You are on page 1of 10

Feature Article

How to Fight Back Against

Antibiotic Resistance
Mapping the exchange of genes between pathogens and nonpathogens offers
new ways to understand and manage the spread of drug-resistant strains.

Gautam Dantas and Morten O. A. Sommer

ot so long ago, it seemed fying the chemical scaffolds of already a record low of one new antibiotic
like the fight against infec- approved classes of antibiotics. in the five-year period from 2008 to
tious diseases was nearly During this innovation gap, bacte- 2012, down from 16 new drugs in the
won. The discovery of rial evolution did not cease. Conse- years from 1983 to 1987 (see the figure
penicillin in 1929 gave clinicians their quently, drugs that were previously on page 44). CDC Director Tom Frieden
first weapon to combat common ail- effective in treating a broad spectrum recently warned, If we dont act now,
ments like pneumonia, gonorrhea, and of infectious bacteria are now useful our medicine cabinet will be empty
rheumatic fever. In the decades that fol- for fewer and fewer infections. Cer- and we wont have the antibiotics we
lowed, medical researchers discovered tain bacteria, including strains of Esch- need to save lives. In reality, the de-
more than 150 other types of antibiotics. erichia coli and Klebsiella pneumonia, are velopment of new antibiotics is only
These widely hailed wonder drugs now resistant to all major antibiotics part of the solution, as pathogens will
were so successful that U.S. Surgeon even carbapenems, which have long inevitably develop resistance to even
General William Stewart announced in been the drug of last resort to treat af- the most promising new compounds.
1967, The time has come to close the flictions such as lung infections. With To save the era of antibiotics, scien-
book on infectious diseases. dwindling treatment options, the mor- tists must figure out what it is about
Stewart and most of his contempo- tality rate from those infections in the bacterial pathogens that makes resis-
raries greatly underestimated the abil- United States is approaching 50 per- tance inevitable. By studying the suite
ity of bacterial pathogens to adapt to cent. In effect, for some diseases we are of genescollectively known as the
these life-saving medicines. Almost as now living in a post-antibiotic age. resistomethat can turn a susceptible
soon as clinical use of penicillin began According to a September 2013 re- pathogen into a superbug, researchers
in 1946, the first drug-resistant patho- port from the U.S. Centers for Disease may be able to uncover the Achilles
gens appeared. During the golden age Control and Prevention (CDC), treat- heel of these multiple drugresistant
of antibiotic development (the 1940s ment of antibiotic-resistant infections strains. Although most studies on drug
to the 1960s), the spread of antibiotic adds $35 billion in health care costs resistance have focused on disease-
resistance was balanced by the con- and 8 million hospital days per year causing pathogens, recent efforts by
tinued discovery and deployment of in the United States. A recent drug- the two of us and by a number of our
new classes of antibiotics. But starting resistant Salmonella outbreak due to colleagues have shifted attention to the
in the 1970s, a dwindling interest and contaminated chicken meat was linked resistomes of nonpathogenic bacteria.
ability of the pharmaceutical indus- to nearly 300 illnesses across 18 states, Importantly, over the past decade ad-
try to develop new antibiotics resulted sickening infants and nonagenar- vances in DNA sequencing have en-
in a 40-year period when virtually no ians alike. At least 23,000 Americans abled us to explore the genomes of
new broad-spectrum classes of anti- die each year from infections, many both pathogenic and non-pathogenic
biotics were brought to the market. caused by the superbug methicillin- bacteria across a variety of different
Instead, companies focused on modi- resistant Staphylococcus aureus (MRSA), natural habitats.
because doctors have run out of drugs This research has led to a growing un-
with which to treat them. derstanding of how antibiotic resistance
Gautam Dantas is an assistant professor of pathol- Government agencies are belatedly
ogy/immunology and biomedical engineering and considering incentives to support re-
member of the Center for Genome Sciences and Clostridium difficile can cause life-threatening
newed antibiotic drug development, diarrhea. In 2000, a strain emerged with re-
Systems Biology at Washington University in St.
Louis, MO. Morten O. A. Sommer is a professor
but these initiatives have not yet had sistance to multiple antibiotics, including
a direct impact on the drug develop-
Science Source

of systems biology and member of the Novo Nord- ciproflaxin and levofloxacin. The bacterium
isk Foundation Center for Biosustainability at the ment pipeline. As a result the number contributed to a 400 percent increase in C.
Technical University of Denmark. E-mail for of antibiotics approved by the Food difficileassociated deaths in the United
Dantas: and Drug Administration (FDA) hit States between 2000 and 2007.

