Abstract
The objectives of this study were to formulate a vaccine based upon the different species/strains of methanogens present in sheep intended to be immunized and to determine if a targeted vaccine could be used to decrease the methane output of the sheep. Two 16S rRNA gene libraries were used to survey the methanogenic archaea in sheep prior to vaccination, and methanogens representing five phylotypes were found to account for >52% of the different species/strains of methanogens detected. A vaccine based on a mixture of these five methanogens was then formulated, and 32 sheep were vaccinated on days 0, 28, and 103 with either a control or the anti-methanogen vaccine. Enzyme-linked immunosorbent assay analysis revealed that each vaccination with the anti-methanogen formulation resulted in higher specific immunoglobulin G titers in plasma, saliva, and rumen fluid. Methane output levels corrected for dry-matter intake for the control and treatment groups were not significantly different, and real-time PCR data also indicated that methanogen numbers were not significantly different for the two groups after the second vaccination. However, clone library data indicated that methanogen diversity was significantly greater in sheep receiving the anti-methanogen vaccine and that the vaccine may have altered the composition of the methanogen population. A correlation between 16S rRNA gene sequence relatedness and cross-reactivity for the methanogens (R2 = 0.90) also exists, which suggests that a highly specific vaccine can be made to target specific strains of methanogens and that a more broad-spectrum approach is needed for success in the rumen. Our data also suggest that methanogens take longer than 4 weeks to adapt to dietary changes and call into question the validity of experimental results based upon a 2- to 4-week acclimatization period normally observed for bacteria.
Livestock are a major source of methane production in Australia, with over half of the total anthropogenic methane emissions attributed to them (1). Methane from livestock is produced as a by-product of enteric fermentation and is vented mostly via the mouth and nose by belching and exhalation, with a small amount expelled via the anus (12). There are many approaches being investigated for reducing the methane production of ruminant livestock (for examples, see references 3 and 22). Our laboratory has been working on a novel immunization approach to decrease the numbers and/or activity of the methanogenic archaea (i.e., methanogens) in the rumen. Wright et al. (30) vaccinated sheep with an anti-methanogen vaccine that was based on three strains (1Y, AK-87, and ZA-10) belonging to the genus Methanobrevibacter and produced a 7.7% decrease in methane production per kg of dry matter (DM) intake (DMI). Wright and his colleagues (28) later discovered that less than 20% of the different species of methanogens detected in those sheep were closely related to the methanogens in the vaccine. On the basis of these findings, it was suggested that greater methane abatement might be possible if a greater proportion of the methanogen species/strains were targeted by the vaccine.
Thus, the objectives of this experiment were to formulate a vaccine based on the key methanogens present in the sheep intended to receive the vaccine and to determine if a targeted vaccine could be used to decrease the methane output of sheep. The hypothesis was that sheep vaccinated with the targeted anti-methanogen vaccine would have decreased methane output compared with control sheep.
MATERIALS AND METHODS
Animal ethics approval and permits from the Australian Quarantine Inspection Service and the Australian Pesticides and Veterinary Medicines Authority were obtained.
Prevaccination.
Thirty-six 2-year-old Merino wethers with a live weight (LW) of 56.1 ± 0.16 kg (average ± standard error of the mean) were sourced from a flock of around 230 sheep grazing at CSIRO's Yalanbee research station (Bakers Hill, Western Australia) in late March 2004. Sheep were randomly allocated to individual pens at the CSIRO animal house facility in Perth. A 14-week prevaccination period allowed sheep to be acclimatized to the pens and experimental diet and to be familiarized with the methane chambers (8).
Sheep were offered a pelletized, oaten hay-based ration as a single feed (1.60 kg day−1) each morning, and they had continual access to water. The ration comprised 63 to 70% chopped oaten hay, 14 to 20% cereal grains, 10 to 12% lupines, 2% Siromin mineral mixture (23), and 2 to 5% binder ingredient. On a DM basis, it contained 93% organic matter, 9% crude protein, 42% neutral detergent fiber, and 22% acid detergent fiber. Methods used for collection and analysis of feed samples have been described by Williams et al. (25). After 4 weeks of acclimatization to the experimental diet, rumen fluid (30 to 50 ml) for construction of 16S rRNA gene libraries was collected from all sheep by insertion of a stomach tube and application of suction 2 to 3 h after they were fed.
Methane output over a 22-h period from each sheep during weeks 11 to 13 was measured, for use as a covariate, using open system chambers and gas chromatography (8). Thirty-two of the 36 sheep that were the most settled, as indicated by both their level of feed intake and their general behavior in the methane chambers, were selected for the experimental period.
Cross-reactivity and sequence identity between vaccine candidates.
Cross-reactivity between Methanobrevibacter ruminantium M1T, Methanobrevibacter strains AK-87 and 1Y, Methanobrevibacter millerae ZA-10T, Methanobrevibacter smithii PST, Methanobrevibacter arboriphilus DH-1T, Methanobacterium formicicum MFT, Methanomicrobium mobile BPT, and Methanosarcina barkeri MST was tested using the indirect fluorescent antibody test, originally described by Gnanasampanthan (4) and modified to resemble the enzyme-linked immunosorbent assay (ELISA) technique outlined by Wright et al. (30). The sources of the methanogen strains M1T, ZA-10T, 1Y, AK-87, PST, DH-1T, MFT, BPT, and MST and their culture conditions have been described previously (26). Cells were prepared for use in the cross-reactivity tests by washing and resuspending them in phosphate-buffered saline (PBS). The cells were then added to wells of microtiter plates, and plasma samples containing antibodies to the different cell types were added to the cells. The plasma samples were obtained from a separate group of sheep that had been vaccinated with the different methanogen cell types. Cross-reactivity was expressed as a percentage of the titer. The results of the clone library and cross-reactivity studies were used to select methanogens for the vaccine used in the experimental period. Pairwise 16S rRNA gene sequence identity between the methanogen strains was also determined, and a linear regression of percent sequence identity and percent cross-reactivity was determined.
