Microbial evolution through horizontal gene transfer by mobile genetic elements

Abstract

Mobile genetic elements (MGEs) are crucial for horizontal gene transfer (HGT) in bacteria and facilitate their rapid evolution and adaptation. MGEs include plasmids, integrative and conjugative elements, transposons, insertion sequences and bacteriophages. Notably, the spread of antimicrobial resistance genes (ARGs), which poses a serious threat to public health, is primarily attributable to HGT through MGEs. This mini-review aims to provide an overview of the mechanisms by which MGEs mediate HGT in microbes. Specifically, the behaviour of conjugative plasmids in different environments and conditions was discussed, and recent methodologies for tracing the dynamics of MGEs were summarised. A comprehensive understanding of the mechanisms underlying HGT and the role of MGEs in bacterial evolution and adaptation is important to develop strategies to combat the spread of ARGs.

INTRODUCTION

Horizontal gene transfer (HGT) plays a pivotal role in the rapid evolution and adaptation of bacteria. Bacteria undergo HGT through various mechanisms including transformation, transduction and conjugation (Aminov, 2011; Arnold et al., 2022; Frost et al., 2005; Heuer & Smalla, 2007). Although HGT via outer membrane vesicles has recently been identified (Abe et al., 2020; Domingues & Nielsen, 2017), other HGTs are mediated by mobile genetic elements (MGEs), which are DNA molecules capable of moving between replicons (intracellular mobility) or between bacterial cells (intercellular mobility). MGEs contain conjugative DNA elements (plasmids and integrative and conjugative elements [ICEs]), transposable DNA elements (transposons and integrons) and bacteriophages (Figure 1). MGEs carry various genes, including antimicrobial and metal resistance, virulence and catabolic genes. Recently, antimicrobial-resistant (AMR) bacteria have become a serious threat to public health. Antimicrobial resistance genes (ARGs) are spread through HGT of ARG-associated MGEs (Rodríguez-Beltrán et al., 2021). In addition, MGEs promote the diversification of AMR bacteria (Che et al., 2021; Shintani, Vestergaard, et al., 2023). Therefore, understanding the mechanisms by which microbes exchange their genes via HGT under different conditions would provide important insights into bacterial evolution and adaptation, as well as the epidemiological basis for the spread of ARGs. This mini review aimed to provide a comprehensive summary of the characteristics of MGEs and recent research findings on microbial evolution through HGT in various environments.

FIGURE 1. Mobile genetic elements (MGEs) and their role in mediating horizontal gene transfer (HGT) within microbial cells and between cells. Bacteriophages infect microbial cells, introducing their DNA into the host chromosome. Conjugative elements, including plasmids and integrative and conjugative elements (ICEs), can be transmitted between different cells. These elements can carry additional MGEs, transposons (or insertion sequences [ISs]) and integrons.

GENERAL FEATURES OF REPRESENTATIVE MGEs

The identification of MGEs in an environment is crucial to understand their potential spread among different bacteria in different environments. Additionally, understanding the extent of the classification of MGEs is essential for estimating their inherent features. This section briefly describes the features of each MGE and the current methods (or tools) employed for identification and classification.

Bacteriophages

Bacteriophages or phages consist of protein capsids that enclose nucleic acids (DNA or RNA) capable of infecting and replicating within bacteria or archaea (Moineau, 2013). Phages are the most abundant biological entities on Earth (Comeau et al., 2008). During integration and excision within the host genome, phage DNA and the host genome can be exchanged along with the co-excision of adjacent bacterial genes, resulting in the HGT of some genes, indicating that phages play a crucial role in promoting microbial evolution (Aylward & Moniruzzaman, 2022; Balcázar, 2018; Brown-Jaque et al., 2015).

Following their discovery, identification and assignment of phages to taxonomic groups are important steps for understanding the diversity of phages. The official taxonomy was established by the International Committee on Taxonomy of Viruses, which organises viruses at several taxonomic levels (Adams et al., 2017). Phages are identified or classified based on several properties, including the composition of the nucleic acids (single-stranded or double-stranded DNA or RNA), morphology, the absence or presence of capsid protein and its structure, the similarity of their nucleotide sequences and their host range (Dion et al., 2020; Simmonds et al., 2017). However, several bacteriophage genome sequences have been determined, and many phages are yet to be identified or classified. PHAST (phage search tool, released in 2011) and its successor PHASTER (phage search tool—enhanced release in 2016) are widely used web tools for identifying prophages in bacterial genomes (Arndt et al., 2016; Zhou et al., 2011). Recently, ‘PHASTEST’ was released as a successor of PHAST and PHASTER (Wishart et al., 2023). Several tools have been developed for the identification and classification of phages based on their genomic sequences. Although creating comprehensive phylogenetic trees of phages is challenging due to the lack of universal marker genes and the mosaic structure of phages, sequence-based phylogenetic analysis using orthologous markers is suitable for classifying closely related phages (Reyes et al., 2017). Additionally, genomic signatures and phage proteomic trees have been employed to classify phages; however, this approach is not suitable for temperate phages due to their amelioration process (Deschavanne et al., 2010; Vernikos et al., 2007). Moreover, vConTACT, based on protein clusters and bipartite network-based distances, is useful for assigning dsDNA phage genomes to a taxon; however, it lacks information about the specific shared genes (Bin Jang et al., 2019; Bolduc et al., 2017). Furthermore, hidden Markov model (HMM)-based search and clustering is efficient for characterising protein families and taxonomic genes, but it requires well-described sample datasets and may not capture all phage diversity (Reyes et al., 2017). Overall, each method has its advantages and disadvantages, and continuous effort is necessary to comprehensively identify, sequence and classify all unknown or undiscovered bacteriophages.

