henhaobio
Sign in

Nature CommunicationsMicrobiologyPeer-reviewed journal article

Gut Bacteria Encode Reductases that Biotransform Steroid Hormones

Gabriela ArpiD, Angela K JiangiD, Keith Dufault‐ThompsoniD, Sophia LevyiD, Aoshu ZhongiD, Jyotsna Talreja WassaniD, Maggie R. GrantiD, Yue Li, Brantley HalliD, Xiaofang JiangiD

PDFDOIPreprintCodeData
In-depth full-text digest available · Jump to digest ↓

Abstract

The metabolism of steroids by the gut microbiome affects hormone homeostasis, impacting host development, mental health, and reproductive functions. In this study, we identify the {Delta}4-3-ketosteroid 5{beta}-reductase, 3{beta}-hydroxysteroid dehydrogenase/{Delta}5-4 isomerase, and {Delta}6-3-ketosteroid reductase enzyme families encoded by common human gut bacteria. Through phylogenetic reconstruction and mutagenesis, We show that 5{beta}-reductase and {Delta}6-3-ketosteroid reductase have evolved to specialize in converting diverse 3-keto steroid hormones into their 5{beta}- and {Delta}6-reduced derivatives. We also find that the novel 3{beta}-hydroxysteroid dehydrogenase/{Delta}5-4 isomerase is fused with 5{beta}-reductase in multiple species, streamlining the multi-step conversion of pregnenolone, a steroid hormone precursor, into epipregnanolone. Through metagenomic analysis, we reveal that these enzymes are prevalent in healthy populations, being enriched in females over males. These findings provide the molecular basis for studying microbial steroid metabolism in the gut, offering insights into its potential impact on hormonal health in hosts, especially in the context of womens health.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (5 Oct 2024) · CC0 1.0

Gut bacteria encode 5β-reductase, fused 3β-HSDH/Δ5-4 isomerase and Δ6-3-ketosteroid reductase enzymes acting on steroid hormones; two of these genes are enriched in female gut metagenomes.

The enzymes that reduce progesterone to 5β-reduced and tetrahydroprogesterone derivatives were unidentified in gut microbes, which limited the study of microbial steroid hormone metabolism. This work identifies the enzyme families and their evolutionary specialization, and it surveys their prevalence across 1549 human gut metagenomes.

Key results

  1. ShownComparative genomics plus heterologous expression identified the Old Yellow Enzyme ci2350 from Clostridium innocuum as the progesterone 5β-reductase. Its paralog ci0705 had the same presence/absence pattern but no detectable activity.

    Seven putative old yellow enzymes in C. innocuum; of 13 orthogroups with the reducer-only distribution, 2 were annotated as old yellow enzymes (ci2350, ci0705). Only ci2350 gave detectable 5β-dihydroprogesterone in E. coli. · Comparative genomics uncovers an Old Yellow…; Identifying Candidate Δ4-3-ketosteroid…

  2. Shownci2350 acts selectively on steroid hormones. It reduced progesterone, cortisone and hydrocortisone, but not the cholesterol derivative cholest-4-en-3-one. Among synthetic progestins, only pregnane and estrane types were reduced.

    Of eight progestins tested, pregnane and estrane types were reduced, while gonanes and spironolactone derivatives were not. Substrate panel: biological triplicates in ci2350 E. coli vs vector-backbone and media controls. · ci2350 is a selective 5β-reductase for steroid…; Figure legends

  3. ShownA Rossmann-fold domain assigned 3β-HSDH/Δ5-4 isomerase function is fused to the 5β-reductase in 61 of 257 homologs. Two fused genes converted pregnenolone to a mixture including epipregnanolone in E. coli, while the separate 3β-HSDH/I gene gave only progesterone.

    The 5β-reductase clade has 318 genes from 282 species; fusion occurs in 61 of 257 homologs. Fused mf2052 and fp1951 gave progesterone, 5β-dihydroprogesterone and epipregnanolone; dw0523 gave only progesterone; dw0526 alone showed no activity in the assay. · Natural fusion of 3β-HSDH/Δ5-4 isomerase and…

  4. SuggestedMutating predicted active-site residues of ci2350 sharply reduced progesterone reduction. Bile-acid-clade-like residues kept partial activity, and C. scindens baiCD showed limited activity, which is consistent with subfunctionalization within a shared steroid reductase family.

