Introduction
Discovery and development of various antibiotics in the past century are a significant milestone and one of the most successful therapeutic strategies to treat bacterial infections in people and animals. However, growing antibiotic resistance (AR) has compromised the efficacy of antibiotics, posing a serious threat to animal health, food safety and public health. In particular, emergence of multidrug-resistant (MDR) bacteria, which usually contain resistance determinants in mobile elements (e.g. transposons and plasmids), have put a severe burden on animal industry and human society (Marshall and Levy, Reference Marshall and Levy2011; Szmolka and Nagy, Reference Szmolka and Nagy2013; Michael et al., Reference Michael, Freitag, Wendlandt, Eidam, Fessler, Lopes, Kadlec and Schwarz2015). In the USA alone, infections caused by MDR organisms are estimated to cost $20 billion annually in direct health care costs, plus an additional $35 billion in costs due to lose of productivity (Centers for Disease Control and Prevention, 2013).
Bacteria have evolved multiple strategies to counteract bactericidal or bacteriostatic effects of antibiotics (Wise, Reference Wise2002; Andersson and Hughes, Reference Andersson and Hughes2010; Fisher et al., Reference Fisher, Gollan and Helaine2017). In past decades, extensive efforts have been placed on the elucidation of the molecular basis of AR and development of innovative mitigation strategies, such as development of β-lactamase inhibitors (Harris et al., Reference Harris, Tambyah and Paterson2015; Bush and Bradford, Reference Bush and Bradford2016). Recently, with the aid of high-throughput sequencing and metagenomic approaches, resistome studies have revealed a much higher level of AR diversity and novelty in different niches than previously anticipated (Pehrsson et al., Reference Pehrsson, Forsberg, Gibson, Ahmadi and Dantas2013; Harris et al., Reference Harris, Tambyah and Paterson2015; Crofts et al., Reference Crofts, Gasparrini and Dantas2017). In addition, extensive in vitro studies also have indicated that the process of horizontal gene transfer (HGT) contributes to the exchange of AR genes between bacterial organisms (same or different species) in different ecological niches, consequently playing a critical role in the dissemination and evolution of AR genes in bacteria (Broaders et al., Reference Broaders, Gahan and Marchesi2013; Huddleston, Reference Huddleston2014). However, to date, there is still a significant knowledge gap regarding in situ HGT events, as well as the factors influencing HGT in different environments (e.g. the intestine), which has impeded development of effective strategies to mitigate AR across the food chain.
In this review, we provide a brief overview of HGT in bacteria, as well as the uniqueness of the intestinal tract for efficient HGT to occur among bacterial organisms. We also identify and summarize recent progress on the factors potentially promoting in vivo HGT in the intestine. In particular, we highlight recent mechanistic studies demonstrating that certain antibiotics, such as fluoroquinolones (FQs) and β-lactams, can enhance HGT in the intestine by serving as DNA-damaging, SOS-inducing agents.
Horizontal gene transfer
The HGT process enables exchange of genetic material between bacterial cells and plays a critical role in bacterial evolution and adaptation to their environment (Ochman et al., Reference Ochman, Lawrence and Groisman2000). As a result, bacteria, even within the same species, could exhibit significant plasticity in genome for successful survival in different ecological niches. It has been widely accepted that there are three major forms of HGT: conjugation, natural transformation, and transduction (Ochman et al., Reference Ochman, Lawrence and Groisman2000). Almost all DNA sequences, including AR and virulence genes, could be transferred between bacterial cells through HGT (Davies, Reference Davies1996; Davies and Davies, Reference Davies and Davies2010). The major features and recent advances of the three types of HGT are briefly summarized below.
