Effect of Camelina sativa seeds on rumen microbiota and fermentation in dairy sheep

Christos Christodoulou; Alexandros Mavrommatis; Marco Severgnini; Paola Cremonesi; Bianca Castiglioni; Panagiota Kyriakaki; Rafaela Andreaki; Basiliki Kotsampasi; Eleni Tsiplakou

PMC · DOI:10.1093/jas/skag033·February 13, 2026

Effect of Camelina sativa seeds on rumen microbiota and fermentation in dairy sheep

Christos Christodoulou, Alexandros Mavrommatis, Marco Severgnini, Paola Cremonesi, Bianca Castiglioni, Panagiota Kyriakaki, Rafaela Andreaki, Basiliki Kotsampasi, Eleni Tsiplakou

PDF

Open Access

TL;DR

This study shows that adding Camelina sativa seeds to dairy sheep diets changes the rumen microbes and fermentation, potentially offering a sustainable feed option.

Contribution

The study reveals specific microbial shifts and fermentation changes in sheep due to varying levels of Camelina sativa seed inclusion.

Findings

01

Higher Camelina inclusion increased acetic and propionic acid concentrations in rumen fluid.

02

Camelina reduced biodiversity in rumen fluid and altered the abundance of key microbial taxa.

03

Lower Camelina levels enriched fibrolytic bacteria, while higher levels favored amylolytic and propionate-associated microbes.

Abstract

This study investigated the effect of three levels of Camelina sativa seeds on ewes’ diet on rumen microbiota using 16S rRNA gene amplicon sequencing and biochemical assays, focusing on rumen fermentation parameters and carbohydrates, proteins, and fats metabolism. Forty-eight dairy ewes were assigned to four homogeneous groups based on the inclusion level of C. sativa seeds in the diet (0, 28, 51.3, and 74.6 g/kg DM; Control, CS6, CS11, and CS16, respectively). Rumen digesta were collected on the 60th day of the trial using an esophageal tube. Rumen fluid was analyzed for volatile fatty acids (VFAs) concentration and rumen enzymatic activity. In addition, rumen microbiota was characterized in both fluid and solid fractions. The acetic and propionic acid concentrations were higher (P < 0.001) in CS11 compared with Control and CS6. Iso-butyric, butyric, iso-valeric, valeric acid, and…

Linked entities

Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.

Species12

Camelina sativa Ovis aries Ruminococcus(genus)Pseudobutyrivibrio(genus)Lachnospiraceae(family)Methanobrevibacter(genus)Succinimonas(genus)Ruminobacter(genus)Fibrobacter(genus)Saccharofermentans(genus)Bacteria Latreille et al. 1825(Bacteria stick insect · genus)Butyrivibrio(genus)

Chemicals7

carbohydrate acetic Iso-butyric, butyric, iso-valeric, valeric acid VFAs fatty acid propionic acid propionate

Figures1

Click any figure to enlarge with its caption.

A) Rarefaction curves of the alpha-diversity of the samples, estimated by Faith’s phylogenetic diversity metric, for the fluid part of the rumen microbiota; each line is the average value over all the samples within the same experimental group (i.e., diet); error bars represent standard errors. B) Principal Coordinate Analysis (PCoA) plot based on the unweighted UniFrac distance among samples for the fluid part of the rumen microbiota; each point represents a sample, colored according to the experimental group (i.e., diet); ellipses are the SEM-based confidence intervals, and centroids are the average coordinate of all the samples in the same group. C) Rarefaction curves of the alpha-diversity of the samples, estimated by Faith’s phylogenetic diversity metric, for the solid part of the rumen microbiota; each line is the average value over all the samples within the same experimental group (i.e., diet); error bars represent standard errors. D) Principal Coordinate Analysis (PCoA) plot based on the unweighted UniFrac distance among samples for the solid part of the rumen microbiota; each point represents a sample, colored according to the experimental group (i.e., diet); ellipses are the SEM-based confidence intervals, and centroids are the average coordinates of all the samples in the same group.

Tables3

Table 1. Dietary treatment components and chemical composition.

	Dietary treatments ¹
	Control	CS6	CS11	CS16
Item²
Diet components, % DM
Forages
Wheat straw	6.13	6.13	6.13	6.13
Alfalfa hay	47.22	47.22	47.22	47.22
Concentrates²
Corn meal	16.05	13.95	12.55	11.15
Barley meal	9.33	9.33	9.33	9.33
Wheat middlings	4.67	4.67	4.67	4.67
Sunflower meal	7.46	8.40	8.40	8.40
Soybean meal	7.23	5.60	4.67	3.73
Camelina sativa seeds	—	2.80	5.13	7.46
Mineral Vitamin premix	1.91	1.91	1.91	1.91
Calculated composition, % DM³
Daily feed intake, g of DM	2,840	2,842	2,844	2,846
Ash	7.81	7.82	7.79	7.77
aNDFom	36.78	37.93	38.68	39.42
ADF	25.21	26.12	26.74	27.37
ADL	5.41	5.77	6.02	6.27
iNDF	12.98	13.85	14.45	15.05
fNDF	23.80	24.08	24.23	24.37
Fibrous carbohydrates	44.44	43.34	42.55	41.76
NFC	35.72	33.81	32.46	31.12
Starch	17.23	15.86	14.94	14.03
Sugars	3.41	3.4	3.38	3.36
Ether extract	1.65	2.63	3.45	4.27
Crude protein	18.12	18.19	18.23	18.27
Soluble crude protein	5.13	5.3	5.4	5.49
RDP3x	11.73	11.76	11.77	11.78
NEl, Mcal/kg	1.54	1.52	1.52	1.51
Methionine	0.29	0.3	0.3	0.31
Lysine	0.83	0.82	0.82	0.82
Lysine: Methionine	3.24:1	3.16:1	3.11:1	3.05:1
Arginine	1.02	1.06	1.1	1.13
Threonine	0.71	0.71	0.72	0.73
Leucine	1.33	1.31	1.3	1.29
Isoleucine	0.72	0.72	0.71	0.71
Valine	0.91	0.92	0.92	0.92
Histidine	0.4	0.39	0.39	0.38
Phenylalanine	0.84	0.84	0.83	0.82
Tryptophane	0.23	0.23	0.23	0.23

Table 2. Ruminal volatile fatty acid accumulation and enzymatic activities in ewes fed four dietary treatments with different levels of Camelina sativa seeds.

	Control n = 6	CS6 n = 6	CS11 n = 6	CS16 n = 6	SEM	P value²
	Dietary treatments¹				SEM	P value²
Volatile fatty acids, mM
Acetic	35.6^b	21.3^c	49.5^a	42.1^ab	2.99	<0.001
Propionic	8.18^bc	4.87^b	13.2^a	11.6^ab	0.885	<0.001
Butyric	7.90^b	3.59^c	11.4^a	9.57^ab	0.874	<0.001
Iso-butyric	1.08^bc	1.00^c	1.80^a	1.47^ab	0.112	<0.001
Valeric	0.51^bc	0.34^c	0.81^a	0.70^ab	0.065	<0.001
Iso-valeric	0.96^b	1.14^ab	2.58^a	1.14^ab	0.613	0.017
Total VFAs	54.2^b	32.2^c	79.2^a	66.5^ab	4.697	<0.001
Acetic: Propionic	4.55^a	4.39^ab	3.81^bc	3.67^c	0.175	0.003
Enzymatic activities
Cellulase, units/mL/h	21.2	20.4	21.0	20.5	0.29	0.291
Amylase, mg maltose release/mL	1.86^a	1.15^b	2.73^a	1.87^a	0.258	<0.001
Xylanase, units/mL/h	16.6	15.6	16.0	16.2	0.28	0.160

Table 3. Relative abundance (%) of bacteria taxa in ewes’ rumen of the four dietary treatments fed different levels of Camelina sativa seeds.