Vaccine preparation.
The five methanogens chosen for use in the vaccine were Methanobrevibacter strains AK-87 and 1Y, Methanobrevibacter millerae ZA-10T, Methanomicrobium mobile BPT, and Methanosphaera stadtmanae MCB-3T. Vaccines were prepared using aseptic techniques described by Wright et al. (30). The concentration of cells in the vaccine was equivalent to 0.4 mg ml−1 protein, or 2.0 × 109 cells ml−1. Each milliliter of vaccine consisted of 0.12 ml of each of the three Methanobrevibacter strains (1 mg protein ml−1 = 5 × 109 cells ml−1) and 0.02 ml of both Methanomicrobium mobile and Methanosphaera stadtmanae (5 × 109 cells ml−1). Prior to in vivo use, vaccines were tested for mycoplasmal, bacterial, and fungal contamination per 9CFR (USDA Code of Federal Regulations) 113.28 and 9CFR 113.27(d). The freshly prepared vaccine was stored at 4°C for less than 2 days before use. Vaccinations (1-ml dose with 18-gauge needle) were administered subcutaneously, high on the neck, behind the ear. Alternate sides of the neck were used for the second and third vaccinations. All sites were liberally swabbed with 70% ethanol prior to injection, and vaccination sites were also clipped of wool.
Experimental design and measurements.
Selected sheep were allocated to treatments on the basis of their methane output prior to the first vaccination (26.1 ± 1.33 liters methane kg DMI−1) and LW (59.6 ± 0.48 kg), using stratified randomization (2). There were two treatments in a randomized complete block design, with 16 sheep per treatment, and the experimental unit was an individual sheep. The control sheep received a vaccine containing only PBS and adjuvant, whereas the methane sheep received a vaccine containing the five-methanogen mixture, PBS, and adjuvant.
The experimental period began with the first vaccination (day 0) and ran for 137 days. A second vaccination occurred on day 28, and methane output was measured approximately 4 weeks after the second vaccination, between days 62 and 70. A third vaccination occurred on day 103, and methane output was again measured approximately 4 weeks after the third vaccination, between days 127 and 136.
Sheep were removed from the individual pens and returned to the research station to graze between day 77 and 97 of the experimental period. During this time, preliminary experimental results were analyzed and it was decided that the sheep would be given a third vaccination to boost specific immunoglobulin G (IgG) titers in the saliva and that the experiment would be continued. Sheep were brought back to their individual pens in the animal house on day 98 for the remainder of the experimental period.
Blood, saliva, and rumen contents were collected from all sheep 2 days prior to primary vaccination and on days 27, 55, 99, and 123 for analysis of specific anti-methanogen antibody titer (IgG). Rumen contents were also used to extract DNA for molecular analyses (real-time PCR and 16S rRNA gene libraries). Methods of sample collection and the ELISA developed to monitor the antibody responses have been described elsewhere (30). Methane output was measured over 22 h using open-system chambers and gas chromatography (8). Four sheep, two from each treatment, were measured on each day during the methane measurement periods. LW was recorded at 3- to 4-week intervals for monitoring purposes.
16S rRNA gene library construction.
Rumen fluid was pooled randomly into two samples, pool A and pool B, with 18 sheep per pool. DNA was extracted from each sample by using a phenol-chloroform/bead-beating method (20) and PCR amplified using archaeon-specific primers Met86F and Met1340R (26). A 16S rRNA gene library was constructed for each sample, following the protocol of Wright et al. (29). Combined results from the two clone libraries were used to identify the dominant methanogens present after 4 weeks of acclimatization to the experimental diet for inclusion in the vaccine formulation. Two additional 16S rRNA gene libraries (control and treatment) were constructed from the pooled PCR product amplified from frozen samples collected from individual sheep on day 55 (4 weeks after the second vaccination).
For all clone libraries, the cloned PCR product was digested with the HaeIII endonuclease and separated on a 4% molecular screening agarose gel (100 V for 2 h). Restriction fragment length polymorphisms were grouped according to their riboprint patterns (26). At least five clones from each riboprint pattern were sequenced in both directions for confirmation. Sequencing was performed with an ABI Prism 3730 48 capillary sequencer using BigDye Terminator and TaqFS with two forward and two reverse methanogen 16S rRNA gene primers (26). The online chimeric detection program Bellerophon (5) was used to identify chimeric sequences from the library.
Distance data were generated for each clone library by using the Kimura-2 parameter model (7) and analyzed using the computer program DOTUR (13) to calculate the Shannon diversity index (14), with 95% confidence limits, at 98% similarity. Clones were given the prefix “CSIRO,” followed by “WA” to indicate the location of collection (i.e., Western Australia) and a number to identify each phylotype.
Real-time PCR analysis.
