
ABBC406	L. brevis	−	−	−	+	−
JCM 1059T	L. brevis	−	−	−	−	−
DSM 1267	L. brevis	−	−	−	−	−
DSM 2647	L. brevis	−	−	−	−	−
DSM 20556	L. brevis	−	−	−	−	−
ABBC44NB	L. brevis	−	−	−	−	−
ABBC45CC	L. brevis	−	−	−	−	−
ABBC46NB	L. brevis	−	−	−	−	−
ABBC64NB	L. brevis	−	−	−	−	−
ABBC104NB	L. brevis	−	−	−	−	−
ABBC400NB	L. brevis	−	−	−	−	−
DSM 15502T	L. paracollinoides	+	+	+	+	+
LA3	L. paracollinoides	+	+	+	+	+
LA4	L. paracollinoides	+	+	+	+	+
LA7	L. paracollinoides	+	+	+	−	+
LA8	L. paracollinoides	+	+	+	−	+
LA9	L. paracollinoides	+	+	+	+	+
LA10	L. paracollinoides	+	+	+	+	+
LA11	L. paracollinoides	+	+	+	+	+
LA12	L. paracollinoides	+	+	+	+	+
LA13	L. paracollinoides	+	+	+	+	+
LA14	L. paracollinoides	+	+	+	+	+
LA15	L. paracollinoides	+	+	+	+	+
ATCC 8291	L. paracollinoides	−	−	−	−	−
DSM 15502NB	L. paracollinoides	−	−	−	−	−
LA9NB	L. paracollinoides	−	−	−	−	−
DSM 20690T	L. lindneri	−	−	−	+	+
DSM 20692	L. lindneri	+	+	−	+	+
HC92	L. lindneri	+	+	−	+	+
HC95	L. lindneri	+	+	−	+	+
HC98	L. lindneri	+	+	−	+	+
LA16	L. lindneri	+	+	−	+	+
DSM 20692NB	L. lindneri	−	−	−	+	−
ABBC478	P. damnosus	+	+	+	+	+
ABBC500	P. damnosus	+	+	+	+	+
VTT E-76065	P. damnosus	+	+	+	+	+
ABBC39	L. casei	−	−	+	−	−
ABBC52	L. casei	−	−	−	−	−
ABBC72	L. casei	−	−	−	−	−
ABBC73	L. casei	−	−	−	−	−
ABBC96	L. casei	−	−	+	+	−
ABBC204	L. casei	−	−	−	−	−
ABBC206	L. casei	−	−	−	−	−
ABBC222	L. casei	−	−	−	−	−
ABBC223	L. casei	−	−	−	−	−
ABBC224	L. casei	−	−	−	−	−
ABBC226	L. casei	−	−	−	−	−
ABBC279	L. casei	−	−	−	−	−
ABBC376	L. casei	−	−	−	−	−
ABBC410	L. casei	−	−	+	−	−
ABBC413	L. casei	−	−	+	−	−
ABBC426	L. casei	−	−	+	−	−
ABBC428	L. casei	−	−	+	−	−
ABBC430	L. casei	−	−	+	−	−
ABBC38	L. plantarum	−	−	−	−	−
ABBC55	L. plantarum	−	−	−	−	−
ABBC58	L. plantarum	−	−	+	−	−
ABBC80	L. plantarum	−	−	+	−	−
ABBC87	L. plantarum	−	−	+	+	−
ABBC97	L. plantarum	−	−	+	−	−
ABBC203	L. plantarum	−	−	−	−	−
ABBC213	L. plantarum	−	−	−	−	−
ABBC235	L. plantarum	−	−	−	−	−
ABBC377	L. plantarum	−	−	−	−	−
ABBC94	L. coryniformis	−	−	−	−	−
ABBC95	L. coryniformis	−	−	−	−	−
ABBC101	L. coryniformis	−	−	−	−	−
ABBC102	L. coryniformis	−	−	−	−	−
ABBC251	L. coryniformis	−	−	−	−	−
ABBC252	L. coryniformis	−	−	−	−	−
JCM 1123T	L. collinoides	−	−	−	−	−
ATCC 27610	L. collinoides	−	−	−	−	−
ATCC 27611	L. collinoides	−	−	−	−	−
ABBC219	L. buchneri	−	−	−	+	−
ABBC222	L. paracasei	−	−	−	−	−
ABBC228	L. fermentum	−	−	−	−	−
ABBC257	L. fructivorans	−	−	−	−	−
ABBC275	L. rhamnosus	−	−	−	−	−
ABBC281	L. delbrueckii	−	−	−	−	−
JCM 8573T	L. parakefiri	−	−	−	−	−
DSM 5707	L. parabuchneri	−	−	−	−	−
HC311	Lactococcus lactis	−	−	−	−	−
HC367	Serratia marcescens	−	−	−	−	−
HC417	Citrobacter freundii	−	−	−	−	−
HC432	Enterobacter cloaceae	−	−	−	−	−
HC437	Staphylococcus warneri	−	−	−	−	−
HC440	Propionibacterium acnes	−	−	−	−	−
HC442	Bacillus thuringiensis	−	−	−	−	−
HC453	Pantoea agglomerans	−	−	−	−	−
HC459	Paenibacillus amylolyticus	−	−	−	−	−
HC466	Paenibacillus jamilae	−	−	−	−	−
HC472	Clostridium beijerinckii	−	−	−	−	−
HC475	Staphylococcus epidermidis	−	−	−	−	−
HC523	Sporolactobacillus racemicus	−	−	−	−	−
HC534	Klebsiella oxytoca	−	−	−	−	−

