Characterization and Identification of Pediococcus Species Isolated from Forage Crops and Their Application for Silage Preparation

Transcription

1 APPLIED AND ENVIRONMENTAL MICROBIOLOGY, July 1999, p Vol. 65, No /99/$ Copyright 1999, American Society for Microbiology. All Rights Reserved. Characterization and Identification of Pediococcus Species Isolated from Forage Crops and Their Application for Silage Preparation YIMIN CAI, 1 * SUMIO KUMAI, 2 MASUHIRO OGAWA, 3 YOSHIMI BENNO, 1 AND TAKASHI NAKASE 1 Japan Collection of Microorganisms, The Institute of Physical and Chemical Research, Wako, Saitama , 1 College of Agriculture, Ehime University, Matsuyama, Ehime , 2 and Kyusyu National Agricultural Experiment Station, Nishigoshi, Kumamoto , 3 Japan Received 16 October 1998/Accepted 7 April 1999 Pediococcus species isolated from forage crops were characterized, and their application to silage preparation was studied. Most isolates were distributed on forage crops at low frequency. These isolates could be divided into three (A, B, and C) groups by their sugar fermentation patterns. Strains LA 3, LA 35, and LS 5 are representative isolates from groups A, B, and C, respectively. Strains LA 3 and LA 35 had intragroup DNA homology values above 93.6%, showing that they belong to the species Pediococcus acidilactici. Strain LS 5 belonged to Pediococcus pentosaceus on the basis of DNA-DNA relatedness. All three of these strains and strain SL 1(Lactobacillus casei, isolated from a commercial inoculant) were used as additives to alfalfa and Italian ryegrass silage preparation at two temperatures (25 and 48 C). When stored at 25 C, all of the inoculated silages were well preserved and exhibited significantly (P < 0.05) reduced fermentation losses compared to that of their control in alfalfa and Italian ryegrass silages. When stored at 48 C, silages inoculated with strains LA 3 and LA 35 were also well preserved, with a significantly (P < 0.05) lower ph, butyric acid and ammonianitrogen content, gas production, and dry matter loss and significantly (P < 0.05) higher lactate content than the control, but silages inoculated with LS 5 and SL 1 were of poor quality. P. acidilactici LA 3 and LA 35 are considered suitable as potential silage inoculants. Pediococci are a major component of the microbial flora which live in various types of forage crops; they commonly grow with other plant-associated lactic acid bacteria (LAB) during silage fermentation, and they may influence the fermentation characteristics of silage. Therefore, Pediococcus species and their effects on silage fermentation require further study. However, the phenotypic procedures to assign isolates to known species are difficult, because it is hard to differentiate readily between species of pediococci (22). This is particularly true of Pediococcus pentosaceus and Pediococcus acidilactici (14, 30). There have been many studies that report changes in numbers of epiphytic lactobacilli, enterococci, and leuconostocs during silage fermentation, but the characterization and identification of pediococci isolated from forage crops and their effect on the silage fermentation have not been reported. In the present study, the characterization of Pediococcus species isolated from forage crops and their application to silage preparation were examined. In order to determine their taxonomic status, representative strains were also studied by 16S rrna sequence analysis and DNA-DNA hybridization. MATERIALS AND METHODS Materials and strains studied. Corn (Zea mays) at ripe stage, sorghum (Sorghum bicolor) at dough stage, alfalfa (Medicago sativa) at flowering stage, Italian ryegrass (Lolium multiflorum) at heading stage, and guinea grass (Panicum maximum) at flowering stage were obtained from the National Grassland Research Institute (Nishinasuno, Tochigi, Japan). Forage crops were chopped