Plasmids

Plasmids, which are either circular or linear extrachromosomal replicons, can be transmitted through conjugation (Coluzzi et al., 2022). Notably, various genes can spread effectively among bacteria via conjugation (Guglielmini et al., 2011), making it an important phenomenon for rapid bacterial evolution and adaptation. Several plasmid identification and classification systems have been proposed based on their replication, maintenance and conjugative transfer features. Incompatibility (Inc) grouping is based on the host phenotype of two different plasmids (Novick, 1987). Different plasmids belonging to the same Inc group are incompatible and cannot be inherited from the same cell line. For example, the plasmids of Enterobacteriaceae, Pseudomonas and Staphylococci and Enterococci contain IncA to IncZ, IncP-1 to P-14 and Inc1 to Inc18 respectively. It should be noted that other factors apart from the replication system, including the partition system, surface exclusion and entrance exclusion systems, can affect the compatibility of two plasmids (Garcillan-Barcia & de la Cruz, 2008; Novick, 1987). Owing to advances in sequencing techniques and the accumulation of plasmid genomic information, genes of replication initiation proteins (RIPs) have been used as universal markers for plasmid classification, known as ‘replicon typing’. ‘GR’ is a plasmid group in Acinetobacter, especially in Acinetobacter baumannii, based on replicon typing. GR1 to GR19 are classified based on polymerase chain reaction (PCR) typing for the gene encoding RIPs (Bertini et al., 2010). Additionally, GR20 to GR61 have been proposed according to the above typing scheme, with at least 80% nucleotide coverage and at least 75% nucleotide identity of RIP genes in the same group (Castro-Jaimes et al., 2022; Chen et al., 2022; Li et al., 2022). Notably, the genetic content encompassing RIPs and their adjacent DNA sequences within plasmids of Acinetobacter are similar, implying that some of the plasmids are within the same incompatibility group (Salgado-Camargo et al., 2020). Another replicon typing method, rep_clusters, has been developed through clustering analyses using mash distance metric with known replicons, providing a novel plasmid classification approach (Douarre et al., 2020; Robertson & Nash, 2018). Replicon typing is a widely used classification system for plasmids; however, it is not easy to classify multiple replicons with more than one RIP gene and oriV region. De la Cruz Barron et al. (2023) proposed that the MOB and MPF families could be used to classify plasmids (Garcillán-Barcia et al., 2009; Smillie et al., 2010). Notably, PlasmidFinder and MOB-suite are widely used tools for plasmid classification based on replicon typing and MOB and MPF typing respectively (Carattoli et al., 2014; Robertson & Nash, 2018). Although these plasmid classification systems are currently being unified in the novel COMPASS database (Douarre et al., 2020), the classifications for incompatibility groups in Pseudomonas (IncP-1 to IncP-14) included critical errors (IncP-7 and IncP-9 plasmids were inaccurately described as ‘IncP’). Recently, plasmid taxonomic unit (PTU) based on the average nucleotide sequence identity among plasmids has been proposed for plasmid identification (Cuartas et al., 2022). COPLA has been developed as a web tool for classifying plasmids based on PTUs (Redondo-Salvo et al., 2021). However, some PTUs show different results from the historical classification, as they are affected by long sized accessory genes in plasmids. For example, pWW0 and NAH7, representative IncP-9 plasmids containing toluene degradative genes (pWW0) or naphthalene degradative genes (NAH7), are not classified to the same PTU. Collectively, these classifications are effective for understanding and predicting the features of unknown plasmids.

Chromosomally integrated MGEs

Recently, several chromosomally integrated MGEs (ciMGEs) have been identified, including ICEs, integrative mobilisable elements (IMEs) and cis-mobilisable elements (CIMEs) (Botelho, 2023). ICEs are self-transmissible conjugative elements that are typically integrated into the host chromosome (Delavat et al., 2017). The transfer of ICEs is initiated by conjugation, which involves the formation of a mating pair with type IV secretion system between donor and recipient cells, followed by the transfer of a single-stranded DNA molecule from the donor to the recipient cell. Once transferred, the ICE integrates into the recipient chromosome, where it can persist and be transferred again in subsequent generations. IMEs are similar to ICEs; however, IMEs cannot transfer themselves. Although IMEs can integrate into host chromosomes, they lack the genes necessary for conjugation and, thus, depend on the presence of other conjugative elements, including conjugative plasmids or ICEs, for mobilisation and transfer to other bacterial cells. CIMEs are flanked by recombination sites (attL and attR), which are recognised by the integrase of an ICE or IME with which they are associated, but CIMEs lack the genes necessary for conjugation or recombination (Bellanger et al., 2014; Guédon et al., 2017). Like plasmids, these elements can carry genes encoding virulence factors and metabolic traits and ARGs, among other functions. Several ICEs have been deposited in the ICEBerg (Bi et al., 2012; Liu et al., 2019) or Integrative and Conjugative Element Ontology (ICEO, https://ontobee.org/ontology/ICEO/) (Liu et al., 2022) databases. Currently, over 13,000 ciMGEs have been identified in more than 20,000 bacterial and archaeal genomes, and IMEs are more abundant than ICEs (Botelho, 2023).

