Introduction
The introduction of antibiotics to clinical therapy is a great advance in human medicine, since Fleming’s great discovery. However, the successful use of antibiotics is gradually compromised by the development of antibiotic resistance [
]. It is estimated that the antibiotic resistance will lead to 10 million annual deaths by 2050 and thus a reduction of 2% to 3.5% in Gross Domestic Product (GDP) [
]. Currently, the problem of antibiotic resistance is receiving unprecedented attention all over the world. In the Communique of G20 Summit held in Hangzhou, China, in 2016, the problem of antibiotic resistance was listed as one of the five factors that greatly impact world economy. Very recently, the UN General Assembly High-Level Meeting of Heads of State agrees on a global collaboration to cope with antibiotic resistance problems [
]. The antibiotic resistance has a long history that is comparable to the discovery of antibiotics, and the resistance even appeared prior to the clinical use of the drugs [
]. After the introduction of antibiotics in clinical and agricultural settings, bacteria have evolved mechanisms to resist nearly all antibiotics discovered so far. Not only that, the worst thing is the continuous emergence of superbugs with multi-drug resistance, leading us to stand on the edge of a post-antibiotic era [
]. It is known that the antibiotic resistance is mediated by the antibiotic resistance genes (ARGs). In fact, ARGs have been verified to be ancient in the natural environments, for example, even in 30 000-year old frozen sediments [
]. In recent years, ARGs have been characterized throughout microbiomes of natural and host-associated environments by metagenomic function-based or sequence-based strategies [
,
]. Accordingly, a new term, antibiotic resistome, has been coined to refer to the collection of ARGs in a specific bacteria or ecological niche [
,
]. The antibiotic resistance is therefore not only a public health concern but also an ecological issue with respect to environment, animal, and human. In this review, we summarize the recent research progress about the antibiotic resistome in microbiomes of natural environments, human and animal guts with respect to the reservoirs of antibiotic resistome (representative studies were summarized in Table 1), the intrinsic and mobile resistome, the forces shaping the resistome and the spread and dissemination pathway of ARGs.
The reservoirs of antibiotic resistome
It is generally accepted that the ARGs are as old as the natural-product antibiotics in bacteria, for the simple reason that the antibiotic producers should equip themselves with resistance genes to protect themselves [
]. As the antibiotic biosynthetic pathways emerged several hundred million years ago, the ARGs probably had circulated for a long time in bacterial communities before their “flourishing” in recent decades. The natural environments, therefore, is reasonably regarded as the first reservoir of ARGs [
,
]. This is indeed the case. The first comprehensive study targeting environmental resistome was published in 2006 [
]. In this study, a total of 480 spore-forming bacteria were isolated from soil samples located in different sites; these strains were screened against 21 antibiotics encompassing all major bacterial targets. It is surprising that each isolate was resistant to seven or eight antibiotics on average, even resistant to the synthetic antibiotic trimethoprim and the new lipopeptide daptomycin. Another study further demonstrated that antibiotics can serve as a sole carbon source to support the growth of soil bacteria, and many of the phylogenetically diverse bacteria are found closely related to human pathogens and resistant to multiple antibiotics [
]. Soil bacteria have also been found to share the same (perfect nucleotide identity) ARGs with diverse human pathogens [
]. Of 110 ARGs identified, 18 showed perfect amino acid identity to those in GenBank, and 55 of the 110 genes wereb-lactamase encoding genes. In addition to terrestrial environment, aquatic environment is another container for ARGs. Unlike soil, some water environments that are easily accessible to human may be more affected by anthropogenic activities. For example, due to the containing of human and animal pathogens and industrial pollutions like antibiotics and disinfectants, sewage wastewater released constantly in human activities contributes greatly to the spread and accumulation of ARGs in water environment [
]. In addition, the heavy use of prophylactic antibiotics in aquiculture, for example, fish rearing results in the frequent occurrence of antibiotic-resistant bacteria and a rapid dissemination of the antibiotic resistance determinants in water environment [
]. The ARGs have been widely explored in water environments including sewage, hospital and animal production wastewaters, ground water, surface water, drinking water, and so on, which has been summarized elsewhere [
]. The most frequently encountered ARGs in these environments are
tet genes encoding resistance to tetracyclines,
aac,
aph, and
ant genes to aminoglycosides, and a variety of
bla genes to b-lactams. Besides, the marine environment is possibly another ARG reservoir. In the work of Hatosya
et al. [
], they found the marine environments host a diversity of ARGs conferring resistance to ampicillin, tetracycline, nitrofurantoin, and sulfadimethoxine, and nearly 30% of these genes are known ARGs, while the majority are new resistance genes that have never been classified. Taken together, these facts suggest the great potential of environmental bacteria as a reservoir of antibiotic resistome that will greatly influence the human pathogens.
