In science, some of the most interesting discoveries happen by chance.
That’s what occurred in the laboratory of Karina Xavier, a GIMM researcher, when her team — while studying how diet and antibiotics affect the gut microbiota — stumbled upon an unexpected bacterial ally.

“We weren’t looking for a live biotherapeutic agent,” recalls Karina. “We were testing how bacterial communication and antibiotic treatment affected the recovery of the gut microbiota. But one bacterium kept showing up — and it had very strong effects.”

Some of the laboratory mice, treated with antibiotics and then exposed to Escherichia coli, were resistant to infection. The difference – between the resistant and the non-resistant -, the PhD student Ana Rita Oliveira found, was the presence of a single, non-pathogenic Klebsiella species that seemed to protect the gut from infections.

This observation led to the first paper on this subject, published in Nature Microbiology (2020), where the team demonstrated that microbial exchange between hosts and nutrient competition were key mechanisms in restoring the natural defenses of the intestinal microbiota after antibiotic use.
The culprit that kept showing up turned out to be a particular Klebsiella strain, which belongs to the same bacterial family as the much feared Klebsiella pneumoniae — known for causing pneumonia and hospital-acquired infections — but in this case, the bacteria is harmless. The researchers named this bacterium Klebsiella sp ARO112 (ARO from the initials of Ana Rita Oliveira).

“It belongs to the Klebsiella group,” explains Xavier. “The pathogenic ones, like K. pneumoniae, cause serious infections. But this one is non-pathogenic — it doesn’t cause disease, and in fact, it protects against it.”

In mice, the researchers found that when Klebsiella ARO112 was present, pathogenic E. coli or Salmonella Typhimurium had trouble colonising the intestine.
The bacteria competed for the same ecological niche and the same nutrients — sugars and simple substrates — effectively blocking the entry of harmful strains.
In some tests, infection levels dropped by a factor of 100.

A new generation of live biotherapeutic agents

Encouraged by these results, the group began to see this bacterium in a new light. Could it be a next-generation biotherapeutic probiotic agent?

“The probiotics we know today mostly come from food sources — like lactic acid bacteria used in yoghurt,” says Karina. “Next-generation probiotics are different: they’re isolated from the gut microbiota of healthy humans and mammals. We’re just starting to tap into the potential of these naturally beneficial bacteria.”

Building on this idea, Ana Rita Oliveira teamed up with Vítor Cabral, to test the potential protective effect of Klebsiella sp. ARO112 in a disease context, which they described in an article just published in Nature Communications.
They decided to test its therapeutic potential in a challenging model: mice with the same genetic mutation commonly found in humans with inflammatory bowel disease (IBD) — a condition marked by chronic inflammation, microbial imbalance (dysbiosis), and high susceptibility to gut infections.

“These animals have an altered microbiota, similar to what we see in patients with IBD,” explains Xavier. “We gave them antibiotics — to simulate a clinical situation where antibiotics are used to control inflammation — and then infected them with pathogenic E. coli. After that, we treated one group with our Klebsiella strain.”

The results were remarkable. In our previous studies, the beneficial Klebsiella strain had reduced E. coli infection by about 100-fold, but some pathogens still lingered. In this disease model, however, the infection was completely cleared in nearly all mice treated. “Here we saw total clearance,” says Karina. “Not only did it eliminate the infection, it also accelerated the recovery of the microbiota and reduced inflammation.”

Healing through balance

The protective effect didn’t come from the Klebsiella directly killing pathogens. Instead, it helped restore the microbial ecosystem — allowing other beneficial bacteria to recover and produce key metabolites, such as butyrate, that protect the intestinal lining and reduce inflammation.

“These good metabolites are not produced by our bacterium itself,” she explains. “They’re produced by other bacteria that recover faster when this Klebsiella is present. It helps the ecosystem regain balance — and that balance is what protects against infection and inflammation.”