Conjugation
Conjugation is a stepwise DNA transfer process through a complex type IV secretion system (Curtiss, Reference Curtiss1969; Alvarez-Martinez and Christie, Reference Alvarez-Martinez and Christie2009). Conjugation needs intimate cell-to-cell contact for bridge formation between the mating pair via a conjugative pilus that belongs to type IV secretion system. Besides plasmids, some specific large mobile elements in the chromosome, such as integrating conjugative elements (ICEs) (Wozniak and Waldor, Reference Wozniak and Waldor2010), can be excised from chromosomes and subsequently transferred through the conjugative apparatus (Burrus et al., Reference Burrus, Pavlovic, Decaris and Guedon2002; Burrus and Waldor, Reference Burrus and Waldor2004). For example, SXT, a large ICE (~100 kb), and its closely related ICEs were not found prevalent among most Vibrio cholerae O1 and O139 clinical isolates until the early 1990s (Amita et al., Reference Amita, Chowdhury, Thungapathra, Ramamurthy, Nair and Ghosh2003). SXT usually bears multiple cassettes conferring resistance to chloramphenicol, sulfamethoxazole, and trimethoprim. SXT was initially found in V. cholerae O139 serogroup in late 1992 on the Indian subcontinent, and spread to most O1 and O139 clinical isolates in Asia (Amita et al., Reference Amita, Chowdhury, Thungapathra, Ramamurthy, Nair and Ghosh2003), most likely via conjugation (Waldor et al., Reference Waldor, Tschape and Mekalanos1996). More in-depth information of conjugation is available in several excellent reviews (Curtiss, Reference Curtiss1969; Smith et al., Reference Smith, Danner and Deich1981; Smith, Reference Smith1991; Christie and Vogel, Reference Christie and Vogel2000; Christie et al., Reference Christie, Atmakuri, Krishnamoorthy, Jakubowski and Cascales2005; Alvarez-Martinez and Christie, Reference Alvarez-Martinez and Christie2009).
Natural transformation
Natural transformation is a phenomenon through which bacterial cells can directly take up extracellular DNA (either linear fragment or circular plasmid), and subsequently maintain them. Since natural transformation was first discovered in Streptococcus pneumoniae in 1928 (Griffith, Reference Griffith1928), over 80 species with high-level natural transformation ability have been identified, which include a panel of gastrointestinal pathogens, such as Campylobacter jejuni, Campylobacter coli, Helicobacter pylori, and V. cholerae (Lorenz and Wackernagel, Reference Lorenz and Wackernagel1994; Johnsborg et al., Reference Johnsborg, Eldholm and Havarstein2007; Johnston et al., Reference Johnston, Martin, Fichant, Polard and Claverys2014). However, many enteric bacterial species do not display natural transformation ability under laboratory conditions. This fact, in combination with the well-known information regarding destruction of foreign DNA via restriction-modification systems in recipient cells (Palmer and Marinus, Reference Palmer and Marinus1994), raises a significant question: Is natural transformation a common and frequent HGT event in the intestine? It is likely that some bacterial organisms may conditionally acquire high-level natural transformation capability in the intestine at a specific growth stage or in response to specific cues in vivo. This speculation is partly supported by recent work showing that the commensal, Escherichia coli, previously not considered as a naturally competent bacterium, could display increased natural transformation ability upon static cultivation after the stationary phase at 37 °C (Sun et al., Reference Sun, Zhang, Mei, Jiang, Xie, Liu, Chen and Shen2006). To better understand the role of natural transformation in HGT in the intestine, examination of specific conditions or cues to trigger bacterial competence for DNA uptake is highly warranted in the future.
Transduction
Transduction is mediated by bacteriophages, viruses with specificity to bacterial hosts. Bacteriophages undergo the life cycle between integration into bacterial genome (prophages) and lytic growth stage. During lytic growth stage, a bacteriophage may be incorrectly excised from its host genome, leading to packaging of some host genetic material (donor DNA) into newly synthesized viral particles, which subsequently transfer donor DNA into another bacterial cell (recipient) through infection. A metagenomic investigation suggested that all functional bacterial genes were distributed in up to 50–60% of bacteriophages (Dinsdale et al., Reference Dinsdale, Edwards, Hall, Angly, Breitbart, Brulc, Furlan, Desnues, Haynes, Li, McDaniel, Moran, Nelson, Nilsson, Olson, Paul, Brito, Ruan, Swan, Stevens, Valentine, Thurber, Wegley, White and Rohwer2008); therefore, the bacteriophages in the gut collectively form a huge gene reservoir and are expected to play a significant role in HGT among intestinal bacteria.