into 10-mm lengths, and three replicates of the same forage crop were used for microbiological analysis. Samples (10 g) were shaken well by hand with 90 ml of sterilized distilled water, and 10 1 to 10 8 * Corresponding author. Mailing address: Japan Collection of Microorganisms, The Institute of Physical and Chemical Research, Wako-shi, Saitama , Japan. Phone: Fax: cai@ulmus.riken.go.jp. serial dilutions were made in 0.85% sodium chloride solution. From each dilution, 0.05 ml of suspension was spread on agar plates. LAB were counted on both plate count agar (Nissui-Seiyaku Ltd., Tokyo, Japan) with bromocresol purple (0.04 g/liter) and glucose, yeast extract, peptone, and CaCO 3 (GYP) agar after incubation in an anaerobic box (TE-HER Hard Anaerobox model ANX-1; Hirosawa Ltd., Tokyo, Japan) at 35 C for 2 to 3 days. The GYP medium contained 10.0 g of glucose, 5.0 g of yeast extract, 5.0 g of peptone, 2.0 g of sodium acetate, 0.25 g of Tween 80, 200 mg of MgSO 4 7H 2 O, 10 mg of MnSO 4 4H 2 O, 10 mg of FeSO 4 7H 2 O, 5.0 g of NaCl, 5gofCaCO 3, and 1,000 ml of distilled water and was adjusted to ph 6.8. Approximately 20 colonies were isolated at random from the agar plates. Each colony was purified twice by streaking on MRS agar (Difco Laboratories, Detroit, Mich.). LAB were detected by the presence of a yellowish colony and a clear zone due to dissolving CaCO 3. Lactobacilli, pediococci, leuconostocs, and enterococci were counted after morphological observation and determination of Gram staining, catalase reaction, spore formation, nitrate reduction, and fermentation type (16). Aerobic bacteria were counted on nutrient agar (Difco), and mold and yeasts were counted on potato dextrose agar (Nissui-seiyaku). The agar plates were incubated at 30 C for 2 to 4 days. Yeasts were distinguished from mold or bacteria by colony appearance and observation of cell morphology. Colonies were counted as viable numbers of microorganisms (CFU g of fresh matter [FM] 1 ). Each colony of LAB was purified twice by streaking on MRS agar. The pure cultures were grown on MRS agar at 30 C for 24 h, and the colonies were collected with nutrient broth (Difco) containing 10% dimethyl sulfoxide and stored as stock cultures at 80 C for further examination. The type strains of Pediococcus were obtained from the Japan Collection of Microorganisms (JCM), The Institute of Physical and Chemical Research, Wako, Saitama, Japan. Morphological, physiological, and biochemical tests. Gram stain, morphology, catalase activity, spore formation, motility, nitrate reduction, and gas production from glucose were determined according to methods for LAB described by Kozaki et al. (16). Growth at different temperatures was observed in MRS broth after incubation at 10 and 15 C for 14 days and at 45 and 50 C for 7 days. Bile tolerance was determined in GYP broth containing bile at 10, 20, 30, and 40%. Salt tolerance was determined in MRS broth containing NaCl at 3.0 and 6.5%. Growth of LAB at ph 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, and 6.0 was determined in MRS broth after incubation at 30 C for 7 days. Carbohydrate fermentation patterns were tested in GYP basal medium (16) containing 1% (wt/vol) carbohydrate. For tests of carbohydrate fermentation, the strains were cultivated on liver broth (20) at 30 C for 24 h and the broth was then diluted 10-fold with sterile saline solution. Carbohydrate fermentation patterns were examined with a semiautomatic system for bacterial identification as described by Benno (2). The isomers of lactate formed from glucose were determined enzymatically with reagents obtained from Boehringer GmbH, Mannheim, Germany. 2901

