PMA-Treated DNA Reveals Hidden Rumen Microbes Under Methane-Mitigating Additives

Rumen microbiologists have long faced a stubborn measurement problem: the DNA they sequence does not always come from living, functioning cells, and the RNA they sequence does not always capture every viable microbe. A new study published in Environmental Microbiome tackles this problem head-on by comparing three nucleic acid approaches—total DNA, propidium monoazide-treated DNA, and RNA-derived cDNA—under three contrasting methane-mitigation conditions in an in vitro rumen fermentation system. The results show that the choice of sample type materially changes how scientists read the rumen microbial community, including which taxa appear to respond to methane-lowering additives.

The research team, led by Woohyung Lee and Tansol Park, collected rumen contents from three Hanwoo cows via stomach tubing before morning feeding, following approval by the Institutional Animal Care and Use Committee of Chung-Ang University. The rumen fluid was pooled, filtered, mixed with a CO2-flushed buffer, and dispensed into serum bottles containing a forage-and-concentrate substrate. Four experimental groups were tested in a completely randomized design with six replicates each: an untreated control, 3-nitrooxypropanol (NOP) at 0.02% of substrate dry matter, monensin (MON) at 0.0346% of substrate dry matter, and the bromoform-containing red seaweed Asparagopsis sp. at 1% of organic matter. Bottles were incubated for 24 hours at 39 degrees Celsius under anaerobic conditions.

The three additives were chosen deliberately as contrasting ecological perturbation models rather than as candidates for efficacy testing. NOP directly inhibits methyl-coenzyme M reductase, the key enzyme of methanogenic archaea. Monensin, an ionophore, modifies bacterial fermentation and generally shifts hydrogen use toward propionate production. Asparagopsis suppresses methanogenesis primarily through bromoform and may exert broader effects on the microbial community. The fermentation data confirmed that distinct methane-mitigated environments had been established: methane production, both in total volume and per gram of degraded dry matter, was lowest in the NOP group and highest in the control, while hydrogen production peaked under NOP. Monensin produced the harshest fermentation profile, with significantly lower dry matter, neutral detergent fiber, and acid detergent fiber digestibility, lower total gas and total volatile fatty acid concentrations, and the lowest total ciliate counts.

To probe the viability question, the researchers applied propidium monoazide, marketed as PMAxx, to aliquots of the fermenting cultures. Propidium monoazide penetrates cells whose membranes are compromised and binds their DNA, preventing that DNA from being amplified in subsequent PCR. The result is a profile enriched for membrane-intact cells—an indirect, membrane-integrity-based estimate of viability that sits conceptually between total DNA, which may include extracellular DNA and DNA from dead or damaged cells, and RNA, which reflects transcriptional activity but can underrepresent viable populations with low RNA abundance. The team then sequenced the V3-V4 region of the 16S rRNA gene from total DNA (tDNA), PMA-treated DNA (vDNA), and RNA-derived cDNA on an Illumina MiSeq platform, processing the reads in QIIME2 with DADA2 and classifying taxa against NCBI reference sequences.

Alpha diversity told a clear and consistent story. Metrics including observed amplicon sequence variants, Chao1 estimates, evenness, and Shannon’s index decreased in the order tDNA, then vDNA, then RNA across all treatments, with P values below 0.001. Faith’s phylogenetic diversity and Simpson’s index did not differ significantly between tDNA and vDNA but were significantly lower in RNA samples. Among treatments, the control group showed the highest alpha diversity on every metric, while monensin consistently produced the lowest. Notably, richness-related metrics showed no significant interaction between sample type and treatment, meaning RNA profiles were uniformly less diverse regardless of the additive, whereas evenness, Shannon’s, and Simpson’s indices did show significant interactions, suggesting the treatments reshaped the RNA-represented fraction of the community more strongly than the DNA-based fractions.

Community structure comparisons revealed a metric-dependent pattern that the authors highlight as one of the study’s key insights. Procrustes analysis, which quantifies the alignment between paired ordinations, showed statistically significant similarity among all three sample type pairs. But which pair looked most similar depended on the distance metric. Using Bray-Curtis distances, which are abundance-based, tDNA and RNA communities were the most similar, with a Procrustes statistic M2 of 0.078. Using Jaccard distances, which consider only presence or absence, vDNA and RNA were the most similar, with an M2 of 0.314. The authors interpret this contrast to mean that the dominant taxa represented in RNA were also abundant in the total community, while the presence-absence composition of RNA more closely resembled the membrane-intact fraction captured by PMA treatment. In other words, the apparent agreement among sample types hinges on whether relative abundance or taxon occurrence is being measured.

