Natural Antibiotic Resistance and Contamination by Antibiotic Resistance DeterminantsThe Two Ages in the Evolution of Resistance to Antimicrobials

José Luis Martínez•View original

OverviewBalancedalloy voice

Imagine discovering that antibiotic resistance didn't begin in hospitals or farms or even with penicillin. It started long before any of that, tucked into the DNA of microbes living in soils, rivers, and the guts of animals. That's the big reframing from José L. Martínez: resistance isn't a modern invention, it's an ancient feature of microbial life. Once you see it that way, many puzzles about today's crisis make more sense. Martínez calls this the natural resistome—the enormous set of genes out in the environment that can, if moved into the right microbe, confer resistance to pretty much any antimicrobial you throw at it. Not just the natural antibiotics that soil bacteria make to compete with each other, but even resistance to synthetic drugs like quinolones shows up in nature. Long before we prescribed antibiotics, bacteria were swapping, mutating, and carrying resistance genes on mobile elements such as plasmids. This is not conjecture. Decades ago, Gardner and colleagues reported resistance plasmids—R-factors—in places with no recorded exposure to antibiotics. That's a historical clue pointing straight at the environment as the original wellspring. Here's what makes this compelling rather than just romantic microbiology. When scientists started sequencing environmental DNA—metagenomics is the umbrella term—they found that wild ecosystems carry vastly more resistance genes than we see in today's pathogens. Wright highlighted this gap, and Davies and Davies underscored that different ecosystems have their own distinct repertoires. A forest soil looks different from a riverbed, which looks different from permafrost. The resistome is not one pool; it’s a mosaic of reservoirs assembled by geology, chemistry, and ecology. Since sequence alone can miss things, functional screens—where you pull genes out of an environmental sample and test what they do—have revealed resistance mechanisms that don't even look like known clinical ones by sequence. That's a sobering thought: we already know our clinical catalog is a small slice of what's possible. If this is the reservoir, how do pieces of it end up haunting hospitals? Mobility and context. Genes don't act in isolation; they live on genetic elements that can move across species, and they function differently depending on where they land. Martínez leans on a concept biologists call exaptation. A gene may have evolved for something else—detoxifying a compound, shuttling molecules across a membrane—and only later, under strong antibiotic pressure, does it show up as a resistance factor in a pathogen. The QnrA gene is the cleanest case study here. Poirel and colleagues traced it back to Shewanella algae, a water-dwelling bacterium that doesn't make antibiotics. Today, QnrA-like determinants are everywhere in human pathogens, typically carried on plasmids. That's the environmental-to-clinical pathway in miniature: a gene with a different life gets mobilized and redeployed. The movement happens at interfaces where bacteria from different worlds mix. Wastewater treatment plants are the poster child. Baquero and colleagues have pointed to them as hotspots where human-associated and environmental microbes meet, trade plasmids, and then disperse. But the same logic holds in agricultural runoff, sediments, and river systems. Knapp and colleagues showed that soils rich in heavy metals—think old mining regions or polluted agricultural lands—are enriched for antibiotic resistance genes. Metals and antibiotics can select for the same efflux pumps and detox systems. So, even if you've never poured ciprofloxacin on a field, contamination with metals can push the resistome in the same direction. You can hear the counterargument forming: if the environment is so loaded with resistance, why wasn't this a problem until modern medicine? Martínez's answer is a two-age model. For billions of years—the natural age—resistance determinants were mostly sitting on chromosomes, woven into the physiology of bacteria and the ecology of mixed microbial communities. They served functions that weren’t about resisting our drugs. Then came the contamination age. We flooded ecosystems with antibiotics and related pollutants. That changed the selection pressures almost overnight and created opportunities for horizontal transfer. Genes that had been passengers became drivers. Mobile genetic elements such as plasmids and transposons started carrying these determinants across species and into human-associated microbes. Same genes, new roles. There's a striking time-lapse that backs up this story. Knapp and colleagues went back to historical soil archives and saw a steady rise in the abundance of antibiotic resistance genes starting around nineteen forty. That's not a single outbreak; that's a long, quiet seep of contamination into the background of our ecosystems. You see a similar breadth if you look geographically. Resistance genes show up in sediments, in deep terrestrial subsurface samples, in ice and permafrost. They're not just hugging hospital pipes. Martínez pointed this out years ago: the same mobile resistance elements we worry about in clinics can be detected in pristine environments and even in wildlife. They persist, sometimes even without ongoing antibiotic pressure because the genetic vehicles that carry them are good at sticking around. Now, there's a real danger of over-reading a metagenomic hit. Martínez is careful about that. A sequence in