Transposons

Transposons are genetic elements that move between replicons via transposase activity (Mahillon & Chandler, 1998). Although transposons are not intercellularly transferable, once they are integrated into plasmids or ICEs, they can be transferred into other cells (Frost et al., 2005). Transposons usually contain a gene that encodes a transposase enzyme that catalyses transposon transposition. Insertion sequences (ISs) are minimal transposons consisting of one or two transposase genes flanked by terminal inverted repeat sequences (TIRs) (Mahillon & Chandler, 1998).

The transposase recognises and binds to the TIRs and then cuts the double-stranded DNA at specific points within the repeats. The DNA-transposase complex can then insert into the target DNA (Bordenstein & Reznikoff, 2005). Some transposons carry genes that contribute to the spread of ARGs and other traits. Multiple copies of the same IS element can promote genomic rearrangements via homologous recombination (Mahillon & Chandler, 1998).

Transposons are classified based on their mechanism of transposition (either conservative or replicative), their TIRs, or the conserved amino acid sequences of transposase, known as DDE, DEDD or HUH (two histidine residues separated by hydrophobic amino acid) motifs (Partridge et al., 2018). Currently available databases of transposons include TnCentral (https://tncentral.ncc.unesp.br/) and ISfinder (http://www-is.biotoul.fr/ISfinder), which contain approximately 400 well-annotated transposons. Apart from serving as a valuable resource to assess transposon diversity in different samples, the databases also provide downloadable software for identification and manual annotation of specific transposons in genomic sequences.

Integrons

Integrons are genetic elements that play a significant role in the spread of ARGs among bacteria, although they are not transposed or transferred like transposons or ciMGEs. All integrons possess an integron-integrase gene (intI1), a recombination site (attI) and a promoter (Pc) in the inner region of intI1 (Escudero et al., 2015; Fonseca & Vicente, 2022). The attI site is where gene cassettes, small DNA fragments containing ARGs or other functional genes, are inserted. IntI1 recognises the attI and attC sites within the gene cassettes and catalyses the site-specific recombination between the attI and attC sites, resulting in the insertion of the cassette into the integron structure. The newly acquired gene cassettes are transcribed from the Pc promoter. Integrons can rapidly accumulate and disseminate various ARGs among bacterial populations by harbouring multiple gene cassettes within a single structure.

Integrons are classified into several categories based on differences in intI1 sequences (Deng et al., 2015). Currently, 1509 integrons with integrase genes have been deposited in the integron database INTEGRALL (http://integrall.bio.ua.pt/) (Moura et al., 2009). Class 1 integron are the most important in the spread of ARGs among bacteria (Gillings, 2017). The insertion of class 1 integrons into Tn402-like transposons enables the Tn402-class1 integrons to transpose different replicons, including conjugative plasmids and spread rapidly. INTEGRALL contains several class 1 integrons with relatively new ARGs, including genes encoding carbapenemases, bla_VIM, bla_GES and bla_IMP (Kaderabkova et al., 2022). IntegronFinder is a highly accurate tool for detecting integrons in DNA sequences using HMM profiles for essential protein detection and covariance models for recombination site detection, while also offering the ability to annotate gene cassettes (Cury et al., 2016; Néron et al., 2022). Recently, the IntegronFinder has been updated to improve efficiency and usability, particularly for large datasets (Néron et al., 2022). Its relevance has been demonstrated by analysing the integron distribution in over 20,000 genomes, revealing a diverse array of integrons in Klebsiella pneumoniae genomes, some of which are associated with antibiotic resistance (Néron et al., 2022).

The four main variants in the cassette promoter region (Pc) of class 1 integrons, namely, PcW (the ancestral and weakest form), PcH1 (stronger than PcW but weaker than PcH2), PcH2 (stronger than PcH1 but weaker than PcS) and PcS (the strongest form), have different transcriptional strengths (Jové et al., 2010). Mutations of these promotor variants that do not change the IntI1 amino acid sequence have been identified in class 1 integrons of clinical isolates. PcS, PcH1 and PcH2 show higher gene cassette expression levels than PcW and are commonly found among clinical integrons but not in environmental samples (Ghaly et al., 2017; Jové et al., 2010). Class 1 integrons can be detected in most ecosystems, indicating extensive contributions to the evolution and adaptation of host strains to different environments.