42 American Scientist, Volume 102 2014 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact 2014 Sigma Xi, The Scientific Research Society. Reproduction 2014 JanuaryFebruary 43
with permission only. Contact
advent of medicinal chemistry
golden age of discovery
innovation gap
1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

antibiotic deployment
antibiotic resistance

Clinical deployment of new antibiotics (blue bars) has quickly been followed by the evolution tance: vertically, through the accumu-
of bacteria able to resist their effects (red). During the golden age of discovery, 150 types of lation of genetic changes during the
antibiotics were developed. Since then, the spread of resistance has greatly outpaced the rate natural process of copying its genome,
of drug development. The Infectious Disease Society of America estimates that 70 percent of and horizontally, by swapping resis-
hospital-acquired infections in the United States are now resistant to one or more antibiotics. tant genes from one microbe to anoth-
er (see box below).
evolves in specific bacteria, as well as ing newer drugs. Understanding factors Vertical transmission is the fun-
how it gets passed between different that influence resistome evolution and damental evolutionary process by
bacteria and between different envi- dissemination may both extend the life which a cell can accumulate errors in
ronments. We are still a long way from of current drugs and point toward new its genome during replication, such
Stewarts old proclamation of victory, disease-fighting strategies. that the resulting progeny differ ge-
but the recent advances are helping netically from their bacterial ancestors.
frame new strategies to complement Origins of Resistance The genome replication or copying er-
the 90-year-old paradigm of just trying There are two ways that pathogenic ror rate is rather low, so typically one
to defeat resistant bacteria by discover- bacteria can develop antibiotic resis- in a thousand growing bacteria will

vertical transmission horizontal transmission

Antibiotic resistance can be ac-
bacterial transformation
resistance quired in two basic ways. In ver-
tical transmission, a bacterium
accumulates errors or mutations
in its genome during replication;
some of those changes (red) give
the ability to resist antibiotics
release of DNA and are passed on to subse-
quent generations. In horizon-
bacterial transduction tal transmission, resistant genes
are swapped from one microbe
to another. This can occur via
three mechanisms: transforma-
tion, when bacteria scavenge
resistance genes from dead bac-
release of phage terial cells and integrate them
into their own genomes; trans-
bacterial conjugation
duction, when resistance genes
are transferred by bacteriophag-
es (viruses that infect bacteria);
or conjugation, when genes are
transferred between bacterial
cells through tubes called pilli.

44 American Scientist, Volume 102 2014 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact
Four mechanisms of resistance: impermeable
barrier (a) blocks antibiotics (blue spheres) a b
because the bacterial cell membrane is now
impermeable to the drug. Target modifica-
tion (b) alters the proteins inhibited by the
antibiotic, so the drug cannot bind properly.
Antibiotic modification (c) produces an en- modified
zyme that inactivates the antibiotic. Efflux cell wall
(d) employs genes coding for enzymes that protein
actively pump the antibiotic out of the cell.

introduce an error (called a mutation)