a river that looks like a known resistance gene isn't enough to declare the river a direct source for hospitals. Even if the match is very high—above ninety percent identity—that often just tells you you're in the same gene family. To convincingly say, "this clinical gene came from that environmental host," you want near-identical sequences found in several strains of the putative source, lined up in the same neighborhood on the chromosome—what geneticists call conserved synteny—and crucially, signs that it's been there a while, not a recent stowaway. Those are tough criteria to meet. This is why really clean origin stories like the Shewanella-to-QnrA link are rare, and why most arrows in this field are sketched with probabilities, not certainties. Chromosomal context matters too. Take the qnr gene family. Sanchez and colleagues found that chromosomally encoded qnr genes are predominantly in water-dwelling bacteria. That supports the environmental provenance of the family, but it doesn't mean there's a single, simple lineage connecting lake microbes to a particular hospital outbreak. It means the raw material is out there, clustered in aquatic ecosystems, waiting for the right combination of pressure and opportunity to move. So, how do we work with all this without getting paralyzed by the complexity? Martínez sees two complementary tracks. One is retrospective: look at the resistance genes already causing problems in pathogens, then scour the environment for related sequences and contexts to map likely reservoirs and plausible routes of dissemination. That’s where case studies like QnrA help anchor intuition. The other is prospective: use functional screens to discover new resistance mechanisms in environmental samples, even when they don't look like anything in our clinical databases. This widens the radar. However, and this is the caveat, novelty in a river does not equal imminent clinical risk. Function proves a gene can do the job in a lab host; it doesn't prove that gene is poised to jump, survive, and spread in clinical bacteria. Bridge those steps carefully. Stepping back, the environmental resistome is not a footnote to the antibiotic era; it's the foundation. Mobile resistance elements didn't spontaneously arise in hospitals. They circulated in nature, sometimes on plasmids, sometimes baked into chromosomes, and the Anthropocene turned up the volume on their mobility and impact. That's why wastewater systems turn up again and again. They’re where our waste—drug residues, metals, human-associated microbes—meets the wild. They're big mixers. And they’re places where interventions are thinkable because they are engineered environments we control. There's also a lesson in persistence. Martínez notes that resistance genes have a frustrating tendency to stick around even when obvious selective pressures fade. Once a plasmid bearing a resistance determinant invades a population, it can carry along addiction systems, compensatory mutations, or simply hitchhike with other useful genes. It doesn't just melt away when a hospital changes a formulary. That stickiness is part biology, part ecology. And it's one reason environmental contamination—antibiotics, metals, disinfectants—matters even when you're focused on clinical outcomes. Selection and co-selection ripple out beyond the immediate target. All of this makes source-tracking hard, but not hopeless. The standards Martínez lays out—near-identical sequences, conserved genomic neighborhoods, multiple strains—set a high bar for claiming origins, and that's good science. It also sets the agenda. If we want to know where the next resistance genes are coming from, we need targeted surveillance in non-clinical habitats, not just more hospital screens. Rivers downstream of cities. Sediments near aquaculture. Soils near metal contamination. We need to pair metagenomic surveys with culture and function, so we can say not just what's there, but what it can do and how it moves. Let's be honest: the field is still young. Analyses of natural ecosystems have focused more on cataloging contamination and reservoirs than on truly quantifying the risk of transfer into pathogens. It's much easier to find a gene than to prove it's on a path into a hospital. But the broad contours are now clear. The environment holds a much larger and more diverse stockpile of resistance genes than medicine has yet seen. Different ecosystems nurture distinct subsets of that stockpile. Mobile genetic elements are the vehicles that move pieces of it into pathogens, helped along by ecological interfaces we've built or polluted. Our use of antibiotics—plus other stressors like metals—has shifted many genes from ancient, quiet roles into loud, clinical ones. If you want one mental image to carry forward, make it this: the resistome is a vast library. Hospitals check out a tiny fraction of the books. Wastewater plants are the interlibrary loan desks. Pollution accelerates the circulation. Every now and then, a volume that once sat quietly on a shelf in a river bacterium gets reissued as a bestseller in a pathogen. The practical takeaways are modest but real. Treat metagenomic hits as clues, not verdicts. Anchor origin stories with the strongest genomic evidence—near-identity, conserved neighborhoods, multiple strains—before drawing arrows on maps. Widen surveillance beyond the clinic because the supply chain of resistance begins far upstream. As Martínez keeps reminding us, the story of antibiotic resistance didn't start with us. But what we've done in the contamination age is change the tempo. The challenge now is to learn the score well enough to slow it down.

