Prophage-Like Gene Transfer Agents—Novel Mechanisms of Gene Exchange for Methanococcus, Desulfovibrio, Brachyspira, and Rhodobacter Species

ARTICLE IN PRESS

Anaerobe 13 (2007) 43–49 www.elsevier.com/locate/anaerobe Mini-review Prophage-like gene transfer agents—Novel mechanisms of gene exchange for Methanococcus, Desulfovibrio, Brachyspira, and Rhodobacter species

Thad B. StantonÃ

Enteric Diseases and Food Safety Research Unit, National Animal Disease Center, Agricultural Research Service, United States Department of Agriculture, 2300 Dayton Road, Ames, IA 50010, USA

Received 17 January 2007; accepted 4 March 2007 Available online 4 April 2007

Abstract

Gene transfer agents (GTAs) are novel mechanisms for bacterial gene transfer. They resemble small, tailed bacteriophages in ultrastructure and act like generalized transducing prophages. In contrast to functional prophages, GTAs package random fragments of bacterial genomes and incomplete copies of their own genomes. The packaged DNA content is characteristic of the GTA and ranges in size from 4.4 to 13.6 kb. GTAs have been reported in species of Brachyspira, Methanococcus, Desulfovibrio, and Rhodobacter. The best studied GTAs are VSH-1 of the anaerobic, pathogenic spirochete Brachyspira hyodysenteriae and RcGTA of the nonsulfur, purple, photosynthetic bacterium Rhodobacter capsulatus. VSH-1 and RcGTA have likely contributed to the ecology and evolution of these bacteria. The existence of GTAs in phylogenetically diverse bacteria suggests GTAs may be more common in nature than is now appreciated. r 2007 Elsevier Ltd. All rights reserved.

Keywords: Bacteriophage; GTA; Horizontal gene transfer

1. Introduction equilibrium tips in favor of the lytic cycle when bacteria are exposed to environmental stressors, such as heat, starva- Genome sequencing projects have revealed a remarkable tion, UV light, or chemicals interacting with DNA, such as diversity and distribution of prophage gene clusters inserted certain antibiotics. into the genomes of both free-living and host-associated Non-functional prophages are also present in bacterial bacteria [1–4]. These clusters frequently contain genes genomes [4,8,9]. The genome of Escherichia coli O157:H7 providing the bacterial strain with properties advantageous EDL933, for example, contains an estimated 18 clusters of for survival in their natural environment [5–7].Thegene prophage genes [1]. Of these, only BP-933W is known to be clusters represent both functional, lysogenizing bacterio- capable of lytic growth, yielding infectious phage particles. phages and non-functional phage-like elements. The other 17 gene clusters apparently lack functional phage A functional prophage in a lysogenic bacterial strain, in genes and have been labeled ‘‘cryptic’’ [1], a label conjunction with host physiological activities, normally synonymous with defective [10]. represses genes for its lytic cycle. Its genome is replicated Cryptic prophages are incapable of self-replication [10]. along with the bacterial genome. The prophage can be They can be clusters of functional prophage genes activated and enter the lytic cycle, killing the bacterial host inactivated by mutations, prophage with incomplete and releasing offspring which are virulent when they genomes, or both. They do not produce infectious phage encounter the appropriate host strain. The lysogeny–lysis particles, that is, they are incapable of forming plaques on a bacterial indicator strain. They do not confer immunity ÃTel.: +1 515 663 7495; fax: +1 515 663 7458. to superinfection, meaning they are incapable of preventing E-mail address: [email protected]. lytic attack by related viruses.

