Accumulation of trans C_18:1 Fatty Acids in the Rumen after Dietary Algal Supplementation Is Associated with Changes in the Butyrivibrio Community

Abstract

Optimization of the fatty acid composition of ruminant milk and meat is desirable. Dietary supplementation of algae was previously shown to inhibit rumen biohydrogenation, resulting in an altered milk fatty acid profile. Bacteria involved in biohydrogenation belong to the Butyrivibrio group. This study was aimed at relating accumulation of biohydrogenation intermediates with shifts in Butyrivibrio spp. in the rumen of dairy cows. Therefore, an experiment was performed with three rumen-fistulated dairy cows receiving a concentrate containing algae (9.35 g/kg total dry matter [DM] intake) for 20 days. Supplementation of the diet with algae inhibited biohydrogenation of C_18:2 omega 6 (n-6) and C_18:3 n-3, resulting in increased concentrations of biohydrogenation intermediates, whereas C_18:0 decreased. Addition of algae increased ruminal C_18:1 trans fatty acid concentrations, mainly due to 6- and 20-fold increases in C_18:1 trans 11 (t11) and C_18:1 t10. The number of ciliates (5.37 log copies/g rumen digesta) and the composition of the ciliate community were unaffected by dietary algae. In contrast, supplementation of the diet with algae changed the composition of the bacterial community. Primers for the Butyrivibrio group, including the genera Butyrivibrio and Pseudobutyrivibrio, were specifically designed. Denaturing gradient gel electrophoresis showed community changes upon addition of algae without affecting the total amount of Butyrivibrio bacteria (7.06 log copies/g rumen DM). Clone libraries showed that algae affected noncultivated species, which cluster taxonomically between the genera Butyrivibrio and Pseudobutyrivibrio and might play a role in biohydrogenation. In addition, 20% of the clones from a randomly selected rumen sample were related to the C_18:0-producing branch, although the associated C_18:0 concentration decreased through supplementation of the diet with algae.

Rumen biohydrogenation is the microbial saturation of dietary unsaturated fatty acids, which limits the availability of health-associated polyunsaturated fatty acids (PUFA) in ruminant meat and milk. Therefore, understanding the ruminal biohydrogenation process is important to generate healthier ruminant products. Interest in research on microorganisms involved in rumen biohydrogenation of linoleic (C_18:2 omega 6 [n-6]) or linolenic (C_18:3 n-3) acid is currently growing. Polan et al. (26) associated rumen biohydrogenation activity with a Butyrivibrio fibrisolvens strain. Since then, several researchers have confirmed the active role of Butyrivibrio species in the partial or complete biohydrogenation of unsaturated C₁₈ fatty acids (19). Bacteria involved in C₁₈ biohydrogenation are grouped in the “Butyrivibrio group,” which includes the genera Butyrivibrio and Pseudobutyrivibrio and the species Clostridium proteoclasticum (24). The last was recently proposed for reclassification as Butyrivibrio proteoclasticus (21), and hereafter, we refer to C. proteoclasticum as B. proteoclasticus. Although isolation and pure-culture studies have provided fundamental insight into the bacteria involved in rumen biohydrogenation, the relative importance of individual strains in the in vivo ruminal lipid metabolism remains largely unclear (25). Indeed, complete in vivo biohydrogenation is thought to be a synergistic process involving a consortium of bacteria, each having a share in the conversion of unsaturated to more saturated fatty acids (15).

Marine products, such as fish oil and algae, proved to possess high effectiveness in the inhibition of rumen biohydrogenation of unsaturated fatty acids (5, 30). The long-chain PUFA eicosapentaenoic acid (EPA) and/or docosahexaenoic acid (DHA) was found to be the active compounds in this process (1, 7). Their supplementation reduced C_18:0 production, resulting in the accumulation of various hydrogenation intermediates, predominantly C_18:1 trans 11 (t11) and C_18:1 t10. Incomplete biohydrogenation, induced by dietary algae, was found to be associated with the disappearance of some ciliates (5). Based on this study, we hypothesized that ciliates and/or their associated bacteria could play a role in rumen biohydrogenation (5). However, further research on rumen biohydrogenation of C_18:2 n-6 by pure Isotricha prostoma and its associated bacteria showed only minor biohydrogenation by the protozoal species, its bacterial symbionts, or a mixture of both (6).

Since the majority of bacteria are as yet unculturable and in vitro findings do not always reflect in vivo mechanisms, the current in vivo research with DHA-enriched microalgae is aimed at studying mutual changes in the rumen C₁₈ biohydrogenation and the composition of the rumen microbial community through cultivation-independent techniques. More specifically, a PCR-denaturing gradient gel electrophoresis (DGGE) and real-time quantitative PCR (qPCR) method was optimized for the Butyrivibrio group to examine the time-dependent shifts in rumen Butyrivibrio species and rumen C_18:1 trans accumulation following supplementation of the diet with algae.