Martínez: resistance isn't a modern invention, it's an ancient feature of microbial life. Once you see it that way, many puzzles about today's crisis make more sense.

Martínez calls this the natural resistome—the enormous set of genes out in the environment that can, if moved into the right microbe, confer resistance to pretty much any antimicrobial you throw at it. Not just the natural antibiotics that soil bacteria make to compete with each other, but even resistance to synthetic drugs like quinolones shows up in nature. Long before we prescribed antibiotics, bacteria were swapping, mutating, and carrying resistance genes on mobile elements such as plasmids.

This is not conjecture. Decades ago, Gardner and colleagues reported resistance plasmids—R-factors—in places with no recorded exposure to antibiotics. That's a historical clue pointing straight at the environment as the original wellspring.

Here's what makes this compelling rather than just romantic microbiology. When scientists started sequencing environmental DNA—metagenomics is the umbrella term—they found that wild ecosystems carry vastly more resistance genes than we see in today's pathogens. Wright highlighted this gap, and Davies and Davies underscored that different ecosystems have their own distinct repertoires.

A forest soil looks different from a riverbed, which looks different from permafrost. The resistome is not one pool; it’s a mosaic of reservoirs assembled by geology, chemistry, and ecology. Since sequence alone can miss things, functional screens—where you pull genes out of an environmental sample and test what they do—have revealed resistance mechanisms that don't even look like known clinical ones by sequence.

That's a sobering thought: we already know our clinical catalog is a small slice of what's possible.

If this is the reservoir, how do pieces of it end up haunting hospitals? Mobility and context. Genes don't act in isolation; they live on genetic elements that can move across species, and they function differently depending on where they land.

Martínez leans on a concept biologists call exaptation. A gene may have evolved for something else—detoxifying a compound, shuttling molecules across a membrane—and only later, under strong antibiotic pressure, does it show up as a resistance factor in a pathogen. The QnrA gene is the cleanest case study here.

Poirel and colleagues traced it back to Shewanella algae, a water-dwelling bacterium that doesn't make antibiotics. Today, QnrA-like determinants are everywhere in human pathogens, typically carried on plasmids. That's the environmental-to-clinical pathway in miniature: a gene with a different life gets mobilized and redeployed.

The movement happens at interfaces where bacteria from different worlds mix. Wastewater treatment plants are the poster child. Baquero and colleagues have pointed to them as hotspots where human-associated and environmental microbes meet, trade plasmids, and then disperse.

But the same logic holds in agricultural runoff, sediments, and river systems. Knapp and colleagues showed that soils rich in heavy metals—think old mining regions or polluted agricultural lands—are enriched for antibiotic resistance genes. Metals and antibiotics can select for the same efflux pumps and detox systems.

So, even if you've never poured ciprofloxacin on a field, contamination with metals can push the resistome in the same direction.

You can hear the counterargument forming: if the environment is so loaded with resistance, why wasn't this a problem until modern medicine? Martínez's answer is a two-age model. For billions of years—the natural age—resistance determinants were mostly sitting on chromosomes, woven into the physiology of bacteria and the ecology of mixed microbial communities.

They served functions that weren’t about resisting our drugs. Then came the contamination age. We flooded ecosystems with antibiotics and related pollutants.

That changed the selection pressures almost overnight and created opportunities for horizontal transfer. Genes that had been passengers became drivers. Mobile genetic elements such as plasmids and transposons started carrying these determinants across species and into human-associated microbes. Same genes, new roles.

There's a striking time-lapse that backs up this story. Knapp and colleagues went back to historical soil archives and saw a steady rise in the abundance of antibiotic resistance genes starting around nineteen forty. That's not a single outbreak; that's a long, quiet seep of contamination into the background of our ecosystems.

You see a similar breadth if you look geographically. Resistance genes show up in sediments, in deep terrestrial subsurface samples, in ice and permafrost. They're not just hugging hospital pipes.

Martínez pointed this out years ago: the same mobile resistance elements we worry about in clinics can be detected in pristine environments and even in wildlife. They persist, sometimes even without ongoing antibiotic pressure because the genetic vehicles that carry them are good at sticking around.

Now, there's a real danger of over-reading a metagenomic hit. Martínez is careful about that. A sequence in a river that looks like a known resistance gene isn't enough to declare the river a direct source for hospitals.