Although they resemble genetic fossils, cryptic pro- 2. RcGTA, GTA of R. capsulatus phages can provide biological functions [10,11]. They may offer integration sites for incoming bacteriophages and R. capsulatus (previously Rhodopseudomonas capsulata) other genetic elements [12]. They might contribute to is a nonsulfur, purple photosynthetic bacterium. R. capsu- bacteriophage evolution, serving as ‘‘genetic spare parts latus is able to grow aerobically as a chemoheterotroph and kits’’ for superinfecting bacteriophages to exchange a tail anaerobically as an anoxygenic (not generating O2) protein and thereby acquire a new host range [13,14]. photosynthetic bacterium. R. capsulatus cells carry the Cryptic prophages may also provide advantages to GTA RcGTA.2 Among GTAs, RcGTA deserves superstar their bacterial hosts [4,7]. Cryptic prophage serve as status. It is the first GTA to be discovered and the most bacteriocins for Pseudomonas aeruginosa [15] and Bacillus extensively studied GTA. RcGTA genes are widespread species [16–18]. Others contain non-phage genes, ‘‘mor- among the a-proteobacteria. ons’’, benefiting their bacterial host.1 The shiga toxin gene First detected as an unknown gene exchange mechanism stx of E. coli O157:H7 EDL933 lies within a cryptic in cell-free filtrates of R. capsulatus [22], RcGTA became prophage gene cluster [1]. Serotype conversion genes the first useful gene transfer mechanism for a photosyn- are part of a cryptic prophage in the Shigella flexneri thetic bacterium [22–24]. RcGTA appears spontaneously, 2457T genome [9]. From a practical viewpoint, the loca- at low levels, in late exponential phase cultures of tion of prophage genes in a bacterial genome can serve as a R. capsulatus [25]. RcGTAs are produced by various road sign to genes important to the ecology of that R. capsulatus strains [26]. RcGTA-mediated gene transfer bacterium. does not occur across Rhodobacter species [26]. RcGTA is Another type of cryptic prophage, the prophage-like the smallest GTA, packaging 4.5 kb of DNA (Table 1) gene transfer agent [3,20], or GTA, is a mechanism of [21,27]. horizontal gene transfer and the subject of this review. The RcGTA genome and control of RcGTA gene When functional prophages become activated from expression have been studied by Lang and Beatty their quiescent state and undergo a lytic cycle, the [20,28–31]. Within the R. capsulatus chromosome, the virions typically package their own genomes and infre- RcGTA genome comprises a 14.1 kb DNA cluster of 15 quently, atypically package a portion of the host bacte- ORFs, many with homologies to head and tail genes of rial genome. Those host genes can then be transmitted known bacteriophages [28,29]. No genes encoding lytic by the virus to another bacterial cell by a small sub- functions (holin, endolysin) have been identified, leaving population of released virions. Allelic exchange within unanswered the question of how RcGTA particles escape that recipient cell completes the process of generalized from R. capsulatus cells [29]. transduction. R. capsulatus genes cckAandctrA regulate production In contrast to functional prophages, GTAs typically of RcGTA [28]. When ctrA is mutated, RcGTA gene package bacterial genome fragments and atypically pack- transcription ceases. Proteins encoded by cckA and ctrA age a portion of their own genome. A population of GTA are predicted, respectively, to be a sensor kinase/response particles contains the entire host genome and appears regulatory system that controls transcription of RcGTA capable of transmitting any bacterial gene in that genome genes in response to environmental signals [29]. The genes between cells of the bacterial host. The DNA content of a are also involved in control of R. capsulatus flagellar gene GTA particle is too small to encode its own genome. They transcription and thus cell motility [30]. A common control are incapable, therefore, of packaging their own complete mechanism for RcGTA and for motility could be genomes into a single particle. Although defective for self- advantageous for R. capsulatus both to escape from a propagation, GTAs are the consummate mechanisms of hostile environment and to benefit from genetic variability generalized transduction. provided by RcGTA in that environment [30]. One Currently known GTAs are produced by the obligate environmental signal activating RcGTA production was anaerobes Methanococcus voltae, Desulfovibrio desulfuri- identified as N-hexadecanoyl-homoserine lactone, a quor- cans, Brachyspira hyodysenteriae, and by Rhodobacter um sensing compound [31]. This finding is consistent with capsulatus which photosynthesizes under anaerobic condi- the production of RcGTA in late exponential phase tions (Table 1). R. capsulatus cultures when cell densities, nutrient limita- tions, and concentrations of quorum sensing compounds are expected to be highest. 1As a prophage traverses from bacterium to bacterium, extra genes can enter its genome through random insertion and imprecise excision events. Recent analyses of bacterial genome sequences have The genes are inessential for phage functions. They can include, however, established that RcGTA gene homologs are widely genes for fitness proteins enhancing the interactions between the bacterial distributed among the a-proteobacteria [20,32]. Thirteen host and its environment. For bacterial pathogens, these include antibiotic resistance genes, and genes encoding toxins (‘lysogenic conversion’ genes), hemolysins, and immunoreactive proteins. These extra, non-phage genes 2The R. capsulatus gene transfer agent was originally called GTA [21]. have affectionately been termed ‘‘morons’’, recognizing that they represent The designation RcGTA was adopted as a way to avoid confusion and ‘‘more DNA on’’ the phage genome than on the genomes of its relatives distinguish this specific R. capsulatus agent from the designation ‘‘GTA’’ [4,19]. applied generically and broadly to bacterial gene transfer agents [3]. ARTICLE IN PRESS T.B. Stanton / Anaerobe 13 (2007) 43–49 45