HOST RANGE OF MGEs

A comprehensive understanding of the host range of MGEs is important for elucidating how MGEs spread among different bacteria and contribute to bacterial evolution via HGT.

Bacteriophages are generally host-specific and infect only a few strains of a species (Koskella et al., 2022). However, some bacteriophages have a broad host range and can infect multiple species of bacteria or multiple strains in the same species (Ross et al., 2016). Bacteriophages with a broad host range may have a larger impact on bacterial ecology and evolution. The relationships between bacteriophages and their hosts are listed in the Virus-Host Database (https://www.genome.jp/virushostdb/note.html), which includes several types of viruses (Mihara et al., 2016).

The host ranges of plasmids are considerably broader than those of bacteriophages (de Jonge et al., 2019; Jain & Srivastava, 2013). Although there is currently no consensus on the definition of the host range of plasmids, plasmids that can be replicated and maintained in different strains belonging to different bacterial classes or phyla are recognised as broad-host-range plasmids, whereas narrow-host-range plasmids are only replicated and maintained in the same genus or species (Thomas & Haines, 2004). Generally, the IncP/P-1, IncW and IncQ plasmids are classified as broad-host-range plasmids, whereas IncF, IncH, IncR and IncX are considered narrow-host-range plasmids (Yano et al., 2019). IncA, IncC, IncI, IncL, IncM, IncN and IncU are mainly found in Gammaproteobacteria, especially in Enterobacterales; therefore, they are not broad-host-range plasmids, but are not necessarily narrow-host-range plasmids either (Yano et al., 2019). The transfer host range of plasmids is sometimes broader than their replication host range (Shintani et al., 2014). Thus, the autonomous replication system is a key factor to determine the host range of plasmids. Broad-host-range plasmids can replicate in phylogenetically distinct taxa, as their replication systems can work in different bacteria. Interactions between RIPs and/or plasmid DNA region (oriV with DnaA box) and host factors, including DnaA (DnaA box binding protein), DnaB (DNA helicase) and DNA polymerase, are important for the replication of the broad-host-range IncP/P-1 plasmid (Konieczny et al., 2015; Wawrzycka et al., 2015). In contrast, IncQ plasmids encode their own DNA helicase and DNA primase (Loftie-Eaton & Rawlings, 2012). Suzuki et al. (2010) reported that the nucleotide compositions of narrow-host-range plasmids and their host chromosomes are more similar than those of broad host-range plasmids and their hosts. Host-specific mutation bias may homogenise the nucleotide composition of the plasmid host chromosome, in which the plasmid remains for a long time, called amelioration. Recently, several IncP/P-1 plasmids that do not replicate in the genus Pseudomonas have been identified (Brown et al., 2013; Hayakawa et al., 2022). Notably, some IncP/P-1 plasmids showed mutual compatibility, suggesting that they might be diversifying their host range (Hayakawa et al., 2022).

A recent comparative study showed that ICEs are more frequently transferred among diverse taxa than conjugative plasmids with an MPF_T type transport channel (Cury et al., 2018), implying that ICEs might have broader host ranges than MPF_T type plasmids. This is probably because the ICEs have an advantage when conjugating to distant hosts because of their integration into the chromosome. In contrast, taxonomically distant hosts do not allow conjugative plasmids to replicate. However, conjugative plasmids exhibit greater variability in their genetic structures, more DNA repeats and more frequent gene exchanges than ICEs. Overall, the conversion from ICE to plasmid enhances plasticity, whereas the transition from plasmid to ICE expands the element's host range (Cury et al., 2018).

The host ranges of integrons and transposons are not necessarily defined because they cannot be transferred between different cells on their own. Although the origin of class 1 integrons is considered to be Burkholderiales in the class Gammaproteobacteria (formerly Betaproteobacteria) because they commonly possess integrons in their chromosomes (Gillings et al., 2008), the host range of integrons probably depends on their vehicles, including plasmids. Conversely, some types of ISs are frequently found in specific genera. One example is IS711, which is exclusive to the intracellular pathogenic species of the genus Brucella (Bounaadja et al., 2009) and plays an important role in the genome plasticity of the genus (Aljanazreh et al., 2023). Another notable example is IS1071, which is a copy-and-paste-type (replicative) transposon that mediates bacterial evolution by homologous recombination or transposition as a composite transposon, frequently with metabolic genes (Dunon et al., 2018). The experimental transposition of IS1071 has been observed in strains of Comamonas or Delftia, but not in Agrobacterium, Escherichia, or Pseudomonas (Sota et al., 2006), suggesting that the transposition of IS1071 usually occurs in Betaproteobacteria.