Real-time PCR was used to determine the number of methanogens per gram wet weight for all 32 sheep. External standards for methanogens were prepared according to the method of Christophersen (3a), using a mixture of pure cultures of Methanobrevibacter ruminantium M1T and Methanobrevibacter smithii PST, ranging from 1.0 × 103 to 1.0 × 108 cells g−1. Real-time PCR amplifications were carried out with a Bio-Rad iCycler in a 25-μl volume containing the following reagents: 12.5 μl SYBR green mixture (QuantiTect SYBR green PCR; Qiagen), 400 nM each of primers Met630F and Met803R (3a), and 1.0 μl template DNA (10 ng). Real-time PCR amplification was initiated by a hot start at 95°C for 15 min, followed by 40 cycles at 95°C for 30 s, 60°C for 30 s, and 72°C for 60 s. Threshold cycles were calculated automatically by the iCycler software program (version 3.5), and PCR efficiency for each extract was calculated from the logarithmic portion of the sigmoid-shaped curve in real-time PCRs according to the methods described by Liu and Saint (9). Three dilutions of DNA were amplified, and the threshold cycle of the most efficient PCR was recorded.
Statistical analyses.
The effects of vaccine treatment on methane output, LW, LW gain, and average feed intake were analyzed according to a completely randomized design using analysis of variance (Genstat V; Lawes Agricultural Trust, Rothamstead Experimental Station, United Kingdom). When a significant F value was detected, means were compared by the least-significant-difference test. Methane outputs, repeatedly measured over time, were analyzed by repeated measures of residual maximum likelihood, using an ante dependence model. Vaccine treatment and time were fitted as fixed effects in the model, while times for each animal were fitted as random effects.
Nucleotide sequence accession numbers.
The new 16S rRNA sequences generated in this study have been deposited in GenBank under accession numbers EU093103 to EU093110.
RESULTS
Prevaccination: clone library and methanogen cross-reactivity.
In the first 16S rRNA gene clone library (pool A), 52 clones were examined, revealing 10 different sequences, or phylotypes. In the second clone library (pool B), 84 clones were examined, also revealing 10 different phylotypes. DOTUR analysis of the two clone libraries (pool A and pool B) indicated that the Shannon diversity indices were not significantly different from each other, regardless of percent similarity cutoff (Table 1). Therefore, it was feasible to combine the data, resulting in a total of 132 clones made up of 12 different phylotypes (Table 2).
TABLE 1.
Shannon indices for methanogen clone libraries comprising rumen fluid samples from sheep before and after vaccination with a control or treatment formulation and also for sheep from other experimentsa
| Clone library | Diet of sheep prior to sampling | Shannon index ± 95% CIc |
|---|---|---|
| Prevaccination (pool A) | Grazing of autumn pasture, followed by pelletized ration (67% oaten hay, 18% lupines, 10% barley) for 4 wk | 1.76 ± 0.20* |
| Prevaccination (pool B) | Same as that for prevaccination (pool A) group | 1.56 ± 0.19* |
| Prevaccination (pools A and B combined) | Same as that for prevaccination (pool A) group | 1.70 ± 0.15* |
| Control (after second vaccination) | Pelletized ration (67% oaten hay, 18% lupines, 10% barley) for 19 wk | 0.69 ± 0.27** |
| Treatment (after second vaccination) | Same as that for control sheep | 1.38 ± 0.24* |
| WA expt 1b | Grazing of autumn and winter pastures | 1.70 ± 0.12* |
| WA expt 2b | Mixed ration (67% oaten hay, 18% lupines, 10% barley) for 12 wk | 0.76 ± 0.13** |
| WA expt 3b | Mixed ration (68% lucerne hay, 20% lupine grain, 10% molasses) for 6 wk | 0.20 ± 0.14*** |
TABLE 2.
Methanogen 16S rRNA gene libraries constructed prior to vaccination and 4 weeks after the second vaccination
| 16S rRNA clone namea | GenBank accession no. | Nearest valid related organism | % Sequence identity | Related vaccine strain | % Sequence identity | Predicted % cross reactivityb | No. of clones per library
|
||
|---|---|---|---|---|---|---|---|---|---|
| Before expt | Postvaccination
|
||||||||
| Control | Treatment | ||||||||
| CSIRO-WA101 | AY351466 | Thermoplasma volcanium | 78.6 | ZA-10 | 76.1 | <0.1 | 31 | 2 | |
| CSIRO-WA04 | AY615202 | Methanobrevibacter ruminantium M1 | 99.0 | AK87 | 100.0 | 100.0 | 24 | ||