2 2902 CAI ET AL. APPL. ENVIRON. MICROBIOL. FIG. 1. Microbiological analysis of fresh forage crops. The values are means standard deviations of three samples., lactobacilli;, leuconostocs;, enterococci;, pediococci;, aerobic bacteria; 2, mold and yeast. 16S rrna sequencing. The 16S rrna sequence coding region was amplified by PCR performed in a PCR ThermalCycler (Takara Shuzo Co., Ltd., Ohtsu, Japan) as described by Suzuki et al. (28). The sequences of the PCR products were determined directly with a sequence kit (ALFexpress AutoCycle; Pharmacia Biotech, Piscataway, N.J.) with the prokaryotic 16S ribosomal DNA universal primers 27F (5 -AGAGTTTGATCCTGGCTCAG-3 ) and 1492R (5 -GGTTA CCTTGTTACGACTT-3 ) (28). Nucleotide substitution rates (K nuc values) were calculated (15), and the phylogenetic tree was constructed by the neighborjoining method (25). The topologies of trees were evaluated by bootstrap analysis of the sequence data with CLUSTAL W software based on 100 random resamplings (31). This sequence was aligned with the following published sequences from DDBJ, GenBank, and EMBL: Enterococcus faecalis JCM5803 (AB012212), Enterococcus faecium JCM 5804 (AB012213), Enterococcus gallinarum (AF039898), Enterococcus casseliflavus (AF039903), Lactobacillus casei JCM 1177 (D16553), Lactobacillus sake DSM (M58829), Lactobacillus alimentarius DSM (58804), Lactobacillus buchneri DSM (M58811), Lactobacillus brevis NCDO 1749 (X61134), Lactobacillus animalis NCD (X61133), Lactobacillus aviarius DSM (M58808), Lactobacillus agilis DSM (M58803), Lactobacillus bifermentans DSM (M58809), Lactobacillus amylophilus DSM (M58806), Lactobacillus amylovorus DSM (M58805), Lactobacillus acidophilus ATCC 4365 (M58802), Lactobacillus acetotolerans DSM (M58801), Pediococcus dextrinicus JCM 5887 (D87679), Pediococcus parvulus JCM 5889 (D88528), P. pentosaceus DSM (58834), P. acidilactici NCDO 2767 (X95976), Pediococcus damnosus JCM 5886 (D87678), Aerococcus urinaeequi IFO (D87677), Lactococcus raffinolactis NCDO 617 (X54261), Lactococcus plantarum NCDO 1869 (X54259), Lactococcus piscium (X53905), Lactococcus lactis subsp. lactis NCDO 2118 (X54260), and Lactococcus garvieae NCDO 2156 (X54262). Bacillus subtilis NCDO 1769 (X60646) was used as an outgroup organism. DNA-DNA hybridization. DNA was extracted from cells harvested from MRS broth incubated for 8 h at 30 C and was purified by the procedure of Saito and Miura (24). DNA base composition was determined by the method of Tamaoka and Komagata (29) with high-performance liquid chromatography following enzymatic digestion of DNA to deoxyribonucleosides. The equimolar mixture of four deoxyribonucleotides in a Yamasa GC kit (Yamasa Shoyu Co., Ltd., Choshi, Japan) was used as the quantitative standard. The DNA-DNA relatedness was determined by the method of Ezaki et al. (11) with photobiotin and microplates. Laboratory silage preparation and chemical analysis. Alfalfa and Italian ryegrass were harvested at the flowering stage. Silages were prepared by using a small-scale system of silage fermentation (4). The isolates and SL 1 (L. casei, isolated from a commercial inoculant [Snow Lact-L; Brand Seed Ltd., Sapporo, Japan]) were used. MRS broth was inoculated with these strains and incubated overnight. After incubation, the optical density of the suspension at 700 nm was adjusted to 0.42 with sterile 0.85% NaCl solution. The inoculum size of LAB was 1 ml of microbial suspension per kg of FM basis. Approximately 100-g portions of forage material, chopped into about 20-mm lengths, were packed into plastic film bags (Hiryu KN type; 180 by 260 cm; Asahikasei), and the bags were sealed with a vacuum sealer (BH 950; Matsushita). The silage treatments were designated untreated control, LA 3, LA 35, LS 5, and SL 1. The film bag silos were kept at 25 and 48 C. Three silos per treatment were used for chemical analysis. The chemical compositions of the forage crops and silages were determined by conventional methods (21). The dry-matter (DM) content of the fresh forage was determined by oven drying at 70 C for 48 h, whereas