into the genome. Not all of these mu-
tations are advantageous, but about plasmid with antibiotic-
resistant genes
one in a billion will generate mutants
that can grow faster or tolerate higher
concentrations of antibiotics than their
predecessors. When such bacterial mu-
tants are exposed to antibiotics, those
possessing antibiotic resistance genes efflux
will increase in prevalence to the point pump
of taking over the entire population.
Multiple cycles of such mutation and
selection are often necessary to evolve
high-level antibiotic resistance. drug-inactivating
Genes that have evolved to confer
antibiotic resistance to one type of bac-
teria can be transferred to another by
a mechanism known as horizontal gene
transfer. While some pathogens acquire antibiotic
resistance via vertical transmission, c d
recent studies have suggested that
horizontal transmission may be the to another is through conjugation, also that contain tens of resistance genes,
dominant force behind growing antibi- known as bacterial sex. The discov- offering the host cells immunity to vir-
otic resistance. During horizontal gene ery of this process, now thought to tually all antibiotics.
transfer, antibiotic resistance genes be the main mechanism of horizontal Recent large-scale efforts sequenc-
catch a ride on mobile genetic elements gene transfer, earned Joshua Leder- ing the genomes of many bacterial
that carry the genetic material between berg the 1958 Nobel Prize in Physiol- pathogens, by such groups as the
different cells. Mobile genetic elements ogy or Medicine. During conjugation, Wellcome Trust Sanger Institute in the
can be linear or circular pieces of DNA,
called plasmids, which are replicated by
the cell along with its genome.
These DNA fragments make their Understanding factors that influence
way into a new cell through three mech-
anisms: transformation, transduction, resistome evolution and dissemination
and conjugation. In transformation,
bacterial cells scavenge DNA remnants may both extend the life of current
from dead bacterial cells and integrate
them into their own genome. In trans-
duction, genetic material is transferred
drugs and point toward new disease-
by bacteriophages (viruses that infect
bacteria). Bacteriophages can insert their
fighting strategies.
DNA into the genome of a host cell,
where they are maintained for many
generations before they pack up their plasmids hijack the cellular machinery United Kingdom and the Broad Insti-
DNA and leave to infect another cell. to create structures called pilli that pro- tute of MIT and Harvard in the United
Along the way, the bacteriophage may trude from the donor cell to penetrate States, have added another complica-
coincidentally integrate a section of the the membrane of the recipient cells, tion to this story. These studies have
bacterial host cell genome into the bacte- enabling the transfer of the conjuga- shown that the genes conferring resis-
riophage genome, enabling genetic ma- tive plasmid and all the functions it tance toward antibiotics in pathogens
terial from one cell to hitchhike a ride to encodes. Many hospital-associated can be acquired via horizontal gene
another cell on the bacteriophage. pathogens, including the carbapenem- transfer from another gene reservoir
The final way antibiotic resistance resistant bacteria mentioned previous- entirely, such as city soil, waste water,
genes can be passed from one microbe ly, harbor large conjugative plasmids or processed meat. Accordingly, there 2014 Sigma Xi, The Scientific Research Society. Reproduction 2014 JanuaryFebruary 45
with permission only. Contact
is a substantial interest in broadly strategies to encode resistance. The in genes that modify the target of the
characterizing these gene reservoirs genes that confer antibiotic resistance antibiotic, dampening its effectiveness.
to improve our overall understand- can be loosely separated into four Each antibiotic is designed to target a
ing of how swapping genes from one groups, each with their own unique well-defined essential bacterial pro-
reservoir to another contributes to the mechanism for combating antibiotic cess. For example, fluoroquinolones
evolution of antibiotic resistance in exposure (see figure on page 45). are a widely used class of antibiotics
bacterial pathogens. Bacteria in the impermeable-barrier for the treatment of skin, lung, or uri-
group are naturally resistant to certain nary tract infections. These antibiot-
How Pathogens Fight Back antibiotics, either because they lack the ics target DNA, disrupting the proper
Just as there are a number of different target of the antibiotic or because their functioning of proteins involved in un-
ways for bacteria to acquire an antibi- membranes are impermeable to the winding DNAs helix during replica-
otic resistance gene, the genes them- drug. In the second group, target mod- tion. Mutations that confer resistance
selves represent a number of different ification, bacteria acquire mutations toward fluoroquinolone antibiotics of-
ten change the conformation of these
proteins, reducing the binding of the
drug to its target and thus increasing
Exploring Resistance Outside the Petri Dish the concentration necessary to block
the process.