% Relative abundance	Dietary treatments, D ¹				SEM	Fraction, F		SEM	P value²
% Relative abundance	Control n = 12	CS6 n = 12	CS11 n = 12	CS16 n = 12	SEM	Fluid n = 24	Solid n = 24	SEM	D	F	D × F
Phylum
Bacteroidota	41.2	41.4	41.5	41.5	0.56	43.5	39.3	0.36	0.978	<0.001	0.214
Firmicutes	28.3	30.0	29.8	29.2	0.76	27.6	31.1	0.47	0.422	<0.001	0.175
Proteobacteria	8.16^b	4.04^c	11.2^ab	12.1^a	0.761	11.8	5.99	0.535	<0.001	<0.001	0.229
Verrucomicrobiota	4.41^a	4.93^a	2.87^b	3.13^b	0.240	2.35	5.32	0.163	<0.001	<0.001	0.455
Fibrobacterota	4.41^a	4.93^a	2.87^b	3.13^b	0.240	2.35	5.32	0.163	<0.001	<0.001	0.065
Patescibacteria	2.05	2.46	2.10	1.81	0.179	2.09	2.12	0.115	0.120	0.833	0.543
Spirochaetota	3.93	3.38	3.90	3.65	0.221	1.85	5.58	0.156	0.281	<0.001	0.236
Euryachaeota	2.17^ab	2.75^a	1.54^b	1.47^b	0.197	1.45	2.52	0.119	<0.001	<0.001	0.777
Chloroflexi	0.96	1.09	0.90	0.69	0.100	0.79	1.03	0.071	0.055	0.019	0.997
Family
Prevotellaceae	26.6^bc	25.2^c	28.0^ab	28.8^a	0.48	29.3	25.0	0.34	<0.001	<0.001	0.894
Succinivibrionaceae	7.13^b	2.85^c	10.3^ab	11.3^a	0.833	10.7	5.15	0.579	<0.001	<0.001	0.162
Lachnospiraceae	6.96^c	7.61^bc	9.47^a	8.45^ab	0.343	7.12	9.12	0.232	<0.001	<0.001	0.069
Rikenellaceae	5.67^ab	5.94^a	5.29^ab	5.04^b	0.188	5.35	5.62	0.127	0.015	0.126	0.683
Uncultured rumen bacterium (WCHB1-41)	4.60^a	5.30^a	2.92^b	3.11^b	0.377	5.17	2.79	0.256	<0.001	<0.001	0.466
Bacteroidales RF16 group	2.92^ab	3.69^a	2.42^b	2.59^ab	0.271	4.09	1.72	0.177	0.016	<0.001	0.056
Ruminococcaceae	3.83	3.52	4.09	4.13	0.190	3.83	3.95	0.130	0.119	0.489	0.906
Acidaminococcaceae	3.35^ab	3.83^a	2.88^b	3.22^ab	0.211	3.44	3.20	0.148	0.033	0.253	0.300
Oscillospiraceae	3.76	4.01	3.63	3.13	0.252	3.09	4.18	0.156	0.126	<0.001	0.064
F082	3.35^ab	4.11^a	3.19^b	2.84^b	0.228	2.98	3.77	0.162	0.003	0.001	0.886
Selenomonadaceae	1.22^b	1.17^b	1.59^ab	2.17^a	0.207	2.52	0.55	0.141	0.010	<0.001	0.258
Fibrobacteraceae	4.41^a	4.93^a	2.87^b	3.13^b	0.240	2.35	5.32	0.163	<0.001	<0.001	0.065
Spirochaetaceae	3.92	3.37	3.89	3.64	0.220	1.84	5.57	0.156	0.274	<0.001	0.232
Methanobacteriaceae	2.17^ab	2.75^a	1.54^b	1.47^b	0.197	1.45	2.52	0.119	<0.001	<0.001	0.777
Uncultured rumen bacterium (Absconditabacteriales [SR1])	0.85	1.28	1.24	1.09	0.120	1.13	1.10	0.075	0.066	0.726	0.164
Clostridia UCG-014 (other)	0.90	0.89	1.20	1.37	0.130	1.11	1.07	0.086	0.040	0.763	0.792
Christensenellaceae	1.56^ab	1.66^a	1.28^bc	1.23^c	0.071	0.97	1.89	0.050	<0.001	<0.001	0.001
Muribaculaceae	1.17	0.82	1.21	1.01	0.167	0.70	1.40	0.107	0.357	<0.001	0.977
Hungateiclostridaceae	1.00^ab	1.09^a	0.83^bc	0.75^c	0.057	0.70	1.13	0.039	0.002	<0.001	0.224
Anaerolineaceae	0.96	1.09	0.90	0.69	0.100	0.79	1.03	0.071	0.055	0.019	0.997
Genus
Prevotella	18.4^b	17.1^b	20.1^a	20.4^a	0.41	21.5	16.6	0.29	<0.001	<0.001	0.693
Succinivibrionaceae UCG-002	4.27^a	1.46^b	5.70^a	5.97^a	0.528	5.69	3.01	0.361	<0.001	<0.001	0.222
Rikenellaceae RC9 gut group	5.59^b	5.86^a	5.23^b	4.98^b	0.188	5.30	5.53	0.126	0.017	0.188	0.716
Uncultured rumen bacterium (WCHB1-41)	4.60^a	5.30^a	2.92^b	3.11^b	0.377	5.17	2.79	0.256	<0.001	<0.001	0.303
Fibrobacter	4.41^ab	4.93^a	2.87^bc	3.13^c	0.240	2.35	5.32	0.163	<0.001	<0.001	0.065
Succiniclasticum	3.34^ab	3.83^a	2.87^b	3.21^ab	0.211	3.44	3.20	0.148	0.011	0.254	0.247
Prevotellaceae UCG-001	2.57	2.36	2.57	2.75	0.213	2.41	2.72	0.151	0.643	0.147	0.683
Treponema	3.73	3.13	3.80	3.53	0.218	1.70	5.39	0.154	0.144	<0.001	0.145
NK4A214 group	2.27	2.68	2.44	2.13	0.143	1.99	2.77	0.098	0.064	<0.001	0.052
Succinivibrionaceae UCG-001	1.84^ab	0.35^b	2.25^ab	3.80^a	0.649	1.08	3.05	0.459	0.006	0.004	0.384
Bacteroidales RF16 group	1.74	2.25	1.42	1.58	0.205	2.52	0.97	0.132	0.052	<0.001	0.081
Methanobrevibacter	2.05^ab	2.61^a	1.46^b	1.38^b	0.162	1.42	2.33	0.115	<0.001	<0.001	0.898
Prevotellaceae (other)	1.77	1.72	1.81	1.90	0.118	2.27	1.33	0.078	0.753	<0.001	0.319
Ruminococcus	1.49^b	1.95^a	1.62^ab	1.79^ab	0.132	1.25	2.17	0.043	<0.001	0.480	0.480
Prevotellaceae UCG-003	1.50	1.74	1.46	1.49	0.088	1.42	1.68	0.062	0.107	0.005	0.429
Ruminobacter	1.45^a	0.47^b	2.25^a	1.54^a	0.306	1.66	1.19	0.207	<0.001	0.168	0.427
Ruminococcaceae (other)	1.54^a	0.86^b	1.60^a	1.45^a	0.134	1.57	1.16	0.089	<0.001	0.004	0.204
Christensenellaceae R-7 group	1.49^a	1.60^a	1.17^b	0.46^c	0.063	0.94	1.42	0.045	<0.001	<0.001	<0.001
F082 (other)	1.37	1.44	0.99	0.07	0.112	0.00	1.39	0.079	0.006	<0.001	0.659
Lachnospiraceae (other)	1.05^b	0.98^b	1.79^a	1.59^a	0.137	1.39	1.31	0.097	<0.001	0.878	0.696
Butyrivibrio	1.01^b	0.95^b	1.45^a	1.19^ab	0.076	1.02	1.27	0.053	<0.001	0.002	0.011
Clostridia UCG-014 (other)	0.90	0.89	1.20	1.37	0.121	1.11	1.08	0.086	0.018	0.795	0.841
Uncultured rumen bacterium (Absconditabacteriales [SR1])	0.85	1.28	1.24	1.09	0.106	1.13	1.10	0.075	0.028	0.775	0.273
Uncultured rumen bacterium (F082)	0.91^b	1.46^a	0.98^b	0.87^b	0.086	1.03	1.08	0.061	<0.001	0.640	0.598
Flexilinea	0.96	1.09	0.90	0.69	0.100	0.79	1.03	0.071	0.055	0.019	0.997
Saccharofermentans	0.86^ab	0.98^a	0.72^bc	0.64^c	0.052	0.60	1.01	0.036	<0.001	<0.001	0.182
Uncultured rumen bacterium (Muribaculaceae)	0.86	0.70	0.95	0.72	0.136	0.53	1.09	0.083	0.522	<0.001	0.867
Prevotellaceae UCG-004	0.86	0.82	0.69	0.84	0.073	0.60	1.01	0.046	0.351	<0.001	0.202
[Eubacterium] ruminantium group	0.70	0.79	0.89	0.78	0.071	0.67	0.91	0.045	0.296	<0.001	0.052
Selenomonas	0.41^b	0.36^b	0.76^ab	1.14^a	1.135	1.13	0.20	0.096	<0.001	<0.001	0.011
Succinivibrio	0.56	0.54	0.76	0.91	0.105	1.09	0.29	0.071	0.064	<0.001	0.884
Lachnospiraceae AC2044 group	0.42^ab	0.57^a	0.52^ab	0.38^b	0.050	0.30	0.64	0.040	0.036	<0.001	0.550
Lachnospiraceae NK3A20 group	0.43^b	0.43^b	0.46^ab	0.57^a	0.032	0.37	0.57	0.022	0.020	<0.001	0.077
Pseudobutyrivibrio	0.35^b	0.30^b	0.68^a	0.51^ab	0.053	0.43	0.48	0.037	<0.001	0.351	0.215
Selenomonadaceae (other)	0.27	0.25	0.25	0.31	0.043	0.469	0.074	0.028	0.725	<0.001	0.657
Prevotellaceae NK3B31 group	0.26	0.26	0.22	0.21	0.032	0.14	0.34	0.022	0.557	<0.001	0.598
Lachnospiraceae ND3007 group	0.23^ab	0.38^a	0.19^b	0.19^b	0.043	0.31	0.19	0.028	0.015	0.002	0.602
Lachnospiraceae XPB1014 group	0.21	0.23	0.26	0.20	0.026	0.22	0.23	0.017	0.316	0.811	0.043
Succinimonas	0.19^a	0.08^b	0.20^a	0.23^a	0.036	0.22	0.13	0.025	0.002	0.046	0.572
Anaerovibrio	0.19^ab	0.18^b	0.19^ab	0.27^a	0.033	0.35	0.07	0.023	0.095	<0.001	0.188
Prevotellaceae YAB2003 group	0.15^ab	0.11^b	0.12^ab	0.26^a	0.039	0.19	0.13	0.027	0.019	0.039	0.100
Lachnospiraceae UCG-008	0.19	0.26	0.21	0.19	0.024	0.18	0.25	0.017	0.153	0.010	0.443
Desulfovibrio	0.12	0.17	0.12	0.09	0.022	0.12	0.13	0.015	0.121	0.469	0.874
Lachnospiraceae NK4A136	0.12	0.16	0.12	0.15	0.016	0.09	0.19	0.012	0.137	<0.001	0.078
Blautia	0.14	0.17	0.14	0.13	0.017	0.06	0.23	0.012	0.303	<0.001	0.379
Methanosphaera	0.11^ab	0.15^a	0.08^b	0.09^ab	0.016	0.03	0.18	0.010	0.040	<0.001	0.019
Prevotellaceae Ga6A1 group	0.09	0.05	0.14	0.07	0.023	0.05	0.12	0.015	0.084	<0.001	0.008
Moraxella	0.07	0.10	0.06	0.08	0.023	0.06	0.09	0.016	0.491	0.248	0.800
Lachnospiraceae UCG-006	0.06	0.05	0.04	0.03	0.009	0.04	0.05	0.007	0.585	0.688	0.250
Lachnospiraceae FCS020 group	0.05	0.08	0.07	0.05	0.010	0.04	0.08	0.007	0.251	<0.001	0.283