Compared with natural environment, host-associated environment especially the gut microbiota is undoubtedly a more complex antibiotic resistome reservoir because of the more frequent exposure to antibiotics of the gut bacteria. One of the host-associated resistome reservoirs of particular importance is the gut bacteria of farm animals. It is estimated that, in the United States, nearly 80% of antibiotics were used in animals for growth promotion, disease prophylaxis and treatment purpose [
]. It is therefore, not unexpected, that the animals and their related environments constitute a huge reservoir of ARGs. For example, in farm samples including manure and compost, Zhu
et al. detected 149 unique ARGs using quantitative PCR, the top 63 ARGs of which were enriched 192-fold (median) up to 28 000-fold (maximum) compared with respective control samples [
]. The three major resistance mechanisms detected in these farms are efflux pumps, antibiotic deactivation, and cellular protection. There are many other studies showing an increased number of resistant bacteria in animal farming environments [
–
], and the ARGs have even been frequently found in a small fecal metagenomic library from organic pigs that were reared in an antibiotic-free environment [
]. In a very recent study, the reference gene catalog of pig gut microbiome was established [
]. The ARG analysis indicated that ARGs encoding resistance to bacitracin, cephalosporin, macrolide, streptogramin B and tetracycline are prevalent among pigs from different countries, and the country-specific farm systems and antibiotic use strategies obviously affect the pig gut antibiotic resistome.
The effect of antibiotics on the alteration of the human gut bacteria as well as the resulting antibiotic resistance has long been recognized [
,
]. However, systematic investigation of the human gut antibiotic resistome just began in the last decade, which can be largely attributed to the limitation of the research methods. The concept of ARG reservoir in human intestinal bacteria was posed in 2004 [
]. It hypothesized that the human intestinal bacteria harbor numerous ARGs; at most time, these ARGs are exchanged mainly among the commensal bacteria in human gut; there will be a problem once the commensals cause post-surgical infections. Another aspect, however, is that some bacteria are merely passing through the gut, but have the chance to become ARG-carriers due to the horizontal gene transfer (HGT) that takes place in the gut flora. These new ARG-carriers may circulate in human microbiome or disseminate to other environments. This ARG reservoir hypothesis was subsequently supported by many efforts, especially through metagenomic strategies, made on exploring the resistome in human gut microbiome. We performed the first large-scale metagenome-wide analysis of human gut antibiotic resistome from 162 individuals of three different populations (Denmark, Spain, and China) [
]. The Chinese individuals were found to harbor the most abundance of ARGs among these three countries. In each population, the tetracycline resistance gene
tet accounted for the highest ratio of total ARGs, and interestingly, the tetracyclines were the most heavily used in animals, suggesting that the antibiotics use for animals potentially impacts the human gut antibiotic resistome [
]. A detailed summary of the research findings on the human gut antibiotic resistome in the metagenomic era, including the ARG databases used for annotation, the potential problems and limitations, etc. can be found in our previous review [
].
The intrinsic and mobile resistome
As the antibiotic resistance can be classified as intrinsic and acquired resistance, the antibiotic resistome is accordingly divided into intrinsic and mobile resistome. Though numerous ARGs are found among natural and host-associated environments, the real risk of these ARGs has been considered to be over-estimated [
]. The mobile ARGs that can be disseminated through HGT among bacteria are generally considered to have a higher risk for the transfer of antibiotic resistance. For example, resistance to polymyxin due to chromosomal mutation has long been recognized, however, only the recent discovery of plasmid-encoded resistance mechanism (MCR-1) arouses worldwide attention [
]. Certainly, relative low risk rank does not mean that the intrinsic ARGs are less important. In fact, the intrinsic ARGs are not merely antibiotic determinants conferring resistance phenotype, but are involved in a complex metabolic network closely related to bacterial physiology [
]. More importantly, there is a possibility that an intrinsic ARG, in certain stage of evolution, can be captured by mobile genetic elements (MGEs) and becomes a mobile ARG that can be easily spread. Understanding the intrinsic resistome will therefore contribute to the prediction of the emergence and evolution of antibiotic resistance in the future [
]. Unlike intrinsic ARGs that are relatively stationary, the mobile ARGs are highly transferable. The transfer of the mobile ARGs is largely mediated by the MGEs that are regarded as the major contributors to bacterial genome innovation and evolution [
]. These ARG-carrying MGEs include plasmids [
], transposons [
], integrons [
], integrative conjugative elements [
], genomic islands, and phages [
]. A special type of genomic island with different ARGs clustered together is called resistance islands (RIs). AbaR1 from
Acinetobacter baumannii is the largest RI identified to date, containing 45 mobile ARGs of different bacteria origins clustered within an 86-kb region [
44]. However, the whole profile of the mobile resistome in bacteria and their interaction among bacteria is not very clear.