MATERIALS AND METHODS

Animals and diets.

Three rumen-cannulated (10-cm inside diameter; Bar-Diamond Inc., Parma, ID) Holstein-Friesian cows (612 ± 32 kg body weight) in mid-lactation (172 ± 45 days in milk) were used to evaluate the effect of feeding a concentrate containing algae on both the rumen fatty acid composition and the rumen microbial population. The experimental setup involved a period without supplementation of the diet with algae, with rumen sampling 2 days before the start of the supplementation (day −2) and 20 days of dietary algal supplementation with samplings on days 6, 13, and 20. The basal diet consisted of a mixed ration based on grass silage, corn silage, wheat, straw, rapeseed meal, soybean meal, and cane molasses. The basal diet was supplemented with 2 kg of concentrate incorporating no or 11% (on a concentrate-product basis) DHA-enriched microalgae (Schizochytrium sp.; DHA Gold; Martek Biosciences Corp., Colombia, MD). The concentrate was offered in two equal portions, after the morning (0900 h) and after the evening (2100 h) milking. Following 15 min of feeding with the concentrate containing algae, refusals were placed into the rumen via the fistula to ensure each cow had its allotment. This was required only on the first day of the algal-concentrate feeding. After concentrate feeding, the basal diet was offered ad libitum (30 kg fresh matter after both morning and evening concentrate feedings). Grass silage, corn silage, and concentrate were sampled four times during the experimental period and stored frozen. Samples were freeze-dried and pooled by feed component prior to fatty acid analysis. The ingredient, chemical, and fatty acid compositions of both diets are shown in Table 1.

TABLE 1.

Ingredient, chemical, and fatty acid composition of the diet before (day −2) and during algal-concentrate feeding

Parameter	Value
	Day −2	ALGa
Ingredients (g/kg DM)
Grass silage	412	401
Corn silage	281	273
Wheat	83.7	81.3
Extracted rapeseed meal	62.7	61.0
Soybean meal	62.7	61.0
Straw	21.6	21.0
Standard concentrate	75.7	16.2
Algal concentrate	−b	85.0
Chemical composition (g/kg DM)c
OM	930	933
Crude protein	158	160
Crude fat	30.9	34.4
Crude fiber	197	189
NDF	398	383
ADF	223	213
Fatty acid composition (g/kg DM)
C12:0	1.15	0.82
C14:0	0.43	0.90
C16:0	3.14	4.54
C18:0	0.45	0.45
C18:1c9	3.88	3.56
C18:2n-6	6.32	6.58
C18:3n-3	4.14	4.30
C22:6n-3	−	2.09
Total fatty acids	21.4	26.4

^aALG, algal-concentrate feeding.

^b−, not present.

^cOM, organic matter; NDF, neutral detergent fiber; ADF, acid detergent fiber.

Sampling.

Ruminal digesta were collected from each animal on days −2, 6, 13, and 20 and sampled just before (0 h) and 1, 2, 4, and 6 h after the morning feeding. Equal amounts of ruminal digesta were collected by hand from the front and middle of the ventral sac and from the cranial sac. After thorough mixing, an aliquot of 10 ml was stored at −80°C until DNA extraction was performed. Another aliquot of approximately 250 g was freeze-dried prior to fatty acid analysis.

Rumen fatty acid analysis.

Fatty acids in freeze-dried ruminal digesta (2.5 g) were extracted with chloroform/methanol (2/1 [vol/vol]) as described by Chow et al. (11). Tridecanoic acid (10 mg; Sigma, Bornem, Belgium) was added as an internal standard. The extracted fatty acids were methylated as described by Raes et al. (27) with NaOH/MeOH (0.5 mol/liter; 3 ml), followed by HCl/MeOH (1/1 [vol/vol]; 2 ml). The FAME were extracted with 3 and 2 ml hexane, pooled, and evaporated to dryness under N₂. The residue was dissolved in 1 ml hexane and analyzed by gas chromatography (HP 6890; Agilent Technologies, Brussels, Belgium) on a CP-Sil88 column for FAME (100 m by 250 μm by 0.2 μm; Chrompack, Middelburg, The Netherlands) (27). FAME were identified using external standards (S37 [Supelco, Poole, Dorset, United Kingdom]; conjugated linoleic acid [CLA] cis 9 [c9] t11, CLA t10c12, and odd- and branched-chain fatty acids [Larodan Fine Chemicals AB, Malmö, Sweden]) and quantified using the internal standard. Some C₁₈ fatty acids were identified according to the elution sequence reported by Ratnayake (28) and Shingfield et al. (31).

Bacterial strains, growth conditions, and DNA extraction.