| CSIRO-WA03 | AY196669 | Methanobrevibacter smithii PS | 100.0 | 1Y | 98.2 | 89.9 | 18 | 5 | 3 |
| CSIRO-WA012 | AJ009958 | Methanobrevibacter millerae ZA-10 | 99.1 | ZA-10 | 99.1 | 94.1 | 18 | 29 | 22 |
| CSIRO-WA02 | AB034185 | Methanobrevibacter millerae ZA-10 | 98.4 | 1Y | 99.3 | 95.0 | 12 | 9 | 4 |
| CSIRO-WA083 | AY995282 | Thermoplasma acidophilum | 74.2 | AK87 | 76.7 | <0.1 | 9 | ||
| CSIRO-WA07 | AY196679 | Methanomicrobium mobile BP | 100.0 | BP | 100.0 | 100.0 | 8 | ||
| CSIRO-WA064 | DQ123883 | Methanosphaera stadtmanae MCB-3 | 95.8 | MCB-3 | 95.8 | 78.9 | 7 | ||
| CSIRO-WA125 | AY995292 | Thermoplasma acidophilum | 74.1 | AK87 | 75.9 | <0.1 | 2 | 1 | |
| CSIRO-WA056 | DQ123877 | Methanimicrococcus blatticola | 89.8 | BP | 79.6 | <5.0 | 1 | ||
| CSIRO-WA097 | AY995280 | Thermoplasma acidophilum | 72.7 | AK87 | 75.3 | <0.1 | 1 | ||
| CSIRO-WA118 | AY995300 | Thermoplasma volcanium | 73.5 | 1Y | 75.2 | <0.1 | 1 | ||
| CSIRO-WA13 | EU093103 | Methanobrevibacter millerae ZA-10 | 98.8 | ZA-10 | 98.8 | 92.7 | 1 | ||
| CSIRO-WA14 | EU093104 | Methanobrevibacter millerae ZA-10 | 99.0 | ZA-10 | 99.0 | 93.6 | 1 | ||
| CSIRO-WA15 | EU093105 | Methanobrevibacter millerae ZA-10 | 98.9 | ZA-10 | 98.9 | 93.2 | 1 | ||
| CSIRO-WA199 | AY351494 | Methanobrevibacter millerae ZA-10 | 99.0 | ZA-10 | 99.0 | 93.6 | 1 | ||
| CSIRO-WA20 | EU093107 | Methanobrevibacter millerae ZA-10 | 98.9 | ZA-10 | 98.9 | 93.2 | 1 | ||
| CSIRO-WA1610 | AY351434 | Methanobrevibacter millerae ZA-10 | 96.3 | AK87 | 97.0 | 84.4 | 2 | 18 | |
| CSIRO-WA1711 | AY351493 | Methanobrevibacter millerae ZA-10 | 99.8 | ZA-10 | 99.8 | 97.3 | 1 | ||
| CSIRO-WA18 | EU093106 | Methanobrevibacter smithii PS | 97.8 | 1Y | 98.7 | 92.2 | 7 | ||
| CSIRO-WA21 | EU093108 | Methanobrevibacter wolinii SH | 95.4 | AK87 | 93.5 | 68.4 | 2 | ||
| CSIRO-WA22 | EU093109 | Methanobrevibacter millerae ZA-10 | 83.6 | ZA-10 | 83.6 | 22.9 | 1 | ||
| CSIRO-WA23 | EU093110 | Methanobrevibacter wolinii SH | 95.3 | AK87 | 93.3 | 67.4 | 1 | ||
| Totalc | 132 | 57 | 55 | ||||||
Alternative designations are represented by superscript numerals as follows: 1, CSIRO1.33; 2, SM9; 3, CSIRO-Qld09; 4, ON-CAN.13; 5, CSIRO-Qld24; 6, ON-CAN.05; 7, CSIRO-Qld07; 8, CSIRO-Qld34; 9, CSIRO3.06; 10, CSIRO3.07; and 11, CSIRO3.05.
Based upon the linear regression equation Y = 360.81284 − 4.59014X (where X equals the percent sequence identity).
Total number of clones examined in each library.
Methanobrevibacter strains accounted for more than two-thirds of the different species/strains of methanogens detected in the rumen (Table 2). Seven phylotypes were identified as new taxa on the basis of a similarity criterion of <97% (19) and accounted for almost 40% of the 132 clones. Of these seven new taxa, five phylotypes (CSIRO-WA08, CSIRO-WA09, CSIRO-WA10, CSIRO-WA11, and CSIRO-WA12) belong to the distant group of uncultured archaea, one phylotype (CSIRO-WA06) is a possible new species within the order Methanobacteriales, and one phylotype (CSIRO-WA05) represents a new genus and a new species within the order Methanosarcinales.
With the exception of Methanobrevibacter ruminantium M1, immunotyping of various methanogen strains by using ELISA indicated that there was high cross-reactivity between the Methanobrevibacter strains, indicating common antigenic epitopes (Table 3). Methanobrevibacter ruminantium M1 had very low cross-reactivity to other closely related strains within its genus. Therefore, Methanobrevibacter strains AK-87 and 1Y and Methanobrevibacter millerae ZA-10T were chosen for the vaccine. Methanosphaera stadtmanae MCB-3T and Methanomicrobium mobile BPT were also chosen for use in the vaccine, as they represented 5% and 6%, respectively, of the combined clone library population (Table 2). Furthermore, the immunotyping data indicated that there was no cross-reactivity between the protein cell walls of Methanomicrobium mobile BPT cells and any of the pseudomurein cell walls of the Methanobrevibacter strains cells (Table 3). A pure culture of Methanosphaera stadtmanae MCB-3T was unavailable when we were conducting our cross-reactivity experiments, so it was not immunotyped. However, it was deduced that this methanogen was likely to have different antigenic sites on its cell wall because it used different energy sources for growth (10).
TABLE 3.