Differential abundance analysis using ANCOM-BC2 with Holm-Bonferroni correction identified taxa represented unequally across sample types. At the phylum level, Bacillota showed significant differences in the control, NOP, and ASP groups, consistently displaying its lowest abundance in RNA samples, while Pseudomonadota showed significant differences in the control and NOP groups and reached its highest abundance in RNA. At the genus level, Ruminobacter and Parabacteroides were the only genera with significant sample-type differences across all four treatment groups: Ruminobacter was consistently enriched in RNA, whereas Parabacteroides was enriched in vDNA, suggesting it belongs to membrane-intact cells that are weakly represented in the RNA pool. Succinivibrio, a succinate- and acetate-producing bacterium, showed significantly reduced RNA-based abundance under all three methane-mitigated conditions while remaining relatively stable in tDNA and vDNA, a pattern the authors suggest may reflect a low-activity or possibly dormant state in which the population persisted without strong transcriptional representation.

Correlation analysis linking microbial genera to methane-related fermentation parameters—methane volume, hydrogen volume, and the acetate-to-propionate ratio—further underscored how sample type shapes biological interpretation. Succiniclasticum, a bacterium that converts succinate to propionate through a vitamin B12-dependent pathway, showed a negative correlation with the acetate-to-propionate ratio in tDNA but a positive correlation in RNA under monensin, with no significant correlation detected in vDNA. The authors propose, cautiously, that monensin’s suppression of Butyrivibrio, a potential cobalamin producer, may have limited vitamin B12 availability, and that DNA from membrane-compromised Succiniclasticum cells contributed to the tDNA-based association—but they stress that vitamin B12 concentrations and pathway-specific gene expression were not directly measured, so this mechanism remains hypothetical. Several correlations emerged exclusively in vDNA, including a significant positive correlation between Desulfovibrio and methane production under NOP (r = 0.8461), which the authors note does not necessarily indicate that Desulfovibrio promoted methane production but may reflect a shared response to hydrogen-rich conditions.

Predicted functional profiles, generated with PICRUSt2 from the 16S data, added another layer with important caveats. Among 13 predicted KEGG modules related to methane metabolism, three—M00345 for formaldehyde assimilation, M00358 for coenzyme M biosynthesis, and M00567 for methanogenesis from CO2 to methane—showed at least one significant sample-type difference in every treatment group. Under monensin, several methane-related modules showed higher predicted abundance in vDNA than in RNA, which the authors suggest may reflect membrane-intact methanogens with low RNA representation, although Methanomassiliicoccus and Methanobrevibacter were only numerically higher in vDNA under monensin, and the broad suppression of other microbes by monensin could have inflated their relative proportions. Because these functions were predicted from 16S rRNA profiles rather than directly measured, the authors label these interpretations hypothetical and in need of validation through shotgun metagenomics, metatranscriptomics, and absolute quantification.

The study’s conclusions are deliberately measured. Membrane integrity, the authors emphasize, does not establish viability in a strict physiological sense, and RNA-based 16S profiling does not directly measure metabolic activity. What the work demonstrates is that tDNA, vDNA, and RNA each offer a distinct and partially complementary window on the rumen community, and that the identification of potential methane-related biomarkers depends partly on which window is chosen. PMA-treated DNA, by stripping away signals from membrane-compromised cells while retaining populations that RNA underrepresents, may serve as a practical additional approach when RNA analysis is constrained by sampling, preservation, or extraction difficulties—though it is not a substitute for RNA in all contexts. For a field racing to understand how methane-lowering feed additives reshape the rumen ecosystem, the message is that the instrument used to look determines, in part, what can be seen.

Subject of Research: Comparative 16S rRNA profiling of total, PMA-treated, and RNA-derived rumen microbial communities under three in vitro methane-mitigation treatments

Article Title: Comparative metataxonomic analysis of total, viable, and active rumen microbial communities under methane-mitigated conditions

Article References: Lee, W., Kim, G., Li, W., & Park, T. (2026). Comparative metataxonomic analysis of total, viable, and active rumen microbial communities under methane-mitigated conditions. Environmental Microbiome, 21(1), Article 122. https://doi.org/10.1186/s40793-026-00963-y

Image Credits: AI Generated

DOI: 10.1186/s40793-026-00963-y

Keywords: rumen microbiome, methane mitigation, propidium monoazide, 16S rRNA sequencing, 3-nitrooxypropanol, monensin, Asparagopsis, metataxonomics, PICRUSt2, microbial viability, in vitro fermentation, ANCOM-BC2