I n the first several decades after the

discovery of antibiotics, research-
ers studied the emerging problem
already known genes but cannot be
used to discover new ones.
Metagenomic sequencing identi-
In antibiotic modification, a resis-
tance gene can encode an enzyme that
helps break down or modify the antibi-
of antibiotic resistance by growing fies all of the DNA from a particu- otic before it can kill the bacteria. This
target microorganisms in the lab. The lar environment, irrespective of its tactic is often used against beta-lactams,
reliance on such methods traces back origin. The sequences are then as- the most widely prescribed and diverse
to the founder of modern bacteriolo- sembled and scanned for new genes chemical class of antibiotics, which in-
gy, Robert Koch, whose work in cul- that are similar to already known re- cludes the well-known drug penicil-
tured bacteria made pure culture sistance genes. lin. Penicillin inhibits enzymes that re-
the gold standard in clinical micro- Metagenomic functional selec- model the bacterial cell wall and are
biology laboratories. We know now tions combine the cultivation-based essential for the cell during growth. Re-
that studying individual organisms methods of old with new, culture- sistance toward penicillin is frequently
grown in pure culture ignores the in- independent techniques. A host or- conferred by beta-lactamases, enzymes
creasing number of diseases caused ganism that is normally susceptible that cleave the penicillin molecule to
not by one pathogenic bacterium but to antibiotics is genetically engineered render it ineffective in inhibiting the cell
by several acting in concert. In addi- to possess various chunks of DNA wall modification enzymes.
tion, most environmental microbes taken from a microbial community of Finally, efflux occurs when a resis-
cannot be readily cultivated in the interest. The modified hosts are then tance gene codes for proteins that ac-
lab. Recent technical breakthroughs exposed to antibiotics. The only sur- tively pump the antibiotic out of the
have led to three established culture- vivors will be those that acquired a cell, keeping its internal concentration
independent strategies for explor- resistance gene. The microbes can then low enough to prevent inhibition. This
ing antibiotic resistance much more be characterized to reveal the sequence resistance mechanism is deployed for
fully, in both pathogenic and non- that confers resistance. all antibiotics that have targets with-
pathogenic bacteria. In addition, we have developed in the cell; in many cases such efflux
The polymerase chain reaction a novel approach called PARFuMS pumps are able to push out several
(PCR), selectively amplifies specific (for Parallel Annotation and Reas- different antibiotics, resulting in mul-
antibiotic resistance genes from com- sembly of Functional Metagenomic tidrug resistance. An example of this
plex microbial communities so they Selections) that integrates culture- has been observed for tetracycline, an
can be easily identified. PCR is well independent functional metagenomic agent used to treat a wide variety of
suited for studying the prevalence of selections, next-generation DNA se- infections. Resistance to the drug can
quencing, and optimized computa- stem from tetracycline efflux genes,
tional sequence assembly and annota- which code for proteins that sit in the
tion algorithms to profile resistomes. bacterial membrane and export the an-
tibiotic out of the cell.
Further complicating matters, resis-
Scientists once relied heavily on meth- tance toward any one drug typically
ods that required culturing bacteria in results from more than one mecha-
the laboratory (usually in petri dishes, as nism. For instance, tetracycline re-
shown at left) to study antibiotic resistance sistance has been observed to occur
in various microbial populations. Today
through target modification, antibiotic
these culture-dependent methods are be-
ing eclipsed by other approaches that en-
modification, and efflux mechanisms.
able researchers to characterize the large Though the term antibiotic resistome
percentageas much as 99 percentof bac- didnt emerge until five years ago, the
teria that cannot be grown in pure culture. concept encapsulates decades of re-
Doncaster and Bassetlaw Hospitals/Science Source search on the transmission and evolu-

46 American Scientist, Volume 102 2014 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact


manure runoff rainfall

agriculture food river/sea/lake wastewater

animals treatment

direct contact drinking water
and meat and swimming
vegetables industry and
and fruit fish households


Interconnections between people, animals, and the environment make it easy for antibiotic- teria Actinomycetes, a major source
resistant bacteria to jump from one species to another. For instance, a resistant strain living of antimicrobials for the commercial
in soil could travel through runoff and get passed on to humans via drinking water or recre- drug industry, must contain elements
ational swimming. Multiple routes of exchange propel the evolution and spread of resistance. protecting against the antibiotics they
produced. These elements were by
tion of antibiotic resistance genes be- resistome evolution and exchange are definition antibiotic resistance genes.
tween different microbes and different the soil and the human body, but there Given that these production capacities
environments. The resistome as it is are other resistomes as well (see the box likely evolved hundreds of millions
currently defined is the entire suite of on page 49). of years ago, the corresponding resis-
genes, from a particular microbe or tance genes are likely just as old.
microbes, which confers antibiotic re- The Soils Reservoir of Resistance The antibiotic resistance genes ob-
sistance. It includes all antibiotic re- Antibiotic resistance is everywhere, served in nonproducer organisms
sistance genes in a group of microbes even in your backyard. Soil microbes (including pathogens) may have been
at any scale, from a single organism to likely represent the evolutionary res- acquired directly from the producers
all of the microbes in an arbitrary en- ervoir of most resistance, and the re- or from their soil-dwelling neighbors
vironmental sample. Viewed this way, sistome of the soil is easily the largest who evolved them in response to the
the resistome from one environment and most diverse of any environment. selection pressure of natural antibiotic
can be evaluated for its capacity to ex- The majority of antibiotics used in the production from other microbes in the
change resistance genes with another clinic were originally discovered as or soil. In fact, Daviess group showed
environment. derived from natural products of soil- the activity of the generally nonpatho-
Mounting evidence indicates that dwelling microbes, primarily of the genic soil bacteria Streptomyces coded
essentially all microbial environments Streptomyces genus of the Actinomy- for resistance enzymes that modify
contain antibiotic resistomes; that cete phylum. specific antibiotics known as amino-
myriad natural and human-driven ac- In 1973, Julian Davies, now at the glycosides identical to those found in
tivities influence their exchange within University of British Columbia, and clinical pathogens.
and between environments; and that a his colleagues first hinted at the idea In the 40 years since Daviess pro-
complex web of interactions connects of a resistome in the Producer Hypoth- posal, a large body of literature has pro-
various resistomes. Two of the most esis on the origin of clinical resistance. vided general support for this hypoth-
important environments for microbial He argued that the producer bac- esis, as well as some key refinements. 2014 Sigma Xi, The Scientific Research Society. Reproduction 2014 JanuaryFebruary 47
with permission only. Contact

