Enumeration of Megasphaera elsdenii in rumen contents by real-time Taq nuclease assay
Queensland Beef Industry Institute, Agency for Food and Fibre Sciences, Department of Primary Industries, Moorooka, Queensland, CRC for the Cattle and Beef Industry (Meat Quality), University of New England, NSW, Australia,
Queensland Beef Industry Institute, Agency for Food and Fibre Sciences, Department of Primary Industries, Moorooka, Queensland, CRC for the Cattle and Beef Industry (Meat Quality), University of New England, NSW, Australia,
Aims: To develop a real-time Taq nuclease assay (TNA) to enable the in vivo enumeration of Megasphaera elsdenii.
Methods and Results:Megasphaera elsdenii YE34 was phenotypically characteristic of the species and had 16S rDNA sequence similarity of 98% to previously described isolates. Calibration of the number of cells of M. elsdenii against the cycle threshold of fluorescent dye release gave a straight-line relationship with a correlation coefficient approximating unity. The specificity of the assay for M. elsdenii was confirmed by performing it against a panel of 24 heterogenous, mainly ruminal bacteria. Megasphaera elsdenii was not detected in ruminal contents from a pasture-fed steer but was readily detected 2 and 50 h after the probiotic introduction of the bacterium into the rumen.
Conclusions: Real-time TNA has provided a sensitive and specific means of enumerating the M. elsdenii population in rumen contents.
Significance and Impact of the Study:Megasphaera elsdenii is an important lactate-degrading ruminal bacterium that has been selected for probiotic use to prevent acidosis and enhance starch utilization in grain-fed cattle. The assay developed in this study provides a tool for determining the ability of probiotically-introduced M. elsdenii to establish useful populations in the rumen.
When cattle are introduced to diets containing a high proportion of cereal grain, lactic acid, an undesirable by-product of anaerobic fermentation of starch by some ruminal bacteria, accumulates in the rumen. Often, this accumulation reduces pH of the rumen contents and decreases the efficiency with which feed is converted to volatile fatty acids (VFAs) and microbial protein, essential for animal production (Strobel and Russell 1986). If the drop in pH is large, a condition known as acute lactic acidosis results, which includes clinical signs of lameness and inappetence.
The rumen bacterium Megasphaera elsdenii can play a major role in preventing or controlling acidosis by removing lactic acid through catabolic action (Stewart and Bryant 1988). Because of this action, M. elsdenii should be a useful probiotic organism to reduce the incidence of acidosis and improve the efficiency of starch utilization in the rumen of grain-fed cattle (Kung and Hession 1995; Wiryawan and Brooker 1995; Owens et al. 1998).
An essential requirement when evaluating the efficacy of M. elsdenii as a probiotic organism is an ability to both enumerate the population of the organism in the rumen reliably and to then monitor establishment and stability of this population.
Specific bacterial populations in the rumen have been enumerated previously using labelled hybridization probes complementary to rRNA genes (Stahl et al. 1988; Briesacher et al. 1992; McSweeney et al. 1993; Krause and Russell 1996) and, more recently, by competitive polymerase chain reaction (cPCR) (Reilly and Attwood 1998). The former method is relatively insensitive with a detection threshold of approximately 106 bacteria ml−1 (Stahl et al. 1988). Both methods are tedious; they involve a number of discrete but complex steps, require separate image analysis and interpretation prior to obtaining data, and processing large numbers of samples is cumbersome. Real-time PCR overcomes many of these disadvantages and was developed to quantify DNA rapidly in a single process (Holland et al. 1991; Higuchi et al. 1993). Quantitative 5′ fluorogenic nuclease assays utilize PCR primers and an internal probe that is 5′-labelled with a fluorescent reporter dye and 3′-labelled with a quencher. The 5′-3′ exonuclease activity of the Taq polymerase is used to cleave the reporter dye during each amplification phase of the PCR. Real-time collection of the fluorescent signal allows an accurate quantification of the specific amplification product. Taq Nuclease assays (TNA) have not been employed previously to enumerate bacteria in the complex rumen microbial ecosystem. This paper reports on the development of a real-time TNA to enumerate M. elsdenii.