Even if the match is very high—above ninety percent identity—that often just tells you you're in the same gene family. To convincingly say, "this clinical gene came from that environmental host," you want near-identical sequences found in several strains of the putative source, lined up in the same neighborhood on the chromosome—what geneticists call conserved synteny—and crucially, signs that it's been there a while, not a recent stowaway. Those are tough criteria to meet.

This is why really clean origin stories like the Shewanella-to-QnrA link are rare, and why most arrows in this field are sketched with probabilities, not certainties.

Chromosomal context matters too. Take the qnr gene family. Sanchez and colleagues found that chromosomally encoded qnr genes are predominantly in water-dwelling bacteria.

That supports the environmental provenance of the family, but it doesn't mean there's a single, simple lineage connecting lake microbes to a particular hospital outbreak. It means the raw material is out there, clustered in aquatic ecosystems, waiting for the right combination of pressure and opportunity to move.

So, how do we work with all this without getting paralyzed by the complexity? Martínez sees two complementary tracks. One is retrospective: look at the resistance genes already causing problems in pathogens, then scour the environment for related sequences and contexts to map likely reservoirs and plausible routes of dissemination.

That’s where case studies like QnrA help anchor intuition. The other is prospective: use functional screens to discover new resistance mechanisms in environmental samples, even when they don't look like anything in our clinical databases. This widens the radar.

However, and this is the caveat, novelty in a river does not equal imminent clinical risk. Function proves a gene can do the job in a lab host; it doesn't prove that gene is poised to jump, survive, and spread in clinical bacteria. Bridge those steps carefully.

Stepping back, the environmental resistome is not a footnote to the antibiotic era; it's the foundation. Mobile resistance elements didn't spontaneously arise in hospitals. They circulated in nature, sometimes on plasmids, sometimes baked into chromosomes, and the Anthropocene turned up the volume on their mobility and impact.

That's why wastewater systems turn up again and again. They’re where our waste—drug residues, metals, human-associated microbes—meets the wild. They're big mixers.

And they’re places where interventions are thinkable because they are engineered environments we control.

There's also a lesson in persistence. Martínez notes that resistance genes have a frustrating tendency to stick around even when obvious selective pressures fade. Once a plasmid bearing a resistance determinant invades a population, it can carry along addiction systems, compensatory mutations, or simply hitchhike with other useful genes.

It doesn't just melt away when a hospital changes a formulary. That stickiness is part biology, part ecology. And it's one reason environmental contamination—antibiotics, metals, disinfectants—matters even when you're focused on clinical outcomes. Selection and co-selection ripple out beyond the immediate target.

All of this makes source-tracking hard, but not hopeless. The standards Martínez lays out—near-identical sequences, conserved genomic neighborhoods, multiple strains—set a high bar for claiming origins, and that's good science. It also sets the agenda.

If we want to know where the next resistance genes are coming from, we need targeted surveillance in non-clinical habitats, not just more hospital screens. Rivers downstream of cities. Sediments near aquaculture.

Soils near metal contamination. We need to pair metagenomic surveys with culture and function, so we can say not just what's there, but what it can do and how it moves.

Let's be honest: the field is still young. Analyses of natural ecosystems have focused more on cataloging contamination and reservoirs than on truly quantifying the risk of transfer into pathogens. It's much easier to find a gene than to prove it's on a path into a hospital.

But the broad contours are now clear. The environment holds a much larger and more diverse stockpile of resistance genes than medicine has yet seen. Different ecosystems nurture distinct subsets of that stockpile.

Mobile genetic elements are the vehicles that move pieces of it into pathogens, helped along by ecological interfaces we've built or polluted. Our use of antibiotics—plus other stressors like metals—has shifted many genes from ancient, quiet roles into loud, clinical ones.

If you want one mental image to carry forward, make it this: the resistome is a vast library. Hospitals check out a tiny fraction of the books. Wastewater plants are the interlibrary loan desks.

Pollution accelerates the circulation. Every now and then, a volume that once sat quietly on a shelf in a river bacterium gets reissued as a bestseller in a pathogen.

The practical takeaways are modest but real. Treat metagenomic hits as clues, not verdicts. Anchor origin stories with the strongest genomic evidence—near-identity, conserved neighborhoods, multiple strains—before drawing arrows on maps.

Widen surveillance beyond the clinic because the supply chain of resistance begins far upstream. As Martínez keeps reminding us, the story of antibiotic resistance didn't start with us. But what we've done in the contamination age is change the tempo. The challenge now is to learn the score well enough to slow it down.

Related lectures