Comprehensive analysis of the association of specific ISs with known ARGs has shown that IS6, Tn3, IS4 and IS1 (and their derivatives) are not only strongly associated with diverse ARGs, but also highly abundant in pathogens (Razavi et al., 2020). Additionally, the IS6 family transposons, including IS26 and IS6100, are replicative transposons (Varani et al., 2021), which are usually found in the Salmonella, Klebsiella, Escherichia, Pseudomonas, Acinetobacter and Enterobacter genera (Razavi et al., 2020). These transposons are important for the dissemination of ARGs because they are found with ARGs in conjugative plasmids (Che et al., 2021). IS6100 was originally isolated as a component of the compound transposon Tn610, which confers sulphonamide resistance in Mycobacterium (Martin et al., 1990). Additionally, IS6100 has also been identified in plasmids isolated from Flavobacterium (Kato et al., 1994) and Sphingomonas (Nagata et al., 2019). Collectively, these findings suggest that IS6100 may have a broad host range.

As described above, the determination of the host range of MGEs relies on reports detailing which MGEs are present in specific hosts and the outcomes of experiments investigating bacteria suitable for hosting these elements. This exploration has revealed that bacteria process defence systems against MGEs, including anti-phage systems such as restriction-modification system (Luria, 1953; Luria & Human, 1952), clustered regularly interspaced short palindromic repeat (CRISPR) (Barrangou et al., 2007) and retrons (Millman et al., 2020). Furthermore, bacteria can hinder bacteriophage infection through surface receptor mutations (Gurney et al., 2019), elucidating the host-specificity observed in bacteriophages, which typically infect only a few strains within a species (Koskella et al., 2022). Conversely, it has been discovered that MGEs have evolved counterpart systems to counteract host defences. Bacteriophages might have evolved in response to host defences, with mutations enabling evasion and changes in receptor-binding proteins for attachment (Broniewski et al., 2020). Notably, anti-CRISPR proteins, capable of directly inhibiting CRISPR-Cas systems, have been identified in numbers of MGEs (Camara-Wilpert et al., 2023; Pinilla-Redondo et al., 2020; Sontheimer & Davidson, 2017; Trasanidou et al., 2019). The investigation of such systems is crucial for a understanding of the host ranges of MGEs and their behaviours.

BEHAVIOURS OF MGEs IN NATURAL ENVIRONMENTS

As highlighted in the previous section, gaining insights into the host range of MGEs is pivotal for comprehending their behaviours in the environment. New-generation sequencing technology can enable prediction of the behaviours of bacteriophages in natural environments. Owing to the expanding availability of microbial genome sequences, comprehensive studies have been performed on bacteriophage-host interactions (Shaidullina & Harms, 2022). Additionally, extracellular viral particles can be separated from microbial cells, and efforts have been made to comprehensively determine the genomic sequences of bacteriophages in the environment (Kleiner et al., 2015; López-Bueno et al., 2009; Pan et al., 2019) to enable the cataloguing of viral DNAs. Moreover, a comprehensive catalogue of viral DNAs is important to predict the phage sequences in metagenomic data (Hasiów-Jaroszewska et al., 2021; Knipe et al., 2022; Urayama et al., 2022).

In contrast to viruses, metagenomic data does not accurately predict the behaviours of plasmids in natural environments, as it is difficult to determine whether DNA fragments originate from a plasmid or a chromosome (Smalla et al., 2015). This makes it particularly challenging to detect plasmids in metagenomic data, as plasmid DNAs are usually physically separate from the host chromosomal DNA. Given that transposons and integrons can be transferred with conjugative plasmids, it is important to understand the behaviours of plasmids in the environment to understand the crucial role that these MGEs play in microbial evolution and ecology. Therefore, this section focuses on the behaviour of plasmids containing transposons and integrons in natural environments (Figure 2). Table S1 presents the lists of several studies on this topic.

FIGURE 2. Key aspects for elucidating the behaviours of plasmids in different environments and corresponding experimental methodologies.

To understand the behaviour of plasmids in the environment, it is important to identify the plasmids that exist in each environment. Until the 1990s, collecting and culturing bacteria from the environment using the phenotype was the classical method of searching for new MGEs, including plasmids. Sobecky et al. (1998) identified various plasmids in isolated marine bacteria, most of which were different from the Inc groups of existing intestinal bacteria-derived plasmids. However, given that most bacteria in the environment cannot be cultured, culture-dependent methods have limited ability to detect plasmids. Currently, PCR using primers designed from specific DNA sequences of known plasmids (and other MGEs) is the simplest method of identifying plasmids in the environment. Total DNA is directly extracted from the target environment and used as a template for PCR, allowing the detection of the presence or absence of plasmids based on amplification. Several primer sequences enabling multiplex PCR have been used to detect plasmids that are particularly important in the spread of ARGs (Carattoli, 2009; Carattoli et al., 2005, 2014; Carloni et al., 2017; Castro-Jaimes et al., 2022; Chen et al., 2022; Clewell, 2007; Li et al., 2022; Lozano et al., 2012). PCR-Southern blot and quantitative PCR analyses using primers or probes specific to plasmid sequences have been used to detect the abundance of specific plasmids and their temporal changes in various environmental samples (Blau et al., 2020). However, a limitation of these methods is that only known plasmid primers and probes can be used.

