Methanogen prevalence throughout the gastrointestinal tract of pre-weaned dairy calves

Abstract

The methanogenic community throughout the gastrointestinal tract (GIT) of pre-weaned calves has not been well studied. The current study firstly investigated the distribution and composition of the methanogenic community in the rumen, ileum, and colon of 3–4 week-old milk-fed dairy calves (n = 4) using 16S rRNA gene clone library analysis. The occurrence of methanogens in the GIT of pre-weaned calves was further validated by using PCR-denaturing gradient gel electrophoresis (PCR-DGGE), and quantitative real-time PCR (qPCR) was applied to quantify the methanogenic community in the rumen, jejunum, ileum, cecum, colon and rectum of 8 3–4 week old animals. Both cloning libraries and PCR-DGGE revealed that phylotypes close to Methanobrevibacter were the main taxon along the GIT in pre-weaned sucking calves. The composition and abundance of methanogens varied significantly among individual animals, suggesting that host conditions may influence the composition of the symbiotic microbiota. Segregation of methanogenic communities throughout the GIT was also observed within individual animals, suggesting possible functional differences among methanogens residing in different GIT regions. This is the first study to analyze methanogenic communities throughout the GIT of milk-fed newborn dairy calves and reveal both their diversity and abundance. The identification of methanogens in the lower GIT of pre-weaned dairy calves warrants further investigation to better define methanogen roles in GIT function and their impact on host metabolism and health.

Introduction

Symbiotic methanogens colonize the rumen of cattle, function as hydrogen (H₂) scavengers during the rumen fermentation process, and produce large amounts of methane (CH₄) through methanogenesis.¹ Consistent with known methanogen function, milk-fed calves produce no (IPCC, 2000) or very little CH₄, even when roughage was included in the diet.^2,3 Current knowledge of methanogens at the pre-ruminant stage is restricted to ovine studies. Methanogens were detected in the rumens of young lambs 2–4 d after birth, and the methanogens proliferated to a level similar to that observed in mature sheep within 7–10 d of age.^4,5 Skillman et al.⁶ examined the methanogenic community in young lambs over a longer interval (1–49 d of age) and found that the species composition of the methanogenic community changed with age, although the overall abundance remained similar. In that study, Methanobrevibacter sp. was consistently present in all animals (n = 5) throughout the study period, while Methanobacterium sp transiently colonized the rumens of only 3 of 5 animals. To our knowledge, the methanogenic community in the immature rumen and gastro-intestinal tract (GIT) of milk-fed calves has not been reported.

Methanogens also reside in the lower GIT of adult ruminants. According to Frey et al., methanogens are as abundant in the ileum as in the rumen of dairy cows, but colonize the duodenum at a much lower density.⁷ The ecology of methanogens in other GIT regions (such as the jejunum, cecum, colon and rectum), however, has not been studied.⁷ Popova et al.⁸ compared the methanogens present in rumen and cecum of lambs fed 2 different diets, and found that diet variation induced changes in the methanogenic ecology of both GIT regions. Furthermore, some studies have investigated the link between methanogens present in feces and those present within the gut community.^9,10 It remains unknown, however, whether the methanogens present in fecal samples are representative of methanogenic ecology only in the distal bovine GIT or other GIT regions. As each region of the GIT performs specific functions, the symbiotic microbiota may also differ in abundance and functions to contribute to these differing bioprocesses. However, no study has examined the existence and diversity of methanogens in the GIT of pre-weaned calves. We speculated that methanogens also colonize the GIT of milk-fed calves, and significant regional differences in methanogen diversity develop throughout the GIT.

Beyond methanogenesis, the potential functions and roles methanogens play in the GIT of ruminants have not yet been investigated. In humans and mice, methanogens were found to interact with different types of bacteria to promote digestion.¹¹ Therefore, it is potentially of great interest to investigate potential interactions between methanogens and other bacteria in the GIT of calves. It was reported that toll-like receptors (TLRs) widely present in the host cells and can detect conserved microbial molecular products, so as to impact the establishment of symbiotic bacteria within the GIT,¹² but whether TLRs also involve in recognizing methanogens in the GIT is not clear. As such, it is also worth to investigate the potential correlations between TLRs and methanogen population.

In the current study, clone library analyses, molecular fingerprinting identification and qPCR were used to characterize the methanogenic ecology in different GIT regions of milk-fed dairy calves and potential methanogen-commensal bacteria correlations were analyzed. We aimed to reveal the methanogens in the entire GIT of pre-weaned calves, to supply foundations for further studies on the functions of GIT methanogens, the interactions between methanogens and other GIT microbes and/or host, and the development of novel methane mitigation methods in the pre-weaned animals.

Results

Methanogenic community comparison among animals and GIT regions

All of the sequences obtained from the 12 clone libraries were assigned to 49 operational taxonomic units (OTUs) at species level (97% sequence similarity) (Fig 1A), which were assigned to 17 known species (Table 1). The most abundant phylotypes were close to Methanobrevibacter ruminantium, followed by Methanobrevibacter wolinii, Methanobrevibacter smithii, and Methanobrevibacter sp. AbM4 (Fig. 1A), accounting for 94.8% of the isolated sequences. Minor phylotypes detected were Methanobacterium species and Haloferax species. All of these sequences fell into 8 genera, among which Methanobrevibacter was the predominant genus, accounted for 95.3% of all retrieved sequences, while Methanobacterium, Methanopyrus, Aciduliprofundum, Haloferax, Halorubrum, and Haloterrigena were identified at very low proportions, plus 2 sequences which could not be classified within a known genus but were similar to a clone sequence (Accession number: JQ952754) identified previously (Table S1).

