Bovis JB1 Research Articles

Expression of Ruminococcus albus xylanase gene ( xynA) in Streptococcus bovis 12-U-1.

The objective of this study was to ligate the xylanase gene A ( xynA) isolated from Ruminococcus albus 7 into the promoter and signal-peptide region of the lichenase [beta-(1,3-1,4)-glucanase] gene of Streptococcus bovis JB1. This fusion gene was inserted into the pSBE11 vector, and the resulting recombinant, plasmid pXA, was used to transform S. bovis 12-U-1 cells. The transformant, S. bovis 12UXA, secreted the xylanase, which was stable against freeze-thaw treatment and long-time incubation at 37 degrees C. The introduction of pXA and production of xylanase did not affect cell growth, and the xylanase produced degraded xylan from oat-spelt and birchwood.

Factors affecting the antibacterial activity of the ruminal bacterium, Streptococcus bovis HC5.

Streptococcus bovis HC5 inhibits a variety of S. bovis strains and other Gram-positive bacteria, but factors affecting this activity had not been defined. Batch culture studies indicated that S. bovis HC5 did not inhibit S. bovis JB1 (a non-bacteriocin-producing strain) until glucose was depleted and cells were entering stationary phase, but slow-dilution-rate, continuous cultures (0.2 h(-1)) had as much antibacterial activity as stationary-phase batch cultures. Because the activity of continuous cultures (0.2-1.2 h(-1)) was inversely related to the glucose consumption rate, it appeared that the antibacterial activity was being catabolite repressed by glucose. When the pH of continuous cultures (0.2 h(-1)) was decreased from 6.7 to 5.4, antibacterial activity doubled, but this activity declined at pH values less than 5.0. Continuous cultures (0.2 h(-1)) that had only ammonia as a nitrogen source had antibacterial activity, and large amounts of Trypticase (10 mg ml(-1)) caused only a 2.0-fold decline in the amount of HC5 cell-associated protein that was needed to prevent S. bovis JB1 growth. Because S. bovis HC5 was able to produce antibacterial activity over a wide range of culture conditions, there is an increased likelihood that this activity could have commercial application.

Bovicin HC5, a bacteriocin from Streptococcus bovis HC5.

Previous work indicated that Streptococcus bovis HC5 had significant antibacterial activity, and even nisin-resistant S. bovis JB1 cells could be strongly inhibited. S. bovis HC5 inhibited a variety of Gram-positive bacteria and the spectrum of activity was similar to monensin, a commonly used feed additive. The crude extracts (ammonium sulfate precipitation) were inactivated by Pronase E and trypsin, but the activity was resistant to heat, proteinase K and alpha-chymotrypsin. Most of the antibacterial activity was cell associated, but it could be liberated by acidic NaCl (100 mM, pH 2.0) without significant cell lysis. When glycolysing S. bovis JB1 cells were treated with either crude or acidic NaCl extracts, intracellular potassium declined and this result indicated the antibacterial activity was mediated by a pore-forming peptide. The peptide could be purified by HPLC and matrix-assisted laser desorption ionization time-of-flight analysis indicated that it had a molecular mass of approximately 2440 Da. The terminal amino acid sequence was VGXRYASXPGXSWKYVXF. The unnamed amino acid residues (designated by X) had approximately the same position as dehydroalanines found in some lantibiotics, but samples that were reduced and alkylated prior to Edman degradation did not have cysteine residues. The only other bacteriocin that had significant similarity was the lantibiotic precursor of Streptococcus pyogenes SF370, but the identity was only 55%. Based on these results, the bacteriocin of S. bovis HC5 appears to be novel and the authors now designate it as bovicin HC5.

The binding and degradation of nisin by mixed ruminal bacteria

We found that nisin catalyzed potassium efflux from glycolyzing Streptococcus bovis JB1 cell suspensions and that the steady state concentration of residual potassium was dependent upon the amount of nisin added. The relationship between nisin concentration and potassium depletion was a saturation function that had considerable cooperativity. By pre-incubating mixed ruminal bacteria with nisin and removing them prior to S. bovis JB1 addition, it was possible to estimate the ability of mixed ruminal bacteria to bind or degrade nisin. Low concentrations of mixed ruminal bacteria did not bind or degrade all of the nisin in 6 h, but little nisin remained if the mixed ruminal bacteria were present at more than 50 μg protein ml −1. Because cell-free ruminal fluid (10% v/v) inactivated the nisin in less than 2 h, and this inactivation could be counteracted by autoclaving, ultra-filtration or proteinase inhibitor, it appeared that there was an enzymatic degradation of nisin. Mixed ruminal bacteria degraded nisin rapidly, but this degradation did not prevent potassium depletion from mixed ruminal bacteria. These latter results indicated that nisin binding was faster than nisin degradation. The idea that nisin binding could protect nisin from degradation was supported by the observation that intact nisin could be extracted from mixed ruminal bacteria. These observations support the hypothesis that bacteriocins can be used to modify ruminal fermentation, but further work will be needed to see if these peptides can be produced economically.