human pathogens


enterica sv. Emek

Salmonella enterica
sv. Typhimurium

typhimurium DT104

non-pathogenic soil
bacteria fragments

resistance gene resistance related mobility elements

Antibiotic resistance genes from nine multidrug-resistant soil bacterial cultures (bottom row) any soil, these natural resistomes are
were genetically identical (shown by gray shading) to genes in several clinical pathogens (top likely being enriched by human activity.
five rows), providing evidence for recent exchange between the resistomes of nonpathogenic David Graham and colleagues at New-
soil bacteria and human pathogens. (Adapted from K. J. Forsberg et al., Science 337:1107.) castle University analyzed a panel of
Dutch soils archived between 1940 and
A landmark paper from Gerry Wright rial world, namely, the Proteobacteria, 2008 for both the presence and the abun-
and his colleagues at McMaster Uni- Bacteroidetes, and Actinomycetes dance of a series of specific resistance
versity in 2006, which also formally that can actually feed and grow on anti- gene types. They found a dramatic in-
introduced the resistome, showed that biotics. These bacteria were on average crease in levels of key beta-lactam, mac-
about 400 randomly isolated soil Strep- resistant to 17 of 18 clinically relevant rolide, and tetracycline resistance genes
tomyces samples were highly multidrug antibiotics profiled, likely displaying over the 70-year study period, closely
matching the era of large-scale human
antibiotic production.
The surveys of known resistance
Just one gram of soil is estimated to genes in the soil resistome are just the
tip of the iceberg. Jo Handelsman, now
contain about one billion bacterial cells, at Yale University, and her colleagues
pioneered the application of culture-
and no current method gets even independent functional metagenomics, a
technique that provides the functional
remotely close to sampling this diversity. gene composition of environmental
samples, to characterize soil resistomes
from both human-affected and pristine
environments. Using this approach,
resistant against a large panel of clini- this incredibly high multidrug resis- they identified a number of novel an-
cally relevant antibiotics. On average, tance because they were cultured un- tibiotic resistance genes, some with
these bacteria were resistant to seven der the selective pressure of extremely never-before-seen mechanisms of re-
to eight drugs, and one superbug was high antibiotic concentrations. Other sistance. Taken together, these studies
resistant to 15 different compounds censuses of cultured soil bacteria popu- indicated that the soil resistome was
of the 21 tested, including drugs that lations have since confirmed these high diverse, ancient, and recently enriched
were entirely synthetic and ones that levels of multidrug resistance. by human activity. Direct support for
were only recently approved for clini- The discovery of ubiquitous multi- the notion that the soil resistome long
cal use. Wrights observation was star- drug resistance in soil microbes suggests predated clinical use of antibiotics
tling, because such high levels of mul- that the soil resistome is immense. Com- comes from recent work from Gerry
tidrug resistance exceed those found in plementary investigations of the genes Wrights group. In 2011, they reported
most pathogens. conferring this multidrug resistance on DNA sequencing of 30,000-year-old
In 2008, our own research further corroborate this prediction. Numerous Beringian permafrost that uncovered
expanded the view of a substantial PCR-based assays (using the polymerase evolutionary relatives of resistance
multidrug soil resistome by describing chain reaction to amplify DNA samples) genes against important modern anti-
about 600 soil bacteriafrom three of have shown that although resistance biotics, including beta-lactams, tetra-
the 60 phyla, or divisions, of the bacte- genes are already present in virtually cyclines, and vancomycin.