Figure 1.

Analyses of the obtained OTUs. (A) Major species represented by the OTUs. (B) Distribution of the OTUs among the 4 animals. (C) Distributions of the OTUs among rumen, ileum and colon.

Table 1.

Phylotype of the represented OTUs (species level)

OTU	Closest known species	Sequence similarity	% of all clones
1	Methanobrevibacter ruminantium str. M1 (NR_074117)	98%	32.3%
2	Methanobrevibacter wolinii str. SH (NR_044790.1)	99%	25.0%
3	Methanobrevibacter sp str. AbM4 (AJ550156.1)	93%	14.3%
4	Methanobrevibacter smithii 35061 str. ATCC 35061 (NR_074235.1)	98%	11.6%
5	Methanobrevibacter ruminantium str. M1(NR_074117)	98%	4.5%
6	Aciduliprofundum boonei str. T469 (NR_074217.1)	79%	2.4%
7	Methanobrevibacter ruminantium str. M1(NR_074117)	99%	1.9%
8	Methanobrevibacter ruminantium str. M1(NR_074117)	98%	0.9%
9	Methanobrevibacter smithii 35061 str. ATCC 35061 (NR_074235.1)	96%	0.7%
10	Haloterrigena sp. str. Thirty–4 (HM747061)	80%	0.6%
11	Haloferax volcanii str. DS 2; NCIMB 2012 (NR_028203)	78%	0.4%
12	Halorubrum sp str. PV6 (FJ685652.1)	79%	0.3%
13	Methanobrevibacter sp str. SM9 (AJ009958)	97%	0.3%
14	Methanobrevibacter sp. str. AbM4 (AJ550156.1)	93%	0.3%
15	Methanobrevibacter sp str. AbM4 (AJ550156.1)	94%	0.3%
16	Methanobrevibacter sp. str. AbM4 (AJ550156.1)	93%	0.3%
17	Methanobrevibacter wolinii str. SH (NR_044790.1)	97%	0.3%
18	Methanobrevibacter smithii 35061 str. ATCC 35061 (NR_074235.1)	97%	0.2%
19	Methanobrevibacter wolinii str. SH (NR_044790.1)	97%	0.2%
20	Methanobrevibacter wolinii str. SH (NR_044790.1)	92%	0.2%
21	Methanogenic archaeon DCM1 (GQ339876.1)	98%	0.2%
22	Uncultured Methanobacteriaceae archaeon clone agric10	99%	0.2%
23	Haloferax sp str. Bej51 (GU361125)	78%	0.1%
24	Haloferax sp. str. Bej51 (GU361125)	79%	0.1%
25	Methanobacterium aarhusense str. Five–4 (DQ649334)	88%	0.1%
26	Methanobacterium formicicum str. DSMZ1535 (AF169245.1)	95%	0.1%
27	Methanobrevibacter gottschalkii str. HO (NR_044789.1)	88%	0.1%
28	Methanobrevibacter gottschalkii str. HO (NR_044789.1)	79%	0.1%
29	Methanobrevibacter ruminantium str. M1(NR_074117)	98%	0.1%
30	Methanobrevibacter ruminantium str. M1(NR_074117)	96%	0.1%
31	Methanobrevibacter ruminantium str. M1(NR_074117)	98%	0.1%
32	Methanobrevibacter ruminantium str. M1(NR_074117)	81%	0.1%
33	Methanobrevibacter ruminantium str. M1(NR_074117)	97%	0.1%
34	Methanobrevibacter ruminantium str. M1(NR_074117)	97%	0.1%
35	Methanobrevibacter ruminantium str. M1(NR_074117)	89%	0.1%
36	Methanobrevibacter ruminantium str. M1(NR_074117)	98%	0.1%
37	Methanobrevibacter smithii 35061 str. ATCC 35061 (NR_074235.1)	92%	0.1%
38	Methanobrevibacter smithii 35061 str. ATCC 35061 (NR_074235.1)	97%	0.1%
39	Methanobrevibacter smithii 35061 str. ATCC 35061 (NR_074235.1)	97%	0.1%
40	Methanobrevibacter smithii 35061 str. ATCC 35061 (NR_074235.1)	90%	0.1%
41	Methanobrevibacter sp str. AbM4 (AJ550156.1)	94%	0.1%
42	Methanobrevibacter wolinii str. SH (NR_044790.1)	89%	0.1%
43	Methanobrevibacter wolinii str. SH (NR_044790.1)	96%	0.1%
44	Methanobrevibacter wolinii str. SH (NR_044790.1)	91%	0.1%
45	Methanobrevibacter wolinii str. SH (NR_044790.1)	99%	0.1%
46	Methanobrevibacter wolinii str. SH (NR_044790.1)	87%	0.1%
47	Methanobrevibacter wolinii str. SH (NR_044790.1)	95%	0.1%
48	Methanopyrus sp. (EU606020.1)	30%	0.1%
49	Natronorubrum aibiense str. Seven–3 (NR_029141.2)	80%	0.1%