Very recently, we presented the profile of the mobile resistome in more than 23 thousand bacterial genomes, about 9.8 million and almost 300 thousand animal gut bacterial genes [
]. We found that the mobile ARGs were mainly harbored by four bacterial phyla: Proteobacteria, Firmicutes, Bacteroidetes, and Actinobacteria, and were highly enriched in Proteobacteria. The bacterial species containing more than 40 mobile ARGs were all from Proteobacteria. In addition, the ARG transfer was the most active in this bacterial phylum. The reasons for this maybe that many human pathogens are from Proteobacteria, which probably encountered more antibiotic selection pressure. We further verified that nearly 70% of the known mobile ARGs co-existed with mobility genes of integrase/recombinase and/or transposase that are key components of MGEs, suggesting the contributing role of MGEs in facilitating mobile ARG transfer. At the species level, tetracycline resistance genes,
tet(M) and
tet(Q), and the sulfonamide resistance gene,
sul1, were the top three widely transferred ARGs, while at the genus level were
tet(C),
tet(W), and
sul1.
Forces shaping the antibiotic resistome
As HGT plays a critical role for the transfer of ARGs, factors influencing HGT undoubtedly impact the ARG exchange among bacteria. The mechanisms of HGT mainly include transformation, transduction, and conjugative transfer [
]. Transformation refers to the naturally uptake and integration of naked DNA from environments. This process depends on a physiological state of competence of the recipient bacterium. Transduction is usually mediated by phage infection to bring the DNA of a previous host to the new one. Conjugative transfer is the transfer of DNA via vehicles such as plasmid, conjugative transposon, etc
.From a macro perspective, phylogeny, ecology, and functional barriers are the major factors impacting HGT [
]. As for the ARGs, both the roles of phylogeny and ecology in shaping the antibiotic resistome have been characterized. Forsberg
et al. constructed metagenomic libraries from 18 soil samples and performed functional selection of the libraries using 18 antibiotics for resistance [
]. A total of 2895 ARGs were discovered; these ARGs represented all major antibiotic resistance mechanisms. The composition of the resistome was found to be correlated with the soil microbial phylogenetic and taxonomic structure, suggesting that the bacterial community composition (bacterial phylogeny) determined the soil resistome content. While, by exploring the recent gene transfer among 2235 bacterial genomes, Smillie
et al. characterized 10 770 unique, recently transferred genes (ARGs are included) [
]. They observed that the exchange of the transferred genes is more frequent among ecologically similar environments. The human-associated bacteria displayed 25-fold more HGT than non-human isolates of different environmental origins. Therefore, an ecology driven gene exchange network in bacteria was proposed. In another study, a new annotation method based on protein families and associated profile hidden Markov models (HMMs) was developed [
]. Using this method, Gibson
et al. analyzed the resistome in more than 600 bacterial genomes and found that the environmental and human-associated microbial communities harbor distinct antibiotic resistome. They also suggested that the antibiotic resistome is clustered by ecology.
In our study targeting the mobile resistome in bacteria, we found that the mobile ARGs were exchanged along the bacterial lineage [
]. At each phylogenetic level, the HGT of the mobile ARGs among bacteria with the same phylogenetic taxa was significantly higher than those with different phylogenetic taxa; while the HGT in bacteria from the same ecological niche only showed slightly higher than those from the different ecological niches. This result was confirmed not only based on known bacterial taxonomy, but also on real bacterial phylogenetic distances (16S rDNA distance). Interestingly, when we interrogated the transfer of the mobile ARGs between animal and human microbiomes, i.e., the transfer at bacterial community level, still we found the contributing role of bacterial phylogeny in the transfer of mobile ARGs. We therefore conclude that the transfer of the mobile resistome in bacteria was mainly controlled by phylogeny, while ecology at a large degree provided a physical barrier.
The resistance gene dissemination
As the ARGs are naturally existent in natural environments, from the evolution perspective, we can speculate that each of the clinical relevant ARGs now we encountered has an environmental origin. However, the function of these original ARGs may be distinct from their roles now serving as antibiotic resistance determinants. Once transferred to a new host, they cannot integrate into the new regulatory and metabolic network and thus became a real ARG [
]. The antibiotic use is undoubtedly the most important factor that provides the selection pressure in this process to facilitate ARG dissemination. In other words, the human activities on the production and consumption of antibiotics in both medicine and agriculture are largely responsible for the accumulation and dissemination of ARGs in modern times.