    N253A lost nearly all activity; N253V kept a small detectable level; N166A and N166S nearly lost activity; N166T, S171A and S171E gave low but detectable activity. baiCD showed limited but detectable activity in E. coli. · Delineating steroid hormone 5β-reductases

  5. ShownIn 1549 human gut metagenomes, 5β-reductase and 3β-HSDH/I were significantly more prevalent and abundant in females. Δ6-RD(c2) was more prevalent in females, but the difference was not significant.

    890 female vs 659 male samples. 5β-reductase 94.61% vs 90.59% (P=3.26e-03); 3β-HSDH/I 82.36% vs 74.51% (P=2.25e-04). Abundance: P=6.65e-03 and P=1.37e-03. Δ6-RD(c2) 57.42% vs 52.50%, not significant. · Differential presence of steroid reductases in…; Profiling of steroid hormone reductases in…; Figure legends

Methods, data and reproducibility

Comparative genomics of progesterone-reducing and non-reducing gut bacteria, E. coli expression with LC-MS product detection, ancestral sequence reconstruction with site-directed mutants, and HMM-based read mapping across human gut metagenomes.

  • Genomes of three experimentally confirmed progesterone reducers and three non-reducers, used for orthogroup comparison · Gut bacteria (C. innocuum, R. gnavus and others) · n = 3 reducers, 3 non-reducers · GCA_000183585.2; GCA_900447045.1; GCA_000507805.1; GCA_001077535.1; GCA_000189595.1; GCA_000469345.1
  • GTDB r214 representative genomes searched for ci2350/ci2349 homologs (318 5β-reductase genes from 282 species; 94 Δ6-RD genes) · bacteria · n = 318 genes; 94 genes
  • Public human gut metagenomes from 13 studies, mapped to HMM-derived gene sets for 5β-reductase, 3β-HSDH/I and Δ6-RD(c2) · human · n = 1549 samples (890 female, 659 male)
  • LC-MS steroid conversion assays in cultures and in E. coli expressing candidate genes and mutants · bacteria (E. coli host) · n = biological triplicates for substrate panel
Methods & tools
Prokka v1.14.5/v1.14.6, OrthoFinder v2.5.5, eggNOG-mapper v2.1.3; BLASTp v2.15.0+, MUSCLE v5.1; IQ-TREE v2.1.2 (LG+I+G4, 1,000 ultrafast bootstraps), AlphaFold v2.2.0, fpocket v4.0, AutoDock Vina v1.2.5, TMalign, HMMER v3.4; ClustalO v1.2.4, Bowtie2 v2.5.3, GTDB r214, UHPLC-MS (Bruker Maxis-II QTOF)
Software & environment
Versions stated: Prokka v1.14.5/v1.14.6, OrthoFinder v2.5.5, MUSCLE v5.1, IQ-TREE v2.1.2, AlphaFold v2.2.0, HMMER v3.4, Bowtie2 v2.5.3, AutoDock Vina v1.2.5, fpocket v4.0.
Reproducibility gaps
HMM profiles are public, but no raw LC-MS data deposit or analysis scripts are stated; gene lists, metagenome metadata, strains and primers are in supplementary tables.
Limitations (5)
  • The authors say it is unclear whether the relatively small female-male abundance differences for 5β-reductase and 3β-HSDH/I affect overall steroid-metabolizing activity of the gut microbiome. · stated by the authors
  • The reasons for enrichment in females remain uncertain; the authors only propose possible factors such as hormonal fluctuation and contraceptive exposure. · stated by the authors
  • Other influences on gut steroid hormones (conjugation and sulfonation, and metabolism by Bacteroidetes and Eggerthellaceae) were not studied, so significant knowledge gaps remain. · stated by the authors
  • The authors attribute the weak progesterone reduction in C. scindens cultures, despite baiCD activity in E. coli, only to possible differences in cellular transport and regulation. · noted in this summary
  • The metagenome comparison pools 13 studies with a CPM>1 presence threshold and no stated adjustment for study, age, diet or contraceptive status; activity rests on heterologous E. coli assays with relative LC-MS intensities, without kinetics. · noted in this summary

Read this if you study gut microbial steroid or hormone metabolism and want the enzyme families, their evolutionary specialization, and sex-linked prevalence in human metagenomes.

Citations
40
FWCI
10.75
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISBeliaeva '26Jacoby '25Binion '26Chen '19Li '24Netter '25
64% semantic matchMicrobiology
Mariia A. Beliaeva, Matthias W. Gross, Dimitrios Papagiannidis+3 more·bioRxiv·Jul 14, 2026