The intestinal tract: a melting pot for HGT
The intestinal tract is a complex ecosystem containing all elements for efficient HGT, which enables bacteria to exchange genetic materials, including but not limited to AR and virulence genes, for adaptation to hostile conditions in the intestine (Capozzi and Spano, Reference Capozzi and Spano2009). Animal gut is increasingly recognized as a ‘melting pot’ for exchange of genetic materials across various phylogenetic distances (Shterzer and Mizrahi, Reference Shterzer and Mizrahi2015). A recent comparative microbiome study supported cross-species HGT in the gut. Genes encoding carbohydrate-active enzymes (porphyranases and agarases), originally identified in the marine bacterium, Zobellia galactanivorans, were observed to be transferred to the gut bacterium, Bacteroides plebeius, in Japanese but not in North American individuals (Hehemann et al., Reference Hehemann, Correc, Barbeyron, Helbert, Czjzek and Michel2010; Sonnenburg, Reference Sonnenburg2010). This interesting observation may be explained by the extensive ingestion of Z. galactanivorans-containing seaweed, e.g. sushi, in the Japanese diet (Hehemann et al., Reference Hehemann, Correc, Barbeyron, Helbert, Czjzek and Michel2010; Sonnenburg, Reference Sonnenburg2010). Another example of a genetic material melting pot via HGT in the intestinal tract is based on genomic examination of the gut archaeon, Methanobrevibacter smithii. Approximately 15% of genes in the genome of M. smithii were speculated to be acquired from co-resident gut bacteria, as evidenced by their GC content as well as adjacent location to mobile genetic elements (Lurie-Weinberger et al., Reference Lurie-Weinberger, Peeri and Gophna2012). Following are several unique features making the intestinal tract a perfect melting pot for HGT.
First, microbial load in the gastrointestinal tract is enormously high, which creates an optimal environment for active microbial interaction. It is estimated that the human gastrointestinal tract is inhabited by more than 1000 bacterial species and 100 billion bacterial cells, which is about ten times the amount of total human cells (Ley et al., Reference Ley, Peterson and Gordon2006). The number of genes in the human gastrointestinal microbiome is more than 100 times that of the human genome (Ley et al., Reference Ley, Peterson and Gordon2006). The bacterial density in the gastrointestinal tracts of various food animals is also as high as 1010– 1011 cells ml-1 (Whitman et al., Reference Whitman, Coleman and Wiebe1998).
Second, the gastrointestinal tract is a hostile environment with multiple levels of stress for intestinal bacteria. These stresses include but are not limited to pH, bacterial and host metabolites (e.g. bile salts), host defense factors (e.g. antimicrobial peptides), nutritional immunity (e.g. iron limitation), respiratory oxygen species, limited oxygen level, and bacterial competition (Kortman et al., Reference Kortman, Raffatellu, Swinkels and Tjalsma2014). Clearly, the antibiotics via oral administration can exert both transient and long-lasting stress to the gastrointestinal bacteria. Some of these stress signals, regardless of whether they are indigenous or exogenous, may induce and promote an in situ HGT process, which will be comprehensively reviewed in a separate section hereinafter.