48 American Scientist, Volume 102 2014 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact
One might expect that this wealth of and culture-independent experimental as the human commensal microbiota
information on the breadth and depth methods, as well as improved com- incorporate the resistome that is most
of the soil resistome would confirm putational tools for their analysis, en- accessible to human pathogens. These
the key prediction of the Producer Hy- able scientists to scan and sequence the microbes outnumber human cells by
pothesis, that the same resistance genes DNA of individual microbes without 10-fold, and their collective genes (the
identified in clinical pathogens would having to culture them first so they can microbiome) outnumber the genes in
also be identified in soil bacteria, sug- fill in these holes. the human genome by over 100-fold.
gesting recent exchange between the Specialized and relatively stable eco-
clinical and soil resistomes. Surpris- The Resistome Inside You systems of microbes inhabit various
ingly, such evidence was lacking until Although the soil resistome is the most parts of the body, with the densest and
very recently. The overwhelming ma- important reservoir of resistance from most diverse community housed in
jority of soil resistome studies revealed an evolutionary perspective, the mi- the human intestine. Nearly every as-
only limited similarity to resistance crobes living in and on usknown pect of the human condition, in health
genes found in pathogens.
To account for this unexpected re-
sult, we hypothesized that a key subset On the Trail of the Other Resistomes
of soil bacteriathe notoriously mul-
tidrug-resistant soil Proteobacteria
may represent a conduit for recent ex-
change with pathogens. We reasoned
I n addition to those in soil and in
the human gut, microbes from
agriculture and aquaculture are be-
a concomitant increase in antibiotic
resistance of bacteria associated with
farmed fish.
that since the greatest current clinical lieved to contribute substantially to Agriculture- and aquaculture-
burden for multidrug resistance comes the exchange of antibiotic resistance associated resistomes most likely act
from multidrug resistant proteobacte- genes. In the United States and Eu- as intermediates between the human
rial pathogens, their closest cousins in rope, antibiotics are used four times microbiota and human pathogens
the soil might show evidence for recent as often in the food industry as in living in more pristine environments
resistome exchanges. human medicine. Frank Aarestrup such as soil, sea, and fresh water. The
We then set out to test this idea using from the Technical University of transfer of antibiotic resistance genes
culture-based selections to selectively Denmark has shown that this high probably goes both ways. Antibiotic-
enrich about 100 highly multidrug- consumption has led to high levels resistant bacteria from farm animals
resistant soil bacterial cultures, com- of antibiotic resistance in gut bacte- are spread through manure onto the
posed primarily of Proteobacteria. ria from farm animals. Furthermore, soil, where they can disseminate re-
We profiled their resistomes using a computational genomics analysis sistance to soil bacteria. Conversely,
novel approach that integrates culture- from Eric Alms group at MIT reveals farm animals are in frequent direct
independent functional metagenomic frequent transfer of resistance genes contact with soil, allowing genes to
selections, next-generation DNA se- between microbes from farm animals be transferred from the soil micro-
quencing, and optimized computa- and those in human microbiota. An- biota back to the animals. Although
tional sequence assembly and anno- tibiotic usage is also growing at an pinpointing specific sources of resis-
tation algorithms (see sidebar on page alarming rate in aquaculture as more tance is difficult, it is clear that heavy
46). Through this method, named PAR- seafood is harvested this way. Felipe use of antibiotics in food production
FuMS (for Parallel Annotation and Re- Cabello of New York Medical Col- is fueling the evolution and spread
assembly of Functional Metagenomic lege and his colleagues have shown of antibiotic resistance genes.
Selections), we uncovered nine differ-
ent antibiotic resistance genes from
diverse U.S. soils that were geneti-
cally identical to a number of globally
distributed clinical pathogens, finally
providing broad support for recent ex-
change between the resistomes of non-
pathogenic soil bacteria and human
pathogens (see figure on page 48).
Despite recent key advances in our
knowledge of soil resistomes, we are
still in the infancy of exploring this in-
Arterra Picture Library/Alamy

credibly diverse ecosystem. Just 1 gram

of soil is estimated to contain about 1
billion bacterial cells, and no current
method gets even remotely close to
sampling this diversity. Approximately
half of the 60 predicted phyla of the
bacterial world cannot be cultured in a
lab, and even the ones that can are still
not completely characterized. Fortu- More than 70 percent of all antibiotics sold in the United States are used to keep pigs, cows,
nately, advances in both culture-based chickens, turkeys, and other meat animals from getting sick and to help them grow larger. 2014 JanuaryFebruary 49