B. fibrisolvens DSM 3071 (10) was purchased from the German collection of microorganisms and cell cultures (DSMZ, Braunschweig, Germany). B. fibrisolvens MDT-5 (13) was kindly donated by T. Hino (Meiji University, Kawasaki, Japan). The bacteria were anaerobically cultured for 64 h at 39°C in a rumen medium containing glucose, cellobiose, maltose, and starch as carbon sources (medium 330; DSMZ, Braunschweig, Germany). Bacterial DNA (2 ml of growth medium) and total DNA of ruminal digesta (0.5 g), sampled before the morning feeding (0 h), were extracted following the method of Boon et al. (9). DNA extracts of Escherichia coli DH5α (Invitrogen, Merelbeke, Belgium), Lactobacillus brevis LMG 7761 (38), and Bifidobacterium animalis LMG 11580 (4) were used as negative controls. The Lactobacillus and Bifidobacterium strains were purchased from the Belgian culture collection (BCCM/LMG, Gent, Belgium).

Ciliate, bacterial, and Butyrivibrio PCRs.

A nested-PCR approach was used to amplify a fragment of the 18S rRNA genes of ciliates for DGGE according to the method of Boeckaert et al. (5). General bacterial PCR for DGGE was performed as described by Boon et al. (8) using the bacterial primers P338F-GC and P518r. It should be noted that research performed by Huws et al. (17) indicated that these primer pairs also amplify nonspecific protozoal 18S rRNA, fungal 18S rRNA, and archaeal 16S rRNA. A third PCR was aimed at amplifying a fragment of the 16S rRNA gene of the Butyrivibrio group. Butyrivibrio-specific primers were designed using PRIMROSE software (3) based on sequences from the genera Butyrivibrio and Pseudobutyrivibrio. After primer design, the primer pair candidates were tested in silico using PRIMROSE (3) and the Ribosomal Database Project (34). The sequences (5′-3′) of the forward and reverse primer were GYG AAG AAG TAT TTC GGT AT (B395f) and CCA ACA CCT AGT ATT CAT C (B812r), respectively. These primers also allowed annealing for other genera within the family Lachnospiraceae. A 40-bp GC clamp (9) was attached to the forward primer for DGGE. The amplification conditions were as follows: initial denaturation at 94°C for 5 min; 35 cycles of denaturation at 95°C for 1 min, annealing at 55°C for 1 min, and extension at 72°C for 2 min; and final extension at 72°C for 10 min. The PCR mixture was prepared according to the manufacturer's instructions (Fermentas, St. Leon-Rot, Germany) and contained 1 μl DNA extract, 0.5 μl of each primer (10 μM stock), 0.5 μl deoxynucleotide triphosphate mixture (10 mM each), 2.5 μl 10× Taq buffer with KCl (500 mM), 1.5 μl MgCl₂ (25 mM), 0.125 μl Taq DNA polymerase (0.6 U), 0.06 μl bovine serum albumin, and DNase-RNase-free filter-sterilized water (Sigma, Bornem, Belgium) to a final volume of 25 μl. The amplicons were visualized by gel electrophoresis with 1% agarose and ethidium bromide.

DGGE analysis.

Ciliate and bacterial DGGE were performed as described by Boeckaert et al. (5) and Boon et al. (8), respectively. For the Butyrivibrio-specific DGGE, Butyrivibrio-specific PCR fragments were loaded onto a 7% (wt/vol) polyacrylamide gel (40% acrylamide, 77.8%; 2% bis-acrylamide, 22.2%) in 1× TAE buffer (40 mM Tris, 20 mM acetate, 2 mM EDTA, pH 8.5) with denaturing gradients ranging from 45% to 60%. The electrophoresis was run for 16 h at 60°C and 45 V. DGGE patterns were visualized by staining with Sybr green I nucleic acid gel stain (Molecular Probes, Eugene, OR). The DGGE patterns obtained were analyzed with BioNumerics software version 3.5 (Applied Maths, Kortrijk, Belgium). Similarities were calculated by the Pearson correlation, taking into account band intensity and band position. The clustering algorithm of Ward (35) was used to calculate dendrograms.

Cloning and identification of Butyrivibrio species.

The Butyrivibrio sp. PCR product of cow 1 on day 6 of algal feeding was cloned using a Topo-TA cloning kit (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions in order to create a clone library. For each clone, an aliquot of 800 μl was stored in 40% (vol/vol) glycerol, while plasmid DNA was isolated from the remaining liquid using the High Pure plasmid isolation kit (Roche Diagnostics GmbH, Mannheim, Germany). Subsequent DGGE analysis excluded identical clones. The resulting 23 exclusive clones were identified by sequencing the partial 16S rRNA gene fragments (ITT Biotech, Bielefeld, Germany). Additionally, Butyrivibrio DGGE bands, which specifically modified upon algal feeding, were excised. After Butyrivibrio-specific PCR and purification of the PCR product (Qiaquick PCR Purification Kit; Qiagen Benelux B.V., Venlo, The Netherlands), these bands were cloned with a Topo-TA cloning kit as described above. Close relatives of the 16S rRNA sequences were identified with the sequence match server of Ribosomal Database Project II (34). Additionally, sequences were aligned with the NAST software (12). The alignment was manually checked using the ARB aligner tool (ARB Software) (37), after which the sequences were added to the original phylogenetic tree (Greengenes database, January 2008 [12]) using Parsimony (ARB Software) (37) without changing the tree topology.

qPCR.