that of the silages was determined by the removal of water by toluene distillation with ethanol correction (9). The organic acid contents were measured by high-performance liquid chromatography (23). The content of ammonia-nitrogen and lactic acid isomers was determined by enzymatic analysis according to the F-Kit UV method (Boehringer GmbH). Gas production and DM loss were determined by a smallscale fermentation loss test for silage as described by Cai et al. (4). Statistical analysis. Data on the chemical composition of the 60-day silages were analyzed by analysis of variance, and the significance of differences among means was tested by the multiple range test (10). RESULTS Counts of microorganisms. The counts of microorganisms in fresh forage are shown in Fig. 1. Overall, there were 10 6 aerobic bacteria; 10 4 to 10 5 enterococci, leuconostoc, mold, and yeast; and 10 3 and less lactobacilli and pediococci in each of the four forage crops (counts are in CFU g of FM 1 ). Physiological and biochemical properties. A total of 41 strains were isolated from the forage crops, these isolates were gram-positive and catalase-negative tetrad cocci that did not produce gas from glucose and formed approximately equal quantities of L-( )- and D-( )-lactic acid. The carbohydrate fermentation patterns of Pediococcus species are shown in Table 1. These strains were divided into three groups on the basis of their carbohydrate fermentation patterns. Strains LA 3, LA 35, and LS 5 are representative isolates from groups A, B, and C, which were originally recovered from forage crops. Group A included 15 strains that produced acid from lactose and did not produce acid from maltose. Group B included eight strains that did not produce acid from lactose and produced acid from maltose. Group C was different from the other groups which produced acid from trehalose and did not produce acid from D-xylose. All of the isolates were easily distinguished from the type strains of Pediococcus species. The characteristics of strains LA 3, LA 35, LS 5, and L. casei SL 1 are shown in Table 2. Strains LA 3, LA 35, and LS 5 were homofermentative gram-positive tetrad cocci that formed L-( )- and D-( )-lactic acid. Strains LA 3 and LA 35 grew under low-ph (3.5) and high-temperature (50 C) conditions. Strain LS 5 did not grow below ph 4.0 or above 45 C. Strain SL 1 was a homofermentative Lactobacillus that formed lactic acid as an L-( ) isomer and grew at ph 3.5 but did not grow above 45 C. 16S rrna sequence. More than 1,500 bases of 16S rrna of LA 3, LA 35, and LS 5 were determined. The phylogenetic tree shown in Fig. 2 was constructed from evolutionary distances by the neighbor-joining method. Following phylogenetic analysis, representative strains LA 3, LA 35, and LS 5 were placed in the cluster making up the genus Pediococcus. This cluster was recovered in 100% of bootstrap analyses. P. acidilactici JCM 8797 T and P. pentosaceus JCM 5890 T were the species most closely related to the strains LA 3, LA 35, and LS 5 in the

3 VOL. 65, 1999 INFLUENCE OF PEDIOCOCCI ON SILAGE FERMENTATION 2903 TABLE 1. Carbohydrate fermentation patterns of Pediococcus and Tetragenococcus species a Results for: Acid source Group A (n 15) Group B (n 8) Group C (n 18) P. acidilactici JCM 8797 T P. dextrinicus JCM 5887 T P. pentosaceus JCM 5890 T T. halophilus JCM 5888 T L-Arabinose D-Ribose D-Xylose Gluconate W W Glucose Fructose Galactose Mannose Rhamnose W W Cellobiose W Lactose Maltose W Melibiose Sucrose Raffinose Salicin Trehalose W Mannitol Sorbitol a All strains were gram-positive tetrad cocci that did not produce gas from glucose. The tests were performed by using the semiautomatic identification system. Reactions were determined at 30 C for 7 days. n, number of strains tested;, positive;, negative; W, weakly positive. phylogenetic tree, and they showed a high sequence