The binding and degradation of nisin by mixed ruminal bacteria1

We found that nisin catalyzed potassium efflux from glycolyzing Streptococcus bovis JB1 cell suspensions and that the steady state concentration of residual potassium was dependent upon the amount of nisin added. The relationship between nisin concentration and potassium depletion was a saturation function that had considerable cooperativity. By pre-incubating mixed ruminal bacteria with nisin and removing them prior to S. bovis JB1 addition, it was possible to estimate the ability of mixed ruminal bacteria to bind or degrade nisin. Low concentrations of mixed ruminal bacteria did not bind or degrade all of the nisin in 6 h, but little nisin remained if the mixed ruminal bacteria were present at more than 50 microg protein ml(-1). Because cell-free ruminal fluid (10% v/v) inactivated the nisin in less than 2 h, and this inactivation could be counteracted by autoclaving, ultra-filtration or proteinase inhibitor, it appeared that there was an enzymatic degradation of nisin. Mixed ruminal bacteria degraded nisin rapidly, but this degradation did not prevent potassium depletion from mixed ruminal bacteria. These latter results indicated that nisin binding was faster than nisin degradation. The idea that nisin binding could protect nisin from degradation was supported by the observation that intact nisin could be extracted from mixed ruminal bacteria. These observations support the hypothesis that bacteriocins can be used to modify ruminal fermentation, but further work will be needed to see if these peptides can be produced economically.

Expression of a cellulase gene, celA, from the rumen fungus Neocallimastix patriciarum in Streptococcus bovis by means of promoter fusions

The aerotolerant rumen bacterium, Streptococcus bovis, has been used as a host for expression of genes of eukaryotic origin. The coding regions of the celA cellulase gene from the rumen fungus, Neocallimastix patriciarum, were fused with bacterial promoter/signal peptide regions from the Ruminococcus flavefaciens xynD and S. bovis β-(1,3-1,4)-glucanase genes. Fusion cassettes were built into shuttle vector constructs based on pIL253 or pTRW10 and constructs carrying celA were transformed into S. bovis JB1. Active N. patriciarum cellulase was produced in S. bovis with either promoter, although better expression levels were obtained with the native S. bovis β-glucanase promoter fragment.

The antibacterial activity and sensitivity of Streptococcus bovis strains isolated from the rumen of cattle1

Approximately 50% of Streptococcus bovis isolates (n=90) from cattle fed hay- or grain-based diets inhibited the growth of S. bovis JB1. The bacteriocin-like inhibitory substance (BLIS) of the most active strain (HC5) was sensitive to Pronase E, and 16S rDNA analyses indicated that HC5 was closely related to S. bovis strains. When repetitive DNA (BOX) sequences were amplified by PCR, blis+ strains could be organized into 16 BOX-PCR groups, but the similarity indexes were as low as 40%. PCR analyses indicated that none of the blis+ strains had the gene for bovicin 255, a bacteriocin produced by Streptococcus gallolyticus LRC0255, and many of the blis+ isolates were sensitive to bovicin 255. Isolates that survived bovicin 255 and were transferred a second time became less sensitive (≤2 log reduction in viability). The isolates were more sensitive to the BLIS of S. bovis HC5 than bovicin 255, the initial average decrease in viability was approximately 3 logs greater (4.8±2.7 versus 1.8±2.1, respectively, P<0.001), and the HC5-sensitive strains did not adapt. Sixteen of the blis+ isolates were highly resistant to S. bovis HC5 (≤1 log reduction in viability), and BOX-PCR indicated that only five of them had the same BOX pattern as S. bovis HC5. The remaining blis+, HC5-resistant isolates (n=11) had distinctly different BOX patterns and were not always highly resistant to bovicin 255. Because blis+ and blis− strains could be isolated from the same animal, it appears that bacteriocin production is not the only factor affecting S. bovis competition in the rumen.