Keywords

Camelina sativaamplicon sequencingrumenmicrobiota

Peer Reviews

No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.

Videos

No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.

Taxonomy

TopicsLipid metabolism and biosynthesis · Ruminant Nutrition and Digestive Physiology · Nitrogen and Sulfur Effects on Brassica

Full text

Introduction

The rumen microbiome is highly dynamic and comprises bacteria, archaea, protozoa, and fungi that work synergistically to degrade dietary components (Henderson et al. 2015; Huws et al. 2018). Modulating the rumen microbial community through nutritional management is a key research area for optimizing feed efficiency and reducing greenhouse gas emissions (Jami and Mizrahi 2012; Huws et al. 2018). Camelina sativa is increasingly recognized as a protein–energy feed ingredient, rather than solely a lipid source (Riaz et al. 2022). In addition to reported effects on ruminant performance, Camelina sativa seeds (CS) and their by-products can potentially influence rumen metabolic processes, including fatty acid biohydrogenation and nitrogen metabolism. Their relatively high crude protein content and polyunsaturated fatty acids (PUFA) profile have been associated with favorable changes, including increased PUFA concentration in milk fatty acid composition, lamb tissue fatty acid profile, and ruminal nitrogen metabolism (Hurtaud and Peyraud 2007; Cieslak et al. 2013; Brandao et al. 2018; Riaz et al. 2022). These metabolic responses are closely linked to rumen microbial activity, as specific microbial groups are involved in biohydrogenation pathways, amino acid degradation, and the production of ammonia. Hence, evaluating the effects of different forms of oilseeds (e.g., whole seeds, oils, meals) in ruminant species is essential for a comprehensive understanding of their impact on rumen fermentation and microbiota populations. Lipid-rich feeds, including oilseeds such as C. sativa, have attracted considerable interest due to their high concentrations of PUFA, which can alter the rumen microbiome, redirect fermentation pathways, and, in some cases, lower methane emissions (Patra 2013). More specifically, the inclusion of 60 g/kg DM of Camelina oil had no significant effect on bacterial population and methane yield in lactating cows (Bayat et al. 2015). Furthermore, in a dual-flow continuous culture system trial, Camelina seed supplementation increased propionate and reduced acetate concentrations in a dose-dependent manner (Dai et al. 2017). It also decreased the relative abundances of Ruminococcus spp., Fibrobacter spp., and Butyrivibrio spp., while increasing Megasphaera and Succinivibrio in the liquid fraction (Dai et al. 2017). Interestingly, the reduction in fibrolysis and populations of fibrolytic bacteria such as Ruminococcus spp. and Fibrobacter spp. can potentially be linked with PUFA toxicity (Maia et al. 2007; Dai et al. 2017). In a previous study, significant changes in targeted microbial populations were found by the quantification of specific microbial species in individual ewes’ rumen fluid and solid particles fed varying levels of CS, using real-time quantitative polymerase chain reaction (RT-qPCR) (Christodoulou et al. 2023). More specifically, in the rumen solid fraction, a reduction in the relative abundance of Ruminococcus flavefaciens, a trend toward reduction in the relative abundance of Fibrobacter succinogenes, which are major fiber-degrading bacteria, as well as reductions in methanogenic archaea were found when 74.6 g/kg DM (16% inclusion level as fed) of CS were included in ewes’ diets (Christodoulou et al. 2023). However, the RT-qPCR analysis limits the assessment of overall rumen microbial diversity and broader community dynamics. High-throughput sequencing provides an untargeted assessment of both abundant and low-abundance taxa, supporting the investigation of microbial groups involved in fermentation, biohydrogenation, and nitrogen metabolism (Klindworth et al. 2013; Poretsky et al. 2014).