and disease, involves some interaction verse resistome to participate in these resistance genes had steadily increased
with the microbiota and microbiome. exchanges. over the course of two decades.
Because one of the microbiotas The earliest insights into the hu- The implication from these studies
main jobs is to keep pathogens from man commensal resistome come from that increased antibiotic use was lead-
invading the gut, and the fact that any culture-based studies of these bacteria. ing to higher levels of resistance
invading pathogen is more likely to Studies in the 1990s by the Universi- was supported by numerous other
studies on other commensal microbes.
Martin Blaser and colleagues at New
The most resounding message that York University found increases in
macrolide resistance in commensal
comes through from every new Enterococci, a bug associated with uri-
nary tract infections and meningitis,
as a collateral response to therapy tar-
resistome study is that the pool of geting another bacterium, Helicobacter
pylori, found in some ulcers. They
resistance genes, and the mechanisms also observed that this enriched resis-
tance persisted for years after therapy
of resisting antibiotics, available to ceased, challenging the conventional
wisdom that antibiotic resistance en-
bacteria are effectively limitless. coded a fitness cost to the host bacte-
rium and hence would evolve away
or be outcompeted by antibiotic-
susceptible strains in the absence of
interact with bacterial cells than hu- ty of Illinoiss Abigail Salyers on the the antibiotic.
man cells, the commensal resistome is normally mutualistic gut microbe Bac- These results are not unique to the
often in a position to exchange resis- teroides established that tetracycline gut ecosystem; for example, persis-
tance genes with nearby pathogens. and macrolide resistance genes were tently antibiotic resistant Staphylococcus
The increasing levels of antibiotic ex- being passed back and forth between epidermis, a major player in hospital-
posure that the microbiota has been the resistomes of pathogenic and non- acquired infections, have been isolated
subjected to since the beginning of the pathogenic forms of the bacteria. Their from the nostrils of antibiotic-treated
antibiotic era provides ample selection analysis of archived Bacteroides samples patients. Numerous studies in animal
pressure to maintain a robust and di- also established that the levels of these models and in humans show the com-

complete resistome cultured resistome subset

30 2 22 78
8 4

shared shared 10
resistance resistance
genes 48 genes

An FDA microbiologist holds tomatoes be-
individual 1
ing tested for salmonella (above). Genetic
sequences of all bacteria (far left) and all
individual 1 genes cultured bacteria (middle left) from two
similar to NCBI healthy individuals are compared with se-
quences cataloged by the National Center
individual 2 for Biotechnology Information (NCBI). The
genes cultured resistome overlaps greatly with the
individual 2 genes NCBI sequences and with resistance genes
individual 1 similar to NCBI individual 2 in common pathogens. (Adapted from M. O.
A. Sommer et al., Virulence 2010 1:299.)