The antibacterial activity and sensitivity of Streptococcus bovis strains isolated from the rumen of cattle

Approximately 50% of Streptococcus bovis isolates ( n=90) from cattle fed hay- or grain-based diets inhibited the growth of S. bovis JB1. The bacteriocin-like inhibitory substance (BLIS) of the most active strain (HC5) was sensitive to Pronase E, and 16S rDNA analyses indicated that HC5 was closely related to S. bovis strains. When repetitive DNA (BOX) sequences were amplified by PCR, blis + strains could be organized into 16 BOX-PCR groups, but the similarity indexes were as low as 40%. PCR analyses indicated that none of the blis + strains had the gene for bovicin 255, a bacteriocin produced by Streptococcus gallolyticus LRC0255, and many of the blis + isolates were sensitive to bovicin 255. Isolates that survived bovicin 255 and were transferred a second time became less sensitive (≤2 log reduction in viability). The isolates were more sensitive to the BLIS of S. bovis HC5 than bovicin 255, the initial average decrease in viability was approximately 3 logs greater (4.8±2.7 versus 1.8±2.1, respectively, P<0.001), and the HC5-sensitive strains did not adapt. Sixteen of the blis + isolates were highly resistant to S. bovis HC5 (≤1 log reduction in viability), and BOX-PCR indicated that only five of them had the same BOX pattern as S. bovis HC5. The remaining blis +, HC5-resistant isolates ( n=11) had distinctly different BOX patterns and were not always highly resistant to bovicin 255. Because blis + and blis − strains could be isolated from the same animal, it appears that bacteriocin production is not the only factor affecting S. bovis competition in the rumen.

In vitro fermentation of polydextrose by bovine ruminal microorganisms

Polydextrose (PD) is a synthetic additive to human foods that has many physical properties of fat, but is poorly metabolized by nonruminant animals. PD was fermented in vitro by mixed ruminal microorganisms, but the kinetics of the fermentation were complex and varied with different ruminal inocula. Gas production during digestion of PD was described by a two-pool digestion model. One pool, which contained 20–47% of the substrate, was fermented without a lag period and at rates similar to that of purified potato starch or corn starch. The second pool was fermented considerably more slowly than was cellulose. The acetate/propionate ratio in products of PD fermentation ranged from 2.7 to 3.5. Growth rate of Selenomonas ruminantium D was higher on PD than on starches, while the reverse was true for Streptococcus bovis JB1, Prevotella ruminicola B 14, and Butyrivibrio fibrisolvens H17c. For all four strains, the low final culture densities obtained suggest that only a fraction of PD was utilized. The fermentation properties of PD suggest that this material, even if available at competitive prices, is a not an effective substitute for starch in ruminant feeds.

Nisin resistance of Streptococcus bovis.

The growth of Streptococcus bovis JB1 was initially inhibited by nisin (1 microM), and nisin caused a more than 3-log decrease in viability. However, some of the cells survived, and these nisin-resistant cells grew as rapidly as untreated ones. To see if the nisin resistance was merely a selection, nisin-sensitive cells were obtained from agar plates lacking nisin. Results indicated that virtually any nisin-sensitive cell could become nisin-resistant if the ratio of nisin to cells was not too high and the incubation period was long enough. Isolates obtained from the rumen were initially nisin sensitive, but they also developed nisin resistance. Nisin-resistant cultures remained nisin resistant even if nisin was not present, but competition studies indicated that nisin-sensitive cells could eventually displace the resistant ones if nisin was not present. Nisin-sensitive, glucose-energized cells lost virtually all of their intracellular potassium if 1 microM nisin was added, but resistant cells retained potassium even after addition of 10 microM nisin. Nisin-resistant cells were less hydrophobic and more lysozyme-resistant than nisin-sensitive cells. Because the nisin-resistant cells bound less cytochrome c, it appeared that nisin was being excluded by a net positive (i.e., less negative) charge. Nisin-resistant cells had more lipoteichoic acid than nisin-sensitive cells, and deesterified lipoteichoic acids from nisin-resistant cells migrated more slowly through a polyacrylamide gel than those from nisin-sensitive cells. These results indicated that lipoteichoic acids could be modified to increase the resistance of S. bovis to nisin. S. bovis JB1 cultures were still sensitive to monensin, tetracycline, vancomycin, and bacitracin, but ampicillin resistance was 1,000-fold greater.