Generally speaking, the ARGs are silently circulated among natural environments, animal, and human beings, i.e., in the one health settings, and increasingly accumulated evidences have supported such an ARG transfer cycle. The first report of antibiotic resistance transfer was published in 1969, demonstrating that the R factors from animal or human
E. coli can be transferred to the resident
E. coli in the alimentary tract of a human being [
]. In recent years, along with the development of sequencing skills, more and more genomic evidence has been provided. For example, a previous study found that more than 40 unique ARGs comprising nine gene families were transferred between human and farm isolates [
]. In our large scale bacterial genome data mining analysis, ARGs were found exchanged the most frequently between animal and human bacteria, followed by between animal and aquatic bacteria and then between animal and terrestrial bacteria, emphasizing the contributing role of animals in disseminating ARGs [
]. Other studies also revealed the specific transfer of ARGs or ARG-carrying bacteria from urban and hospital wastewater to aquatic environment [
,
], manure to soil [
], between food animals and humans [
,
], etc.
The dissemination pathway of ARGs or ARG-carrying bacteria involves different aspects including direct pathway and indirect pathway, comprising various sectors and settings [
]. The sectors may include food animals, aquaculture, industrial and household antibacterial chemicals, humans, among others, and the core settings consist of farm effluents and manure spreading, commercial abattoirs/processing plants, rivers and streams, handling, preparation, and consumption of food, etc. All these sectors and settings are interconnected with each other, contributing to the propagation and dissemination of ARGs. A particularly important pathway among these is the food chain dissemination pathway. For one thing, in the food production process, antimicrobials may be used for improving the efficiency of the system, thus leading to selective pressure that promotes resistance [
]. For another, resistance bacteria may be transferred from farm to fork due to contamination or inappropriate preparation of foods like meat, fish, and vegetables. A typical example of ARG dissemination through this manner is perhaps the recently discovered
mcr-1 colistin resistance gene. The
mcr-1 gene was first found in animals and retail meats, and then in food samples and human gut microbiome, suggesting a complete food chain dissemination pathway [
].
Another important biological force implicated in the spread of antibiotic resistome among different environments is wild animal. The use of antibiotics in human and farm animals to a large degree leads to the contamination of natural environments with resistant human pathogens and antimicrobials, for example, by physical forces like wind and river flow [
]. Therefore, the wildlife in proximity to the contaminated ecosystems is increasingly exposed to resistant bacteria and ARGs. Once the wild animals acquired ARGs, there is a potential for them, especially the highly mobile species, to spread the ARGs around the world. For example, wild birds, particularly waterfowl and birds of prey, can travel a long distance and inhabit a wide range of environments, with a high potential to disseminate ARGs in their life activities [
]. The investigation of the antibiotic resistome in wildlife, culture-dependent or-independent, has been performed in many wild animals. Very recently, Vittecoq
et al. reviewed more than 210 studies concerning antibiotic resistance in wildlife [
]. They concluded that a strong link exists between human activities and the carriage of antibiotic resistant bacteria in wildlife; omnivorous, anthropophilic, and carnivorous species are at high risk of acquiring and disseminating resistance bacteria; aquatic environments are hotspots for resistance bacteria exchange and are potential sites that we can intervene by reducing contamination to control the spread of antibiotic resistance.
Conclusions
The antibiotic resistome in natural environments, animal and human microbiomes is more complex than we expected; the exchange of ARGs among the bacterial communities in these habitats can be considered as dynamic lateral gene flow. It is known that the ARGs are from natural environments, however, the human activities using antibiotics lead to a significant amplification of the original ARGs in human clinical settings. Then, antibiotics as well as the ARGs, which are now considered to be new pollutants [
,
], are released to the environments by various physical or biological forces during anthropogenic activities. Therefore, from an ecological point of view, the scenario for the resistance gene flow is that the ARGs are “from the natural environments” and “to the natural environments” (Fig. 1). Human and animals, as intermediate recipients and disseminators, contribute greatly in such a circulation from the beginning to the end.
To limit the gene flow and reduce the risk of antibiotic resistance, urgent global actions are needed, including prudent use of antibiotics in both human and animals, development of new drugs and new diagnostic methods, and international collaboration on antibiotic-resistance prevention, surveillance and control [
,
]. In addition, large-scale sampling and sequencing efforts are also necessary to fully understand the antibiotic resistome in nature and in human-associated settings. Accordingly, accurate and reliable methods for ARG prediction or annotation should be further developed. For example, revealing the structure of the ARG-carrying MGEs is vitally important for understanding the transfer, evolution, and dissemination of antibiotic resistance, however, it is still a challenge for bioinformatic algorithms to discriminate the MGEs from the next generation sequencing data. Beyond these, a particularly important and long-term task is to raise the public awareness of the problem of antibiotic resistance. According to the recent systematic review on the public’s knowledge and beliefs about antibiotic resistance, the authors concluded that the public does not completely understand antibiotic resistance and its causes; many peoples believe that they do not contribute to the development of antibiotic resistance [
].
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(173KB)