50 American Scientist, Volume 102 2014 Sigma Xi, The Scientific Research Society. Reproduction
with permission only. Contact
mensal resistome can readily partici- research groups led by Peer Bork at every new resistome study is that the
pate in exchanges within and between the European Molecular Biology Labo- pool of resistance genes, and the mech-
ecosystems. Anette Hammerum and ratory in Heidelberg and Baoli Zhu anisms of resisting antibiotics, avail-
colleagues at Statens Serum Institut in at the Institute of Microbiology in able to bacteria are effectively limitless.
Denmark recently demonstrated the Beijing reported on computationally To stay ahead of the game we must
transfer of vancomycin resistance genes predicted resistomes from sequencing take a multipronged approach, look-
between human and swine hosts. data of human gut microbiomes from ing for new ways to keep pathogens
207 and 162 individuals, respectively, in check, while also searching for even
Just Scratching the Surface representing multiple nationalities newer antipathogen strategies. Even
All of this evidence indicates that anti- and cultural traditions. Collectively, when therapies appear to be highly
biotic treatment selects for genes con- these studies predict the existence of effective during their initial deploy-
ferring antibiotic resistance, that these thousands of resistance genes across ment, it is only a matter of time before
increases in resistance can persist for the analyzed commensal microbiome. pathogens tap into the enormous re-
many years, and that these resistance The abundance of resistance genes cor- sistomes of their many neighbors and
genes can be exchanged within the relates with the available data on the once again thwart our very best drugs.
commensal microbiota as well as with amount of human and agricultural an-
foreign microbes. As with the soil, we tibiotic use, as well as with how long References
and others in our field have also begun ago those antibiotics were introduced. Aarestrup, F. M. 1999. Association between the
to appreciate that this portrait of the Although we are beginning to consumption of antimicrobial agents in ani-
mal husbandry and the occurrence of resis-
human commensal resistome is a vast gain a glimpse into the genetics of
tant bacteria among food animals. Internation-
underestimate due to an over-reliance such complex ecosystems, we will al Journal of Antimicrobial Agents 12: 279285.
on culture-based methods. need improved culture-based and Cabello, F. C. 2006. Heavy use of prophylactic
In 2009 we reported on the first ap- culture-independent techniques to antibiotics in aquaculture: A growing prob-
plication of culture-independent func- better understand these reservoirs of lem for human and animal health and for
tional metagenomic selections to study antibiotic resistance. Long-term stud- the environment. Environmental Microbiol-
ogy 8:11371144.
the commensal gut resistomes of two ies of various healthy and perturbed
Dantas, G., and M. O. A. Sommer. 2012. Con-
healthy, unrelated individuals, who groups of human microbiota will en- text mattersthe complex interplay be-
had been antibiotic therapyfree for able us to transition from snapshots tween resistome genotypes and resistance
at least one year. We characterized the to movies of commensal resistomes. phenotypes. Current Opinion in Microbiology
resistomes of both a subset of bacte- Where possible, matched samples 15:577582.
ria cultured from the subjects fecal are being collected from microbial Dantas, G., M. O. A. Sommer, P. H. Degnan,
and A. L. Goodman. 2013. Experimental
samples (the cultured resistome) as habitats touched by human activity
approaches for defining functional roles of
well as that of the entire uncultured to map out the ecology and transmis- microbes in the human gut. Annual Reviews
bacteria (the complete resistome) from sion dynamics of resistome exchange. of Microbiology 67:459475.
the same samples. We found that the Community-wide microbiome sur- Dantas, G., M. O. A. Sommer, R. D. Oluwase-
cultured resistome largely consisted veys, such as the Earth Microbiome gun, and G. M. Church. 2008. Bacteria sub-
of genes found in public sequence da- Project and the Hospital Microbiome sisting on antibiotics.Science 320:100103.
tabases, and that most of these genes Project, are profiling the genetic se- Forsberg, K. J., A. Reyes, B. Wang, E. M. Selleck,
M. O. A. Sommer, and G. Dantas. 2012. The
were identical to resistance genes quences of specialized environmental shared antibiotic resistome of soil bacteria
found in common pathogens. microbiomes and providing a frame- and human pathogens. Science 337:11071111.
This result confirmed that the re- work for mapping resistome inter- Imamovic, L., and M. O. A. Sommer. 2013. Use
sistome of cultured bacteria from the actions. Methods such as PARFuMS of collateral sensitivity networks to design
human gut have been in recent ex- will enhance the ability of functional drug cycling protocols that avoid resistance
development. Science Translational Medicine
change with pathogenic microbes. In metagenomics to uncover novel resis-
stark contrast, the majority of genes we tance mechanisms.
AP Photo/Kevork Djansezian

Moore, A.M., et al. 2013. Pediatric fecal micro-

identified in the complete resistomes The investigation of resistomes in biota harbor diverse and novel antibiotic
from these same fecal samples were diverse microbial habitats, including resistance genes. PLoS One 8(11): e78822.
novel, meaning they had little similar- the many not discussed here, serves Sommer, M. O. A., and G. Dantas. 2011. Antibi-
ity to any genes in public sequence multiple purposes. There is clearly the otics and the resistant microbiome. Current
databases. Note that every one of these basic science interest in understanding Opinion in Microbiology 14:556563.
novel genes enabled the model patho- the principles that govern the ecology, Sommer, M. O. A., G. Dantas, and G. M.
Church. 2009. Functional characterization
gen E. coli to resist clinically relevant evolution, and dynamics of antibiotic of the antibiotic resistance reservoir in the
antibiotics, highlighting that this un- resistance. Such knowledge will also human microflora. Science 325:11281131.
characterized genetic reservoir is fully assist the critical goal of slowing down
functional and must be considered the spread of multidrug resistance in
when evaluating resistome exchanges. the clinic and extending the therapeu-
A couple of recent microbial cen- tic lives of antibiotics. But no amount
For relevant Web links, consult this
suses indicate we have just scratched of resistome characterization will end issue of American Scientist Online:
the surface; deeper interrogation of the the fight against infectious diseases.
resistomes of many more individu- What is most desperately needed
als is required to begin to approach are new antibiotics, and as many of issues/id.106/past.aspx
a comprehensive view of the human them as possible. The most resound-
commensal resistome. Earlier this year, ing message that comes through from 2014 JanuaryFebruary 51