Third, the gastrointestinal tract serves as an immense reservoir for AR genes which can be acquired by other gastrointestinal bacteria via HGT (Salyers et al., Reference Salyers, Gupta and Wang2004). According to the Antibiotic Resistance Genes Database (ARDB, http://ardb.cbcb.umd.edu/), there are 23,137 known AR genes against 249 different antibiotics (Liu and Pop, Reference Liu and Pop2009). A metagenome study identified 1093 AR genes from the gastrointestinal samples of 162 individuals (Sommer et al., Reference Sommer, Dantas and Church2009). A recent metagenomic analysis of gastrointestinal microbiomes from 275 individuals revealed AR genes conferring resistance to 53 antibiotics (Ghosh et al., Reference Ghosh, Gupta, Nair and Mande2013). This study also found that multiple AR genes are clustered and linked to integrase and transposase, suggesting the AR genes are part of mobile genetic elements (Ghosh et al., Reference Ghosh, Gupta, Nair and Mande2013). In food animals, the intestinal microbiome also serves as an immense reservoir for AR genes. In a metagenomic study using intestinal samples from conventionally raised beef cattle with no exposure to therapeutic antibiotics, approximately 3.7% of the sequences encoded AR genes to antibiotic and toxic compounds (Durso et al., Reference Durso, Harhay, Bono and Smith2011). In a large-scale swine study, high-capacity quantitative PCR arrays detected 149 unique AR genes among all swine fecal samples tested (Zhu et al., Reference Zhu, Johnson, Su, Qiao, Guo, Stedtfeld, Hashsham and Tiedje2013).
Notably, to date, our understanding of HGT in the gut primarily build on extensive in vitro studies. It is still largely unknown how HGT occurs in the intestine, particularly in terms of the in vivo factors influencing HGT efficiency in the intestine. The limited in vivo studies using rodent models only provided evidence of plasmid-mediated HGT transfer of AR genes, with a very narrow scope of HGT events in the gut (Schlundt et al., Reference Schlundt, Saadbye, Lohmann, Jacobsen and Nielsen1994; Feld et al., Reference Feld, Schjorring, Hammer, Licht, Danielsen, Krogfelt and Wilcks2008; Garcia-Quintanilla et al., Reference Garcia-Quintanilla, Ramos-Morales and Casadesus2008). Thus, research on in vivo HGT is still in its infancy. A better understanding of the factors influencing HGT in the intestine would help to develop practical and innovative strategies to reduce the threat and risk of AR in animal production as well as in agricultural ecosystems.
Factors potentially influencing HGT processes in the intestine
The efficiency of HGT can be influenced by various environmental factors, which have been identified and characterized by recent mechanistic, molecular, and microbiological studies. Despite the in vitro nature of these studies, many factors could be present in the intestine and potentially influence HGT efficiency in vivo. Thus, in this section, we comprehensively review and discuss the factors potentially influencing HGT processes in the intestine.
SOS response and SOS-inducing antibiotics
The SOS response, regulated by lexA and recA genes, is a global stress response triggered by DNA damage, in which DNA repair and mutagenesis are induced (Erill et al., Reference Erill, Campoy and Barbe2007). The transcription of SOS response genes is normally repressed. Upon chromosomal damage, the exposed single-stranded DNA can attract RecA to form nucleoprotein filaments, which in turn activates the expression of SOS genes by facilitating the cleavage of the LexA repressor (Beaber et al., Reference Beaber, Hochhut and Waldor2004; Schlacher and Goodman, Reference Schlacher and Goodman2007). Activation of SOS response by UV irradiation or DNA damaging agents can influence multiple aspects of cellular functions including DNA repair-mediated, mutation-based AR development (Radman, Reference Radman1975; Matic et al., Reference Matic, Rayssiguier and Radman1995). The findings from recent extensive studies further provided compelling evidence that induction of the SOS response by other agents could promote HGT in bacteria. In particular, a panel of antibiotics, such as FQs and β-lactams, can serve as DNA-damaging agents to promote HGT of virulence as well as AR genes in bacteria.