^aPrimer pairs specific to horB, horC, ORF5, and horA were used. All positive amplicons were subjected to Southern blot analysis to determine whether the PCR products were homologs of the genetic markers. All PCR results were found to be consistent with those of Southern blot analysis. The superscripts NB and CC indicate hop-sensitve variants obtained from beer spoilage wild-type strains with the same strain number (31, 33, 34).

^bABBC, HC, and LA: our culture collections, consisting principally of brewery isolates; ATCC, American Type Culture Collection; DSM, Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH; JCM, Japan Collection of Microorganisms; VTT, culture collection obtained from the Technical Research Center of Finland.

As for two false-negative beer spoilage strains, L. brevis ABBC44 and L. lindneri DSM 20690^T, Southern blot analysis was performed with genomic DNA and their EcoRI digests extracted from ABBC44 and DSM 20690^T. As a consequence, the probes specific to horB and horC were unable to detect the presence of horB and horC homologs, even under less stringent hybridization and washing conditions, indicating these strains are deficient in homologs of horB and horC (data not shown). Interestingly, hop-sensitive variants obtained from L. brevis and L. paracollinoides strains lost reactivity with primers specific to horB and horC. Southern blot analysis with horC-specific probes further confirmed the loss of horC homologs in these variants, as revealed by the lack of hybridization with EcoRI-digested genomic DNA extracted from the variants (data not shown), suggesting these DNA regions are unstable.

For comparative purpose, ORF 5 and horA were evaluated as genetic markers for differentiating the beer spoilage ability of the bacteria detected. As the results showed, ORF 5 homologs were detectable in 45 out of 51 beer spoilage strains of LAB (Table 3). The relatively high rates of false positive reactions were observed with nonspoilage strains of L. casei and L. plantarum. The horA homologs were detectable in 48 out of 51 beer spoilage strains of LAB. The false-positive results were observed with nine nonspoilage strains. The comparison of these genetic markers revealed that horB and horC are superior to ORF 5 and horA in differentiating the beer spoilage ability of LAB examined in this study. It was also found that by assaying for both horB/horC and horA, false negative results for beer spoilage activity could be eliminated (Table 3).

DNA sequencing of horB_L and horC_L identified in L. lindneri DSM 20692.

The presence of horB and horC homologs in L. lindneri DSM 20692 prompted us to perform sequence analysis of this DNA region. On the basis of the nucleotide sequence previously reported for L. brevis ABBC45^C, we designed primers to amplify this region and obtained the nucleotide sequence of the corresponding region of L. lindneri DSM 20692. BLASTP analysis indicates that the deduced product of the horB_L exhibited homology with the regulators belonging to the AcrR family (Table 4). The deduced product of horB identified in L. brevis ABBC45^C was also found to display a similar level of identity with AcrR family regulators (data not shown). This insight suggests that horB and its homologs potentially modulate the transcription of a gene encoding a multidrug transporter that belongs to the resistance-nodulation-cell division (RND) superfamily. As the results of hydropathy analysis showed, the deduced protein product of horC_L was found to possess six putative transmembrane domains with a large hydrophilic domain between the first transmembrane domain and the second transmembrane domain (Fig. 3), which is characteristic of multidrug transporters belonging to the RND superfamily. The deduced protein product of horC showed an essentially identical hydropathy profile (data not shown).

TABLE 4.