homology value ( 98%) with each other. DNA-DNA hybridization. The results of DNA base composition and DNA-DNA hybridization analyses are shown in Table 3. Representative strains LA 3 and LA 35 had a G C TABLE 2. Characteristics of strains LA 3, LA 35, LS 5, and L. casei SL 1 Characteristic Value a LA 3 LA 35 LS 5 SL 1 Shape Tetrad Tetrad Tetrad Rod Gram stain Fermentation type Homo Homo Homo Homo Optical form of lactate DL DL DL L-( ) Growth ( C) Growth (ph) a, positive;, negative; Homo, homofermentative. FIG. 2. Phylogenetic tree derived from 16S ribosomal DNA sequence. The tree was created by the neighbor-joining method with K nuc values. The numbers indicate bootstrap values for the branch points. content range of 40.5 to 41.0 mol%. Strain LS 5 had a G C content of 38.6 mol%. Strains LA 3 and LA 35 were 93.6 to 97.6% homologous with their representative strains, showing that they belonged to a single species which had the highest levels of DNA relatedness ( 91.8%) to the type strains of P. acidilactici. Strain LS 5 was 88.8 or 97.5% homologous with the type strains of P. pentosaceus. Silage quality and DM loss. Silage quality and DM loss are shown in Table 4. When stored at 25 C, the alfalfa and Italian ryegrass silages that were treated with LA 3, LA 35, LS 5, and SL 1 were well preserved; the ph values, butyric acid, propionic acid, and ammonia-nitrogen contents, gas production, and DM loss were significantly (P 0.05) lower and the lactic acid contents were significantly (P 0.05) higher than those of the control. When stored at 48 C, silages inoculated with LA 3 and LA 35 were also well preserved and had significantly (P 0.05) higher lactic acid contents and significantly (P 0.05) lower ph values and butyric acid, propionic acid, and ammonianitrogen contents than the control. However, in silages inoculated with LS 5 and SL 1, these values were similar to those of the control in alfalfa and Italian ryegrass silages. Compared with the control, silages inoculated with LA 3 and LA 35 had significantly (P 0.05) reduced DM loss and gas production, but silages inoculated with LS 5 and SL 1 resulted in similar levels of these contents in alfalfa and Italian ryegrass silages. Nucleotide sequence accession numbers. The 16S rrna sequences of LA 3, LA 35, and LS 5 have been deposited in the DDBJ database under accession no. AB018213, AB018214, and AB018215, respectively.

4 2904 CAI ET AL. APPL. ENVIRON. MICROBIOL. TABLE 3. DNA base compositions and levels of DNA-DNA homology for Pediococcus species % DNA-DNA reassociation with strain: Strain G C content (mol%) LA 3 LA 35 LS 5 JCM 8797 T JCM 5890 T LA LA LS P. acidilactici JCM 8797 T P. dextrinicus JCM 5887 T P. parvulus JCM 5889 T P. pentosaceus JCM 5890 T DISCUSSION Pediococci are often found living in association with plant material, dairy products, and foods produced by LAB (5, 12, 13, 17, 18), and several papers have reported pediococci as the dominant microbial population on forage crops and silage. Some isolates from forage crops and silage have been identified as P. acidilactici and P. pentosaceus (17, 18). However, available phenotypic procedures to assign isolates to known species are difficult because it is not easy to differentiate clearly between species of pediococci (14, 22, 30). In the present study, the isolates were gram-positive and catalase-negative tetrad cocci that did not produce gas from glucose and formed approximately equal quantities of L-( )- and D-( )-lactic acid. These properties show that these strains belong to the genus Pediococcus. Strains in groups A, B, and C were different from the type strains of P. acidilactici and P. pentosaceus in some carbohydrate fermentation patterns, such as those of lactose, maltose, trehalose, and D-xylose, and could not be identified to the species level on the basis of phenotypic characteristics. Parameter The