16S rRNA gene sequence identities and percents cross-reactivity for some methanogen strainsa
| Strain | % Identity or cross-reactivityb
|
||||||||
|---|---|---|---|---|---|---|---|---|---|
| M1 | AK-87* | 1Y* | ZA-10* | PS | DH-1 | MF | BP* | MS | |
| Methanobrevibacter ruminantium M1T | 100.0 | 98.9 | 94.1 | 93.5 | 94.2 | 94.5 | 91.9 | 78.7 | 78.9 |
| Methanobrevibacter sp. strain AK-87* | 47.5 | 100.0 | 94.4 | 94.2 | 94.8 | 94.9 | 91.8 | 78.6 | 79.4 |
| Methanobrevibacter sp. strain 1Y* | 41.4 | 87.1 | 100.0 | 98.4 | 98.2 | 94.9 | 90.5 | 78.8 | 78.7 |
| Methanobrevibacter millerae ZA-10T* | 39.2 | 82.6 | 94.8 | 100.0 | 97.9 | 94.4 | 90.5 | 78.5 | 78.5 |
| Methanobrevibacter smithii PST | 34.6 | 72.9 | 83.6 | 88.2 | 100.0 | 95.0 | 90.9 | 78.4 | 78.2 |
| Methanobrevibacter arboriphilus DH-1T | 28.3 | 59.6 | 68.4 | 72.1 | 81.8 | 100.0 | 92.0 | 78.3 | 79.6 |
| Methanobacterium formicicum MFT | 21.0 | 44.2 | 50.8 | 53.5 | 60.7 | 74.2 | 100.0 | 80.8 | 78.8 |
| Methanomicrobium mobile BPT* | 0.1 | 0.2 | 0.3 | 0.3 | 0.3 | 0.4 | 0.5 | 100.0 | 82.2 |
| Methanosarcina barkeri MST | 0.1 | 0.2 | 0.2 | 0.2 | 0.2 | 0.3 | 0.4 | 0.6 | 100.0 |
Asterisks indicate four of the five methanogen strains used in the vaccine formulation.
Values in the upper triangle (to the right of the bolded values) are 16S rRNA gene sequence identities, and values in the lower triangle (to the left of the bolded values) are percents cross-reactivity.
The five methanogen strains selected for the vaccine formulation accounted for 52% of the different methanogen sequences recovered from a 16S rRNA gene clone library during the prevaccination period. The remaining 48% of the clone sequences were either from uncultivated strains, from strains that we did not have available, or from strains not approved for use in vaccines. Linear regression analysis of percent sequence identity and percent cross-reactivity (Y = 360.81284 − 4.59014X) was significant (P < 0.001) and produced an R2 value of 0.90 (Table 3).
Experimental period: immune response, methanogen numbers, and diversity.
Titers of methanogen-specific IgG in plasma, saliva, and rumen fluid increased after each vaccination in methanogen-vaccinated sheep but not in control-vaccinated sheep (Fig. 1).
FIG. 1.
Specific IgG titers over the experimental period in plasma, saliva, and rumen fluid samples of sheep vaccinated with a control (○) or anti-methanogen vaccines (•). Timings of first (1°), second (2°), and third (3°) vaccinations are indicated by arrows.
Mean densities of methanogens (number of cells g wet weight−1 ± standard error of the mean) in the rumen digesta of individual sheep before vaccination, as determined by real-time PCR, were 1.58 × 107 ± 6.86 × 106 for the control group and 2.33 × 107 ± 1.96 × 107 for the treatment group, whereas after the second vaccination, the methanogen numbers were 1.50 × 106 ± 5.77 × 105 and 1.33 × 106 ± 3.39 × 105 for the control and treatment animals, respectively. There was no difference (P > 0.05) between the control and treatment groups in mean density of methanogens before or after vaccination.
The 16S rRNA gene clone library for control sheep generated 57 clones, comprising 10 different phylotypes (Table 2). The 16S rRNA gene library for the treatment sheep generated 55 clones, made up of 10 phylotypes. Overall, 244 clones were examined, representing 23 different phylotypes. The Shannon diversity indices (14) for the methanogen clone libraries differed (P < 0.05) between the control and treatment sheep after the second vaccination (Table 1). Of the 10 phylotypes from the treatment animals, four were identified as new taxa and were not detected in the other clone libraries (i.e., during the prevaccination period and in the control animals after the second vaccination). Of these four new taxa, three phylotypes (CSIRO-WA16, CSIRO-WA21, and CSIRO-WA23) represent possible new species within the order Methanobacteriales, and one phylotype (CSIRO-WA22) represents a new genus and a new species. With the exception of CSIRO-WA16, all clones recovered from the control animals were markedly decreased (CSIRO-WA01, CSIRO-WA02, and CSIRO-WA03) or undetectable (CSIRO-WA12, CSIRO-WA13, CSIRO-WA14, CSIRO-WA15, CSIRO-WA17, and CSIRO-WA18) in the treatment animals.
Overall, Methanobrevibacter strains made up approximately 98% and 95% of the rumen methanogen clones recovered from the control and treatment sheep, respectively (Table 2). The five methanogens used in the vaccine formulation accounted for 67% of the methanogen clones recovered from control sheep but only 47% of the methanogen clones recovered from sheep vaccinated against these same five methanogens.
The Shannon index for the clone library prior to the experiment was not different (P > 0.05) from that for the clone library of treatment sheep after the second vaccination but was different (P < 0.05) from that for control sheep after the second vaccination (Table 1). With the exception of clone CSIRO-WA01, all clones found in animals prior to the first vaccination were either markedly decreased in number (CSIRO-WA02, CSIRO-WA03, and CSIRO-WA12) or undetectable (CSIRO-WA04, CSIRO-WA05, CSIRO-WA06, CSIRO-WA07, CSIRO-WA08, CSIRO-WA09, CSIRO-WA10, and CSIRO-WA11) in the control animals (Table 2). CSIRO-WA01 increased in number by 61% and was the predominant clone recovered from the control animals. Six additional clones (CSIRO-WA13, CSIRO-WA14, CSIRO-WA15, CSIRO-WA16, CSIRO-WA17, and CSIRO-WA18), which were not detected during prevaccination, were recovered from the control animals (Table 2).