ORF analysis of the DNA region containing horB_L and horC_L^a

ORF	Strand	Range (bp)	Size (aa)	Homologous protein	Identityb	Accession no.
horBL	−	56-628	190	Transcriptional regulators (AcrR family)	30% (53/175)	NP_346843.1
horCL	+	724-2028	434	No homology	NDc

^aThe nucleotide sequence of this DNA region has been deposited in DDBJ under accession number AB191156.

^bIdentity is expressed as the percentage of the number of matched amino acids in the aligned amino acid sequences, as indicated in parentheses.

^cND, not determined.

FIG. 3.

Hydropathy profile of the deduced protein product of horC_L determined by the algorithm of Kyte and Doolittle. The bars indicate putative transmembrane domains.

We have also compared nucleotide sequence identities among the horB and horC homologs found in four distinct beer spoilage species. As the results showed, horB_L found in L. lindneri DSM 20692 exhibited 99.3 to 99.5% identity with the corresponding ORFs identified in L. brevis ABBC45^C, L. paracollinoides DSM 15502^T and Pediococcus damnosus ABBC478 (Table 5). The horC_L was also found to be 96.6 to 99.4% identical with horC and the horC homologs of the other strains.

TABLE 5.

Nucleotide sequence comparison of horB and horC homologs identified in four distinct species of beer spoilage lactic acid bacteria^a

ORF	L. lindneri DSM 20692		L. brevis ABBC45C		L. paracollinoides DSM 15502T		P. damnosus ABBC478
	Size (bp)	Identity (%)b	Size (bp)	Identity (%)	Size (bp)	Identity (%)	Size (bp)	Identity (%)
horB	573	NDc	573	99.3 (569/573)	573	99.3 (569/573)	573	99.5 (570/573)
horC	1,305	ND	1,287	96.7 (1,244/1,287)	1,263	99.4 (1,252/1,260)	1,335	96.6 (1,260/1,305)

^aThe sequences of DSM 20692, ABBC45^C, DSM15502^T, and ABBC478 have been deposited in DDBJ under accession numbers AB191156, AB118106, AB159716, and AB159715, respectively.

^bThe nucleotide sequence identities with horB_L and horC_L are expressed as the percentage of the number of matched nucleotides in the aligned sequences shown in parentheses.

^cND, not determined.

G+C content analysis of ORFs identified in the ORF 1 to 7 regions was carried out with beer spoilage strains belonging to distinct species. As the results showed, the G+C contents of horB homologs and horC homologs were found to be approximately 30 mol% and 35 mol%, respectively, which is rather low for L. brevis and L. paracollinoides (Table 6). In contrast, the G+C contents of ORF 4, ORF 5, and ORF 6 homologs ranged between 53 mol% and 59 mol%, which appear to be high for all three beer spoilage species which possess this region. Notably, the differences between the horB-horC regions and ORF 4 to 6 regions exceeded 20 mol%, suggesting these two DNA regions have distinct origins.

TABLE 6.

G+C content analysis of ORFs 1 to 7 found in four beer spoilage LAB strains^a

Strain	G+C content (mol%)
	horB	horC	ORF4	ORF5	ORF6	Species
L. brevis ABBC45C	30.4	35.5	55.7	59.1	53.1	44-47
L. paracollinoides DSM 15502T	30.5	35.6	55.8	58.9	52.8	44.8
P. damnosus ABBC478	30.4	35.6	55.8	59.0	53.1	37-42
L. lindneri DSM 20692	29.8	35.7	ND	ND	ND	35

^aG+C contents for the species carrying these regions are listed for reference (4, 11, 12, 39). ND, not detected.

DISCUSSION

L. lindneri has been reported to be the second most frequent cause of microbiological incidents by beer spoilage LAB (2, 3). So far, nonspoilage strains of L. lindneri have not been found (17, 29), leading to the widespread belief that beer spoilage ability is an innate character of L. lindneri. In this study, we attempted to isolate hop-sensitive variants of L. lindneri DSM 20690^T and DSM 20692. Under the subculturing conditions examined, a hop-sensitive variant was isolated only from L. lindneri DSM 20692. Although different subculturing conditions may have led to the isolation of a hop-sensitive variant from DSM 20690^T, the acquisition of one hop-sensitive variant of L. lindneri, the first successful case for this species, was considered significant for brewing microbiology. Therefore we decided to conduct biochemical and genetic characterization of this variant. This strain, designated DSM 20692^NB, exhibited considerably diminished hop resistance ability and completely lost beer spoilage ability. The loss of hop resistance ability by DSM 20692 was accompanied by the loss of resistance to other structurally unrelated drugs, such as ethidium bromide, novobiocin, and CTAB (Table 2), suggesting the hop resistance ability of DSM 20692 is conferred by a multidrug transporter (15, 16).