genetic interrelationships of members of the LAB have been studied extensively in 16S rrna sequence and DNA- DNA hybridization experiments, and new genera and species have been added (3, 6 8). Recent results have clearly indicated that the genera Pediococcus, Enterococcus, Leuconostoc, Weissella, and Lactococcus exhibit a high degree of sequence similarity to each other and form a phylogenetically coherent group that is separate from other bacteria (3, 6, 8). In the present study, the representative strains LA 3, LA 35, and LS 5 were placed in the genus Pediococcus in the phylogenetic tree, confirming that these strains belong to the genus Pediococcus and that they are the species most closely related to P. acidilactici and P. pentosaceus. The DNA-DNA hybridization results demonstrated that strains LA 3 and LA 35 could be assigned to P. acidilactici, and LS 5 could be assigned to P. pentosaceus. The addition of LAB inoculants at ensiling is intended to ensure rapid and vigorous fermentation that results in faster accumulation of lactic acid, lower ph values at earlier stages of ensiling, and inhibition of growth of some pathogenic bacteria TABLE 4. Fermentation quality of silage ensiled at 25 or 48 C for 60 days a Alfalfa silage Value for: Italian ryegrass silage Ut LA 3 LA 35 LS 5 SL 1 Ut LA 3 LA 35 LS 5 SL 1 25 C ph DM (% FM) Lactic acid (% FM) 0.42B 0.94A 1.01A 0.89A 1.05A 0.85B 1.49A 1.65A 1.42A 1.60A Acetic acid (% FM) Butyric acid (% FM) 0.63A 0.06B ND 0.10B ND 0.1A 0.04B 0.02B 0.06B 0.01B Propionic acid (% FM) ND 0.00 ND ND Ammonia-N (% DM) 0.66A 0.45B 0.38B 0.46B 0.40B 0.47A 0.28B 0.32B 0.36AB 0.25B Gas production (liters/kg of FM) 5.98A 1.81B 1.96B 2.01B 1.50B 5.38A 1.18B 1.01B 1.33B 1.14B DM loss (% DM) 7.27A 4.25B 4.05B 4.33B 4.01B 6.23A 3.86B 3.57B 4.20B 3.64B 48 C ph DM (% FM) Lactic acid (% FM) 0.31B 0.74A 0.80A 0.35B 0.38B 0.50B 1.03A 0.98A 0.53B 0.62B Acetic acid (% FM) Butyric acid (% FM) 0.78A 0.13B 0.12B 0.70A 0.75A 0.45A 0.02B 0.04B 0.31A 0.39A Propionic acid (% FM) 0.02 ND ND 0.02 ND Ammonia-N (% DM) 0.72A 0.56B 0.48B 0.78A 0.75A 0.32A 0.17B 0.19B 0.27A 0.30A Gas production (liters/kg of FM) 7.10A 1.98B 1.90B 5.50A 5.68A 6.05A 2.13B 1.98B 4.23A 5.03A DM loss (% DM) 9.78A 4.33B 4.20B 8.24A 9.35A 8.00A 4.40B 4.26B 7.84A 7.91A a Values are means of three silage samples. Means in the same row within a silage type with different letters differ (P 0.05). Ut, control; ND, not detected.

5 VOL. 65, 1999 INFLUENCE OF PEDIOCOCCI ON SILAGE FERMENTATION 2905 (19). Many studies (19, 26, 27) have shown the advantage of such inoculants. Generally, moist dairy farm silage is based on a natural lactic acid fermentation. The epiphytic LAB transform the water-soluble carbohydrates into organic acid in the ensiling process. As a result, the ph is reduced and the forage is preserved (19). However, LAB, especially lactobacilli, are present in forage in very low numbers (5, 17, 18). When LAB fail to produce sufficient lactic acid during fermentation to reduce the ph and inhibit the growth of clostridia, the resulting silage will be of poor quality. As shown in Fig. 1 and Table 4, low numbers of lactobacilli ( 10 3 CFUgofFM 1 ) and high numbers of aerobic bacteria ( 10 5 CFUgofFM 1 ) were present in the material, and poor-quality silage resulted. The factors involved in assessing fermentation quality include the chemical composition of the silage material and the physiological properties of epiphytic bacteria. Generally, alfalfa and Italian ryegrass have relatively low water-soluble-carbohydrate content and low numbers of lactobacilli. During silage fermentation, the lactobacilli could not produce sufficient lactic acid to inhibit the growth of harmful bacteria, and the resulting silage was of poor quality. Therefore, it is necessary to use some bacterial inoculants to control microbes in silage fermentation. The changes