Ciliate and total bacterial rRNA gene copies present in the DNA extract of each ruminal-digesta sample were quantified as described by Boeckaert et al. (5) and Boon et al. (9), respectively. Butyrivibrio rRNA gene copies present in the DNA extract of each sample were quantified using an ABI Prism SDS 7000 instrument (Applied Biosystems, Lennik, Belgium) following the principle of Heid et al. (16). Dilutions (1:20) of DNA from all samples were added to amplification reaction mixtures (25 μl) containing 12.5 μl Sybr green PCR Master Mix (Applied Biosystems, Warrington, United Kingdom), 6 μl RNA-free water, 0.75 μl B395f primer (10 μM stock), 0.75 μl B812r primer (10 μM stock), and 5 μl DNA. The cycling conditions were 1 cycle of 50°C for 2 min and 95°C for 10 min and 40 cycles of 95°C for 1 min, 54°C for 30 s, and 60°C for 1 min. Measurements were done in triplicate for each run. A standard curve for qPCR was constructed using six different DNA concentrations (n = 3) ranging from 2.67 copies to 2.67 × 10⁸ copies of DNA per μl. A Butyrivibrio 417-bp PCR fragment inserted in a Topo vector (see above) was used as a template for the standard curve. The slope of the standard curve was −3.42 (R² = 0.99).

Statistical analysis.

Rumen fatty acid and qPCR data were analyzed using the Mixed procedure of the SAS Institute (29). The model for rumen fatty acid data included the fixed effects of day and time of sampling and their interaction and the random effect of cow, assuming an autoregressive order one covariance structure fitted on the basis of Akaike information and Schwarz Bayesian model fit criteria. The time of sampling was treated as a repeated measure. The statistical model for qPCR data included the fixed effect of day and the random effect of cow, assuming the covariance structure as described above. Least squares means were reported, and significance was declared at a P value of <0.05.

Nucleotide sequence accession numbers.

The nucleotide sequences for clones NC1 through NC23 and B1-1 through B4-3 have been deposited in the GenBank database under accession numbers EU839861 to EU839892.

RESULTS

Rumen fatty acid composition.

Supplementation of the diet with algae reduced saturated fatty acids (P < 0.001), while monounsaturated fatty acid (P < 0.001) increased (P < 0.001) (Table 2). PUFA (P = 0.001) and odd- and branched-chain fatty acids (P < 0.001) also increased upon algal feeding, whereas the amounts of C_18:0 strongly decreased (P < 0.001). All C_18:1 fatty acids increased (P < 0.001) after algal feeding, except for C_18:1 c12. Dietary algae resulted in 6- and 20-fold increases in C_18:1 t11 and C_18:1 t10, respectively. C_18:1 t4 (P < 0.001), C_18:1 t5 (P < 0.001), and C_18:1 t11 (P < 0.001) showed maximum concentrations on day 6, after which they decreased, whereas the greatest concentration of C_18:1 t10 (P < 0.001) was observed on day 13. C_18:1 c14 + t16 (P < 0.001), C_18:1 c15 (P < 0.001), and C_18:1 t12 (P < 0.001) progressively increased with time of algal feeding. Besides increased monounsaturated fatty acid proportions, dietary algae increased PUFA proportions due to an increase in C_18:3 c9t11c15 (P = 0.008), C_18:2 t11c15 (P < 0.001), CLA isomers (P < 0.001), and C_22:6 n-3 (P < 0.001). CLA c9t11 reached its maximum concentration on day 6 after algal feeding and significantly decreased thereafter. Concentrations of CLA c9t11 on day 20 did not differ from the values before algal feeding, whereas the increased concentrations of CLA t10c12 and CLA c9c11 + t11c13 after dietary algal supplementation were stable until at least day 20 after algal feeding. Bacterial synthesized odd- and branched-chain fatty acids increased through algal feeding. This was mainly due to higher proportions of iso C_15:0 (P < 0.001), iso C_17:0 (P < 0.001), anteiso C_15:0 (P < 0.001), and anteiso C_17:0 (P < 0.001).

TABLE 2.