Feed intake, methane output, and LW.
There was no effect (P > 0.05) of vaccination treatment on feed intake, which averaged 1.39 kg DM day−1 over the experimental period (Table 4). There was no effect (P > 0.05) of vaccination treatment on total methane output by sheep (data not shown) or on methane output corrected for DMI (Table 4). Methane outputs measured after the second and third vaccinations were 20% and 18% higher, respectively, than those for controls, but the difference was not significant (Table 4). Furthermore, there was no significant effect (P > 0.05) of vaccination on LW at the end of the experiment (72.0 ± 0.70 kg [mean ± standard error]) or LW gain between the primary vaccination and the end of the experiment (81 ± 4.2 g day−1).
TABLE 4.
Total methane outputs (corrected for DMI) and feed intakes for sheep during the prevaccination period and after the second and third vaccinationsa
| Vaccine formulation | Methane output (liters kg DMI−1)
|
Feed intake (kg DM)
|
||||
|---|---|---|---|---|---|---|
| Prevaccination | After second vaccination | After third vaccination | Prevaccination | After second vaccination | After third vaccination | |
| Control | 23.0 ± 1.65 | 25.0 ± 2.83 | 23.9 ± 2.22 | 1.41 ± 0.016 | 1.30 ± 0.069 | 1.50 ± 0.011 |
| Anti-methanogen | 24.2 ± 1.67 | 30.2 ± 3.00 | 28.2 ± 2.58 | 1.44 ± 0.007 | 1.25 ± 0.079 | 1.44 ± 0.052 |
Values shown are means ± standard errors.
DISCUSSION
We used a variation of a vaccination protocol aimed at inducing a humoral response against methanogens (30) that had previously resulted in small but significant reductions in methane emissions from vaccinated sheep where only 20% of the different species/strains of methanogens in the sheep were targeted by the vaccine formulation (28). Our vaccine formulation targeted approximately 52% of the different species/strains of methanogens present in the sheep and also contained three methanogens which were used in the previous successful formulation. We therefore expected a greater impact of the vaccination procedure on methane output in vaccinated sheep. However, we did not observe a decrease in methane output in sheep that received the anti-methanogen vaccine in comparison to the level in the controls. In fact, methane output was nearly 18% higher in treatment sheep vaccinated three times than in the controls (P > 0.05).
Vaccination did induce a humoral immune response, as indicated by the specific IgG titers in plasma and saliva, and specific anti-methanogen IgG was also delivered to the rumen, as indicated by the titers in the rumen fluid. However, the specific IgG titer levels observed in sheep plasma after the second vaccination were lower than those recorded in our previous study (30), whereas after the third vaccination, plasma IgG-specific titers increased to levels similar to those observed in our previous study (30). However, saliva and rumen IgG-specific antibodies increased only moderately, and no methane abatement was observed. This strongly highlights the variable outcomes of vaccination for decreasing enteric methane and suggests that methane abatement is obtained only when certain undefined conditions are met.
The methanogen populations in these sheep were dominated by Methanobrevibacter phylotypes, which is consistent with other studies of rumen methanogen populations (6, 10, 11, 15-17, 21, 24, 27, 29). Although the total numbers of methanogens were not significantly different between control and anti-methanogen sheep after vaccinations, our results suggest that some of the targeted methanogens were affected by the vaccine. The majority of clones recovered from the control animals were markedly decreased or undetectable in the treatment animals and had at least 98.7% sequence identity to Methanobrevibacter millerae ZA-10T or Methanobrevibacter sp. strain 1Y, two of the key methanogens in the vaccine formulation. Another indication of changed populations was the difference in the Shannon indices of the clone libraries for the control and anti-methanogen-vaccinated sheep. We believe that the presence of six clones (CSIRO-WA10, CSIRO-WA19, CSIRO-WA20, CSIRO-WA21, CSIRO-WA22, and CSIRO-WA23) in the treatment animals that were not detectable in the control animals was due to the niche created by the decrease in numbers or elimination of the dominant strains targeted with the vaccine. The increase in population of these other clones has masked the effect of the vaccine regimen.
Our results also suggest that the vaccine formulation is very specific. The predicted percents cross-reactivity between the most closely related methanogen strain in the vaccine formulation and the clones whose numbers decreased in the treatment animals ranged from 89.9 to 97.3%. The clones in treatment animals that increased in number or appeared as additional clones had lower predicted percents cross-reactivity, which ranged from <0.1 to 84.4%. There was a small exception to this observation, however, with two clones (CSIRO-WA19 and CSIRO-WA20) that appeared for the first time in treatment animals, despite having predicted cross-reactivities of 93.2% and 93.6%, respectively, to Methanobrevibacter millerae ZA-10T. Even with these anomalies, the data still demonstrate that a highly specific vaccine can be made to target specific strains of methanogens and also confirm that a more broad-spectrum approach is needed for success in the rumen.
The clone library results show a difference in composition of clones between sheep prior to the experiment and control group animals. A comparison of the diversity indexes of these sheep prior to vaccination, 4 weeks after the start of the pelletized diet, with another group of grazing sheep from the same farm (29) showed they were nearly identical (Table 1). Furthermore, the Shannon indices (14) for our control sheep after 18 weeks on the experimental diet were very similar to those for the sheep used by Wright et al. (29) after more than 3 months on a similar oaten hay-based animal house diet (Table 1). We suggest that the rumen methanogen population had not fully adapted from the grazing diet to the pelletized experimental diet after only 4 weeks. If the clone library constructed postvaccination from the control sheep represents the true acclimatized methanogen population on the experimental diet, then acclimatization occurred somewhere between 4 and 18 weeks and therefore our vaccine was formulated against methanogens in grazing sheep and not those in sheep on the experimental diet.