Thus, we hypothesized that applying 16S rRNA gene amplicon next-generation sequencing would reveal broader and fraction-specific shifts in rumen microbial diversity and community composition beyond those identified by targeted RT-qPCR analysis. Therefore, the objective of this study was to apply a high-throughput sequencing approach to expand upon previous findings from the same experimental material described in Christodoulou et al. (2023), providing a deeper insight into rumen microbiota in both rumen fluid and solid fractions, as well as associated rumen fermentation characteristics in response to graded levels of PUFA-rich and protein–energy CS.

Materials and methods

Experimental design and diets

The experimental procedures used in the present study were previously described by Christodoulou et al. (2023) and were carried out in accordance with protocols approved by the Ethics Committee for Research of the Agricultural University of Athens for animal handling, housing, and care (approval no. 000007/22-01-2017). Forty-eight dairy Chios breed ewes were assigned to four homogeneous groups (n = 12 per treatment), balanced by age (2–4 years), body weight (55.0 ± 6.5 kg), fat-corrected (6%) milk yield (1.85 ± 0.30 kg/d), and days in milk (67 ± 8 days). Ewes were separated into different groups based on their diets and managed separately to ensure each group only experienced their assigned treatment. During feeding, within each block, ewes were moved into individual feeders to enable precise feed intake for each animal. The experimental period spanned 60 days. The diets included a daily average of 1.5 kg concentrate mix, 1.5 kg alfalfa hay, and 0.2 kg wheat straw per ewe into two meals. Fresh water was available to all ewes ad libitum. Concentrates were formulated with three levels of CS by partially replacing soybean meal and maize grain. Specifically, CS were included at levels of 0% (0 g/kg DM; Control), 6% (28.0 g/kg DM; CS6), 11% (51.3 g/kg DM; CS11), and 16% (74.6 g/kg DM; CS16), respectively. The CS were incorporated as whole seeds with the rest of the concentrate ingredients into the unpelleted concentrate blends, to ensure uniform distribution at the abovementioned inclusion levels. Dietary treatment components and chemical composition are presented in Table 1. The fatty acid composition of the diets has been previously presented in Christodoulou et al. (2023).

Sample collection

Rumen digesta samples were collected on day 60 of the experiment before feeding. Samples were collected using an electric vacuum pump at 2 mbar (MZ2CNT, Vacuubrand GmbH & Co KG, Wertheim, Germany) and a stomach tube (flexible PVC tube of 1.5 mm thickness and 10 mm I.D.) as described for sheep and goats by Ramos-Morales et al. (2014) and considering the protocols of Muizelaar et al. (2020). The stomach tube was placed at a depth up to 120 cm, while the first quantity of fluid (approx. 20–30 mL) was discarded to reduce the effect of the saliva contamination. Rumen digesta was collected in pre-warmed 1-L glass bottles and hand-shaken. Solid particles were separated from the rumen fluid using four layers of bleached cheesecloth. Immediately after collection, ensuring no saliva contamination and after quality control, 9 out of 12 samples per group were snap-frozen in liquid nitrogen and stored at −80 °C for later analysis.

Ruminal volatile fatty acid and enzymatic activity analyses

Ten mL of the rumen fluid was centrifuged at 13,000 × g at 4 °C for 5 min, and then the supernatant was filtered under natural pressure through a polytetrafluoroethylene 0.45 μm syringe filter (Macherey-Nagel GmbH & Co., KG, Düren, Germany) and stored in three aliquots at −80 °C until the analysis of enzymatic activities. Each aliquot was defrosted only once to ensure enzyme functionality. Alpha-amylase activity was measured using a UV/Vis spectrophotometer (GENESYS 180, Thermo Fisher Scientific, Waltham, MA, USA) as described by Mavrommatis et al. (2021). Cellulase and xylanase activities were determined using the Petri dish method according to previously described protocols (Mavrommatis et al. 2025). The ImageJ densitometry software (version 1.6, National Institute of Health, Bethesda, MD, USA) was used for clearance zone quantitative analysis. Rumen samples were also used to measure volatile fatty acid (VFA) concentrations. More specifically, 0.8 mL of rumen fluid supernatant from the previous centrifugation (without the filtration) was acidified with 0.2 mL of 25% metaphosphoric acid. After 30 min of incubation at 4 °C, samples were centrifuged at 13,000 × g at 4 °C for 5 min and then the supernatant was diluted with cold extra pure water (1:2). Samples were then injected in an Agilent 6890 N gas chromatograph equipped with an HP-FFAP capillary column (30 m × 0.25 mm i.d. with 0.20 µm film thickness, Agilent, Santa Clara, CA, USA) and a flame ionization detector according to previously described protocols (Mavrommatis et al. 2025).

DNA extraction and 16S amplicon sequencing

DNA was extracted from 72 samples (36 rumen fluid and 36 solid fraction samples) following the protocol by Mavrommatis et al. (2021). Briefly, 1 g of each rumen fluid or solid sample was ground into a fine powder in a mortar with liquid nitrogen. This powdered material was immediately transferred to a Falcon tube containing preheated lysis buffer and incubated at 57 °C. RNase A was then added, followed by a 37 °C incubation. DNA extraction involved three rounds of chloroform-alcohol treatment, followed by isopropanol precipitation. Overnight, samples were centrifuged at 7,500 × g for 15 min at 4 °C, the supernatant was discarded, and the pellet was washed twice with ethanol. The DNA pellet was then resuspended in ultrapure water and further purified using a NucleoSpin Tissue spin column (Macherey-Nagel), following the manufacturer’s protocol. The extracted DNA quality was assessed based on abundance and purity (260/230 and 260/280 ratios), using an ND-1000 spectrophotometer (Nanodrop, Wilmington, DE, USA), and its integrity was also verified on a 0.7% agarose gel. Following quality control assessment, six of the nine rumen samples per dietary treatment and fraction were retained for downstream analysis, resulting in a total of 48 samples (24 rumen fluid and 24 rumen solid).