This possibility led us to reexamine the ORF 1 to 7 region that was implicated in the hop resistance ability of LAB and considered to carry a gene that encodes a proton motive force-dependent multidrug transporter (30, 33, 35). Southern blot analysis indicates that L. lindneri DSM 20692 possesses homologs of ORF 1 and ORF 2, renamed horB and horC, respectively, on a plasmid and this DNA region was found to be lost, concomitant with the loss of hop resistance ability of DSM 20692 (Fig. 1). On the other hand, the ORF 4 to 6 region was not found in DSM 20692 (Fig. 2). These results indicate that horB and horC homologs exist in DSM 20692 without the presence of the ORF 4 to 6 region and these two ORFs may play a vital role in the hop resistance of DSM 20692.

Our previous study showed that the internal organization of the ORF 1 to 7 regions was well conserved among three beer spoilage strains, L. brevis ABBC45^C, L. paracollinoides DSM 15502^T and P. damnosus ABBC478 and ca. 99% nucleotide sequence identities were found in these approximately 8.2-kb DNA regions (30). Therefore we assumed that the ORF 1 to 7 region transfers between strains as a composite transposon with ORF 3 and ORF 7, encoding putative transposase and nicking enzyme, being the mediators of horizontal gene transfer. But the findings of our present study suggest that the horB-horC region has a distinct origin from the ORF 4 to 6 region and these two regions were subsequently united into one DNA unit. This hypothesis is also supported by the general lack of ORF 5 homologs among horB- and horC-positive L. lindneri strains (Table 3). It is also notable that the ORF 5-positive strains of L. casei and L. plantarum appear to lack horB and horC homologs. Furthermore, wide differences in G+C contents were observed between the horB-horC regions and ORF 4 to 6 regions identified in three species of beer spoilage LAB (Table 6). These findings further reinforce the hypothesis that the horB-horC region and the ORF 4 to 6 region have distinct origins.

In our previous articles, we reported that the products of the horB and horC homologs have no homologous proteins (30, 33) but in this study we have more closely examined these two ORFs. The sequencing analysis of horB_L indicates the deduced product of horB_L is homologous to the regulators belonging to the AcrR family (Table 4). The regulators of the AcrR family modulate the transcription of genes that encode multidrug transporters belonging to the RND superfamily (15). Hydropathy analysis indicates that the deduced products of horC_L and its homologs have multiple putative transmembrane domains with a large hydrophilic domain between transmembrane domain 1 (TMD1) and TMD2, a feature characteristic of multidrug transporters belonging to the RND superfamily (Fig. 3). It was also shown that DSM 20692^NB which lost the horB and horC homologs exhibited diminished resistance to several structurally unrelated drugs (Table 2). These results suggest that the products of horC homologs may act as a multidrug transporter. Nonetheless deduced horC products possess only six putative transmembrane domains. This is in contrast to transporters belonging to the RND superfamily, which typically possess 12 transmembrane domains (5, 15, 16).

Nevertheless, transporters belonging to the RND superfamily have been presumed to have arisen from the duplication of ancestral six transmembrane domain units (13, 18). It is therefore conceivable that horC homologs have evolved to retain six transmembrane domain structures. Our preliminary data indicate that the transformation of a hop-sensitive L. brevis strain with the horB and horC region leads to enhanced hop resistance coupled with increased resistance to other structurally unrelated drugs (unpublished data). This insight also supports that horC encodes a multidrug transporter that mediates hop resistance in LAB.

From our present study, the horB and horC regions appear to play a more important role than we previously anticipated. This prompted us to determine whether these ORFs are useful genetic markers to determine the beer spoilage ability of bacteria, using 130 strains of LAB and frequent brewery isolates. The PCR and Southern blot analyses indicate that horB and horC homologs were detectable in 49 out of 51 beer spoilage LAB (Table 3). No false positive results were observed with nonspoilage strains of LAB and frequent brewery isolates, showing that horB and horC are excellent trans-species genetic markers for determining the beer spoilage ability of LAB. Nonetheless the presence of false negative results requires brewers to use multiple trans-species genetic markers.