The construction of clone libraries before diet adaptation resulted in lower target rates for the vaccine than we originally calculated. Three of the five methanogens (Methanobrevibacter sp. strain AK-87, Methanomicrobium mobile BPT, and Methanosphaera stadtmanae MCB-3T) used in the vaccine formulation, due to their presence in clone libraries during prevaccination, were not detected in the clone library from the acclimatized control sheep. However, the remaining two methanogens (Methanobrevibacter millerae ZA-10T and Methanobrevibacter sp. strain 1Y) accounted for two-thirds of the methanogenic archaea in the acclimatized control sheep. If the vaccine had been formulated on the basis of the acclimatized data, then it would have contained approximately 50% more protein of ZA-10T and 1Y than the current vaccine and may have been more effective. In light of this, it is questionable whether the administered load was sufficient to induce an effective response, which could explain why the two clones with predicted cross-reactivities of 93.2 to 93.6% with Methanobrevibacter millerae ZA-10T appeared in the treatment animals.
The trend toward greater methane output by treatment vaccinated sheep, even though they had lower numbers of methanogens, is contrary to the general belief that methanogenesis is coupled to growth (18). It could be suggested that the other methanogens that were detected in the treatment vaccinated sheep may be more-potent producers of methane, therefore increasing methane output. This highlights the need to measure both cell numbers and methane output in this type of work.
In conclusion, we describe a vaccination regimen which induced a substantial serum antibody response against methanogens in sheep but failed to significantly affect the methane emission by these sheep and the density of methanogens. However, upon closer inspection, it would appear that the vaccine may have affected the diversity and composition of the methanogen population. Furthermore, compared to what was found for the control sheep, some Methanobrevibacter strains were either missing or decreased in the treatment sheep and replaced by other methanogens not detected previously before vaccination (Table 2). In the treatment sheep, the appearance of these other methanogens accounted for 44% of the different species/strains of methanogens detected and may explain why methanogen numbers were not significantly different between the control and treatment sheep but methanogen populations were significantly different between the two groups. Another complicating factor is that there does not appear to be a definite correlation between methanogen numbers and methane production, even though methanogenesis is an energetic reaction coupled to growth. Our data indicate that there is a correlation between 16S rRNA gene sequence relatedness and cross-reactivity for the methanogens and also suggest that adaptation of the rumen methanogen populations to differing diets may be much slower than expected (i.e., more than 4 weeks). This now calls into question the validity of experimental results based upon a 2- to 4-week acclimatization period normally observed for bacteria. Further studies are warranted to properly assess the acclimatization period for the methanogenic archaea.
Acknowledgments
We acknowledge the capable technical support provided by Louis Klein, Carolyn Pimm, Andrew Williams, Roslyn Owen, Claire Auckland, Qingmin Li, and Mike Carthew. Statistical advice was provided by Kevin Murray of the University of Western Australia. The contributions of Michel Denis and Bryce Buddle of AgResearch, New Zealand, and Mark Morrison (CSIRO Livestock Industries, Brisbane, Queensland, Australia) to the final version of the manuscript are gratefully acknowledged.
Footnotes
Published ahead of print on 6 February 2009.
REFERENCES
- 1.Australian Greenhouse Office. 2008. National Greenhouse Gas Inventory 2006. Australian Greenhouse Office, Canberra, Australia.
- 2.Baird, D. B. 1994. The design of experiments with covariates. Ph.D. thesis. University of Otago, Dunedin, New Zealand.
- 3.Boadi, D., C. Benchaar, J. Chiquette, and D. Massé. 2004. Mitigation strategies to reduce enteric methane emissions from dairy cows: update review. Can. J. Anim. Sci. 84:319-335. [Google Scholar]
- 3a.Christophersen, C. T. 2007. Grain and artificial stimulation of the rumen change the abundance and diversity of methanogens and their association with ciliates. Ph.D. thesis. University of Western Australia, Perth, Western Australia, Australia.
- 4.Gnanasampanthan, G. 1993. Immune responses of sheep to rumen ciliates and the survival and activity of antibodies in the rumen fluid. Ph.D. thesis. University of Adelaide, Adelaide, Australia.