To identify the actively propagating plasmids in the environment, the exogenous plasmid capture method proves valuable for isolating conjugative plasmids from environmental samples. In this method, the type of the recipient strain and mobilisable plasmid used can influence the plasmids obtained. There are two mating methods: biparental and triparental. In the biparental mating method, a culturable recipient strain is used to collect conjugative plasmids with specific accessory gene(s), including antimicrobial resistance, heavy-metal resistance and metabolic genes (Blau et al., 2020; Smalla et al., 2015). The triparental mating method uses an intermediate donor carrying a mobilisable plasmid (Figure 3). It relies on the ability of a self-transmissible plasmid to mobilise a prepared mobilisable plasmid, which is more efficient for collecting conjugative plasmids from the environment because it does not require any marker gene(s) in the plasmid. The advantage of these methods is that they enable the direct acquisition of conjugative plasmids independent of the culturability of the original host bacteria in the environment. Several previously unknown conjugative plasmids have been identified using these methods, including PromA group plasmids, novel IncP/P-1 subgroup plasmids and unclassified conjugative plasmids (Blau et al., 2018; González-Plaza et al., 2019; Hayakawa et al., 2022; Li et al., 2014; Oliveira et al., 2013; Smalla et al., 2000; Werner et al., 2020; Wolters et al., 2014; Yanagiya et al., 2018). Some disadvantages of these methods are that they cannot identify the original host bacteria in environmental samples, and only conjugative (or mobilisable) plasmids that can replicate in the recipient cell can be collected. With the triparental method, the type of plasmid obtained is biased by the type of mobilisable plasmid used.

FIGURE 3. Exogenous plasmid capture methods: triparental mating and biparental mating.

To identify which plasmids are transferred into the microbial communities (‘plasmid destination’), fluorescent proteins, including GFP, DsRed, mCherry, GusA and LuxAB, and flow cytometry have been used to identify bacterial strains capable of acquiring conjugative plasmids (Amann & Fuchs, 2008; Sørensen et al., 2005). Researchers have constructed a system in which fluorescent proteins downstream of a lac-modified promoter in the MGE are inserted into a gene that expresses the inhibitory factor LacI in the chromosomal DNA of the donor bacterium, enabling the detection of cells that have received the MGE at the single-cell level in various environments (Christensen et al., 1998; Geisenberger et al., 1999; Haagensen et al., 2002; Mølbak et al., 2003; Normander et al., 1998).

Additionally, it is important to understand which bacteria in the environment contain specific plasmids (genes). Various methods have been used to address this issue. Fluorescence in situ hybridisation (FISH), targeting the 16S rRNA gene sequences of bacteria, or in situ PCR, detecting functional genes, can detect and identify cells that contain plasmids at the single-cell level (Laflamme et al., 2009; Ma et al., 2013; Shintani et al., 2014). Recently, direct-geneFISH was developed to detect specific genes and rRNA genes to identify the ‘owner’ of the target gene(s) using fluorochrome-labelled polynucleotide gene probes (Barrero-Canosa et al., 2017).

Other methods for detecting plasmids in microbial communities include emulsion, paired isolation and concatenation PCR (epicPCR), a type of fusion PCR that links a targeted sequence to the 16S rRNA gene fragment of its carrier. The carrier can be taxonomically assigned (Roman et al., 2021; Spencer et al., 2016). Additionally, droplet digital PCR (ddPCR) was developed based on digital PCR using a water–oil emulsion droplet system, which can simultaneously generate millions of PCR reactions in a single tube (Hindson et al., 2011). Each droplet has a picolitre-to-nanolitre-sized volume and is separated by oil; therefore, each PCR reaction is conducted independently. ddPCR can quantify the precise copy number of targeted genes; thus, it is used to detect the targeted bacteria in the microbial community (Aung et al., 2023). High-throughput chromosome conformation capture (Hi-C) is another method used to identify bacteria that carry a target gene, including an ARG and a plasmid-specific gene. Overall, these culture-independent methods have the potential to facilitate future studies on the behaviour of MGEs in microbial communities.

epicPCR has been used to identify ARG-carrying bacteria in bacterial communities in the influent and effluent of wastewater treatment plants (Hultman et al., 2018). Additionally, epicPCR has been used for the host cell assignment of the conjugative plasmid RP4 in a synthetic bacterial community (Heß et al., 2021). De la Cruz Barron et al. (2023) reported that multiple target genes can be detected using multiplex ddPCR. Stalder et al. (2019) identified the host ranges of ARGs, plasmids and integrons, validating the approach and demonstrating that IncQ plasmids and class 1 integrons had the broadest host range in the wastewater community examined.

One of the pivotal discoveries of these studies is the identification of large reservoirs of previously unknown plasmids in the environment, suggesting that plasmids may play a major role in microbial communities. Moreover, recent single-cell analyses have identified the transfer ranges of conjugative plasmids. The finding challenges existing assumptions and indicates that the mobility of genetic material between bacterial cells is more extensive and complex than previously realised. Additionally, these findings indicate that conjugative plasmids exhibit a remarkable capacity to recruit a diverse array of genes in different environments. Overall, these findings contribute to the understanding of microbial genetics, with potential applications in the fields of biotechnology, ecology and medicine.