Bacterial DNA was amplified according to the “16S Metagenomic Sequencing Library Preparation” protocol (https://support.illumina.com/documents/documentation/chemistry_documentation/16s/16s-metagenomic-library-prep-guide-15044223-b.pdf) by Illumina (San Diego, CA, USA) using the primers described previously (Klindworth et al. 2013), which target the V3–V4 hypervariable regions of the 16S rRNA gene. All polymerase chain reaction (PCR) amplifications were performed in 25 μL volumes per sample. A total of 12.5 μL Phusion high-fidelity master mix 2 × (Thermo Fisher Scientific) and 0.2 μL of each primer (100 μM) was added to 2 μL genomic DNA as template (5 ng/μL). Blank controls (i.e., no DNA template added to the reaction) were also included. A first amplification step was performed in an Applied Biosystems 2700 thermal cycler (Thermo Fisher Scientific). The samples were denatured at 98 °C for 30 s, followed by 25 cycles with a denaturing step at 98 °C for 30 s, annealing at 56 °C for 1 min, and extension at 72 °C for 1 min, with a final extension at 72 °C for 7 min. The amplicons were cleaned using Agencourt AMPure XP beads (Beckman Coulter, Brea, CA, USA), and libraries were prepared following the 16S Metagenomic Sequencing Library Preparation Protocol (Illumina). The libraries obtained were quantified by real-time PCR using KAPA library quantification kits (Kapa Biosystems Inc., Wilmington, MA, USA), pooled in equimolar proportions, and sequenced in one MiSeq (Illumina) run with 2 × 250-bp paired-end reads. The raw 16S rRNA sequences were processed through a pipeline, including fragment rebuilding by PANDAseq (Masella et al. 2012) and quality filtering aimed at removing low-quality reads (i.e., showing stretches of bases with a Q score of <3 for more than 25% of their length). Bioinformatic analyses were conducted using the QIIME pipeline release 1.9.0 suite (Caporaso et al. 2011), clustering filtered reads into zero-radius operational taxonomic units (zOTUs) at the 97% identity level (Edgar 2016). In order to sort out putative chimeras, zOTUs supported by fewer than 5 reads across all samples were removed. Taxonomic assignment was performed by the RDP classifier (Wang et al. 2007) against the SILVA 138 database (Quast et al. 2013) using 0.8 as the confidence threshold.

Statistical analysis

The alpha diversity, which estimates the microbial species diversity on a single sample scale, was measured using the Chao1, Shannon’s diversity, observed species, and Faith’s phylogenetic diversity (“PD whole tree”) indexes. The dataset was downsampled to the least sequenced sample to have a comparable picture of the taxonomic composition after checking the alpha-diversity rarefaction curves. A non-parametric permutation-based t-test (equivalent to Mann–Whitney U-test), with 999 random permutations, was used to assess the difference in the alpha-diversity. Weighted and unweighted UniFrac distances (Lozupone et al. 2011) and Principal Coordinates Analysis (PCoA) was used to represent the microbial community structure for beta-diversity, which measures the variation of microbial communities between samples (Whittaker 1960). The “Adonis” test function (Permutational Multivariate Analysis of Variance Using Distance Matrices, using pseudo-F ratios) was used to define whether there was a significant difference among the experimental groups.

For the statistical analysis of the ruminal volatile fatty acid concentration, ruminal enzymatic activity, and the ruminal relative abundances of bacterial taxa, IBM SPSS v29.0 was used based on a linear mixed model. For the ruminal volatile fatty acid concentration and ruminal enzymatic activity, the dietary treatment (Control, CS6, CS11, CS16) was the fixed factor, and the ewe’s ID nested within dietary treatment was the random factor. Regarding the ruminal relative abundances of bacterial taxa, the dietary treatment (Control, CS6, CS11, CS16) and the fraction (rumen fluid and rumen solid) were the fixed factors, and the ewe’s ID nested within dietary treatment was the random factor. Normality was tested using the Kolmogorov–Smirnov test, and data violating the normality assumption were log-transformed. Pairwise comparisons of estimated marginal means for fixed effects were performed, with P values adjusted using the Bonferroni correction. A Pearson correlation analysis was performed to assess relationships among VFAs, ruminal enzyme activity, and microbial genera in the rumen fluid dataset. The significance level of all tests was set at *P *< 0.05.

Results

Volatile fatty acids and enzymatic activity

The VFAs’ accumulations are summarized in Table 2. Acetic and propionic acids were higher (*P *< 0.001) in CS11 than in Control and CS6. The iso-butyric, butyric, valeric acid, and total VFAs concentrations were higher (*P *< 0.001) in CS11 and CS16 compared with CS6, and in CS11 compared with Control. Iso-valeric acid was higher (*P *= 0.017) in CS11 than in Control. The acetic: propionic ratio was higher (*P *= 0.003) in Control than in CS11 and CS16.

The ruminal enzymatic activity is presented in Table 2. Amylase activity was lower (*P *< 0.001) in CS6 compared to the other dietary groups. Cellulase and xylanase activities did not differ (*P *> 0.05).

Rumen microbial communities

The amplicon-based 16S rRNA next-generation sequencing allowed us to characterize the main constituents of the rumen microbiota (Figures S1 and S2; see online supplementary material for a color version of these figure).

At the phylum level, Bacteroidota was the most abundant taxa, both for the fluid and the solid fractions (average rel. ab: 43.5% vs. 39.3%, respectively), followed by Firmicutes (27.6% vs.31.0%, respectively), Proteobacteria (11.5% vs. 6.0%, respectively), Spirochaetota (1.8% vs. 5.6%), Verrucomicrobiota (5.7% vs. 3.3%), and Fibrobacterota (2.4% vs. 5.3%, respectively).

At family level, the samples all showed a similar composition, with Prevotellaceae (29.3% vs. 30.0%), Succinivibrionaceae (10.7% vs. 5.2%), Lachnospiraceae (7.1% vs. 9.1%), and Rikenellaceae (5.3% vs. 5.6%) as the main groups of the rumen microbiota; uncultured rumen bacteria from WCHB1-41 (Kirimatellae), Bacteroidales RF16 group, Acidaminococcacea, Spirochaetaceae, Fibrobacteraceae, Oscillospiraceae, and Ruminococcaceae were also consistently present.

At genus level, among the main genera, we observed Prevotella, Succiniclasticum, Fibrobacter, and Treponema. In contrast, many other taxa remained unresolved at genus level (e.g., Succinivibrionaceae UCG-002, Rikenellaceae RC9 gut group, uncultured rumen bacteria from WCHB1-41, Bacteroidales RF16 group, Prevotellaceae UCG-001), due to the relatively worse characterization of the rumen in the reference databases.

The analysis of the rarefaction curves for the Chao1 metric determined that all the samples (both fluid and solid fractions) showed a tendency toward reaching a plateau at around 5,000–10,000 reads, suggesting that this number of sequences was sufficient to capture the majority of the ecosystem composition. Samples’ biodiversity (alpha-diversity) had a trend toward a reduction with increasing CS inclusion levels, with the CS11 and CS16 having a lower diversity than the Control (chao1, Shannon, PD whole tree, observed species metrics; *P *≤ 0.036 for all metrics). Furthermore, CS11 and CS16 composition (unweighted UniFrac distance) differed (*P *= 0.039) compared with the Control and CS6 (Figure 1, panels A–B).