The number of OTUs detected within each sample ranged from 4 to 15, and coverage (calculation details in Materials and Methods) of the obtained sequences ranged from 88.4% to 98.9% (Table 2). The distribution of the OTUs within individual animals and GIT regions were shown in Figure 1. Only one OTU (OTU1, Mbb. ruminantium str. M1, Accession No. NR_074117) was shared by all of the 4 animals and OTUs shared by 2 or 3 animals were also limited with the majority of the OTUs specific to individual animals (Fig 1B). The rumen hosted the most diversified methanogenic community (33 OTUs), whereas methanogens were much less diverse in ileum (18 OTUs) and colon (21 OTUs) (Fig. 1C). Among all the phylotypes, 9 OTUs were common to rumen, ileum and colon (OTUs 1–5, 7–9, and 15), while other OTUs were either shared by only 2 GIT regions or specific to a single GIT region (Fig. 1C). Detailed analysis of OTU distribution for each animal showed that a unique combination of the OTUs was presented in each individual (Fig. S1). The rumen hosted the most diverse community in 3 out of 4 animals (Animals 1, 5, and 8), while the ileum collected from 3 of 4 animals (Animals 1, 6, and 8) hosted the least diverse community but unique OTUs were observed within each GIT region (Fig. S1).

Table 2.

Summary of methanogenic community of each library

Animal	GIT Region	OTU numbera	Chao1 (lci-hci)b	Shannon (lci-hci)	Simpson (lci-hci)	Good's coverage	Predominant phylotype
1	RU	15	30 (18—79)	1.6 (1.3–1.9)	0.3 (0.2–0.4)	88.4%	Methanobrevibacter sp str. AbM4
	IL	7	8 (7–15)	1.3 (1.1–1.5)	0.4 (0.3–0.4)	97.5%	Methanobrevibacter sp. str. AbM4
	CO	9	12 (9–32)	1.2 (1.0–1.5)	0.4 (0.3–0.5)	95.8%	Methanobrevibacter sp str. AbM4
5	RU	7	13 (8–44)	1.1 (0.9–1.3)	0.4 (0.3–0.5)	95.2%	Methanobrevibacter wolinii str. SH
	IL	6	7 (6–17)	0.4 (0.2–0.6)	0.3 (0.7–0.9)	96.6%	Methanobrevibacter wolinii str. SH
	CO	4	4 (4–4)	0.4 (0.2–0.6)	0.8 (0.7–0.9)	98.9%	Methanobrevibacter wolinii str. SH
6	RU	9	14 (10–41)	1.3 (1.0–1.6)	0.3 (0.3–0.4)	91.9%	Methanobrevibacter ruminantium str. M1
	IL	8	14 (9–45)	1.1 (0.9–1.3)	0.5(0.4–0.6)	95.7%	Methanobrevibacter ruminantium str. M1
	CO	10	18 (11–52)	1.3 (1.0–1.5)	0.4 (0.3–0.5)	92.9%	Methanobrevibacter ruminantium str. M1
8	RU	11	17 (12–49)	1.6(1.4–1.9)	0.3 (0.2–0.4)	95.5%	Methanobrevibacter ruminantium str. M1
	IL	7	8 (7–15)	1.1 (0.9–1.3)	0.5 (0.4–0.5)	97.8%	Methanobrevibacter ruminantium str. M1
	CO	8	11 (8–33)	1.3 (1.1–1.5)	0.4 (0.3–0.5)	96.4%	Methanobrevibacter ruminantium str. M1

^aAll data were calculated with Mothur Program (V.1.31.0) at 0.03 cutoff.

^blci: low confident interval; hci: high confident interval.

Community similarity (calculation details in Materials and Methods) was compared among the 12 clone libraries (Fig. 2). Communities within the 3 GIT regions of individual animals displayed a much higher similarity than when comparing the same GIT region among different animals. Animals 1 and 5 hosted specific communities with very low similarity when compared to all other animals but animals 6 and 8 hosted communities with substantial similarity.

Figure 2.

Comparison of similarity of methanogenic communities among 4 animals. Similarities were indicated as 1-community distance as calculated in UniFrac. Methanogenic communities within colon (CO), ileum (IL), and rumen (RU) were compared among individual animals (Animal 1, 5, 6, and 8) using data generated from16S rRNA gene clone analyses. Red color indicates higher similarity; green color indicates lower similarity.

PCR-DGGE profiling of methanogens throughout the GIT of 8 animals

To further analyze the diversity and distribution of the methanogenic community throughout the GIT, samples collected from 8 animals were subjected to PCR-DGGE profiling. Methanogenic profiles clearly separated according to animal (Fig. 3A), with an R value among animals of 0.900 (P = 0.001) and pattern similarities were high within individual animals (64.1% to 95.2%). The number of bands detected within samples also differed significantly (P < 0.01) among animals (Fig. 3B) with almost twice as many methanogen phylotypes detected in Animal 6 as Animal 2. Most of the bands detected were also distributed unevenly among animals (Table 3).

Figure 3.

Fingerprinting analysis of methanogenic communities throughout the GIT of 8 animals. (A) Clustering of PCR-DGGE profiles of Animals 1 through 8 using multi-dimensional-scale (MDS) plot. (B) The number of retrieved DGGE bands was compared among animals 1 through 8 and significant differences (P < 0.05) are indicated by letters (a-d) above each bar.

Table 3.