Effect of time of algal feeding on the ruminal fatty acid composition^a

Fatty acid	Sampling day				SEMb	Significance
	−2	6	13	20		Dc	Td	D*Te
C12:0	0.50a	0.10b	0.11b	0.11b	0.014	<0.001	0.043	0.007
C14:0	0.44b	0.54ab	0.54ab	0.64a	0.039	0.003	0.021	0.298
C16:0	4.81b	6.28a	6.21a	6.73a	0.217	<0.001	0.598	0.652
C14:1c9	0.01ab	0.01ab	0.01b	0.01a	0.001	0.051	0.468	0.624
C16:1c9	0.03c	0.07b	0.07a	0.07ab	0.006	<0.001	0.036	0.796
C15:0	0.35	0.37	0.33	0.37	0.030	0.684	0.572	0.701
C17:0	0.14	0.16	0.16	0.17	0.011	0.144	0.064	0.568
iso C13:0	0.01b	0.02a	0.02a	0.02a	0.001	<0.001	0.016	0.440
iso C14:0	0.03	0.03	0.03	0.03	0.003	0.101	0.003	0.417
iso C15:0	0.11b	0.18a	0.15a	0.17a	0.010	<0.001	0.004	0.372
iso C16:0	0.04c	0.08a	0.06b	0.08a	0.009	<0.001	0.027	0.769
iso C17:0	0.07b	0.14a	0.15a	0.16a	0.008	<0.001	0.003	0.128
anteiso C13:0	0.03a	0.03b	0.03b	0.03ab	0.003	0.014	0.001	0.431
anteiso C15:0	0.22c	0.30a	0.27b	0.30ab	0.011	<0.001	0.001	0.258
anteiso C17:0	0.06c	0.10b	0.10b	0.11a	0.003	<0.001	<0.001	0.310
C18:3n-3	1.41ab	1.29b	1.65a	1.47ab	0.085	0.019	<0.001	0.201
C18:2n-6	2.37	2.21	2.39	2.08	0.131	0.254	<0.001	0.497
C18:3c9t11c15	0.08b	0.16a	0.15a	0.15a	0.020	0.008	0.788	0.527
C18:2t11c15	0.22b	0.65a	0.63a	0.67a	0.043	<0.001	<0.001	0.142
CLA c9t11	0.05c	0.13a	0.09b	0.08bc	0.013	<0.001	0.616	0.380
CLA t10c12	0.04b	0.06a	0.07a	0.07a	0.005	<0.001	0.033	0.265
CLA c9c11 + t11c13	0.02c	0.03bc	0.03b	0.04a	0.002	<0.001	0.009	0.740
C18:1c9	1.78b	2.59a	2.45a	2.32a	0.123	<0.001	<0.001	0.233
C18:1c11	0.35b	0.69a	0.67a	0.68a	0.024	<0.001	0.319	0.698
C18:1c12	0.21a	0.12b	0.12b	0.14b	0.013	<0.001	0.384	0.390
C18:1c13	0.02c	0.09b	0.09ab	0.10a	0.006	<0.001	0.124	0.257
C18:1c14 + t16	0.26c	0.40b	0.43ab	0.52a	0.055	<0.001	0.044	0.634
C18:1c15	0.08c	0.25b	0.28a	0.31a	0.014	<0.001	0.002	0.134
C18:1t4	0.02c	0.05a	0.04b	0.04b	0.003	<0.001	0.013	0.534
C18:1t5	0.02c	0.07a	0.06b	0.05b	0.002	<0.001	0.005	0.388
C18:1t6 to t8	0.13b	0.38a	0.37a	0.34a	0.023	<0.001	0.013	0.506
C18:1t9	0.10b	0.57a	0.56a	0.63a	0.054	<0.001	0.009	0.621
C18:1t10	0.23c	3.60b	5.19a	5.08a	0.368	<0.001	0.129	0.392
C18:1t11	0.92d	6.38a	4.32c	5.42b	0.201	<0.001	0.001	0.117
C18:1t12	0.23c	1.17ab	1.05b	1.21a	0.056	<0.001	0.002	0.227
C18:1t13 to t14	0.44b	1.45a	1.41a	1.69a	0.118	<0.001	<0.001	0.090
C18:1t15	0.27b	0.67a	0.69a	0.81a	0.060	<0.001	0.007	0.384
C18:0	16.0a	1.27b	1.39b	1.22b	0.354	<0.001	0.534	0.762
C22:6n-3f	NDl	0.72a	0.71a	0.91a	0.088	0.225	<0.001	0.904
SFAg	22.8a	9.61b	9.54b	10.2b	0.536	<0.001	0.553	0.542
MUFAh	5.10b	18.5a	17.8a	19.4a	0.460	<0.001	0.042	0.399
PUFAi	4.19b	5.25a	5.72a	5.48a	0.248	0.001	<0.001	0.405
OBCFAj	1.08b	1.42a	1.29a	1.45a	0.051	<0.001	0.007	0.242
n-3k	1.41b	2.01a	2.36a	2.38a	0.107	<0.001	<0.001	0.178
n-3/n-6	0.59c	0.93b	0.99b	1.14a	0.033	<0.001	0.261	0.444

^aIn mg/g ruminal digesta DM (n = 60). The data represent the means of fatty acids present in ruminal digesta collected at five different time points from three different cows. Means within a row with different lowercase letters following (a to c) differ (P ≤ 0.05).