A) Rarefaction curves of the alpha-diversity of the samples, estimated by Faith’s phylogenetic diversity metric, for the fluid part of the rumen microbiota; each line is the average value over all the samples within the same experimental group (i.e., diet); error bars represent standard errors. B) Principal Coordinate Analysis (PCoA) plot based on the unweighted UniFrac distance among samples for the fluid part of the rumen microbiota; each point represents a sample, colored according to the experimental group (i.e., diet); ellipses are the SEM-based confidence intervals, and centroids are the average coordinate of all the samples in the same group. C) Rarefaction curves of the alpha-diversity of the samples, estimated by Faith’s phylogenetic diversity metric, for the solid part of the rumen microbiota; each line is the average value over all the samples within the same experimental group (i.e., diet); error bars represent standard errors. D) Principal Coordinate Analysis (PCoA) plot based on the unweighted UniFrac distance among samples for the solid part of the rumen microbiota; each point represents a sample, colored according to the experimental group (i.e., diet); ellipses are the SEM-based confidence intervals, and centroids are the average coordinates of all the samples in the same group.

Rumen solid sample biodiversity (alpha diversity) did not differ among the dietary treatments. Regarding the microbial composition, this was different for both unweighted and weighted UniFrac distances for Control (*P *≤ 0.007) and CS6 (*P *≤ 0.004), and these two groups were separated from CS11 and CS16 (Figure 1, panels C–D).

The results for the relative abundances at the phylum level are presented in Table 3. More specifically, Proteobacteria relative abundance was higher (*P *< 0.001) in CS16 compared with the Control and CS6, and in CS11 than in CS6. Verrucomicrobiota, Fibrobacterota, and Euryachaeota relative abundances were higher (*P *< 0.001) in CS6 compared with CS11 and CS16. Verrucomicrobiota relative abundance was also higher (*P *< 0.001) in Control than in CS11 and CS16. Furthermore, the fraction effect was significant (P < 0.05) for every phylum except for Patescibacteria. More specifically, Bacteroidota and Proteobacteria were more abundant in the fluid than in the solid, while the remaining phyla showed the opposite behaviour (Table 3). The dietary treatment × fraction interaction was not significant (Table 3).

At the family level, we detected a significant (*P *< 0.05) increase of Prevotellaceae and Succinivibrionaceae in both CS11 and CS16, of Lachnospiraceae in CS11 only, and of Selenomonadaceae in CS16 only. At the same time, Rikenellaceae, Bacteroidales RF16 group, Acidaminococcaceae, F082, and Methanobacteriaceae were higher in CS6, as well as Uncultured rumen bacterium (WCHB1-41) Fibrobacteraceae, Christensenellaceae, and Hungateiclostridaceae were higher in Control and CS6 compared with samples from ewes fed a higher camelina concentration. Regarding the fraction effect, it was found to be significant for most of the families (Table 3). The dietary treatment × fraction interaction was significant (*P *= 0.001) for Christensenellaceae.

Significant (*P *< 0.05) effects of treatment, fraction, and their interaction at the genus level are presented in Table 3. In detail, Prevotella and Lachnospiraceae (other) were higher (*P *< 0.001) in CS11 and CS16 than in Control and CS6. Furthermore, Ruminococcaceae (other), Succinivibrionaceae UCG-002, Ruminobacter, and Succinimonas relative abundances were lower in CS6. Rikenellaceae RC9 gut group and uncultured rumen bacterium (F082) relative abundances were higher in CS6. Uncultured rumen bacterium (WCHB1-41) and Christensenellaceae R-7 group were higher (*P *< 0.001) in Control and CS6. Lachnospiraceae AC2044 group was higher in CS6 than in CS16. Fibrobacter, Methanobrevibacter, Saccharofermentans, and Lachnospiraceae ND3007 group relative abundances were higher in CS6 than in CS11 and CS16. Fibrobacter and Saccharofermentans were also higher in the Control compared with the CS16. Succiniclasticum and Methanosphaera were higher in CS6 than in CS11. Ruminococcus was higher (*P *< 0.001) in CS6 than in Control. Succinivibrionaceae UCG-001, Anaerovibrio, and Prevotellaceae YAB2003 group were higher in CS16 compared with CS6. Butyrivibrio and Pseudobutyrivibrio were higher (*P *< 0.001) in CS11 compared with Control and CS6. Selenomonas and Lachnospiraceae NK3A20 group relative abundances were higher in CS16 than in Control and CS6. The fraction effect was found significant for most of the genus (Table 3), while the dietary treatment × fraction interaction was found significant (*P *< 0.05) for Christensenellaceae R-7 group, Butyrivibrio, Selenomonas, Lachnospiraceae XPB1014 group, Prevotellaceae Ga6A1 group, and Methanosphaera (Table 3).

Correlation analysis

Core genera, including Prevotella, Succinivibrionaceae UCG-002, Selenomonas, Ruminococcaceae (other), and Succinimonas were positively correlated with acetic and propionic acids (*P *< 0.05). In contrast, Rikenellaceae RC9 gut group, Fibrobacter, Methanobrevibacter, F082 (other), Lachnospiraceae ND3007 group, Desulfovibrio, and Lachnospiraceae NK4A136 group were negatively correlated with acetic, propionic, butyric, and total VFAs and other VFAs. In addition, amylase was positively correlated with Succinivibrionaceae UCG-002, Ruminobacter, Ruminococcaceae (other), Pserudobutyrivibrio, Lachnospiraceae AC2044 group, and Lachnospiraceae XPB1014 group, while it was negatively correlated with F082 (other). Cellulase was negatively correlated with Treponema (File S2).

Discussion

Studying rumen microbial structure is crucial for understanding the complex interactions between diet, microbial populations, and host physiology, which directly influence fermentation efficiency, nutrient utilization, and overall animal homeostasis. This study, which shares the same experimental conditions as Christodoulou et al. (2023), builds upon those findings, which were based on RT-qPCR analysis of selected microbial taxa, by applying an untargeted and more detailed high-throughput sequencing approach. By employing 16S rRNA gene amplicon sequencing, this study provides a more comprehensive understanding of microbial dynamics, capturing the broader diversity and interactions within the rumen microbiota in both the fluid and solid fractions, as well as the distinct microbial profiles of each fraction. The alpha-diversity analysis indicated a shift toward less diverse microbial communities in the rumen fluid with increasing levels of CS, particularly in CS11 and CS16. The distinct beta-diversity observed between these higher levels of inclusion and the lower ones suggests significant compositional changes driven by diet. Regarding the rumen solid fraction, the stability in alpha-diversity contrasted with the compositional shifts between the dietary treatments. The different clustering of the Control and CS6 from CS11 and CS16 highlighted the diet-specific impact on the solid-associated microbiota.

As expected, the dominant phyla were Bacteroidota, Firmicutes, and Proteobacteria, a finding consistent with the literature (Morgavi et al. 2015). In parallel, the reduction in Fibrobacterota and Verrucomicrobiota relative abundances found in the higher C. sativa seed inclusion levels highlight a potentially crucial trade-off between fiber degradation and carbohydrate utilization, especially since microbes belonging to Fibrobacterota are critical for fiber degradation.

