Bacteria can make testosterone. What could that mean for prostate cancer?
Auburn physicists help reveal how a urinary microbe carries out hormone chemistry, one atom at a time.
Auburn University Department of Physics
image:
A bacterium isolated from the human urinary tract can convert DHEA into testosterone. Molecular simulations helped researchers explain how differences in the shape and motion of two bacterial enzymes determine which steroid reactions they can perform.
view moreCredit: Auburn University Physics.
AUBURN, AL. Many prostate cancers depend on androgen hormones, including testosterone, to grow. That is why treatments for advanced disease often try to reduce testosterone production or block its effects. Now, scientists have identified an unexpected participant in androgen chemistry: a bacterium from the human urinary tract that can convert the steroid precursor DHEA into testosterone under laboratory conditions.
The finding does not mean that the bacterium causes prostate cancer, nor does it show that testosterone produced by bacteria reaches tumors or changes the course of disease. It does, however, reveal a previously unknown microbial route to a hormone that is central to prostate biology, raising new questions about whether the microorganisms living near the prostate could help shape its chemical environment.
Published in Nature Communications, the study brought together experts in microbiology, cancer biology, chemistry, genomics and computational physics. Researchers in Auburn University’s Department of Physics contributed computational structural biology that helped explain, at the atomic level, how the newly discovered bacterial machinery works.
“We are not saying that these bacteria cause cancer,” said Rafael Bernardi, associate professor in Auburn University’s Department of Physics and a co-author of the study. “What we now know is that they possess the molecular machinery to produce testosterone. Because prostate cancer is so closely connected to androgen signaling, that is something worth understanding.”
An unexpected ability in the urinary microbiome
For decades, urine was commonly assumed to be sterile, and the urinary tract was not included in the original Human Microbiome Project. Scientists now know that it contains its own community of microorganisms, often called the urinary microbiome or urobiome. Much of the research in this young field has focused on determining which microbes are present and whether those communities differ between healthy people and people with disease. The new study asked a more difficult question: What are those microbes actually doing?
The research team examined bacteria isolated from urine samples collected from men before prostate biopsy. To find organisms capable of transforming steroids, the researchers developed a rapid screening method called the Human Sterolbiome Discovery High-throughput assay, or HSDH assay. Among the organisms they identified was Actinobaculum massiliense. When supplied with DHEA, the bacterium produced intermediate steroid molecules and ultimately testosterone.
DHEA, or dehydroepiandrosterone, is a steroid naturally produced by the human body. Human tissues can use it as a starting material to make more potent hormones, including testosterone. The new work shows that a bacterium living in the urinary tract can carry out similar chemistry using its own enzymes.
“The urinary microbiome has often been studied by asking which organisms are there,” said Jason M. Ridlon, who conceptualized and supervised the study. “We wanted to understand what those organisms are capable of doing. The discovery of this androgen-producing pathway gives us specific genes and enzymes that can now be investigated in the context of urinary health and disease.”
After confirming the bacterium’s testosterone-producing ability, the researchers searched its genome for the genes responsible. They identified two candidates, which they named dirA and dirB, for DHEA isomerase reductase. Laboratory experiments showed that DirA was unusually versatile, performing several different steroid transformations and allowing the bacterium to reach testosterone through more than one route. DirB could perform only part of that chemistry.
That difference presented a puzzle. The two enzymes appeared broadly similar and contained many of the same chemical components, yet one behaved like a multipurpose steroid-processing machine while the other had much narrower abilities. Solving that puzzle required looking beyond the genes and watching the molecules move.
Watching the chemistry in motion
At Auburn, Bernardi’s group generated three-dimensional models of DirA and DirB and used molecular simulations to follow what happened when steroid molecules entered each enzyme. Rather than treating the proteins as frozen structures, the researchers created atom-by-atom movies showing the enzymes flexing, the steroids rotating and the surrounding molecules shifting over time.
That motion is essential to understanding how enzymes work. A molecule does not react simply because it fits inside a protein. The correct part of the molecule must also face the correct part of the enzyme at the right distance and angle. A steroid can remain tightly bound inside an enzyme and still be chemically useless if it is pointing the wrong way.