MATERIALS AND METHODS
Bacterial strains and growth
A pure culture of M. elsdenii strain YE34 was obtained from a grain-adapted steer, identified by phenotypic characteristics and confirmed by 16S rDNA sequence comparison as M. elsdenii. This isolate was used throughout the study. In addition, the following panel of bacteria was used to confirm specificity of the PCR primers and probe set: Bacteroides fragilis 683, Butyrivibrio fibrisolvens AR12, AR27, AR73, ATCC 19171, YE44, Clostridium butyricum YE12, YE15, Escherichia coli K13 (ATCC 15766), Eubacterium ruminantium AR2, Fusobacterium necrophorum AR4, Lactobacillus sp., YE07, YE08, YE16, Prevotella ruminicola AR20, AR29, Ruminococcus flavefaciens AR45, Selenomonas ruminantium AR55, Streptococcus bovis AR25, Sb15, YE01, 2B, and Streptococcus intermedius AR36. The origins of all strains prefixed by AR and Bact. fragilis 683, streptococcal strains Sb15, 2B and YE01, and Lactobacillus and Cl. butyricum strains have been reported previously by Klieve et al. (1989), Klieve et al. (1999a) and Ouwerkerk and Klieve (2001), respectively. ATCC strains were from the American Type Culture Collection. Butyrivibrio fibrisolvens YE44 was isolated in this laboratory and characterized in a manner similar to that used for M. elsdenii YE34.
Bacterial strains were cultured in a rumen fluid-based medium, as previously reported (Klieve et al. 1989). For M. elsdenii YE34, the medium was modified whereby lactic acid replaced both glucose and cellobiose as the major substrate for growth.
Rumen fluid samples
Rumen fluid was collected from a rumen-cannulated steer at pasture and four rumen-cannulated steers in pens. Two of the latter animals were sampled prior to, 2 h and 50 h after being intra-ruminally inoculated with 5·5 × 1012 colony-forming units (cfu) of M. elsdenii YE34. In the week prior to inoculation these animals were fed Paspalum (Paspalum dilatatum) hay supplemented with Molafos (molasses-based mineral supplement; Ridley AgriProducts (Aust) Pty. Ltd, Wacol, Queensland, Australia; 8% w/w of total ration), protein meal (4·5%), urea (1%), lime (1%) and bicarbonate (0·5%). At inoculation, the cattle had been fed twice with 45% rolled barley (balance of ration as above), once 24 h prior to inoculation and the second time at inoculation, i.e. 2 h prior to sampling. Twenty-four and 48 hours after inoculation (2 h prior to the final sampling) the proportion of barley in the ration was increased to 60% (balance of ration as above). Rumen fluid samples were collected from the two non-inoculated control animals at the same time intervals. These animals received the same diet as the inoculated animals.
The rumen fluid samples were strained through nylon gauze and 1 ml aliquots were centrifuged at 15 000 g. The supernatant liquid was discarded and the bacterial pellet frozen at −20°C prior to DNA extraction.
DNA from rumen fluid samples and from bacterial cultures was extracted by physical disruption using a bead beater (Mini-Beater; BioSpec Products, Bartlesville, OK, USA) following the protocol described by Whitford et al. (1998).
PCR amplification, sequencing and sequence analysis of 16S rRNA genes
The 16S rRNA gene from M. elsdenii YE34 was enzymatically amplified from genomic DNA using PCR methods previously described by Ouwerkerk and Klieve (2001).
Sequencing was performed using the ABI PrismTM Dye Terminator Cycle Sequencing Ready Reaction Kit with Amplitaq® DNA Polymerase FS and a model 373 A DNA sequencing system (PE Applied Biosystems Inc., Foster City, CA, USA) following the manufacturer's protocols. The primers used for sequencing were 27F, 530F, 926F, 1114F, 342R, 787R, 1100R and 1525R (Lane 1991). Sequence fragments were assembled using Sequence Navigator (PE Applied Biosystems Inc). The Gapped BLAST database search programme (Altschul et al. 1990) at the National Centre for Biotechnology Information (NCBI) was used to compare sequences.