Considering the family level, Prevotellaceae, which increased at higher C. sativa levels, is one of the most dominant families in the rumen microbiome, including bacteria of the genus Prevotella, which play a significant role in carbohydrate and protein degradation (Griswold et al. 1999). Microbes within this family are involved in the degradation of starch, simple sugars, and protein, producing VFAs and organic acids such as acetate and succinate, respectively, as their principal fermentation products, with lesser production of branched-chain VFAs like iso-butyric and iso-valeric, as well as lactic acid (Tett et al. 2021; Cheng et al. 2022). Therefore, their increase in CS16 could potentially be linked with facilitated carbohydrate and protein degradation. Although the decline in Fibrobacteraceae (key fibrolytic taxa), including the genus Fibrobacter, which confirms previous data (Dai et al. 2017) may suggest reduced fibrolytic potential, the increase in Lachnospiraceae, some of which exhibit fibrolytic traits (Palevich et al. 2019) may represent a shift in fiber degradation rather than an overall suppression of fibrolysis. Microbes of the Lachnospiraceae family produce hydrogen during fermentation (Kaminsky et al. 2023), creating a natural competition for hydrogen between methanogens and other hydrogen-consuming microorganisms in the rumen. In addition, the increased abundance of Succinivibrionaceae in CS16 suggests enhanced succinate fermentation (Russell and Rychlik 2001), likely driven by increased availability of fermentable substrates, as members of this family produce succinate as the principal fermentation end product (Lee et al. 1999).

Several notable changes in microbial abundance were observed at the genus level, reflecting the influence of C. sativa seed inclusion levels on the rumen microbiota in both fluid and solid compartments. Among the main genera observed in both rumen fluid and solid particles were Prevotella, Succiniclasticum, and Fibrobacter, with Treponema additionaly present in the solid fraction. However, many other taxa remained unresolved at the genus level in both compartments (e.g., Succinivibrionaceae UCG-002, Rikenellaceae RC9 gut group, uncultured rumen bacteria from WCHB1-41, Bacteroidales RF16 group, Oscillospiraceae NK4A214 group, Prevotellaceae UCG-001), due to the relatively more limited characterization of the rumen microbiota in the reference databases. The increased abundance of Prevotella in the highest inclusion level (CS16) could indicate enhanced fermentation of readily fermentable substrates, which may contribute to altered fermentation patterns (Betancur-Murillo et al. 2022). While Prevotella relative abundance was not significantly altered in the targeted RT-qPCR analysis of Christodoulou et al. (2023), the 16S amplicon sequencing approach of this study revealed increased Prevotella relative abundance with higher CS inclusion. Considering the correlation analysis, the positive correlations between Prevotella and propionic acids may suggest an active role of Prevotella in carbohydrate fermentation and VFA production. Similar associations have been reported previously, particularly for Prevotella and Selenomonas, which are known to contribute to propionate formation and efficient energy use in the rumen. This increase may also reflect an inhibitory fibrolytic activity, in line with the reduced Fibrobacter abundance, suggesting a potential decrease in fiber degradation efficiency at higher inclusion levels of CS. This reduction may be attributed to the altered nutrient profile associated with increased Camelina inclusion, which could selectively favor or inhibit specific microbial groups. For instance, dietary PUFA can influence ruminal fatty acid biohydrogenation by altering the abundance of microbial groups involved in lipid metabolism. It is essential to link the results of this study and interpret the ruminal lipid metabolism with the ruminal fatty acid profiles reported by Christodoulou et al. (2023), which were generated from the same experimental animals and dietary treatments. Interestingly, increasing the inclusion of CS increased the ruminal availability of unsaturated fatty acids, particularly cis-9 C_18:1_. While biohydrogenation pathways remained active, as indicated by the presence of trans-C_18:1_ intermediates and responsive biohydrogenation-associated taxa at moderate inclusion levels, the fatty acid profile may suggest that the capacity for complete biohydrogenation was exceeded under higher PUFA supply. Specifically, cis-9 C_18:1_ was likely subjected to biohydrogenation but not fully converted to C_18:0_, consistent with a limitation at the terminal reduction step. The absence of a linear increase in Butyrivibrio abundance at the highest CS level further supports the interpretation of functional saturation rather than enhanced biohydrogenation capacity. Furthermore, PUFA toxicity has been linked to reduced cellulolytic activity (Maia et al. 2007; Dai et al. 2017). The reduction in Fibrobacter at higher C. sativa inclusion levels is consistent with the RT-qPCR results of Christodoulou et al. (2023), supporting the sensitivity of fiber-associated microbial populations to PUFA-rich oilseed supplementation. The potential PUFA-toxicity toward cellulolytic taxa such as Fibrobacter has been reported in goats fed oilseeds such as linseed (Abuelfatah et al. 2016) or dairy cows fed whole or ground flaxseed (Huang et al. 2021). Despite the observed reduction in such a key fibrolytic microbe, in this study, cellulase enzymatic activity remained unaffected, suggesting that fiber degradation may have been preserved, or that the sample collection time was not indicative of cellulase and xylanase activity, which typically occurs several hours after feeding. Plant cell wall polysaccharides are degraded during the later stages of ruminal fermentation, after soluble carbohydrates have been utilized and fibrolytic microbes have grown (Martin and Michalet-Doreau 1995). Interestingly, unclassified members of the Ruminococcaceae had a higher abundance in CS11 and CS16 compared to CS6, contrasting with the previously discussed decline in fiber-degrading microbes, particularly Fibrobacteraceae. This suggests the competition within the rumen microbiota (niche expansion theory), which could help balance fermentation pathways and the rate of carbohydrate degradation (Nagaraja 2016; Owens and Basalan 2016). Together, these results suggest that the extent of PUFA release in the rumen, rather than the specific oilseed source, plays a central role in shaping rumen microbial responses. In this study, the reduction in Ruminobacter, with potent amylolytic activity, alongside an increase in fiber-degrading groups like Ruminococcus and Fibrobacter, might be linked to the combined effects of plant compounds from CS and the lower ether extract content in the CS6 diet compared with CS11 and CS16. Moreover, CS contain phenolic compounds and glucosinolates. These compounds can bind to microbial enzymes and interact with nutrients such as starch and protein, exerting antimicrobial effects (Giuberti et al. 2020). The reduced amylase activity observed in the CS6 supports the hypothesis of inhibited amylolytic function. In CS6, which is a lower-fat diet compared with CS11 and CS16, phenolic compounds were likely more bioavailable and able to interact directly with rumen microbes and enzymes such as amylase, thereby limiting starch degradation (Sun et al. 2019). A wide variety of plant-derived feed additives have been investigated as rumen habitat modifiers due to their ability to selectively influence rumen microbial activity and fermentation patterns (Tsiplakou et al. 2023). Essential oils, which are secondary metabolites extracted from the volatile fraction of plants, have been reported to exert inhibitory effects on ruminal microbial activity through their action on functional bacterial groups involved in deamination and starch utilization, including hyper-ammonia-producing bacteria such as Prevotella spp. and Ruminobacter amylophilus (Tsiplakou et al. 2023). Therefore, aromatic plants and essential oils rich in phenolic compounds can theoretically selectively inhibit starch-degrading microbes without completely suppressing overall rumen fermentation (Calsamiglia et al. 2007). In addition, starch-degrading microbes may have reduced competition for fiber, potentially promoting fiber-degraders to grow. In the higher fat content diets (CS11 and CS16), fats may coat feed particles or interact with phenols, reducing their antimicrobial activity and allowing starch-degrading microbes to thrive. This could potentially promote starch-degrading microbes. Dietary fats can shift rumen microbial populations, potentially promoting microbes less sensitive to phenols or more efficient at starch digestion despite phenolic presence, although excessive fat may also suppress fiber degraders such as Ruminococcus (Maia et al. 2007). Considering the abovementioned, in CS6, phenolic compounds may have a stronger inhibitory effect on starch-degrading while supporting fiber-degrading microbes, whereas in CS11 and CS16, the higher fat content levels likely masked the phenolic effect, favoring starch-utilizing microbes and suppressing fiber breakdown. In addition, the lowest VFA concentrations were observed in the CS6, potentially due to enhanced phenolic bioavailability and reduced starch fermentation, while CS11 and CS16, with higher fat and greater Camelina seed inclusion levels, supported higher VFA production than both CS6 and the Control. Overall, these findings highlight the importance of considering both phenolic compounds and dietary fat when evaluating the effects of Camelina seeds on rumen fermentation and microbial populations.