CO-EVOLUTION OF PLASMIDS AND THEIR HOSTS

Plasmids not only deliver genes but also accelerate host evolution by promoting the adaptation of plasmid–host interactions. A major question in microbiology is how MGEs, including plasmids, coexist and evolve in their hosts. In this section, we summarise research reports on how MGEs have co-evolved with their hosts, with a focus on plasmids.

Compensatory mutations can resolve the ‘plasmid paradox’

This section focuses on the question of the ‘plasmid paradox.’ Although the fitness of the host increases in certain environments, plasmids are eliminated by selection during the long process of microbial evolution once the necessary genes are incorporated into the chromosome. However, several plasmids have been identified, including those without accessory genes (cryptic plasmids), and it is difficult to explain why plasmids still paradoxically exist. Some recent reviews have summarised how this paradox can be resolved based on experimental data and theories regarding the ecological and evolutionary coexistence mechanisms of plasmids and their hosts (Brockhurst & Harrison, 2022; Rodríguez-Beltrán et al., 2021). Fitness cost by the plasmid carriage can be ameliorated by mutations in genes of plasmids or the host chromosome, which is known as ‘compensatory mutation.’ Previous studies have shown that mutations in the plasmid replication initiation protein (RepA) of pPS10 or its host DNA replication initiation protein (DnaA) can enhance plasmid persistence and change its host range (Fernández-Tresguerres et al., 1995; Maestro et al., 2002, 2003). Interaction between the replication initiation protein of the IncP/P-1 plasmid TrfA and the host helicase DnaB reduces the fitness of Shwenella, an inappropriate host for the IncP/P-1 plasmid (Yano et al., 2012, 2016). TrfA has two forms: TrfA44 (the larger form) and TrfA33 (the smaller form), which differ in terms of the location of the initial codons (Pansegrau et al., 1994). Mutations in TrfA44 can inhibit its interaction with DnaB, resulting in mitigation of the fitness cost (Yano et al., 2016). Several studies have shown that compensatory mutations can ameliorate the cost of carriage of the plasmid pQBR103 in Pseudomonas spp., not only under laboratory conditions (Harrison et al., 2015, 2016), but also in the rhizosphere (Bird et al., 2023). Mutations in the GacA/GacS system encoded on the host chromosome can reduce the fitness cost of plasmid carriage (Harrison et al., 2015). San Millan et al. (2018) investigated the interaction between Pseudomonas aeruginosa and different plasmids and found that the acquisition of the plasmids resulted in altered gene expression and reduced the host fitness. Mutations that decrease the expression of the plasmid replication gene of the small plasmid pNUK73 recover fitness costs (San Millan et al., 2015). The fitness cost imposed on the host is thus dependent on the combination of the plasmid and host and can be ameliorated by compensatory mutations in either the plasmid or chromosomal genes.

Fitness costs of plasmids and the spread of ARGs

The fitness cost of the plasmid is an important factor in understanding how the various accessory genes including ARGs spread via plasmids. The fitness cost by plasmid carriage varies between host species, depending on the specific genetic interaction between the plasmid and host chromosome (Kottara et al., 2018; Shintani et al., 2010; Takahashi et al., 2015). Thus, there is an optimal combination between a plasmid and its host species, which may be an important factor in the spread of accessory genes. In addition to the carriage of plasmids, the carriage of ARGs can also reduce the fitness of the host. Yang et al. (2021) showed that even a single copy of mcr-1, a colistin-resistance gene, exerts a fitness cost on its host Escherichia coli and that copy number control of the plasmid carrying mcr-1 is important for the persistence of the gene. Additionally, a novel activator of the IncX4 plasmid, PixR, mitigated the fitness cost of mcr-1 in E. coli (Yi et al., 2022). Moreover, mcr-1 has also been identified in IncP/P-1 plasmids that carry only trfA2 (Hayakawa et al., 2022; Zhao et al., 2017). TrfA1 is more detrimental to the persistence of plasmids than TrfA2 (Yano et al., 2016). This is because TrfA1 activates the host helicase DnaB and increases the copy number of plasmids, which is important for the promiscuity of IncP/P-1 plasmids but is not advantageous to fitness cost for some bacteria (Yano et al., 2016). In contrast, IncP/P-1 possessing only TrfA2 (narrow-host-range-type IncP/P-1 plasmids) show lower fitness cost than those with TrfA1. This could explain why only narrow-host-range-type IncP/P-1 plasmids with low fitness cost for the host cells carry the colistin-resistance gene (Hayakawa et al., 2022).