in temperature during silage fermentation are well known. Generally, fermentation heating in the silage was consistently correlated with microorganism development and plant respiration. The temperature rises rapidly in the early stage of the ensiling processes and reaches over 45 C (1). In addition, the growth of some lactobacilli would be inhibited by the high-temperature conditions. In our study, when stored at 25 C, silages inoculated with P. acidilactici LA 3 and LA 35, P. pentosaceus LS 5, and L. casei SL 1 were well preserved, with significantly (P 0.05) reduced fermentation loss compared with the control in alfalfa and Italian ryegrass silages. The most plausible explanation lies in the physiological properties of LAB. The strains LA 3, LA 35, LS 5, and SL 1 used in this study were homofermentative LAB which grew well at 25 C and under low-ph (3.5) conditions. Therefore, inoculation with these LAB may result in beneficial effects by promoting the propagation of LAB and by inhibiting the growth of clostridia and aerobic bacteria, as well as by decreasing the amount of gas production and DM loss. On the other hand, when stored at 48 C, silages inoculated with strains LA 3 and LA 35 were also well preserved, with significantly (P 0.05) lower ph, butyric acid and ammonia-nitrogen content, gas production, and DM loss and significantly (P 0.05) higher lactate content than the control. However, silages inoculated with LS 5 and SL 1 were of poor quality and were of quality similar to the control in the two kinds of silage. These results reflect the observation that strains LA 3 and LA 35 could grow at 50 C but strains LS 5 and SL 1 did not grow at that temperature and may die above 45 C. Therefore, during silage fermentation, isolates LA 3 and LA 35 improved silage quality and reduced fermentation loss at high temperatures while strains LS 5 and SL 1 were unable to grow and ferment WSC to produce sufficient lactic acid, resulting in the ph value of silage not falling to less than 4.2 and allowing butyric acid fermentation by clostridia to occur. The results confirmed that P. acidilactici LA 3 and LA 35 were suitable as potential silage inoculants and that they were more effective in improving silage quality than P. pentosaceus LS 5 and the L. casei inoculant strain SL-1 under high-temperature (48 C) conditions. ACKNOWLEDGMENT We thank J. A. Hudson (Environmental Science and Research Ltd., Christchurch Science Centre, Christchurch, New Zealand) for reading the manuscript. REFERENCES 1. Barnett, A. J. G Silage fermentation, p Butterworth Scientific Publication, London, England. 2. Benno, Y A semi-automatic system for bacterial identification. Riken Rev. 12: Cai, J., and M. D. Collins Evidence for a close phylogenetic relationship between Melissococcus pluton, the causative agent of European foulbrood disease, and the genus Enterococcus. Int. J. Syst. Bacteriol. 44: Cai, Y., S. Ohmomo, M. Ogawa, and S. Kumai Effect of NaCl-tolerant lactic acid bacteria and NaCl on the fermentation characteristics and aerobic stability of silage. J. Appl. Microbiol. 83: Cai, Y., Y. Benno, M. Ogawa, S. Ohmomo, S. Kumai, and T. Nakase Influence of Lactobacillus spp. from an inoculant and of Weissella and Leuconostoc spp. from forage crops on silage fermentation. Appl. Environ. Microbiol. 64: Collins, M. D., A. M. Williams, and S. Wallbanks The phylogeny of Aerococcus and Pediococcus as determined by 16S rrna sequence analysis: description of Tetragenococcus gen. nov. FEMS Microbiol. Lett. 70: Collins, M. D., C. Ash, J. A. Farrow, S. Wallbank, and A. M. Williams S ribosomal ribonucleic acid sequence analyses of lactobacilli and related taxa. Description of Vagococcus fluvialis gen. nov., sp. nov. J. Appl. Bacteriol. 67: Collins, M. D., J. Metaxopoulos, and S. Wallbanks Taxonomic study on some leuconostoc-like organisms from fermented sausages: description of a new genus Weissella for the Leuconostoc paramesenteroides group of species. J. Appl. Bacteriol. 