Using conjugation and gene expression assays, Beaber et al. (Reference Beaber, Hochhut and Waldor2004) demonstrated that induction of the SOS response using a SOS-inducing agent, mitomycin C, and ciprofloxacin, a FQ antibiotic, markedly enhanced ICE transfer (>300-fold) in both E. coli and V. cholerae. Mechanistic work showed that the SOS response increased the expression of genes required for ICE transfer by inactivating the repressor, SetR, consequently enhancing HGT frequency (Beaber et al., Reference Beaber, Hochhut and Waldor2004).
The FQs and β-lactams can also serve as SOS-inducing agents to trigger competence and enhance natural transformation ability of bacterial cells (Charpentier et al., Reference Charpentier, Polard and Claverys2012). For example, FQ antibiotics, which inhibit DNA gyrase, can break the DNA double strand, consequently inducing competence and enhancing the transformation in S. pneumoniae (Prudhomme et al., Reference Prudhomme, Attaiech, Sanchez, Martin and Claverys2006).
Recently, increasing evidence also indicated that SOS-inducing agents, particularly the SOS-inducing antibiotics, can trigger bacterial prophage induction, consequently enhancing transduction frequency (Comeau et al., Reference Comeau, Tetart, Trojet, Prere and Krisch2007; Allen et al., Reference Allen, Looft, Bayles, Humphrey, Levine, Alt and Stanton2011). For example, upon ciprofloxacin treatment, the genes associated with the SOS response as well as those for a viable bacteriophage were induced in Burkholderia thailandensis (Ulrich et al., Reference Ulrich, Deshazer, Kenny, Ulrich, Moravusova, Opperman, Bavari, Bowlin, Moir and Panchal2013). Modi et al. (Reference Modi, Lee, Spina and Collins2013) evaluated the effect of oral administration of ciprofloxacin and ampicillin on the resistome with a focus on the phage metagenome. They observed that the antibiotic treatment induced the SOS response, leading to the elevated abundance of AR genes in released viral particles (Modi et al., Reference Modi, Lee, Spina and Collins2013). The virome loaded with expanded AR genes potentially could transduce other resident or transient bacteria upon subsequent bacteriophage infection. Kim et al. (Reference Kim, Chui, Wang, Shen and Jeon2016) examined the effects of bovine antibiotic growth promoters (bAGPs) on the propagation and spread of Shiga toxin (Stx)-encoding phages in E. coli. Co-culture of E. coli O157:H7 and other E. coli isolated from cattle in the presence of sub-lethal concentrations of bAGPs significantly increase the emergence of non-O157 but Stx-producing E. coli by triggering the SOS response system in E. coli O157:H7. Of a panel of bAGPs tested, ciprofloxacin, chlortetracycline, and oxytetracycline induced the most significant propagation of Stx phages (Kim et al., Reference Kim, Chui, Wang, Shen and Jeon2016).
Stress hormone
Stress-related neurotransmitter hormones in the gut, such as norepinephrine (NE), can influence both the growth and virulence-associated features of a number of bacterial species (Lyte, Reference Lyte2011; Barrett et al., Reference Barrett, Ross, O'Toole, Fitzgerald and Stanton2012; Cryan and Dinan, Reference Cryan and Dinan2012). Interestingly, a recent report also showed that NE could increase conjugative transfer of AR genes between a clinical strain of Salmonella typhimurium and an E. coli recipient strain in vitro; the greatest effect was observed at the physiologically relevant concentration of 5 mM of NE during acute host stress (Peterson et al., Reference Peterson, Kumar, Gart and Narayanan2011). Phentolamine, an α-adrenergic receptor antagonist, negated the effect of NE on conjugation more strongly than propranolol, a β-adrenergic receptor antagonist. This NE-mediated enhancement in conjugation is likely due to the significantly upregulated expression of plasmid-encoded transfer (tra) genes, which is necessary for conjugation, in the presence of NE (Peterson et al., Reference Peterson, Kumar, Gart and Narayanan2011).