Streptococcus bovis as a host for the expression of cloned polysaccharidase genes

The shuttle vector, pVA838, was used to introduce the xynD gene from the cellulolytic rumen anaerobe, Ruminococcus flavefaciens 17, into Streptococcus bovis JB1. Expression of xylanase and β-(1,3-1,4)-glucanase activity due to xynD was demonstrated in culture supernatants. SDS-PAGE zymograms revealed the full-length xynD 90 kDa product, together with some proteolytic products. Activities due to the cloned xynD gene, and to the R. flavefaciens 17 endA endoglucanase gene present in the construct pVACMC1, decreased after early growth stages in batch cultures of S. bovis JB1, reflecting the sensitivity of the cloned R. flavefaciens 17 enzymes to inactivation in the presence of accumulated lactic acid. Of xynD activity, 40–80% was detected in the culture supernatant, indicating recognition of the xynD signal peptide by S. bovis JB1.

Effects of thymol on ruminal microorganisms.

Thymol (5-methyl-2-isopropylphenol) is a phenolic compound that is used to inhibit oral bacteria. Because little is known regarding the effects of this compound on ruminal microorganisms, the objective of this study was to determine the effects of thymol on growth and lactate production by the ruminal bacteria Streptococcus bovis JB1 and Selenomonas ruminantium HD4. In addition, the effect of thymol on the in vitro fermentation of glucose by mixed ruminal microorganisms was investigated. Neither 45 nor 90 microg/ml of thymol had any significant effect on growth or lactate production by S. bovis JB1, but 180 microg/ml of thymol completely inhibited growth and lactate production. In the case of S. ruminantium HD4, 45 microg/ml of thymol had little effect on growth and lactate production; however, 90 microg/ml of thymol completely inhibited growth of S. ruminantium HD4. Thymol also decreased glucose uptake by whole cells of both bacteria. When mixed ruminal microorganisms were incubated in medium that contained glucose, 400 microg/ml of thymol increased final pH and the acetate to propionate ratio and decreased concentrations of methane, acetate, propionate, and lactate. In conclusion, thymol was a potent inhibitor of glucose fermentation by S. bovis JB1 and S. ruminantium HD4. Even though thymol treatment decreased methane and lactate concentrations and increased final pH in mixed ruminal microorganism fermentations of glucose, concentrations of acetate and propionate were also reduced.

Deoxyribonuclease activity in Selenomonas ruminantium, Streptococcus bovis, and Bacteroides ovatus.

Six Selenomonas ruminantium strains (132c, JW13, SRK1, 179f, 5521c1, and 5934e), Streptococcus bovis JB1, and Bacteroides ovatus V975 were examined for nuclease activity as well as the ability to utilize nucleic acids, ribose, and 2-deoxyribose. Nuclease activity was detected in sonicated cells and culture supernatants for all bacteria except S. ruminantium JW13 and 179f sonicated cells. S. ruminantium strains were able to utilize several deoxyribonucleosides, while S. bovis JB1 and B. ovatus V975 showed little or no growth on all deoxyribonucleosides. When S. ruminantium strains 5934e, 132c, JW13, and SRK1 were incubated in medium that contained 15 mm ribose, the major end products were acetate, propionate, and lactate. S. ruminantium 5521c1 and S. bovis JB1 did not grow on ribose, and none of the S. ruminantium strains or S. bovis JB1 grew on 15 mm 2-deoxyribose. In contrast, B. ovatus V975 was able to grow on ribose and 2-deoxyribose. In conclusion, all S. ruminantium strains, S. bovis JB1, and B. ovatus V975 had nuclease activity. However, not all bacteria were able to utilize deoxyribonucleosides, ribose, or 2-deoxyribose.

Effects of ammonia and amino acids on the growth and proteolytic activity of three species of rumen bacteria: Prevotella albensis, Butyrivibrio fibrisolvens, and Streptococcus bovis.