75: Dewar, W. A., and P. McDonald Determination of dry matter in silage by distillation with toluene. J. Sci. Food Agric. 12: Duncan, D. B Multiple range and multiple F test. Biometrics 11: Ezaki, T., Y. Hashimoto, and E. Yabuuchi Fluorometric deoxyribonucleic acid-deoxyribonucleic acid hybridization in microdilution wells as an alternative to membrane filter hybridization in which radioisotopes are used to determine genetic relatedness among bacteria strains. Int. J. Syst. Bacteriol. 39: Gashe, B. A Involvement of lactic acid bacteria in the fermentation of Tef (Eragrostis tef), an Ethiopian fermented food. J. Food. Sci. 50: Gashe, B. A Kocho fermentation. J. Appl. Bacteriol. 62: Judicial Commission Opinion 68. Designation of strain B213c (DSM 20284) in place of strain NCDO 1859 as the type strain of Pediococcus acidilactici Lindner Int. J. Syst. Bacteriol. 46: Kimura, M., and T. Ohta On the stochastic model for estimation of mutation distance between homologous proteins. J. Mol. Evol. 2: Kozaki, M., T. Uchimura, and S. Okada Experimental manual of lactic acid bacteria, p Asakurasyoten, Tokyo, Japan. 17. Lin, C., K. K. Bolsen, B. E. Brent, and D. Y. C. Fung Epiphytic lactic acid bacteria succession during the pre-ensiling and ensiling periods of alfalfa and maize. J. Appl. Bacteriol. 73: Lin, C., K. K. Bolsen, B. E. Brent, R. A. Hart, J. T. Dickerson, A. M. Feyerherm, and W. R. Aimutis Epiphytic microflora on alfalfa and whole-plant corn. J. Dairy Sci. 75: McDonald, P., N. Henderson, and S. Heron The biochemistry of silage, 2nd ed., p Chalcombe Publications, Marlow, Berkshire, United Kingdom. 20. Mitsuoka, T Vergleichende Untersuchungen uber die Laktobazillen aus dem Faeces von Menschen, Schweinen und Huhnern. Zentbl. Bakteriol. Parasitenkd. Infektionskr. Hyg. Abt. 1 Orig. 210: Morimoto, H Experimental method of animal nutrition, p Youkendo, Tokyo, Japan. (In Japanese.) 22. Nigatu, A., S. Ahrne, B. A. Gashe, and G. Molin Randomly amplified polymorphic DNA (RAPD) for discrimination of Pediococcus pentosaceus and Pediococcus acidilactici and rapid grouping of Pediococcus isolates. Lett. Appl. Microbiol. 26: Ohmomo, S., O. Tanaka, and H. Kitamoto Analysis of organic acids in silage by high-performance liquid chromatography. Bull. Natl. Grassl. Res. Inst. 48: Saitou, H., and K. Miura Preparation of transforming deoxyribonucleic acid by phenol treatment. Biochim. Biophys. Acta 72: Saitou, N., and M. Nei The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4: Sebastian, S., L. E. Phillip, V. Fellner, and E. S. Idziak Comparative assessment of bacterial inoculated corn and sorghum silages. J. Anim. Sci. 71: Sharp, R., P. G. Hooper, and D. J. Armstrong The digestion of grass silages produced using inoculants of lactic acid bacteria. Grass Forage Sci. 49:42 53.

6 2906 CAI ET AL. APPL. ENVIRON. MICROBIOL. 28. Suzuki, K., J. Sasaki, M. Uramoto, T. Nakase, and K. Komagata Agromyces mediolanus sp. nov., nom. rev., comb. nov., a species for Corynebacterium mediolanum Mamoli 1939 and for some aniline-assimilating bacteria which contain 2,4-diaminobutyric acid in the cell wall peptidoglycan. Int. J. Syst. Bacteriol. 46: Tamaoka, J., and K. Komagata Determination of DNA base composition by reversed-phase high-performance liquid chromatography. FEMS Microbiol. Lett. 124: Tanasupawat, S., S. Okada, M. Kozaki, and K. Komagata Characterization of Pediococcus pentosaceus and Pediococcus acidilactici strains and replacement of the type strains of P. acidilactici with the proposed neotype DSM Request for an opinion. Int. J. Syst. Bacteriol. 43: Thompson, J. D., D. G. Higgins, and T. J. Gibson CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting position specific gap penalties and weight matrix choice. Nucleic Acids Res. 22: