Enteric methane emissions in response to ruminal inoculation of Propionibacterium strains in beef cattle fed a mixed diet

Additional keywords: beef, methane, Propionibacterium.

Some strains of Propionibacteria are natural propionate producers that inhabit the rumen and comprise 1.4% to 4.3% of the total microbial population (Mead and Jones 1981). The development of Propionibacterium strains as direct-fed microbials could offer an effective means of increasing ruminal propionate production and reducing enteric methane (CH₄) emissions from cattle fed forage-based diets (Jeyanathan et al. 2014). Previous studies explored the potential to use Propionibacterium strains to mitigate CH₄ emissions in beef cattle fed high-forage (Vyas et al. 2014a) and high-grain diets (Vyas et al. 2014b). However, no effects were observed on total CH₄ emissions due to low persistency of the inoculated strains. Contrary to the previous in vivo studies (Vyas et al. 2014a, 2014b), in vitro batch culture studies with Propionibacterium strains showed promising results with significant reduction in CH₄ production using both high-forage and high-grain diets (Alazzeh et al. 2013). The discrepancy between studies might be related to different strains of Propionibacteria used in the in vitro batch culture experiment compared with the in vivo study. The efficacy of Propionibacterium strains identified using an in vitro batch culture experiment as having CH₄ mitigation potential (Alazzeh et al. 2013) needs to be validated in vivo before such a strategy can be recommended for lowering CH₄ emissions from beef cattle. Hence, the primary objective of this study was to confirm in vivo the efficacy of Propionibacterium strains previously identified in vitro as having the potential to mitigate CH₄ emissions in beef cattle.

The protocol for the study was approved by Lethbridge Research Centre Animal Care Committee before the experiment began and animals were cared for according to the guidelines of the Canadian Council on Animal Care (1997). Sixteen ruminally cannulated crossbred beef heifers were used in this study. The heifers were grouped on the basis of pre-experimental bodyweight (mean ± s.d.: Group 1 = 602 ± 31 kg, Group 2 = 570 ± 60 kg, Group 3 = 590 ± 50 kg and Group 4 = 620 ± 27 kg). Dietary treatments included: (1) Control, (2) Propionibacterium freudenreichii T114, (3) P. thoenii T159, and (4) P. freudenreichii T54. Strains were grown daily in sodium lactate broth under anaerobic conditions according to the method described earlier (Alazzeh et al. 2013) and were administered daily (1 × 10¹¹ colony forming units) at the time of feeding directly into the rumen. Treatments were randomly allotted within each group. All heifers were fed the basal diet (60 : 40 forage to concentrate [(dry matter (DM) basis); Table 1] formulated to provide adequate metabolisable energy and protein for 600 kg growing beef cattle with an average daily gain of 1 kg/day (NRC 2000). Heifers were fed for ad libitum intake once daily at 1300 hours, housed in a ventilated tie-stall barn, and exercised daily in an open dry lot. Diets were supplemented with melengesterol acetate (1.3 mg/head.day; MGA-100 premix, Pfizer Animal Health, Pfizer Canada Inc., Kirkland, QC, Canada) to suppress oestrus and prevent ovulation in the beef heifers.

Table 1.  Ingredient and chemical composition of the total mixed ration and the melengestrol acetate (MGA) supplement

During the experiment, Days 1 to Day 14 were used to adapt heifers to their treatments. Ruminal contents were collected on Day 15 and Day 18 at 0, 3, 6 and 9 h post feeding. Ruminal pH was measured continuously from Day 15 to Day 21, and enteric CH₄ emissions were measured from Day 19 to Day 21. Daily intakes and orts of the diets for individual heifers were recorded. Diets and orts were sampled daily during days of CH₄ measurement and were pooled for each animal at the end of the period of CH₄ measurement. Dietary ingredients were sampled once weekly and analysed for DM by drying at 55°C for 72 h. Ingredients and total mixed ration samples were stored at −20°C until analysed.

Methane emissions were measured from individual heifers for 3 days using environmental chambers as described earlier (Beauchemin and McGinn 2006). Briefly, chambers were calibrated before and after each period by sequentially releasing 0, 0.2, and 0.4 L/min of CH₄ (Praxair Canada Inc., Mississauga, ON, Canada) separately into each empty chamber using a mass-flow meter (Omega Engineering, Stamford, CT, USA). A 3-point regression was developed by plotting actual against calculated CH₄ emission (R² = 0.99). The slopes of these best fit linear relationships were used to correct for between-chamber variability. Conditions of air circulation and sampling procedures were as described by Avila-Stagno et al. (2013).

To determine the effect of Propionibacteria on ruminal pH, daily pH profiles were measured starting at feeding on Day 15 using an indwelling pH data acquisition system (LRC pH dataloggers, Dascor, Escondido, CA, USA) that was retained in the rumen for 7 days (includes the period of CH₄ measurement). The system was standardised using pH 4 and 7 buffers before insertion on the first day and then upon removal on the last day as described earlier (Penner et al. 2006). On Days 15 and 18, at 0, 3, 6, and 9 h post feeding, ruminal contents were sampled from four different sites (cranial, caudal, ventral, and dorsal sacs), composited and strained through a double layer of polyester monofilament fabric (Pecap 7–1180/59, mesh opening 1180 µm, Tetko Inc., Scarborough, ON, Canada). Two samples of filtered ruminal fluid (5 mL) were preserved by adding 1 mL of 25% (wt/vol) phosphoric acid for volatile fatty acids (VFA) and lactate determination, and 1 mL of 1% (wt/vol) sulfuric acid for ammonia-N (NH₃-N) determination. The samples were stored at −20°C until analysed.

Rumen samples collected at 0, 3, and 9 h were processed for microbial analysis, separately for each heifer. Microbial pellet was extracted based on a method described earlier (Vyas et al. 2014a). Quantitative real-time PCR assays were performed with a 7900 HT Fast Real-time PCR system (Applied Biosystems, Foster City, CA, USA) using POWER SYBR Green PCR Master Mix (Applied Biosystems, Warrington, UK), forward and reverse primers (500 nM of each primer/reaction), and ~20 ng of template DNA in a final volume of 25 μL per reaction. The primers for universal bacteria (forward primer: 5ʹ-TCCTACGGGAGGCAGCAGT-3ʹ; reverse primer: 5ʹ-GGACTACCAGGGTATCTAATCCTGTT-3ʹ; Nadkarni et al. 2002) and total Propionibacteria (forward primer: 5ʹ-RGTGGCGAAGGCGGTTCTCTGGA-3ʹ; reverse primer: 5ʹ-TGRGGTCGAGTTGCAGACCCCAAT-3ʹ; Rossi et al. 1999) were used. Amplifications were performed under the conditions described earlier (Vyas et al. 2014a). The relative population size of total Propionibacteria was determined as the ratio of the amplification of total Propionibacteria 16S rRNA to the amplification of the universal bacteria. PCR efficiency was calculated using the formula E = [10(–1/slope) – 1]. The slopes ranged from –3.37 to –3.40 for total bacterial primer and –3.29 to –3.30 for total Propionibacteria.

Dry matter for all samples was determined by oven drying at 55°C for 72 h. Dried samples were ground in a Wiley mill (A. H. Thomas, Philadelphia, PA, USA) through a 1-mm screen. Analytical DM content of the ground sample was determined by drying at 135°C for 2 h (method 930.15; AOAC 2005), followed by hot weighing. The organic matter content was calculated as the difference between 100 and the percentage ash (method 942.05; AOAC 2005). The neutral detergent fibre and acid detergent fibre contents were determined according to Van Soest et al. (1991) with heat stable amylase and sodium sulfite used in the neutral detergent fibre procedure. Samples were ground using a ball mill (Mixer Mill MM2000, Tetsch, Haan, Germany) for the determination of crude protein. Total N was quantified by flash combustion and thermal conductivity detection (Carlo Erba Instuments, Milan, Italy). Ruminal VFA, lactate and NH₃-N concentration were quantified as described earlier (Vyas et al. 2014a). Gross energy concentration was determined using a bomb calorimeter (model E2k, CAL2k, Johannesburg, South Africa).

The data were analysed using the MIXED procedure of SAS with heifer as the experimental unit. For data that were collected serially [DM intake (DMI), CH₄, and ruminal fermentation] the model included the fixed effect of treatment, sampling time and their interaction, with sampling time considered as a REPEATED effect in the model. Group was used in the RANDOM statement. Variance components were estimated by the restricted maximum likelihood method. Kenward–Roger’s option was used in the MODEL statement to estimate denominator degrees of freedom. Time-series covariance structure was modelled using the options of autoregressive order-one, compound symmetry, and unstructured order-one. Best time-series covariance structure for each variable was selected based on lowest Akaike and Bayesian information criteria. CONTRAST statement was used to evaluate differences between means of Control and Propionibacterium treatments. Data are presented as least-squares means ± standard error of the means. Statistical significance was declared at P ≤ 0.05 and trends are discussed at P ≤ 0.10.

No treatment effects were observed on DMI (P = 0.90; Table 2) or ruminal pH variables. Likewise, total VFA production was similar across all treatments (P = 0.44; Table 3). Ruminal acetate proportion was reduced with Propionibacterium T159 (P = 0.02) whereas no effects were observed with other strains. Correspondingly, no treatment effects were observed on proportion of ruminal propionate (P = 0.12). The proportion of ruminal isobutyrate (P < 0.01) and acetate : propionate ratio was increased (P = 0.04) with Propionibacterium T114. Ruminal NH₃-N concentration was similar for all the treatments (P = 0.79).

Table 2.  Dry matter intake (DMI) and ruminal pH for beef heifers fed a mixed diet supplemented with Control or Propionibacterium strains T114, T159 or T54^A

Table 3.  Ruminal fermentation characteristics of beef heifers fed a mixed diet supplemented with Control or Propionibacterium strains T114, T159 or T54^A
a, b values within a row with different letters differ (P ≤ 0.05)

No treatment differences were observed for DMI on the days of CH₄ measurement in chambers (P = 0.65; Table 4). However, DMI in chambers was reduced by 5–21% compared with DMI measured during the metabolism experiment, with a greater decline observed in the animals receiving Propionibacterium strains, primarily Propionibacterium T159. Total enteric CH₄ production was not affected by treatments and averaged 188 g/day (P = 0.51). Methane yield adjusted for DMI (P = 0.19) and gross energy intake (P = 0.17) were similar across all the treatments. However, contrary to our hypothesis, CH₄ yield adjusted for DMI and gross energy intake tended to increase when means were compared between Control and Propionibacterium treatments (P = 0.08). The numerical differences in total CH₄ emissions with the inoculation of Propionibacterium strains were driven by changes observed during the initial 0–10 h post feeding (Fig. 1).

Table 4.  Enteric methane (CH₄) emissions from beef heifers housed in chambers and fed a mixed diet supplemented with Control or Propionibacterium strains T114, T159 or T54^A
DMI, dry matter intake; GE, gross energy

Fig. 1.  Total methane emissions post feeding in beef heifers fed a mixed diet supplemented with Control or Propionibacterium strains T114, T159 or T54. Propionibacterium strains T114, T159 and T54 (1 × 10¹¹ colony forming units) were administered in the rumen daily at the time of feeding.

Inoculation of Propionibacterium T159 increased the relative abundance of total Propionibacteria probably due to the greater prevalence of the respective strain (P = 0.04; Fig. 2). However, no effects were observed on the relative abundance of total Propionibacteria with the inoculation of other strains. Relative abundance of total Propionibacteria was not affected by sampling time (P = 0.27).

Fig. 2.  Relative abundance^A of total ruminal Propionibacteria at 0, 3 and 9 h post feeding in beef heifers fed a mixed diet supplemented with Control or Propionibacterium strains T114, T159 or T54^B,C. ^ARelative abundance was determined as the ratio of copies of total Propionibacteria to copies of total bacteria and expressed as percentage. ^BTreatment effect: Control, 0.37^a; T114, 0.87^a; T159, 1.96^b; T54, 0.84^a; s.e.m. 0.36 (^a,b Values with different letters differ; P ≤ 0.05). ^CPropionibacterium strains T114, T159 and T54 (1 × 10¹¹ colony forming units) were administered in the rumen daily at the time of feeding.

Strategies to mitigate CH₄ emissions in cattle fed a mixed diet are desirable as emissions are higher from feedlot cattle during the growing, as compared with the finishing, phase of beef production (Beauchemin and McGinn 2005). Recently, the role of Propionibacterium species in reducing CH₄ emissions was explored in beef cattle fed a high-forage diet (Vyas et al. 2014a) and a high-grain diet (Vyas et al. 2014b); however, inoculated strains failed to increase ruminal propionate proportion and mitigate total CH₄ emissions. It is possible that the lack of effect of Propionibacterium species in those studies was due to the strains selected; thus, the present study examined additional Propionibacterium strains. The Propionibacterium strains used in the present study were previously screened for their CH₄ mitigation potential in vitro using both forage- and grain-based diets (Alazzeh et al. 2013), unlike in the studies of Vyas et al. (2014a, 2014b). Animals used in the present study had no previous exposure to Propionibacterium strains thereby ruling out the possibility of any carry-over effects that might have confounded results in the present study.

The present in vivo study showed no significant treatment effects on total CH₄ production for any of the strains used, in contrast to observations from a previous in vitro study (Alazzeh et al. 2013). Moreover, total enteric CH₄ emissions corrected for DM and gross energy intake tended to be greater with the inoculation of Propionibacterium strains. The results observed in the present study are contrary to the suppression of CH₄ yield observed earlier with Propionibacterium strains inoculated under similar dietary conditions (Vyas et al. 2014a). The inconsistency between results might be attributed to the use of different strains or species of Propionibacterium across the two studies and their differential effects on DMI under stressful conditions when animals were in chambers. For some unknown reason, in the present study the drop in the DMI in chambers was more prominent for animals inoculated with Propionibacterium strains than Control cattle. Given that intake affects ruminal passage rate of digesta (Sniffen et al. 1992), reduced intake with Propionibacterium strains might have increased retention time of substrates in the rumen resulting in greater fermentation and thereby greater CH₄ emissions. A similar inverse relationship between CH₄ production and ruminal passage rates was observed previously where CH₄ production was decreased by 29% with 63% increase in the fractional passage rate of particulate matter in steers (Okine et al. 1989).

The absence of treatment effects on total CH₄ emissions observed in a previous study by Vyas et al. (2014a) was attributed to the lack of survival and persistence of inoculated Propionibacterium strains as the abundance of the inoculated bacteria returned to pre-treatment levels within 9 h post inoculation. In contrast, in the present study, the relative abundance of total Propionibacteria was greater with the inoculation of Propionibacterium T159, with greater levels of abundance at every time point post inoculation, relative to the Control. Discrepancy between studies might be due to the use of different Propionibacterium strains and better adaptability of Propionibacterium T159 to ruminal conditions in animals fed mixed diets. The lack of survival and persistence of Propionibacterium strains in the rumen observed previously (Vyas et al. 2014a) was attributed to absence of ruminal lactate, a preferred substrate for the growth of Propionibacterium spp. Although ruminal lactate was not detectable in the present study, better persistence of Propionibacterium T159 could have been due to utilisation of alternative substrates including glucose as well as amino acids to produce propionate and acetate (Piveteau 1999). The presence of metabolically active Propionibacterium T159 might have accounted for the reduced molar proportion of ruminal acetate; however, lack of significant effect on total CH₄ emissions might be attributed to the inefficacy of Propionibacterium T159 to increase ruminal propionate.

Contrary to the effects on VFA profile observed with Propionibacterium T159, inoculation of Propionibacterium T114 increased acetate : propionate ratio. The relative abundance of total Propionibacteria in the rumen of heifers inoculated with Propionibacterium T114 suggested a lack of persistence of the inoculated strain making it difficult to explain the induced changes in ruminal VFA profile. It is possible that the relative abundance of total Propionibacteria with the inoculation of Propionibacterium T114 was below the detection limit, yet sufficient enough to influence and induce corresponding changes in ruminal VFA profile.

The lack of response on ruminal fermentation and CH₄ emissions with the inoculation of Propionibacterium T54 might be attributed to the absence of metabolically active Propionibacterium T54 given that there was no significant increase in relative abundance of total Propionibacteria post inoculation of the respective strain. The effects on ruminal fermentation and CH₄ emissions with Propionibacterium T114 and T54 are contrary to the in vitro results observed earlier (Alazzeh et al. 2013). The discrepancy between studies could be attributed to the use of different methods for studying ruminal fermentation. The in vivo method used in the present experiment is more representative of the biological system as compared with the in vitro batch culture experiment used earlier (Alazzeh et al. 2013).

It should also be acknowledged that failure to detect numerical differences observed on total CH₄ emissions and CH₄ yield could also be attributed to the lack of statistical power of the experiment. Hence, results observed from the present study might definitively dismiss the role of Propionibacterium strains on mitigating CH₄ emissions; however, future studies with greater replication are required to validate the results observed in this study.

In conclusion, Propionibacterium strains induced changes in ruminal VFA profile but failed to elicit significant treatment differences in total CH₄ emissions probably due to the inability of Propionibacterium strains to significantly alter ruminal propionate concentrations. When results of this study are examined together with previous in vivo studies, it appears that supplementing cattle diets with Propionibacterium has limited potential to mitigate enteric CH₄ emissions.