The addition of increasing physiological concentrations of ammonia or amino acids had distinct effects on the growth and proteolytic activity of Streptococcus bovis JB1, Prevotella albensis, and Butyrivibrio fibrisolvens DSM3071. The growth of S. bovis and B. fibrisolvens was enhanced by NH(3) and AA, and that of P. albensis was reduced compared with a control with protein as the sole source of nitrogen. The proteolytic activity of S. bovis and P. albensis was reduced, but that of B. fibrisolvens was improved. NH(3) seemed to act mainly on the cell-associated fraction of the proteolytic activity, while the action of AA was not specific. In the rumen the proteolytic activity of S. bovis and P. albensis would be optimal at low concentrations of NH(3) or AA (<0.05 and <0.27 g/L respectively). In contrast, B. fibrisolvens would need higher concentrations (0.5 g/L of NH(3) or 2.7 g/L of AA). It can be assumed that these bacteria will grow in different ecological niches.

Protonmotive force regulates the membrane conductance of Streptococcus bovis in a non-ohmic fashion.

Because the DCCD (dicyclohexylcarbodiimide)-sensitive, F-ATPase-mediated, futile ATP hydrolysis of non-growing Streptococcus bovis JB1 cells was not affected by sodium or potassium, ATP hydrolysis appeared to be dependent only upon the rate of proton flux across the cell membrane. However, available estimates of bacterial proton conductance were too low to account for the rate of ATP turnover observed in S. bovis. When de-energized cells were subjected to large pH gradients (2.75 units, or -170 mV), internal pH declined at a rate of 0.15 pH units s(-1). Based on an estimated cellular buffering capacity of 200 nmol H+ (mg protein)(-1) per pH unit, H+ flux across the cell membrane (at -170 mV) was 108 mmol (g protein)(-1) h(-1). When potassium-loaded cells were treated with valinomycin in low-potassium buffers, initial K+ efflux generated membrane potentials in close agreement with values predicted by the Nernst equation. These artificial membrane potentials drove H+ uptake, and H+ influx was counterbalanced by a further loss of cellular K+. Flame photometry indicated that the rate of K+ loss was 215 (+/-26) mmol K+ (g protein)(-1) h(-1) at -170 mV, but the potassium-sensitive fluorescent compound CD222 indicated that this rate was only 110 (+/-44) mmol K+ (g protein)(-1) h(-1). As pH gradients or membrane potentials were reduced, the rate of H+ flux declined in a non-ohmic fashion, and all rates were <25 mmol (g protein)(-1) h(-1) at a driving force of -80 mV. Previous estimates of bacterial proton flux were based on low and unphysiological protonmotive forces, and the assumption that H+ influx rate would be ohmic. Rates of H+ influx into S. bovis cells [as high as 9x10(-11) mol H+ (cm membrane)(-2) s(-1)] were similar to rates reported for respiring mitochondria, but were at least 20-fold greater than any rate previously reported in lactic acid bacteria.

Natural genetic transformation in the rumen bacterium Streptococcus bovis JB1

Natural transformation of Streptococcus bovis JB1 was demonstrated after development of competence in normal culture medium. Transformation efficiencies were not significantly increased when heat-inactivated horse serum was added to the medium before growth. This is the first time that a resident rumen bacterial species has been shown to be naturally transformable. Transformation allowed the acquisition of plasmids or integration of sequences into the chromosome. No transformation was observed in the presence of undiluted autoclaved or filter-sterilised ovine rumen fluid or filter-sterilised ovine saliva, suggesting that transformation in the ruminant digestive tract is a rare event, although transformation was observed in the presence of 1% and 0.5% filter-sterilised rumen fluid. The use of natural transformation of S. bovis should facilitate further molecular biological studies on this species.

Physiological characterization of Streptococcus bovis mutants that can resist 2-deoxyglucose-induced lysis.