Frequency of DGGE bands observed in individual animals and GIT regions

	Animal									GIT regions
Band	1	2	3	4	5	6	7	8	P	RUa	JE	IL	CE	CO	RE	P
1	— b	16.7c	—	—	—	—	—	—	NS	14.3d	—	—	—	—	—	NS
2	—	—	20.0	100.0	50.0	100.0	—	100	‡	57.1	28.6	42.9	50.0	50.0	57.1	NS
3	100.0	83.3	100.0	60.0	100.0	33.3	100.0	16.7	‡	100.0	85.6	71.4	75.0	50.0	57.1	NS
4	—	—	—	—	16.7	—	20.0	—	NS	28.6	—	—	—	—	—	NS
5	—	—	100.0	60.0	—	—	80.0	100.0	NS	42.9	42.9	42.9	50.0	37.5	28.6	NS
6	100.0	—	—	20.0	100.0	100.0	80.0	—	‡	42.9	57.1	57.1	50.0	50.0	42.9	NS
7	20.0	—	—	—	100.0	100.0	100.0	—	‡	42.9	57.1	42.9	37.5	37.5	28.6	NS
8	—	—	—	—	100.0	100.0	100.0	—	‡	42.9	42.9	42.9	37.5	37.5	28.6	NS
9	20.0	16.7	—	20.0	—	83.3	80.0	—	‡	28.6	42.9	28.6	25.0	25.0	14.3	NS
10	100.0	100.0	20.0	20.0	100.0	100.0	40.0	16.7	‡	85.7	71.4	71.4	50.0	50.0	57.1	NS
11	—	16.7	—	100.0	—	—	—	—	NS	28.6	—	14.3	12.5	12.5	14.3	NS
12	60.0	16.7	100.0	20.0	—	—	—	—	‡	42.9	28.6	14.3	25.0	12.5	14.3	NS
13	—	—	—	20.0	—	—	—	100.0	‡	28.6	14.3	14.3	12.5	12.5	14.3	NS
14	80.0	16.7	100.0	20.0	100.0	100.0	40.0	—	‡	71.4	71.4	57.1	50.0	50.0	42.9	NS
15	—	16.7	—	—	—	—	—	—	NS	14.3	—	—	—	—	—	NS
16	80.0	—	20.0	—	—	—	—	100.0	‡	28.6	28.6	28.6	25.0	12.5	28.6	NS
17	—	16.7	100.0	100.0	—	—	—	—	‡	42.9	14.3	14.3	25.0	25.0	28.6	NS
18	100.0	100.0	—	80.0	100.0	83.3	100.0	—	*	42.9	71.4	85.7	75.0	75.0	71.4	NS
19	—	100.0	—	20.0	—	—	—	100.0	‡	28.6	28.6	28.6	25.0	37.5	28.6	NS
20	—	—	100.0	20.0	—	—	—	—	‡	28.6	14.3	—	12.5	12.5	14.3	NS
21	—	—	100.0	20.0	100.0	100.0	100.0	83.3	‡	71.4	71.4	57.1	62.5	62.5	57.1	NS
22	—	—	—	—	100.0	100.0	—	—	‡	28.6	28.6	28.6	25.0	25.0	28.6	NS
23	—	—	—	20.0	—	—	—	—	NS	14.3	—	—	—	—	—	NS
24	—	—	100.0	20.0	—	—	—	—	‡	28.6	14.3	—	12.5	12.5	14.3	NS

^aRU: rumen; JE: jejunum; IL: ileum; CE: cecum; CO: colon; RE: rectum.

^bband not observed for the sample.

^cdnumber represents the percentage of a band observed using the current classification for all GIT regions.

^‡P < 0.01; NS Not Significant.

In contrast to the distinct grouping pattern for individual animals, methanogenic profiles did not group significantly (R value = −0.127; P = 1) among GIT regions (Fig. 4A). The similarity of profiles within each GIT regions was also low, ranging from 32.5% to 41.5%. The number of retrieved phylotypes gradually decreased from proximal to distal GIT, with significant (P < 0.05) differences identified between rumen and colon and between rumen and rectum (Fig. 4B). None of the observed bands segregated to a specific GIT region (Table 3).

Figure 4.

Finger printing analysis of methanogenic communities throughout the GIT. (A) Clustering of PCR-DGGE profiles for cecum (CE), colon (CO), ileum (IL), jejunum (JE), rectum (RE), and rumen (RU) were compared using multi-dimensional-scale (MDS) plot (N = 8). (B) The number of retrieved DGGE bands was compared among individual GIT locations and significant differences (P < 0.05) are indicated by letters (a, b) above each bar.

To further clarify the phylotypes represented by individual DGGE bands, sequencing analysis was performed with isolated DGGE bands. Among the 24 distinctive DGGE bands identified, only 9 bands were successfully sequenced (Table S3) due to low DNA recovery rates for the remaining bands. Among the 9 identified bands, 6 were classified as Methanobrevibacter species, 2 were classified as Methanosphaera species, while one was similar to an uncultured clone (Accession number: DQ445723).