^bSEM, standard error of the mean.

^cD, effect of days on algal-concentrate feeding.

^dT, effect of time of rumen sampling.

^eD*T, interaction between days and sampling time.

^fY_ijk = μ + A_i + B_j + C_k + AB_ij + ɛ_ijk, where Y_ijk is the response, A_i the effect of day (day 6, day 13, and day 20), B_j the effect of time of sampling (0, 1, 2, 4, and 6 h), C_k the random effect of cow, AB_ij the interaction between day and time of sampling, and ɛ_ijk the residual error. Means with letters following differ from 0.

^gSFA, saturated fatty acids. Σ (C_12:0; C_14:0; C_15:0; C_16:0; C_17:0; C_18:0; iso C_13:0; iso C_14:0; iso C_15:0; iso C_16:0; iso C_17:0; anteiso C_13:0; anteiso C_15:0; anteiso C_17:0).

^hMUFA, monounsaturated fatty acids. Σ (C_14:1 c9; C_16:1 c9; C_18:1 c9; C_18:1 c11; C_18:1 c12; C_18:1 c13; C_18:1 c14 + t16; C_18:1 c15; C_18:1 t4; C_18:1 t5; C_18:1 t6 to t8; C_18:1 t9; C_18:1 t10; C1_8:1 t11; C_18:1 t12; C_18:1 t13 to t14; C_18:1 t15).

ⁱPUFA, Σ (C_18:3 n-3; C_18:2 n-6; C_18:3 c9t11c15; CLA c9c11 + t11c13; C_18:2 t11c13; CLA c9t11; CLA t10c12; C_22:6 n-3).

^jOBCFA, odd- and branched-chain fatty acids. Σ (C_15:0; C_17:0; iso C_13:0; iso C_14:0; iso C_15:0; iso C_16:0; iso C_17:0; anteiso C_13:0; anteiso C_15:0; anteiso C_17:0).

^kn-3, Σ (C_18:3 n-3; C_22:6 n-3).

^lND, not detected.

Ciliate community analysis.

DGGE profiles of rumen ciliates before and during algal feeding were similar for all cows (data not shown). Cluster analysis according to Pearson correlation grouped 9 of the 12 DGGE profiles in one cluster independent of cow or days on algal feeding. This group showed four to six bands. Ciliate DGGE profiles of cow 1 on day 6 and day 20 and of cow 2 on day −2 showed greater diversity, with the number of bands varying between 7 and 10. Therefore, these DGGE patterns clustered separately.

The number of ciliates present in the bovine rumen content before and during algal feeding was estimated by amplifying the 18S rRNA gene fragment of the ciliates and measuring the increasing amounts of amplification products. Before algal feeding, ciliates numbered 5.68 ± 0.52 log copies per g ruminal digesta (mean ± standard deviation). The numbers of ciliates were unaffected (P = 0.359) by algal feeding (5.14 ± 0.10, 5.17 ± 0.66, and 5.48 ± 0.07 log copies per g ruminal digesta on day 6, day 13, and day 20, respectively).

Bacterial community analysis.

Rumen bacterial DGGE profiles showed high complexity, with more than 30 bands both before and during algal feeding. Cluster analysis (Pearson correlation) resulted in two major clusters corresponding to the dietary treatments (Fig. 1). Nevertheless, these shifts in the bacterial community were difficult to link with the biohydrogenation activity, since no (dis)appearance of specific bands could be identified.

FIG. 1.

Cluster analysis of the DGGE profile of total bacteria present in the rumen of cows fed algae. d, day.

Algal feeding had no influence on total bacterial numbers (P = 0.091), and the numbers were on average 10.5 ± 0.15 log copies per g dry ruminal digesta (mean ± standard deviation).

Butyrivibrio community analysis.

The primers designed to detect Butyrivibrio spp. and relatives were tested for their specificity with B. fibrisolvens DSM 3071 and B. fibrisolvens MDT-5 as positive controls, whereas E. coli DH5α, L. brevis LMG 7761, and B. animalis LMG 11580 were used as negative controls. The primer set produced PCR products for both B. fibrisolvens strains of the expected size of 417 bp, whereas no amplification was observed for the negative control strains.

Butyrivibrio-specific DGGE generated around 15 bands, which clearly changed with time on dietary algae. Cluster analysis separated the DGGE profiles of rumen contents taken on day 13 and day 20 from cows 1 and 2 from the other DGGE profiles (Fig. 2). For these two cows, a further distinction was possible between DGGE profiles of rumen contents prior to and after the 6 days of algal feeding. For the third cow, algal feeding did not result in major shifts in the Butyrivibrio population, as DGGE profiles clustered together. Some bands disappeared from day 6 onward, whereas other bands became more pronounced on day 13 and day 20 (Fig. 2).

FIG. 2.