The horA gene is considered useful for eliminating false-negative results observed with the horB/horC PCR method, because horA homologs were detectable in L. brevis ABBC44 and L. lindneri DSM 20690^T, in which horB and horC homologs are absent (Table 3). Although false positive reactions with horA were observed with nine nonspoilage strains, four L. brevis strains and one L. casei strain were able to grow in beers adjusted to pH 4.6, which represent Japanese beers with weak microbiological stability (38). Depending on the beers manufactured, these false-positive strains should be considered spoilage strains. The remaining strains, L. brevis ABBC4, L. buchneri ABBC219, L. plantarum ABBC87, and L. lindneri DSM 20692^NB, exhibit no beer spoilage ability even in beers with weak microbiological stability. The reason for the inconsistency is currently unknown, but the horA homologs may not be functionally expressed in these false-positive strains, a feature that cannot be detected by the present method. Therefore the false-positive results observed in these strains represent the limits of trans-species genetic markers for differentiating the beer spoilage ability of LAB. Despite these shortcomings, the combined applications of the trans-species genetic markers horB/horC and horA to microbiological control in breweries are attractive, since species-specific identification methods inevitably produce higher rates of false positive results.

Our hypothesis that beer spoilage LAB have arisen from innocuous LAB through horizontal transfer of hop resistance genes also provides a theoretical basis for the practical applications of trans-species genetic markers to the microbiological control of breweries. The comparative study of horB and horC homologs shows the nucleotide sequences of horB_L and horC_L identified in L. lindneri DSM 20692 exhibited 99.3 to 99.5% and 96.6 to 99.4% identity, respectively, with those of L. brevis ABBC45^C, L. paracollinoides DSM15502^T and P. damnosus ABBC478 (Table 5). This level of identity among strains belonging to four distinct beer spoilage species suggests that horB and horC homologs were acquired through horizontal gene transfer rather than evolved with the speciation of these species. The wide and exclusive distribution of horB and horC homologs among various species of beer spoilage strains of LAB also supports our hypothesis. Another important finding is that the horB and horC homologs were lost concomitant with the loss of hop resistance ability of DSM 20692 (Fig. 1). Interestingly, other previously reported nonspoilage variants obtained from L. brevis and L. paracollinoides also lost horB and horC homologs with the loss of beer spoilage ability, except for L. brevis ABBC44, which is deficient in these homologs (Table 3). The unstable nature of the DNA regions containing horB and horC homologs suggests that these regions are not intrinsic genetic factors for beer spoilage LAB.

The same features were also reported for another trans-species genetic marker, horA. The nucleotide sequences of ca. 5.5 kb DNA regions surrounding horA homologs were found to be essentially identical among L. brevis ABBC45, L. lindneri DSM 20690^T and L. paracollinoides DSM15502^T (38). The full sequencing of horA-carrying plasmids pRH45 and pRH20690, identified in L. brevis ABBC45 and L. lindneri DSM 20690^T, revealed that the ORF structures and nucleotide sequences of the backbone regions of the two plasmids are virtually identical, indicating that the horA homologs have been acquired by LAB using plasmid-mediated horizontal gene transfer as one mechanism (36). The horA homologs were also shown to be unstable and the spontaneous loss of the horA homologs occurred concomitantly with the loss of beer spoilage ability of LAB (37, 38), indicating horA homologs are not innate DNA regions of beer spoilage LAB. The wide and almost exclusive distribution of horA homologs among beer spoilage LAB strains further supports that the horA homologs were also acquired by horizontal gene transfer.

Coupled with the results obtained in this study, it is tempting to hypothesize that beer spoilage LAB have acquired hop resistance genes, such as horA and horB/horC, through horizontal gene transfer. The observation that most of the beer spoilage LAB strains possessed both genetic markers simultaneously further suggests that beer spoilage LAB inhabit the environments that foster the exchanges of hop resistance genes among themselves. If this hypothesis is true, the combined use of horA and horB/horC is useful for species-independent detection of beer spoilage LAB, including those that have not yet been characterized.

In conclusion, we successfully isolated a hop-sensitive variant from L. lindneri, which has been considered an innate beer spoilage species. On the basis of genetic characterization of the hop-sensitive variant, new trans-species genetic markers, horB and horC, were identified for differentiating the beer spoilage ability of LAB. The results obtained in this study indicate that the combined use of horA and horB/horC allows the species-independent detection of hop-resistant LAB, showing the concept of trans-species genetic markers has reached a level that can be practically applied to the microbiological control of breweries. The newly proposed detection system is advantageous over the conventional species identification method in that it yields fewer false positive results among established beer spoilage species. In addition to this advantage, trans-species genetic markers are presumably capable of dealing with unencountered beer spoilage species, which is not the case with the species-specific approaches. It is also possible that horA and horB/horC can be used in combination with the conventional species identification method to develop more comprehensive microbiological control systems in breweries.