Methanogen abundance among animals and GIT regions

The abundance of total methanogens varied significantly (P <0.01) among calves: Animal 2 hosted the lowest number of methanogens at 7.9 × 10⁵ copies/gram of content, while methanogen abundance in Animals 3, 5, 6, 7, and 8 ranged between 4.2 × 10⁷ to 3.9 × 10⁸ copies/gram of rumen content and was similar to that reported for adult animals (Table 4). Abundance of Mbb. sp AbM4 and Msp. stadtmanae also differed among individual animals, with Animal 1 hosting very few Mbb.sp AbM4 and Animal 2 had no detectable Msp. stadtmanae. In contrast, there were no significant differences in either total methanogen populations, or Mbb. sp AbM4, or Msp. stadtmanae when comparing GIT regions (Table 4). Correlation between methanogens and total bacteria/TLR correlations was observed for particular animals and/or specific GIT regions (Tables S3 and S4).

Table 4.

Microbial densities in GIT ingesta of individual animals and GIT regions

		Total methanogen	Mbb. sp. AbM4	Msp stadtmanae
Animal	1	(3.8±3 .2)×106a	(8.9±3 .2)× 103	(5.8±2 .0)× 104
	2	(7.9±7 .7)×105	(7.7±7 .5)× 105	— b
	3	(4.2±3 .6)×107	(1.4±1 .2)× 107	(1.9±1 .6)× 107
	4	(6.5±5 .9)×106	(4.8±2 .5)× 105	7.1×105c
	5	(1.6±0 .8)×108	(2.2±1 .0)× 107	(4.0±1 .6)× 106
	6	(2.5±0 .8)×108	(3.8±3 .2)× 107	(4.1±1 .2)× 107
	7	(3.9±1 .3)×108	(5.2±1 .6)× 107	(2.8±1 .3)× 108
	8	(7.4±2 .2)×107	(2.8±0 .8)× 107	(6.0±2 .1)× 107
	P	‡	‡	‡
Locationd	RU	(1.5±1 .2)×108	(1.0±0 .4)×107	(1.5±1 .0)×107
	JE	(8.7±4 .6)×107	(4.0±2 .2)×107	(4.4±1 .6)×107
	IL	(1.1±0 .5)×108	(1.3±0 .5)×107	(4.3±2 .2)×107
	CE	(4.3±2 .7)×107	(9.8±3 .8)×106	(4.7±3 .7)×107
	CO	(7.4±3 .3)×107	(1.9±0 .9)×107	(2.3±1 .2)×107
	RE	(2.3±1 .0)×108	(4.0±1 .4)×107	(1.7±1 .0)×108
	P	NS	NS	NS

^aLog10 transformation of 16S rRNA gene copy numbers per gram of ingesta.

^bMsp. stadtmanae were not detectable in all samples from Animal 4.

^cMsp stadtmanae was only detected in one sample from Animal 6.

^dRU: rumen; JE: jejunum; IL: ileum; CE: cecum; CO: colon; RE: rectum.

^‡P < 0.01; NS Not Significant.

Discussion

Current understanding of the functions of symbiotic methanogens and their impact on host performance in ruminants is based primarily on identifying the composition of this community in the rumen. Rumen methanogenic ecology has been extensively studied in adult dairy^13–15 and beef^16–18 cattle, sheep,^18–21 and adult deer.¹⁸ The main objective of those studies was to understand the composition of the methanogen community and identify factors altering methanogen populations, so that methane production could be reduced and animal production and/or feed efficiency improved. Research on methanogens in the undeveloped rumen has been restricted to milk-fed sheep^4–6 and fiber-fed young lambs¹ but methanogens in the undeveloped rumen of dairy or beef cattle have not been investigated. Inadequate information regarding methanogens in pre-weaned ruminants has limited further investigation of their possible roles in the immature rumen. Furthermore, little is known about methanogens in the GIT of ruminants, their functions in the GIT and their interactions with other microorganisms and host animals are not clear. A single study reported the existence of methanogens in the duodenum and ileum of adult dairy cows,⁷ while cecum methanogens in lambs have only been recently reported by Popova et al.⁸ There is no evidence concerning methanogen colonization throughout the entire GIT of young suckling calves and whether this colonization occurs as efficient as in the rumen. Thus, the current study is the first to analyze methanogenic communities throughout the GIT of milk-fed newborn dairy calves and reveal both their diversity and abundance.

To keep our analyses consistent, the same fragment of 16S rRNA gene (∼180 bp, V3 region) was targeted in both clone library and fingerprinting analyses. The high sequence coverage (88.4%-98.9%) of the clone libraries, and the comparable phylotype identification between the current study (33 OTUs and 24 DGGE bands) and previous studies (31 OTUs in Zhou et al.¹⁷; 28 DGGE bands in Zhou et al.²²; 30–46 T-RFs in Frey et al.⁷) suggested that the methods applied in the current study was sufficient to reveal the GIT methanogenic communities. The similar numbers of phylotypes identified for each examined location (OTUs vs. bands: RU 10 vs.10; IL 7 vs.7; CO 8 vs. Seven) suggested that PCR-DGGE profiling was adequate for pioneer identification, and warranted further metagenomic analysis to avoid amplication failure and to provide more comprehensive description of the methanogenic ecology throughout the entire GIT of milk-fed calves.