Cluster analysis of the DGGE profile of Butyrivibrio spp. present in the rumen of cows fed algae. The numbers in the profile indicate disappearing (1 and 2) and more pronounced (3 and 4) bands following algal-concentrate feeding. d, day.

The number of Butyrivibrio species in the rumen content was determined before and during algal feeding by qPCR. Before algal feeding, Butyrivibrio species numbered 7.21 ± 0.05 log copies per g ruminal digesta (mean ± standard deviation). The numbers decreased numerically (P = 0.246) during algal feeding (6.96 ± 0.18, 7.04 ± 0.02, and 7.02 ± 0.28 log copies per g ruminal digesta on day 6, day 13, and day 20, respectively).

Cloning and sequence analysis of the Butyrivibrio community.

A clone library was constructed with the PCR product of cow 1 on day 6 in order to identify the bacterial species detected with the Butyrivibrio-specific primers. Phylogenetic clustering (Fig. 3) based on 16S rRNA sequence (417 bp) analysis indicated that the 23 clones were related to bacterial species within the genus Butyrivibrio, the genus Lachnospiraceae incertae sedis, and other, so far unknown genera within the family Lachnospiraceae and were closely related to the genera Butyrivibrio and Pseudobutyrivibrio.

FIG. 3.

Phylogenetic tree representing the classification of 23 clones (denoted as clone NCi; 417 bp), obtained from rumen fluid of a cow fed a diet supplemented with algae (cow 1 on day 6), within the genus Butyrivibrio, the genus Lachnospiraceae incertae sedis, and other (unknown) genera within the family Lachnospiraceae. Clones Bi to Bj are derived from changing DGGE bands upon algal-concentrate feeding (i, band excised and cloned from Butyrivibrio DGGE [Fig. 1]; j, clone number).

The nine clones originating from four specific, changing DGGE bands upon supplementation of the diet with algae were phylogenetically classified (Fig. 3). The majority of these clones (eight) clustered on branches between the genus Butyrivibrio and the genus Pseudobutyrivibrio. They include the four clones (B1-1, B1-2, B2-1, and B2-2) originating from the two disappearing bands (bands 1 and 2 in Fig. 2) and four of the five clones (B3-1, B3-2, B4-1, and B4-3) originating from the more pronounced bands (bands 3 and 4 in Fig. 2). Disappearing bands did not cluster apart from more pronounced bands and did not cluster in the B. proteoclasticus branch. Clone B4-2 clustered distinctly and showed genetic relationship with butyrate-producing bacteria.

DISCUSSION

The aim of the current study was to relate inhibition of rumen biohydrogenation with changes in ciliates, total bacteria, and Butyrivibrio species. In order to induce accumulation of biohydrogenation intermediates in the rumen, DHA-enriched algae were used (5). A unique feature of this study was the identification of altering Butyrivibrio DGGE bands by means of cloning and their possible relationship with decreased C_18:0 and increased C_18:1 trans fatty acid concentrations.

Ciliate DGGE and qPCR (on average 5.37 log copies/g ruminal digesta) analysis showed no major changes in the ciliate community structure after supplementation of the diet with algae. This is in contrast with a previous experiment in which some ciliates, especially I. prostoma, disappeared after algal feeding (5). However, the daily amount of DHA supplied in the latter experiment (76.1 g DHA/day) was almost twice as high as in the current experiment (43.7 g DHA/day), probably explaining these results. In addition, further research on the specific role of I. prostoma in rumen biohydrogenation indicated that this ciliate and its associated bacteria are not directly involved (6).