In other words, the bacterium worked indirectly but powerfully: by helping the microbiota heal, it allowed the host’s natural defenses to take over.
The effect was not only stronger than in previous models, but also better than that of E. coli Nissle 1917, a widely used commercial probiotic in Europe.

“We compared our strain with E. coli Nissle, which is often used in Germany,” says Karina. “Nissle didn’t work in this inflammatory model. It’s not that it’s a bad probiotic — it just doesn’t act in this kind of ecosystem. The gut is complex; what works in one context might not work in another.”

Safe by nature

Given that many Klebsiella species are known for antibiotic resistance, safety was a central concern. Before moving forward, the team ran extensive tests comparing ARO112 with pathogenic relatives and clinical isolates, in collaboration with Carles Ubeda laboratory in FISABIO (Valencia). The results were reassuring: It does not produce biofilms; it does not acquire antibiotic resistance plasmids easily and, surprisingly, even when artificially given a resistance plasmid, it lost it naturally within days.

“We were expecting it to behave like other Klebsiella, to easily acquire resistance genes,” Karina says. “But it didn’t — it seems to actively reject foreign DNA. That was a surprise, and it makes it much safer, because it decreases the risk of it to acquire antibiotic resistance”, a major concern in hospital settings.

The bacterium also does not persist indefinitely in the gut: as the microbiota recovers, its numbers naturally drop. It fills an ecological gap temporarily, helping other microbes return to balance, and then fades away. In evolutionary terms, it is a transient helper — the microbial equivalent of a volunteer firefighter.

The findings highlight the ecological nature of the microbiota.
As Xavier explains, our intestines host a vast and diverse bacterial community — with different species occupying different niches.
When that balance is disrupted by antibiotics, diet, or disease, harmful bacteria can take over. Filling those empty niches with good competitors like Klebsiella ARO112 could prevent infections by more dangerous species such as the pathogenic K. pneumoniae, which are a growing concern in hospitals, especially in Southern Europe.

“The microbiota works like an ecosystem with many small niches,” says Karina. “These non-pathogenic Klebsiella and E. coli strains compete for the same space as the bad ones. If we lose the good ones, the bad ones can move in. That’s why maintaining balance is so important.”

From lab to application

Could this protective bacterium one day be available as a pill? Possibly — but it will take time and careful testing.
Because ARO112 is a natural gut inhabitant, it could be used to restore the microbiota after antibiotic treatments, reducing the risk of infection and inflammation.
Unlike fecal transplants — currently used in some clinical settings but complex and costly — a defined live biotherapeutic agent like this would be precise, safe, and easy to administer.

“We’ll always need antibiotics — they’re essential,” says Karina. “But they disrupt the microbiota. One day, we might take an antibiotic and then follow it with a bacterium like this to restore balance. That’s the vision for next-generation biotherapeutic probiotic agents: using bacteria that already live naturally in our intestines.”

The story of Klebsiella ARO112 mirrors a broader transformation in microbiome research.
For years, gut bacteria were thought to be too difficult to culture and study. But advances in DNA sequencing and microbial cultivation techniques over the last decade have made it possible to identify and isolate new beneficial strains directly from the human microbiota.

“Fifteen years ago, it was almost impossible to grow these bacteria in the lab,” Xavier recalls. “Now, we can not only detect them by sequencing but also culture and study them in detail. That’s why microbiome research is exploding — we’re finally learning how these invisible ecosystems work.”

From an unexpected observation in mice to a candidate for clinical biotherapeutic development, Klebsiella sp. ARO112 represents a paradigm shift: a bacterium once overlooked because of its infamous relatives may soon help heal the gut, reduce inflammation, and protect against infection — not through eradication, but through cooperation.

“It’s all about ecology,” concludes Karina. “A healthy microbiota isn’t a battlefield — it’s a balanced community. Our job is to understand how to keep that balance, and sometimes, all it takes is the right bacterium at the right time.”

Sara Sá
GIMM Science Communication