The Auburn simulations revealed that DirA has a broad, open pocket that gives the steroid room to move and reposition itself. This flexibility allows different parts of the steroid to approach the enzyme’s catalytic machinery during different stages of the pathway from DHEA to testosterone. DirB, by contrast, has a much narrower internal tunnel. The steroid can enter and remain inside, but the restricted space frequently leaves it flipped or misaligned, with the portion that needs to react facing away from the catalytic machinery.
“At this scale, chemistry depends on choreography,” Bernardi said. “The steroid has to be in the right place, facing the right way, at the right moment. One enzyme gives it room to do that. The other does not.”
That relatively simple difference in molecular architecture explained the experimental results. DirA could support a broad range of reactions because the steroid had enough room to turn and present different parts of itself to the enzyme. DirB could perform only those reactions compatible with its more confined pocket.
Raissa Rosa, a doctoral candidate in Bernardi’s group, performed the computational studies and developed the catalytic models in collaboration with Bernardi. The Auburn researchers combined AI-based protein modeling with atom-by-atom simulations, using software that Bernardi’s group has helped develop.
“A static structure can show us that a molecule fits inside a protein,” Rosa said. “The simulations tell us whether it can reach the precise orientation needed for the reaction and whether that arrangement remains stable. In this case, those movements helped explain why the two enzymes behave so differently.”
A new question for prostate cancer research
Testosterone and other androgens are essential for normal prostate development and function, but they can also promote the growth of many prostate cancers by activating the androgen receptor. For that reason, therapies for advanced prostate cancer frequently aim to suppress androgen production or block androgen signaling.
The discovery that a urinary bacterium can generate testosterone introduces a new question: Could microbial metabolism contribute to the androgen environment near the prostate? The present study does not answer that question. It demonstrates that the bacterium has the biochemical ability to produce testosterone under laboratory conditions. Researchers still need to determine whether the pathway is active inside the human body, how much androgen the bacteria could produce there, whether those molecules reach nearby tissue and whether they have any measurable effect on prostate biology.
Still, the discovery provides scientists with specific genes and enzymes to investigate. Researchers can now search urinary microbiome datasets for dirA and dirB, determine how common they are and examine whether their presence is associated with differences in urinary hormones, prostate conditions or responses to treatment.
“The next step is to move from molecular capability to physiological relevance,” Bernardi said. “We now understand how the bacterial enzymes can perform the chemistry. The larger question is whether that chemistry has a meaningful effect in the complex environment of the human body.”
The work may also have implications beyond prostate research. Steroids found in urine are measured in medical research and diagnostics, as well as in testing for performance-enhancing drugs. If microorganisms can transform those compounds within the urinary tract, microbial metabolism may eventually become another factor scientists need to consider when interpreting urinary steroid profiles.
The study brought together patient-derived samples, bacterial culturing, genomic sequencing, analytical chemistry, protein biochemistry and computational biophysics. The experimental work established that the bacterium could produce testosterone and identified the genes involved. The computational work then revealed why the enzymes encoded by those genes had different capabilities.
“No single technique could have provided the complete story,” Bernardi said. “The experiments established the biological pathway, while the simulations revealed the physical basis for the enzyme activities. Together, they allowed us to move from observing testosterone production to understanding how it happens.”
The study brought together researchers from the University of Illinois Urbana-Champaign, Auburn University, Virginia Commonwealth University and Carle Foundation Hospital. Bernardi led Auburn’s computational structural biology contribution, while Rosa carried out the computational studies and developed the catalytic models. Bernardi leads Auburn’s Computational Biophysics Group, and both researchers are affiliated with the university’s Department of Physics and Department of Chemistry and Biochemistry.
Journal
Nature Communications
Method of Research
Experimental study
Subject of Research
People
Article Title
The urinary pathobiont Actinobaculum massiliense can generate androgens via the dirAB pathway
Article Publication Date
3-Sep-2026