Rearrangements and deletions

Rearrangements of the DNA regions of plasmids and host chromosomes are usually observed during the adaptation of host cells to plasmids under specific conditions. pCAR1 is an IncP-7 plasmid that endows carbazole metabolism to its host Pseudomonas strain, as it carries degradative genes (car and ant operons) involved in the bioconversion of carbazole to catechol via anthranilate (Nojiri, 2012). However, one of the hosts of pCAR1, Pseudomonas fluorescens Pf0-1(pCAR1), showed a significantly slower growth rate than did other Pseudomonas host strains (Takahashi et al., 2009). This is because the host Pf0-1(pCAR1) accumulates catechol, a compound that is toxic to cells whose metabolism is mediated by enzymes encoded on the host chromosome. Notably, the dominant strains at the end of the growth period of Pf0-1(pCAR1) carried pCAR1 lacking the carbazole degradative operons (Takahashi et al., 2009). Deletion of these genes prevented the dominant strains from metabolising carbazole or anthranilate and suppressed the accumulation of catechol. The loss of these DNA regions is mediated by the homologous recombination of two of the four copies of ISPre1 (Shintani, Horisaki, et al., 2011; Shintani, Matsumoto, et al., 2011). Such DNA rearrangements have also been reported in the multidrug-resistant IncHI2 plasmid pJXP9 (Zhang et al., 2022). Deletion of specific DNA regions of the plasmid can compensate for the fitness cost to its host, and adaptive evolution of chromosomal gene mutations and altered mRNA expression correlate with changes in the physiological functions of the host (Zhang et al., 2022).

PERSPECTIVES

Innovations in sequencing technology and advances in analytical methods at the bacterial cellular level have enhanced understanding of the mechanisms of MGEs in the evolution and adaptation of bacteria and have improved understanding of the significance of MGEs.

Identification of MGEs and hosts

A future challenge is the development of methods to easily identify various MGE sequences and their corresponding hosts from metagenomic data. Although in silico identification of MGEs has been developed, new types of MGEs have recently been identified, including phage satellites, phage plasmids and tycheposons (Hackl et al., 2023; Ibarra-Chávez et al., 2021; Pfeifer et al., 2022). These elements should not be ignored, although comprehensive analyses are currently not available. Additionally, the host information of bacteriophages and plasmids is particularly inadequate. Several attempts have been made to predict hosts by analysing the dissimilarities in sequence information between the plasmid and its host (Suzuki et al., 2008, 2010; Tokuda et al., 2020, 2023). Recently, machine learning-based tools, including PlasmidHostFinder, have been developed to predict the host of a plasmid based on its sequence (Aytan-Aktug et al., 2022). These developments bridge the gap between our understanding of MGEs and their intricate relationships with host bacteria.

Standardisation of MGE annotation and classification

Another challenge is the uniformity of the annotation and classification of MGEs, especially for plasmids. Despite the efforts of various researchers to develop reference databases for plasmids, there is currently no unified index for their classification and gene names. The discrepancies between the plasmid ‘taxonomy’ classified in the early phase of plasmid research and the classification method based on their nucleotide sequences have caused several confusions (Shintani et al., 2022; Shintani, Suzuki, et al., 2023). To avoid such confusion, uniformity in the annotation (Thomas et al., 2017) and classification of plasmids (Garcillán-Barcia et al., 2023) is crucial. Moreover, it is essential to regularly update comprehensive databases that summarise the properties and classification of each plasmid.

Utilising single-cell genomics

Furthermore, single-cell genomics of environmental bacteria has the potential to provide valuable information (Hosokawa et al., 2022). Although such methods are time-consuming, they can help prevent misidentification of the hosts of MGEs or mis-assembly of genomic sequences. An incorporation and utilisation of these technologies and methods may facilitate the real-time monitoring of MGE-induced bacterial evolution in the near future.

As DNA sequencing has become increasingly accessible, there is a rapid accumulation of information from metagenomic sequencing. Moreover, decoding samples through time series analysis has become more straightforward. However, the availability of information regarding the presence and distribution of MGEs, including plasmids and transposons, is limited. This scarcity arises from the unfulfilled requirements of the aforementioned breakthroughs. Upon achieving these advancements, it will be feasible to discern the types of MGEs existing in a specific environment and identify the microorganisms propagating among them based on nucleotide sequence information. This will lead to an understanding of eco-evolutionary plasmid trajectories proposed by Castañeda-Barba et al. (2024) and to the elucidation of the mechanisms of gene propagation and microbial evolution.

AUTHOR CONTRIBUTIONS

Maho Tokuda: Writing – original draft (supporting); writing – review and editing (supporting). Masaki Shintani: Conceptualization (lead); funding acquisition (lead); supervision (lead); writing – original draft (lead); writing – review and editing (lead).

ACKNOWLEDGEMENTS

We thank Editage (www.editage.jp) for English language editing. This work was supported by grants (JP19H02869, JP19H05686 and JP23H02124 to M. Shintani; JP22J12723 to M. Tokuda) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan; grants (JP21wm0325022 and JP21wm0225008 to M. Shintani) from the Japan Agency for Medical Research and Development (AMED); a grant to M. Shintani from the Consortium for the Exploration of Microbial Functions of Ohsumi Frontier Science Foundation, Japan, and Institute for Fermentation Osaka (IFO), Japan (L-2023-1-002); a grant to M. Shintani from the Research Institute of Green Science and Technology Fund for Research Project Support (2023-RIGST-23104), National University Corporation Shizuoka University, Japan.

FUNDING INFORMATION

No funding information provided.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no conflict of interest.