Real-time TNA primers, probes and operating conditions
Primers and probes were initially selected from 16S rRNA gene sequence data using HYBsimulator v4 (RNAture Inc, West Irvine, CA, USA), then refined using ABI PRISM, Primer Express software to fit the specifications of the Taqman Universal Master Mix Protocol (AB Applied Biosystems, Foster City, CA, USA). Identified primers and probe were checked for specificity using the programmes Probe Match version 2·1 from the Ribosomal Database Project (Maidak et al. 2000) and the BLAST programme (Altschul et al. 1990) at the NCBI site. The primers and probe sequences used were: forward primer MelsF: 5′-GACCGAAACTGCGATGCTAGA-3′ (E. coli numbering system 635–655); reverse primer MelsR: 5′-CGCCTCAGCGTCAGTTGTC-3′ (745–763); and probe MelsP: 5′-TCCAGAAAGCCGCTTTCGCCACT-3′ (721–743). The probe was end-labelled at the 5′ end with the fluorescent reporter dye 6-carboxyfluorescein (6FAM) and at the 3′ end with the quencher dye 6-carboxy tetramethylrhodamine (TAMRA).
Real-time TNA was performed with the ABI PRISM 7700 Sequence Detection System (AB Applied Biosystems), and was undertaken using the universal procedures and conditions as outlined by the manufacturer (Taqman Universal PCR Master Mix Protocol, AB Applied Biosystems). A threshold value for the fluorescence of all samples was set manually and the PCR reaction cycle at which the reaction exceeded this was identified as the cycle threshold (CT).
PCR calibration and verification of specificity
Calibration of the real-time TNA of M. elsdenii YE34 used DNA from two standardized series of samples. For both sets of standards, M. elsdenii YE34 was grown in broth culture at 39°C overnight. The number of M. elsdenii YE34 cells was determined using a Petroff-Hauser Bacteria Counter (Arthur H. Thomas Company, Philadelphia, PA, USA), as per the manufacturer's instructions, at a magnification of 400× with an Olympus BH-2 microscope. For series 1, genomic DNA (gDNA) was extracted from a known number of bacterial cells and used in a dilution series. For series 2, known numbers of M. elsdenii YE34 cells from an overnight culture were added in a dilution series (from 100 cells to 1 × 108 cells) to 1 ml aliquots of rumen fluid (from a pasture-fed steer) prior to DNA extraction.
The specificity of the real-time TNA for M. elsdenii YE34 was verified using genomic DNA, extracted from 5·0 ml pure cultures grown overnight at 39°C, of the panel of bacteria listed above as template in the real-time TNA reaction instead of M. elsdenii YE34. The amount of DNA used was standardized at 10 nanograms of genomic DNA template per reaction. All of the extracted bacterial DNA used had been previously confirmed as amplifiable by a standard 16S rRNA gene PCR reaction (Klieve et al. 1999b).
A common slope, different intercepts, simple linear regression model was fitted to the data obtained for the two dilution series in which y=cycle threshold (CT) and x=log10 number of M. elsdenii cells ml−1. The horizontal distance between the parallel lines and its approximate 95% confidence interval was calculated (Finney 1964).
Megasphaera elsdenii YE34
Megasphaera elsdenii YE34 was typical of the species M. elsdenii based on phenotype (Holdeman et al. 1977). The isolate grew as a large coccus and mostly as diplococci. It was able to use lactic acid as a sole source of carbon and energy. N-butyric, iso-valeric, iso-butyric, n-valeric and n-caproic acids were the by-products of fermentation, with n-butyric and iso-valeric acids being the major by-products.
Genetically, the DNA sequence of the 16S rRNA gene of isolate YE34 was 98% similar, over 1495 bp, to that of M. elsdenii Accession M26493 (Zhao et al. 1989). The 16S rRNA gene sequence of M. elsdenii YE34 has been submitted to the GenBank database with accession number AF283705.
Real-time TNA calibration
The cycle threshold (CT) values obtained following the real-time TNA of the duplicates of each standard from the pure culture of M. elsdenii YE34 (series 1) and the standard using a dilution series of known numbers of M. elsdenii cells added to rumen fluid (series 2) were averaged and plotted against the log of the initial cell numbers ml−1. Results are presented in Fig. 1. There was no significant difference between residual variances or slopes of the two regressions (P > 0·05). However, the intercepts of the two lines were significantly different (P < 0·01). The difference in intercepts divided by the common slope provides an estimate of the horizontal distance (log10 cell numbers ml−1) between the parallel lines. It was estimated to be, with 95% confidence limits, 1·01 (0·73, 1·99). This equates to a 10-fold over-estimation of cell numbers for unknown samples if gDNA from pure culture is used as standard.