The change in the total bacterial community structure with dietary supplementation of marine products was observed before (5, 20). Cluster analysis of the DGGE profile showed two clusters based on the absence or presence of algae in the diet. However, due to the complexity of the DGGE pattern of the total bacterial community, we examined in more detail bacterial groups of importance in rumen biohydrogenation. As C₁₈ biohydrogenation is thought to be mainly performed by strains of the Butyrivibrio group, including the genera Butyrivibrio and Pseudobutyrivibrio (24), a primer set to study bacteria within this Butyrivibrio group was developed. In spite of major shifts in rumen biohydrogenation, supplementation of the diet with algae did not affect the total amount of Butyrivibrio spp. However, Butyrivibrio-specific DGGE profiles showed a shift in this group with time on a diet supplemented with algae. Some DGGE bands disappeared upon supplementation of the diet with algae, which was associated with a decrease in rumen C_18:0 concentrations (Fig. 2, bands 1 and 2). We hypothesize that bacteria associated with these disappearing bands might play a role in the conversion of C_18:1 trans to C_18:0. Until now, C_18:0 production was linked with isolates belonging to the B. proteoclasticus group (19) and the Butyrivibrio sp. branch (21), which all possess a close phylogenetic relationship. However, none of the species present in the disappearing bands belonged to this B. proteoclasticus group. On the other hand, 20% of the clones from the clone library of cow 1 on day 6 were classified in the B. proteoclasticus branch despite the reduced ruminal C_18:0 concentration (17.0 versus 1.65 mg/g ruminal digesta DM before and on day 6 of dietary algal supplementation, respectively). The presence of these bacteria in the B. proteoclasticus branch was unexpected, as the bacteria are highly sensitive to PUFA (19) and C_18:0 production was significantly decreased. This might indicate that the presence of DHA decreases the capacity of these bacteria to hydrogenate C_18:1 trans fatty acids rather than the bacteria as such. Competitive inhibition of bacterial isomerases and reductases and/or competition for hydrogen used in the simultaneous biohydrogenation of DHA and unsaturated C₁₈ fatty acids (1, 36) might also explain the limited conversion of C_18:1 trans fatty acids to C_18:0 by bacteria belonging to the B. proteoclasticus group. Alternatively, these results could indicate that bacteria within the B. proteoclasticus branch have a limited contribution to in vivo C_18:0 formation. Kim et al. (20) found that the decreased duodenal C_18:0 flow in steers fed fish oil was not associated with C. proteoclasticum 16S rRNA gene concentrations in strained rumen fluid. Similarly, Huws et al. (18) reported that DNA concentrations from the Butyrivibrio C_18:0-producing group did not correlate with the C_18:0 concentrations of rumen planktonic and biofilm samples. This suggests that other, yet-uncultivated microbial species might be involved in C_18:0 production and might fulfill a more important role in the final step of the biohydrogenation process. Wallace et al. (33) stated that isolation of more C_18:0 producers might have been hampered by the fact that these bacteria must be growing in order to carry out biohydrogenation and that PUFA themselves have a strong tendency to inhibit this growth. It is generally recognized that only a minor part of the diversity of microorganisms in nature is presently known (2). Hence, it is not unlikely that reduced C_18:0 production is associated with the disappearance of uncultivated species. Identification of the disappearing bands through cloning and subsequent phylogenetic classification indicated that these noncultivated species were located on a separate branch between the genera Butyrivibrio and Pseudobutyrivibrio, belonging to unknown Lachnospiraceae strains genetically more distant from the B. proteoclasticus group. The concomitant disappearance of these species and decrease in the rumen C_18:0 concentration might suggest that they play a role in in vivo C_18:0 production. Further research is under way to evaluate the disappearance of these unknown Lachnospiraceae strains and reduced C_18:0 production under various rumen conditions.

Besides reduced C_18:0 concentrations, increased C_18:1 t11 and C_18:1 t10 concentrations were observed after feeding with the algal concentrate. This was associated with some more pronounced bands in the Butyrivibrio DGGE profile (Fig. 2, bands 3 and 4). Identification through cloning and subsequent phylogenetic classification indicated that these bands also represented noncultivated species located on a separate branch between the genera Butyrivibrio and Pseudobutyrivibrio. This might indicate that some Butyrivibrio-like bacteria are associated with changes in the rumen resulting in C_18:1 t11 and C_18:1 t10 accumulation. It remains to be determined whether the more pronounced appearance of these bands and accumulation of C_18:1 t10 or C_18:1 t11 is a causal relationship. In addition, the origin of these increased amounts of C_18:1 t10, either isomerization of C_18:1 c9 (23) or C_18:1 t11 (22) or biohydrogenation of CLA t10c12 (14), needs to be clarified.

Although this research mainly focused on molecular techniques to describe shifts within the rumen microbial community, other bacterial markers, such as odd- and branched-chain fatty acids (32), also showed shifts upon algal feeding. In the present study, supplementation of the diet with algae increased rumen iso C_17:0, anteiso C_15:0, and anteiso C_17:0 concentrations. In milk, these fatty acids showed a negative correlation with C_18:0 (r_pearson = −0.807, −0.207, and −0.446, respectively), whereas milk iso C_17:0 was positively correlated with milk C_18:1 t11 (r_pearson = 0.983) and, to a lesser extent, with C_18:1 t10 (r_pearson = 0.522) (32).

Conclusion.

Supplementation of the diet with algae inhibited rumen C₁₈ biohydrogenation, resulting in decreased C_18:0 concentrations, whereas C_18:1 t11 and C_18:1 t10 concentrations increased. Changes in the rumen fatty acid profile were associated with changes in the structure of the bacterial community and, more specifically, with changes in the Butyrivibrio group. Clones associated with altered DGGE bands indicated that dietary algae affected noncultivated species, which cluster between the genus Butyrivibrio and the genus Pseudobutyrivibrio. Additionally, 20% of the clone library, from a randomly selected rumen sample after the start of supplementation of the diet with algae, clustered within the C_18:0-producing B. proteoclasticus branch, although C_18:0 production was reduced. This suggests that other, as-yet-uncultivated bacteria are involved in C_18:0 production, possibly being more important than B. proteoclasticus.