The phylotypic analyses showed that the majority of the OTUs resembled to Methanobrevibacter species such as Methanobrevibacter ruminantium, Methanobrevibacter sp AbM4, Methanobrevibacter smithii, and Methanobrevibacter wolinii, which were widely reported previously to be the major phylotypes present in the rumen of adult cattle using clone library analyses.^6,16,17 In addition, although the retrieved sequences from PCR-DGGE were low, the major phylotypes (Methanobrevibacter species) recovered were close to the phylotypes observed from clone libraries, which provided further evidence for the methanogenic composition within the GIT of newborn calves. Calves in the present study were provided free access to supplemental feed containing crude fiber (5.7% dry matter). This dietary supplement may have contributed to the establishment of Methanobrevibacter species throughout the GIT since Methanobrevibacter species have also been detected with a high frequency in the GIT of adult animals fed diets rich in fiber.⁸ Noticeable difference in the methanogen phylotypes recovered between clone library and PCR-DGGE was observed for Methanosphaera stadtmanae, which was absent in the clone libraries but present in the DGGE profiles of almost all animals (except Animal 2, Table 4). Since the same primers were used for both analyses, the discrepancy results for this species between the 2 assays may due to the small number of selected clones. Besides of the known methanogenic species, a small proportion of the OTUs were similar to phylotypes reported from ocean samples (e.g. Haloferax volcanii and Haloterrigena sp. Thirty–4).^23,24 However, the similarity of the representative OTUs to these type strains was lower than 80% and only a short fragment (180bp) of the gene was covered, these OTUs may belong to as yet unidentified archaea that reside in the GIT with a low prevalence. The phylogenic analyses using conventional molecular identification have revealed some technical limits and pointed out the necessity to develop better deep sequencing method to cope with these shortages.

The OTU distribution varied significantly among animals. The marked differences in methanogenic communities when comparing among different hosts are consistent with a previous report that each animal hosted a specific methanogenic community despite being maintained within the same environment.²⁵ This individual animal variation in methanogenic communities was further supported by the PCR-DGGE profiles with band patterns for each animal forming a clear cluster segregated from other animals (Fig. 4). A similar result was reported in a previous study, where distinctive methanogen PCR-DGGE profiles for individual hosts also formed distinct clusters.²⁶ Furthermore, the total methanogen populations, as well as the abundance of Methanobrevibacter sp AbM4 and Methanosphaera stadtmanae also varied markedly among different animals (Table 5). These observations are also consistent with those reported for adult beef cattle,²⁶ suggesting that distinction of methanogenic communities establishes very early in life. Diet can impose further compositional and quantitative variations on the methanogenic community,²⁶ but the recurrence of significant differences in OTU distribution, fingerprinting profiling, and quantitative measurements in the current study emphasize the importance of a host effect. The host selects symbionts that are mutually beneficial²⁷ and this appears to occur regardless of age (calves in the current study vs. adult cattle)²⁶ or diet (milk + starter in the current study vs. fiber + dried distillers grains with solubles).²⁶ Furthermore, the GIT developing speed of the animals varied noticeably (visual observation during dissection, Malmuthuge et al., unpublished data), the alteration in the symbiotic microbiota may also concurrent with the GIT development, which partially explain the high variation in methanogenic ecology. Additional validation of the host effect on selection of symbiotic methanogens may be obtained if samples were collected during the first few days of life when both environmental and dietary effects could be minimized.

Methanogenic communities present in the 3 GIT regions of newborn calves in this study were less diverse than previously reported for adult animals, suggesting that the symbiotic methanogens alter along age and become more diversified during dietary transition from milk-fed to fiber-fed.⁷Although no consistent composition was observed for individual GIT region among different hosts, the OTUs still presented differently when comparing rumen, ileum, and colon: only 9 OTUs were common to all 3 locations and these may be the core phylotypes of GIT methanogenic communities. Each GIT region hosted distinctive OTUs (Fig. 1C; Fig. S1), which may be related to the distinct digestion and absorption functions performed in each GIT region. A similar regional segregation of the microbial ecology was also reported for adult animals.⁷ The phylotypes presented in the same GIT regions of each animal may conduct similar functions, and are functionally segregated from the ones present in the other GIT regions.

As the rumen is the major organ where methanogenesis takes place, it was not surprising that the most complex methanogenic community resided in the rumen, including OTUs closer to marine phylotypes than known GIT isolates and not previously reported in adult cattle.^16,17,28 The OTU numbers retrieved from the current study of 3-month old pre-weaned calves were comparable to those reported in the rumen of 3-month old goats.²⁹ In their study, applying bromochloromethane to diets of does and kids was found to modify the methanogens in early life of goats, and the effects on reducing host methane emission persisted 3 months later in the treated kids of treated does. Their results suggest that redirecting rumen archaeal community during early age may be an effective way for methane mitigation in goats, which also deliver solid validation of the current study, which aims to supply detailed description of the archaeal communities in the pre-weaned dairy calves and to provide comprehensive information for methane mitigation in bovine.