Using Probe Match (Maidak et al. 2000) and BLAST (Altschul et al. 1990), the DNA sequence of the forward primer matched, with 100% similarity, DNA from M. elsdenii and M. cerevisiae. The reverse primer was 100% similar to M. elsdenii, M. cerevisiae, Veillonella criceti and V. ratti DNA. In addition, the primers were 100% similar to DNA sequences within 16S rDNA clones from two unidentified rumen bacteria, 3C3d-18 and 3C3d-7 (Tajima et al. 2001). The probe was less specific than the primers, with a sequence similarity identical to sequences in a range of bacteria. However, the combined primers and probe set provided sufficient specificity to differentiate the target bacterium, M. elsdenii.
The gDNA extracted from members of the bacterial panel produced product of the correct size when used as template in a 16S rRNA gene PCR (Klieve et al. 1999b). The M. elsdenii YE34 gDNA used in the bacterial panel was treated in the same manner as the rest of the bacterial panel, and produced a real-time PCR signal equivalent to 2 × 109 cells of M. elsdenii. With the exception of R. flavifaciens AR45 and Bact. fragilis 683, none of the other bacterial gDNAs tested gave a positive result in the real-time PCR (Table 1). However, R. flavifaciens AR45 and Bact. fragilis 683 did produce signal late in the 40-cycle PCR programme, with cycle thresholds at cycles 30 and 34, respectively. These signals would be equivalent to 2187 and 1127 M. elsdenii cells.
Table 1. Cycle thresholds and the average Megasphaera elsdenii cell equivalent ml−1 of the bacterial panel in the M. elsdenii real-time TNAZ
Megasphaera elsdenii was not detectable in the rumen contents from the pasture-fed steer. However, the addition of cells of M. elsdenii YE34 to this rumen fluid gave a positive result, with as few as 100 cells being detectable.
Detection of M. elsdenii in rumen contents
Megasphaera elsdenii cells were not detected by real-time TNA in the rumen fluid samples obtained from the rumen-cannulated steer at pasture, or in the initial samples from the four penned cattle prior to inoculation with M. elsdenii YE34. Two hours after inoculation, M. elsdenii cells were detected at 3·3 × 106 cells ml−1 and 3·1 × 106 cells ml−1 in the samples taken from the two inoculated steers. After 50 h, the numbers of M. elsdenii cells detected were 2·7 × 106 cells ml−1 and 7·4 × 106 cells ml−1, respectively, in the inoculated animals. No M. elsdenii cells were detected in the uninoculated control animals over this period.
The use of real-time TNA has enabled specific enumeration of the lactate-degrading ruminal bacterium. M. elsdenii in rumen contents.
Megasphaera elsdenii YE34 was typical of the species, with phenotypical and biochemical characteristics the same as previously identified strains (Holdeman et al. 1977). The DNA sequence of the 16S rRNA gene also compared favourably with previously published data, with 98% similarity to the type strain (Zhao et al. 1989).
Calibration of the number of M. elsdenii cells with the CT value obtained in the PCR reaction correlated well with the CT values obtained for duplicate standards, showing high levels of reproducibility. There was a straight line relationship between the parameters with a correlation coefficient close to unity.
Standard curves for the diluted DNA series, and the dilution series of known numbers of M. elsdenii cells placed in samples of rumen contents prior to DNA extraction, were parallel. The statistical difference between intercepts of the two standard curves equates to a 10-fold over-estimation of cell numbers for unknown samples if gDNA from pure culture is used as standard in the real-time PCR. This effect is probably due to a carryover of an inhibitory compound present in the rumen fluid during the DNA extraction process. This inhibition of the PCR reaction by components in ruminal fluid has previously been observed in a competitive PCR assay (Reilly and Attwood 1998). However, the single log difference in estimated numbers between the calibration curves was uniform across the series. The calibration curve of cells added to rumen contents prior to DNA extraction was used for the enumeration of rumen fluid samples where the population size of M. elsdenii was unknown.