Streptococcus bovis JB1 does not normally lyse, but stationary phase lysis can be induced by including 2-deoxyglucose (2DG) in the growth medium. Isolates deficient in glucose/2DG phosphotransferase activity (PTS-) also lysed when 2DG was present (Lys+) and this result indicated that 2DG phosphorylation via the PTS was not an obligate requirement for 2DG-induced lysis. Cells and cell walls from 2DG-grown cultures lysed faster when proteinase K was added, but glucose-grown cultures and cell walls were not affected. A lipoteichoic acid (LTA) extract (aqueous phase from hot phenol treatment) from glucose-grown cells inhibited the lysis of 2DG-grown cultures, but a similar extract prepared from 2DG-grown cells was without effect. Thin-layer chromatography and differential staining indicated that wild-type and Lys+ PTS- cells incorporated 2DG into LTA, but lysis-resistant cultures (Lys- PTS+ and Lys- PTS-) did not. LTA from lysis-resistant (Lys- PTS+ and Lys- PTS-) cells grown with glucose and 2DG also prevented 2DG-dependent lysis of the wild-type. LTA could not inhibit degradation of cell walls isolated from 2DG-grown cultures, but LTA inhibited the lysis of Micrococcus lysodeikticus (Micrococcus luteus) cells that were exposed to supernatants from 2DG-grown S. bovis cultures. Group D streptococci (including S. bovis) normally have an alpha-1,2 linked glucose disaccharide (kojibiose) in their LTA, but kojibiose cannot be synthesized from 2DG. This observation suggested that the kojibiose moiety of LTA was involved in autolysin inactivation. Wild-type S. bovis had ATP- as well as PEP-dependent mechanisms of 2DG phosphorylation and one lysis-resistant phenotype (Lys- PTS-) had reduced levels of both activities. However, the Lys- PTS+ phenotype was still able to phosphorylate 2DG via ATP and PEP and this result indicated that some other step of 2DG incorporation into LTA was being inhibited. Based on these results, growth in the presence of 2DG appears to prevent synthesis of normal LTA, which is involved in the regulation of autolytic enzymes.

The ability of "low G + C gram-positive" ruminal bacteria to resist monensin and counteract potassium depletion.

Gram-negative ruminal bacteria with an outer membrane are generally more resistant to the feed additive, monensin, than Gram-positive species, but some bacteria can adapt and increase their resistance. 16S rRNA sequencing indicates that a variety of ruminal bacteria are found in the "low G + C Gram-positive group," but some of these bacteria are monensin resistant and were previously described as Gram-negative species (e.g., Selenomonas ruminantium and Megasphaera elsdenii). The activity of monensin can be assayed by its ability to cause potassium loss, and results indicated that the amount of monensin needed to catalyze half maximal potassium depletion (K(d)) from low G + C gram-positive ruminal bacteria varied by as much as 130-fold. The K(d) values for Butyrivibrio fibrisolvens 49, Streptococcus bovis JB1, Clostridium aminophilum F, S. ruminantium HD4, and M. elsdenii B159 were 10, 65, 100, 1020, and 1330 nM monensin, respectively. B. fibrisolvens was very sensitive to monensin, and it did not adapt. S. bovis and C. aminophilum cultures that were transferred repeatedly with sub-lethal doses of monensin had higher K(d) values than unadapted cultures, but the K(d) was always less than 800 nM. S. ruminantium and M. elsdenii cells were highly resistant (K(d) > 1000 nM), and this resistance could be explained by the ability of these low G + C Gram-positive bacteria to synthesize outer membranes.

Natural genetic transformation in the rumen bacterium Streptococcus bovis JB1.

Natural transformation of Streptococcus bovis JB1 was demonstrated after development of competence in normal culture medium. Transformation efficiencies were not significantly increased when heat-inactivated horse serum was added to the medium before growth. This is the first time that a resident rumen bacterial species has been shown to be naturally transformable. Transformation allowed the acquisition of plasmids or integration of sequences into the chromosome. No transformation was observed in the presence of undiluted autoclaved or filter-sterilised ovine rumen fluid or filter-sterilised ovine saliva, suggesting that transformation in the ruminant digestive tract is a rare event, although transformation was observed in the presence of 1% and 0.5% filter-sterilised rumen fluid. The use of natural transformation of S. bovis should facilitate further molecular biological studies on this species.

Hydration of linoleic acid by bacteria isolated from ruminants.

Two strains of Enterococcus faecalis isolated from the ovine rumen and known to hydrate oleic acid were shown to transform linoleic acid by hydration into two products. The products, identified as 10-hydroxy-12-octadecenoic acid and 13-hydroxy-9-octadecenoic acid, were formed during stationary phase in yields of 13% and 6% respectively. Yields increased to 22% and 14% when culture conditions were optimised. To our knowledge, this is the first report of 13-hydroxy-9-octadecenoic acid production by bacteria. During a search for further linoleic-acid-hydrating bacteria, a strain of Streptococcus bovis isolated from bovine faeces and the ruminal strain S. bovis JB1 were found to hydrate linoleic acid. Both strains formed only one product and the most rapid appearance occurred during exponential growth. The S. bovis product, identified as 13-hydroxy-9-octadecenoic acid, formed in a yield of 28%. This study provides the first information on linoleic acid hydration by ruminal bacteria.