All of the OTUs discovered in the ileum fell into known methanogenic phylotypes. Species richness of the methanogenic communities in the colon was similar to that of ileum, but diversity was different, including some non-methanogenic phylotypes (Fig. 1; Fig. S1). The lack of sequence information in previous studies of the GIT in adult dairy cattle and sheep precludes a comparison of compositional difference in adult and neonatal calf gut methanogenic communities. Results of the current study would be valuable for understanding age-related changes in these communities.⁸ Currently, the functions of methanogens in the lower GIT of ruminants are unknown. In a human study, Mbb. smithii was found to enhance Bacteroides thetaiotaomicron-mediated degradation of polyfructose-containing glycans.³⁰ This observation provided evidence that methanogens in re may be involved in host nutrient metabolism. A similar activity may also take place in the GIT of ruminants, but it remains to be determined which methanogen species may be involved in this process and how they interact with other GIT bacteria. In addition, it is not clear how the host responds to methanogen colonization and how the host selects ‘permanent’ symbionts in the GIT. Identifying the mechanisms by which cattle can select specific GIT microbiota may help develop effective methane mitigation methods. The present study suggests that this selection process begins very early in life and to identify selection mechanisms it may be important to study this early development period. Previous studies of host pathogen-recognition genes, including the TLR family, found that TLR2, in combination with either TLR6 or TLR10 recognized primarily gram-positive bacteria, while TLR4 recognized gram-negative bacteria.^31–33 Archaea containing pseudomurein in their cell walls are gram positive (Methanobacterium, Methanosarcina), whereas those containing s-layer proteins are gram negative (Methanosarcina, Methanococcus).³⁴ Therefore, a primary correlation analysis was conducted for potential interactions between methanogenic communities and TLR gene expression levels. Potential correlations between Methanobrevibacter sp. AbM4 and TLRs 1/2/6 and between Methanosphaera stadtmanae and TLR1 was observed within specific GIT regions and individual animals (Tables S3 and S4). These results suggest one possible mechanism by which the host may recognize the GIT symbiotic archaeal communities. However, there is no evidence that TLRs are directly involved in recognition of the methanogen community and whether the archaeal community plays a role in shaping development of the host innate mucosal immune system. Further investigations are necessary to explore the potential correlations revealed in the current study.

In conclusion, this is the first study to demonstrate that the methanogenic community is established throughout the GIT of milk-fed newborn calves. The methanogen diversity and abundance at 3–4 weeks of age varied markedly among individual animal, suggesting there is host-specific selection of the methanogenic community. The methanogen community in different GIT regions was also significantly different, indicating that functional differences among GIT regions may also be an important factor in selection of the symbiotic methanogen community. The identification of abundant methanogens within young animals, even in the absence of methane production, may reveal a new direction for identifying methane mitigation practices. Finally, these results also suggest that further investigations are warranted to better define the function and impact of methanogens throughout the GIT of ruminants.

Material and Methods

Animal experiment and GIT content sample collection

All animal experiment protocols were approved by the University Committee on Animal Care (University of Saskatchewan; Animal Protocol: 20020105), and all procedures were performed under the guidance of the Canadian Council on Animal Care. Healthy male Holstein calves (n = 8, Animals #1–8) were purchased from a commercial dairy farm when they were 7- to 10-days old and then reared for 2 weeks at the Vaccine and Infectious Disease Organization (VIDO, University of Saskatchewan). Calves were fed fresh, pasteurized, whole milk and had access to Blue Medallion 20% calf starter (maximum 5.7% crude fiber) (Feed-Rite, Saskatoon, SK, Canada).

Calves were humanely euthanized using an intravenous injection of Euthanyl® (240 mg mL⁻¹; Bimeda-MTC Animal Health Inc.., Cambridge, ON, Canada) and tissues were collected within 15 minutes. Tissue samples from each GIT location were cut into 5 mm² pieces and placed in 5× volume of RNAlater®. Ingesta adjacent to the tissue sample site in rumen, jejunum, ileum, cecum, colon, and rectum were also collected and mixed with a 5× volume of RNAlater®. All samples were stored at −80°C until processed.

DNA extraction

Total DNA was extracted from each ingesta sample (∼100 mg) using the bead-beating method as described by Li et al.³⁵ Briefly, samples were physically disrupted in a BioSpec Mini Beads Beater 8 (BioSpec, Bartlesville, OK, USA) at 4800 rpm for 3 min, followed by a phenol/chloroform/isoamyl alcohol (25:24:1) extraction. DNA was precipitated with cold ethanol and dissolved in 30 μL of nuclease-free water. The DNA concentration and quality were measured with ND1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA).

Construction of 16S rRNA gene libraries, sequencing, and phylogenic analysis

Although pyrosequencing has been widely used to study the symbiotic microbiota in the GIT of ruminants, the existing available primer sets^17,36–38 for pyrosequncing failed to generate products for some lower GIT samples. Long primers with barcodes or linkers may interfere with the binding efficiency to the template DNAs, particularly the lower GIT samples. Therefore, the current study used clone library and fingerprinting analyses to verify our hypothesis that methanogens exist throughout the GIT of milk-fed calves. Four animals (Animals 1, 5, 6, and 8) were randomly selected, and DNA samples from rumen, ileum (representing small intestine), and colon (representing large intestine) were subjected to 16S rRNA gene clone library analyses to assess methanogen diversity. The partial 16S rRNA gene (∼180 bp, V3 region) was amplified with the archaeal universal primer pair 344F/519R (344F, 5′-ACGGGGYGCAGCAGGCGCGA-3′; 519R, 5′-GWATTACCGCGGCKGCTG-3′)),³⁹ using the following program: primary initiation at 94°C for 5 min; 30 cycles of 94°C for 30 s, 57°C for 30 s, 68°C for 30 s; and a final elongation at 68°C for 7 min. Each reaction mixture (50 μl) contained 20 pmol of each primer, 10 mM deoxynucleoside triphosphate, 2.5 U Taq polymerase (Invitrogen, Calsbad, CA), 1× PCR buffer, 50 mM MgCl₂, and 50 ng of DNA template. The amplicons were cloned into pCR®2.1-TOPO® vector (TOPO TA cloning kit; Invitrogen) using chemical transformation. Colony selection, plasmid extraction and sequencing reaction were performed following Zhou et al.¹⁷ Approximately 100–150 clones were selected from each library, and the obtained sequences were analyzed using the Mothur program (Mothur v. One.31.0, http://www.mothur.org/wiki/Main_Page) by comparing the operational taxonomic units (OTUs) determined at 97% sequence similarity.⁴⁰ Methanogenic community diversity estimation such as Chao1, Shannon, and Simpson indices were calculated, and Good's coverage estimator was used to estimate what percent of the total species is represented in a sample. Distance of the communities were estimated using UniFrac and similarity between communities = 1 - distance.⁴¹