The assay was very sensitive and as few as 10–100 cells ml−1, using pure cultures, could be accurately enumerated. However, the threshold for enumeration was set arbitrarily at 104 cells ml−1 of sample. With total ruminal bacteria numbering 109–1010 cells ml−1 (Hungate 1966) it could be expected that a ruminal population below 104 cells ml−1 would have a negligible effect on the metabolism of major substrates in feed material. In addition, this threshold would exclude any ambiguity caused by CT values obtained from unrelated bacteria due to non-specific amplification in the last few cycles of the PCR reaction, as observed with the pure culture gDNA panel. These CT values were produced from gDNA template obtained from bacterial cultures with cell numbers in the range of 108–109.
The primers and probe were designed to be, in combination, specific to M. elsdenii. This was confirmed by comparison of the DNA sequences for the primers and probes with other sequences in international databases. Although the DNA sequence of the primers and probe were identical to DNA sequences present in two clones, 3C3d-18 and 3C3d-7, from unidentified ruminal bacteria (Tajima et al. 2001), these clones had a sequence similarity of 98% to M. elsdenii 16S rDNA and are most likely clones of M. elsdenii. Further support for this conclusion arises from the fact that these clones were from a 16S rDNA clone library of ruminal bacteria from cattle being adapted to a high grain diet (Tajima et al. 2001). It could be expected that the density of the M. elsdenii population in these animals would be rapidly increasing (Stewart and Bryant 1988).
The only other bacterium that this set of primers and probe could recognize was the closely related species M. cerevisiae. However, this was not considered problematic as this bacterium is a food spoilage organism (Doyle et al. 1995) that has never been isolated from rumen contents, nor has it ever been identified in rumen contents by any culture-independent, DNA-based methodology.
The specificity of the real-time TNA for M. elsdenii was confirmed by challenging the assay with a panel of common but diverse bacteria, with an emphasis on ruminal isolates. The amount of DNA used as a template was standardized to 10 ng for each bacterium. The known positive in the panel produced a strong signal, indicating in excess of 109 cells to be present. Except for two unrelated bacteria, all others were negative in the assay. The two bacteria in the panel that did produce a CT value did so very late in the PCR reaction and suggested the presence of approximately 103M. elsdenii cell equivalents in the culture. This was 106 times fewer than for the known positive with the same amount of template DNA. These bacteria (Ruminococcus and Bacteroides) are not closely related to M. elsdenii and do not have sequences in common with the primers or probe. Therefore, it is assumed that this product is an inefficiently produced non-specific product. It is highly unlikely that these non-specific products would interfere with the assay, as the rumen would have to contain a pure culture of one or more of these bacteria, at densities > 1010 cells ml−1, for them to register at the detection threshold (104 cells ml−1) set for M. elsdenii. The lack of any interference would appear to be confirmed by negative results obtained from the rumen contents of cattle that had not been inoculated with M. elsdenii.
Where populations of M. elsdenii can be increased, it will be an important probiotic addition in preventing lactic acid accumulation in cattle introduced to a high grain diet, as a reasonable population can metabolize an average of 74% of the lactate produced in the rumen (Counotte et al. 1983a, b). In adult animals, M. elsdenii is mainly found in those adapted to a high grain diet (Stewart and Bryant 1988) and therefore, few cells would be expected to be present in pasture/roughage-fed cattle in which little lactate is produced. The resident population of M. elsdenii in these animals appears very small since as few as 100 cells added to a sample of rumen contents were detectable. The real-time TNA successfully detected M. elsdenii cells in samples of rumen contents 2 and 50 h after inoculation of the bacteria into the rumen. The real-time TNA will be used to evaluate M. elsdenii YE34 as a probiotic to be used to prevent acidosis by permitting an accurate determination of the ability of M. elsdenii to establish and survive in the rumen environment.
The authors thank Dr David Hennessy, Grafton Research Station, NSW Agriculture, for provision of the samples from the pen trial with rumen-cannulated beef cattle. In addition, they thank Andrea Turner, Ian Brock and Gary Blight at the Animal Research Institute, QDPI; David Rutter, Dr Andrew Masel and Dr Brant Bassam of AB Applied Biosystems Pty, Ltd and the staff at the Real Time PCR Facility of the Department of Biochemistry at the University of Queensland for technical assistance. This work was funded through the Meat Quality CRC, Armidale, NSW, and the Queensland Beef Industry Institute, QDPI.