Microbiota and derived factors
Recently, enhanced exchange of genetic material has been observed among the bacteria engulfed by amoebae (Moliner et al., Reference Moliner, Fournier and Raoult2010). Protozoans can serve as a survival niche and protective shelter for high levels of foodborne pathogens (Tezcan-Merdol et al., Reference Tezcan-Merdol, Ljungstrom, Winiecka-Krusnell, Linder, Engstrand and Rhen2004; Olofsson et al., Reference Olofsson, Axelsson-Olsson, Brudin, Olsen and Ellstrom2013; Lambrecht et al., Reference Lambrecht, Bare, Chavatte, Bert, Sabbe and Houf2015). Given its abundance in the gastrointestinal tract, particularly in the rumen, protozoans may provide a unique niche for efficient dissemination of AR genes between bacterial cells. Notably, the rumen ciliates have been shown to boost HGT between Klebsiella and Salmonella, both in vitro and in vivo (in the rumen), most likely via conjugation (McCuddin et al., Reference McCuddin, Carlson, Rasmussen and Franklin2006). It was speculated that the close proximity of the donor and recipient, together with other stress conditions inside the ciliate, may contribute to the enhanced HGT between the different bacterial species (McCuddin et al., Reference McCuddin, Carlson, Rasmussen and Franklin2006).
Recent studies also indicated that some microbiota-derived factors that involve quorum sensing could promote HGT. Quorum sensing is a bacterial density-dependent phenomenon mediated through production and release signal molecules (autoinducers) from bacteria to the extracellular environment (Fuqua et al., Reference Fuqua, Winans and Greenberg1994; Waters and Bassler, Reference Waters and Bassler2005). Once the concentration of autoinducer reaches the minimal threshold, bacteria respond with significantly altered gene expression profiles, leading to significant changes in behavior, physiology, and even virulence (Waters and Bassler, Reference Waters and Bassler2005). It has been reported that quorum sensing induced synchronous development of competence in Pneumococcus and S pneumoniae (Tomasz, Reference Tomasz1965; Havarstein et al., Reference Havarstein, Coomaraswamy and Morrison1995). Bacterial pheromones, secreted peptides associated with quorum-sensing signaling pathway, were also observed to regulate conjugative plasmid transfer through intercellular signaling system (Dunny, Reference Dunny2013). Pheromone-responsive plasmids also have been shown to promote genome plasticity in antibiotic-resistant Enterococcus faecalis (Clewell, Reference Clewell2007; Dunny, Reference Dunny2007) as well as Enterococcus faecium (Heaton and Handwerger, Reference Heaton and Handwerger1995). Given that pheromone was prevalent in intestinal enterococci, quorum sensing may contribute significantly to HGT between Enterococcus spp. in the gut.
Conclusions
The gastrointestinal tract is a unique and ideal environment for HGT of AR genes, a programmed process playing a critical role in the development, transmission, and evolution of AR genes among bacterial organisms (same or different species). However, the process of HGT in the intestine is still largely unknown, particularly in terms of the specific factors promoting in situ HGT, which greatly impedes the development of effective AR mitigation strategies. This review summarized the findings from extensive in vitro studies and discussed the factors that potentially influence HGT processes in the intestinal tract, such as SOS response, stress hormone, microbiota and microbiota-derived factors. In the future, well-controlled animal studies are highly warranted to examine in vivo HGT processes and to evaluate the role of factors contributing to HGT in the intestine; such studies will generate new and important information for risk assessment and risk management of AR resistance, consequently developing practical and effective strategies to mitigate AR in animal production systems. For example, if a specific SOS-inducing antibiotic is demonstrated to greatly enhance HGT efficiency in the intestine, particularly via multiple HGT pathways (conjugation, transduction, or transformation) with respect to diverse AR genes, this antibiotic may be recommended for restricted use in food animals to reduce the risks of dissemination of AR genes and the emergence of AR pathogens.
Acknowledgment
The authors thank Sarah Gillespie for editing this manuscript. The authors are supported by the University of Tennessee AgResearch and NIH R21AI119462-01A.