PCR-DGGE and sequencing analysis

To further explore methanogen diversity throughout the GIT rapid finger printing analysis was employed. DNA isolated from tissue samples collected from all 8 animals was diluted to a final concentration of 50 ng μL⁻¹. Touchdown PCR was conducted to amplify the V3 region of methanogen 16S rRNA gene with the same primer pair as clone library analysis, with a GC-clamp (5′-CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCACGGGGGG-3′) added to the 5′ end of 344F using the following program: primary initiation at 94°C for 4 min; 20 cycles of 94°C for 30 s, 60°C (−0.5°C cycle⁻¹) for 30 s, 72°C for 30 s; 25 cycles of 94°C for 30 s, 50°C for 30 s, 72°C for 30 s; and a final elongation at 72°C for 7 min.

The PCR products were separated on a 6% acrylamide gel with a 35–45% denaturing gradient using the Bio-Rad DCode Universal Mutation Detection System (Bio-Rad laboratories, Inc.., Hercules, CA, USA) at 130 V at 60°C for 4 h. The gels were stained with ethidium bromide for 15 min, washed with water, and inspected with the FluorChem^TM SP imaging system (Alpha Innotech, San Leandro, CA, USA).

DGGE patterns were analyzed using the Bionumerics® software package version 6.1 (Applied Maths, Austin, TX, USA) with 0.8% position tolerance and 0.5% optimization for all generated profiles. These parameters were calculated by the software and are specific to this particular experiment. A reference ladder including 8 different previously identified methanogen phylotypes was loaded into each gel for normalization among gels.²⁵ The similarity of the DGGE patterns was analyzed using the unweighted pair group method with arithmetic mean (UPGMA) and indexed using the average Dice coefficient (D_sc). The Dice coefficient was based on the presence and/or absence of the band using the formula S_D = 2N_AB/(N_A⁺N_B). All of the DGGE profiles were further plotted using the multi-dimensional scaling (MDS) method within BioNumerics.

All of the obtained DGGE bands were excised from the gel, and DNA retrieved from each band was subjected to cloning and sequencing analyses using the TOPO TA cloning kit (Invitrogen, Carlsbad, CA) and the ABI PRISM BigDye Terminator v3.1 cycle sequencing kit (Applied Biosystems, Foster City, CA). Sequences (∼180 bp) were recovered with the ABI 3730 sequencing system, and similarities of retrieved sequences and known taxons were compared using BLAST from the NCBI database.

Quantitative real-time PCR (qPCR)

Densities of total methanogens, Mbb. sp AbM4, and Msp. stadtmanae from each sample were estimated with qPCR amplification of 16S rRNA gene copies using a Fast SYBR® Green Master Mix kit (Applied Biosystems, Foster City, CA, USA) in a StepOnePlus^TM Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Msp stadtmanae was selected due to its presence in both human and mouse GIT, and Mbb. sp. AbM4 was selected due to a reported role in the variation of host feed efficiency.¹⁷ All experimental protocols and calculations followed previous study.¹⁵

The total bacteria population of each sample of the current study was measured by Malmuthuge et al.⁴² and toll-like receptor (TLRs) expressions of the same samples were measured by Malmuthuge et al.⁴³ These data were subjected to the correlation analysis of the current study.

Statistical analysis

All statistical analyses were completed with SAS software (SAS System version 9.2, SAS Institute, Cary, NC) and R software package (URL: http://www.R-project.org, 2013).⁴⁴ All PCR-DGGE patterns were converted to ‘1/0’ categories and analyzed using a PROC CATMOD model in SAS to identify the host effect on band distributions. The similarities of the patterns were analyzed using the program ‘analysis of similarity’ (ANOSIM) in R. The DGGE band counts and methanogen populations (log10 transformation) were compared to determine both regional and animal effects using analysis of variance (ANOVA) model in SAS. The correlations among methanogens, bacteria, and TLRs expressions were analyzed using PROC CORR model in SAS. Significance was determined at P <0.05, and tendency was determined when 0.05 > P < 0.1.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors would like to thank P. Fries for supplying ingesta samples and G. Vavaak for conducting all the DNA extractions. The authors would also like to extend acknowledgments to N. Malmuthuge for providing bacterial population data.

Funding

This work was supported by Grant #2008F079R through the Agriculture and Food Council, Alberta Livestock Industry Development Fund, Alberta Milk, Alberta Livestock and Meat Agency Ltd, and a National Science Engineering Research Council Discovery Grant. Dr. Philip Griebel is funded by a Tier I CRC in Neonatal Mucosal Immunology provided by the Canada Institutes for Health Research (CIHR).

Supplemental Material

Supplemental data for this article can be accessed on the publisher's website.