Another key finding was the increased abundance of Selenomonas, known for its ability to utilize lactate and produce propionate, in CS16, which assists in maintaining rumen pH stability by preventing lactate accumulation (Millen et al. 2016). This result is consistent with our previous findings from targeted RT-qPCR analysis of Selenomonas ruminantium, as reported in Christodoulou et al. (2023). The positive correlation of Selenomonas with acetic and propionic acids further supports its potential involvement in carbohydrate fermentation and VFA production under this dietary treatment. The increased abundance of Selenomonas in CS16 could suggest its adaptability to different ruminal niches. It may be attributed to the nutrient profile of CS, particularly its PUFA and fermentable substrates. The presence of PUFA, especially omega-3 fatty acids, could modulate rumen microbial populations by favoring propionate-producing pathways, while potentially reducing acetate production, as in previous studies with dietary flaxseed inclusion (Vargas et al. 2020; Huang et al. 2021). This shift may also have implications for methane mitigation, as propionate production competes with methanogenesis for hydrogen, thereby reducing methane emissions (Wang et al. 2023). This is further supported by the decreased (numerically) abundance of Methanobrevibacter at higher Camelina levels, although the studied region (V3–V4) is not representative of these taxa (Zhou et al. 2021). In compliance with the present findings, reported reductions in methanogenic archaea with increasing CS inclusion levels were also observed through the targeted RT-qPCR analysis in Christodoulou et al. (2023), suggesting that PUFA-rich diets potentially suppress methanogen-associated communities, despite differences in analytical resolution between RT-qPCR and 16S amplicon sequencing. In addition, the reduced acetic: propionic ratio in CS11 and CS16 highlights a shift toward pathways favoring propionate production, and competing with methanogenesis for hydrogen utilization (Hook et al. 2010). Moreover, the negative correlation with Methanobrevibacter could also point to competition for hydrogen between methanogenesis and propionate-producing pathways. In accordance with this finding, the Succinivibrionaceae family, which was in higher abundances in CS16, could also be linked with the higher propionate production in these treatments (Ren et al. 2019; Han et al. 2021).

The different responses between rumen fluid and solid microbiota are critical to understanding rumen functionality. The results of the present study highlight the complex microbial shifts occurring in both rumen fluid and solid particles in response to varying levels of C. sativa seed inclusion. Considering the abovementioned, the present 16S amplicon sequencing-based findings extend the observations of Christodoulou et al. (2023) by demonstrating that PUFA-rich CS inclusion induces consistent shifts in rumen microbial communities across multiple taxonomic levels, while also revealing additional taxa and patterns that are not detectable using targeted approaches. The rumen fluid microbiota, which helps break down rapidly fermentable nutrients, showed reduced diversity and shifts in composition. In contrast, the fiber-attached microbes in the solid phase had a more stable diversity, but their composition changed in ways that suggest a shift in function. Holistically, profiling both compartments with 16S amplicon sequencing provides valuable insights into how dietary modifications influence microbial ecology and rumen function. Nevertheless, dairy ewes’ performance in this trial was not significantly affected (Christodoulou et al. 2021). This could suggest that despite the microbial shifts and the different fermentation patterns, ewes could efficiently digest their feed and extract sufficient nutrients for maintenance and production. Overall, additional key parameters, such as pH and ammonia, should be considered to better link the rumen microbiome with rumen function and metabolism.

Conclusion

This study highlights that the complex interaction between fat content and plant compounds such as glucosinolates in the diet can modify the rumen microbiome. Inclusion of C. sativa seeds at a lower fat content level (CS6) appeared to favor fiber-digesting bacteria like Fibrobacter, likely due to the enhanced activity of bioavailable phenolics suppressing starch-degrading microbes. In contrast, higher fat content diets (CS11 and CS16) reduced the microbiota diversity (for the rumen fluid compartment) and showed potentially reduced phenolic impact, most likely by altering their availability or through direct fat effects, favoring amylolytic populations. Therefore, balancing both secondary metabolites and nutrients in ruminant diets is essential to optimize rumen fermentation and microbial function.

Supplementary Material

skag033_Supplementary_Data

Bibliography51

The reference list from the paper itself. Each links out to its DOI / PubMed record.

1Abuelfatah K , Zuki AB, Goh YM, Sazili AQ, Abubakr A. 2016. Effects of feeding whole linseed on ruminal fatty acid composition and microbial population in goats. Anim Nutr. 2:323–328. 10.1016/j.aninu.2016.10.00429767110 PMC 5941056 · doi ↗ · pubmed ↗
2Bayat AR et al 2015. Effect of camelina oil or live yeasts (Saccharomyces cerevisiae) on ruminal methane production, rumen fermentation, and milk fatty acid composition in lactating cows fed grass silage diets. J Dairy Sci. 98:3166–3181. 10.3168/jds.2014-797625726099 · doi ↗ · pubmed ↗
3Betancur-Murillo CL , Aguilar-Marín SB, Jovel J. 2022. Prevotella: a key player in ruminal metabolism. Microorganisms. 11:1. 10.3390/microorganisms 1101000136677293 PMC 9866204 · doi ↗ · pubmed ↗
4Brandao VLN et al 2018. Effect of replacing calcium salts of palm oil with camelina seed at 2 dietary ether extract levels on digestion, ruminal fermentation, and nutrient flow in a dual-flow continuous culture system. J Dairy Sci. 101:5046–5059. 10.3168/jds.2017-1355829525304 · doi ↗ · pubmed ↗
5Calsamiglia S , Busquet M, Cardozo PW, Castillejos L, Ferret A. 2007. Invited review: essential oils as modifiers of rumen microbial fermentation. J Dairy Sci. 90:2580–2595. 10.3168/jds.2006-64417517698 · doi ↗ · pubmed ↗
6Caporaso JG et al 2011. Global patterns of 16S r RNA diversity at a depth of millions of sequences per sample. Proc Natl Acad Sci USA. 108(Suppl 1):4516–4522. 10.1073/pnas.100008010720534432 PMC 3063599 · doi ↗ · pubmed ↗
7Cheng Z et al 2022. Effects of supplementation of sodium acetate on rumen fermentation and microbiota in postpartum dairy cows. Front Microbiol. 13:1053503. 10.3389/fmicb.2022.105350336478854 PMC 9720668 · doi ↗ · pubmed ↗
8Christodoulou C et al 2023. Inclusion of camelina sativa seeds in ewes’ diet modifies rumen microbiota. Animals (Basel). 13: 377. 10.3390/ani 13030377 PMC 991382536766266 · doi ↗ · pubmed ↗