- 5.Huber, T., G. Faulkner, and P. Hugenholtz. 2004. Bellerophon: a program to detect chimeric sequences in multiple sequence alignments. Bioinformatics 20:2317-2319. [DOI] [PubMed] [Google Scholar]
- 6.Irbis, C., and K. Ushida. 2004. Detection of methanogens and proteobacteria from a single cell of rumen ciliate protozoa. J. Gen. Appl. Microbiol. 50:203-212. [DOI] [PubMed] [Google Scholar]
- 7.Kimura, M. 1980. A simple method of estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16:111-120. [DOI] [PubMed] [Google Scholar]
- 8.Klein, L., and A.-D. G. Wright. 2006. Construction and operation of open-circuit methane chambers for small ruminants. Aust. J. Exp. Agric. 46:1257-1262. [Google Scholar]
- 9.Liu, W., and D. A. Saint. 2002. Validation of a quantitative method for real time PCR kinetics. Biochem. Biophys. Res. Commun. 294:347-353. [DOI] [PubMed] [Google Scholar]
- 10.Miller, T. L. 2001. Genus II. Methanobrevibacter, p. 218-226. In D. R. Boone and R. W. Castenholz (ed.), Bergey's manual of systematic bacteriology, 2nd ed., vol. I. Springer, New York, NY. [Google Scholar]
- 11.Miller, T. L., and M. J. Wolin. 1986. Methanogens in human and animal digestive tracts. Syst. Appl. Microbiol. 7:223-229. [Google Scholar]
- 12.Murray, R. M., A. M. Bryant, and R. A. Leng. 1976. Rates of production of methane in the rumen and large intestine of sheep. Br. J. Nutr. 36:1-14. [DOI] [PubMed] [Google Scholar]
- 13.Schloss, P. D., and J. Handelsman. 2005. Introducing DOTUR, a computer program for defining operational taxonomic units and estimating species richness. Appl. Environ. Microbiol. 71:1501-1506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Shannon, C. E., and W. Weaver. 1949. The mathematical theory of communication. University of Illinois Press, Urbana.
- 15.Sharp, R., C. J. Ziemer, D. S. Marshall, and D. A. Stahl. 1998. Taxon-specific associations between protozoal and methanogen populations in the rumen and a model system. FEMS Microbiol. Ecol. 26:71-78. [Google Scholar]
- 16.Skillman, L. C., P. N. Evans, C. Strompl, and K. N. Joblin. 2006. 16S rDNA directed PCR primers and detection of methanogens in the bovine rumen. Lett. Appl. Microbiol. 42:222-228. [DOI] [PubMed] [Google Scholar]
- 17.Skillman, L. C., P. N. Evans, G. E. Naylor, B. Morvan, G. N. Jarvis, and K. N. Joblin. 2004. 16S ribosomal DNA-directed PCR primers for ruminal methanogens and identification of methanogens colonising young lambs. Anaerobe 10:277-285. [DOI] [PubMed] [Google Scholar]
- 18.Soliva, C. R., H. R. Hess, L. Meile, M. Kreuzer, and A. Machmüller. 2003. Suppression of ruminal methanogenesis by dietary means: apparent inconsistency between methane release and counts of microbes involved in methanogenesis. Trop. Subtrop. Agroecosyst. 3:209-213. [Google Scholar]
- 19.Stackebrandt, E., and B. M. Goebel. 1994. Taxonomic note: a place for DNA:DNA reassociation and 16S rRNA sequence analysis in the present species definition in bacteriology. Int. J. Syst. Bacteriol. 44:846-849. [Google Scholar]
- 20.Stahl, D. A., B. Flesher, H. R. Mansfield, and L. Montgomery. 1988. Use of phylogenetically based hybridization probes for studies of ruminal microbial ecology. Appl. Environ. Microbiol. 54:1079-1084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Tokura, M., I. Chagan, K. Ushida, and Y. Kojima. 1999. Phylogenetic study of methanogens associated with rumen ciliates. Curr. Microbiol. 39:123-128. [DOI] [PubMed] [Google Scholar]
- 22.Ulyatt, M. J., H. Clarke, and D. K. R. Lassey. 2002. Methane and climate change. Proc. N. Z. Grassland Assoc. 64:153-157. [Google Scholar]
- 23.White, C. L., D. G. Masters, D. W. Peter, D. B. Purser, S. P. Roe, and M. J. Barnes. 1992. A multi element supplement for grazing sheep. I. Intake, mineral status and production responses. Aust. J. Agric. Res. 43:795-808. [Google Scholar]
- 24.Whitford, M. F., R. M. Teather, and R. J. Forster. 2001. Phylogenetic analysis of methanogens from the bovine rumen. BMC Microbiol. 1:5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Williams, Y. J., S. Rea, S. Popovski, A. F. Toovey, L. Skillman, and A.-D. G. Wright. 2008. Responses of sheep to a vaccination of entodinial or mixed rumen protozoal antigens to reduce rumen protozoal numbers. Br. J. Nutr. 99:100-109. [DOI] [PubMed] [Google Scholar]
- 26.Wright, A.-D. G., and C. Pimm. 2003. Improved strategy for presumptive identification of methanogens using 16S riboprinting. J. Microbiol. Methods 55:337-349. [DOI] [PubMed] [Google Scholar]
- 27.Wright, A.-D. G., X. Ma, and N. E. Obispo. 2008. Methanobrevibacter phylotypes are the dominant methanogens in sheep from Venezuela. Microb. Ecol. 56:390-394. [DOI] [PubMed] [Google Scholar]
- 28.Wright, A.-D. G., A. F. Toovey, and C. L. Pimm. 2006. Molecular identification of methanogenic archaea from sheep in Queensland, Australia reveal more uncultured novel archaea. Anaerobe 12:134-139. [DOI] [PubMed] [Google Scholar]
- 29.Wright, A.-D. G., A. J. Williams, B. Winder, C. T. Christophersen, S. L. Rodgers, and K. D. Smith. 2004. Molecular diversity of rumen methanogens from sheep in Western Australia. Appl. Environ. Microbiol. 70:1263-1270. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Wright, A.-D. G., P. Kennedy, C. O'Neill, A. F. Toovey, S. Popovski, S. M. Rea, C. L. Pimm, and L. Klein. 2004. Reducing methane emissions in sheep by immunization against rumen methanogens. Vaccine 22:3976-3985. [DOI] [PubMed] [Google Scholar]