Table 1 Characteristics of prophage-like gene transfer agents (GTA)a

GTA/phage Size (nm) Capsid DNA Bacterial host Ref. size (kb) Head diameter Tail length Species Habitat/niche

Lambda 60 150 48.5 Escherichia coli Intestinal tract, facultative anaerobe [54] P1 87 226 97 Escherichia coli Intestinal tract, facultative anaerobe [54] RcGTA 30 50 4.5 Rhodobacter capsulatus Free-living, aquatic; nonsulfur, purple photosynthetic [27] VSH-1 45 64 7.5 Brachyspira hyodysenteriae Swine intestine; anaerobic spirochete pathogen [34] Dd-1 43 7 13.6 Desulfovibrio desulfuricans Free-living, soil, aquatic; anaerobic sulfate reducer [46] VTA 40 61 4.4 Methanococcus voltae Free-living, aquatic; archaebacterium, methanogen [45]

aFor the sake of comparison with GTAs, properties of functional prophages l and P1 are included in the table. l is a specialized transducing prophage and P1, a generalized transducing prophage of E. coli. of 15 species of Rhodobacterales contain the same RcGTA [36]. VSH-1 was the first natural gene transfer mechanism gene cluster as R. capsulatus and thus have the potential of to be described for any spirochete and has been used as a producing this GTA [20]. RcGTA genes are also present in tool in designing mutant strains to investigate B. hyody- the Sargasso Sea metagenome, although their bacterial senteriae biology [37]. origin is unknown [20]. A phylogeny of the a-proteobac- The amino acid sequences of VSH-1 capsid and tail teria based on the gene encoding the RcGTA capsid proteins were used to design probes and primers to detect protein resembles the phylogeny based on 16S rRNA and eventually sequence the VSH-1 genome [38]. The sequences. These findings suggest that an RcGTA ancestor genome, 16.3 kb of DNA, is over twice the size of the DNA arose early in the a-proteobacteria lineage and has evolved fragments carried by VSH-1 virions and is inserted at a along with these bacteria [20]. The wide distribution of the unique site in the chromosome of B. hyodysenteriae B78T a-proteobacteria in natural environments suggests [39]. In addition to structural proteins, the genome encodes RcGTA-like elements are active in diverse bacterial pore-forming and peptidoglycan-degrading proteins en- ecosystems. abling VSH-1 virions to escape from bacterial cells [38]. Unfortunately, only one VSH-1 gene, encoding endolysin, 3. VSH-1, GTA of B. hyodysenteriae has significant sequence similarity with any gene in the NCBI gene database. Eleven genes have been identified and Brachyspira species (previously designated Treponema seven ORFs are unidentified. The genes are arranged in and Serpulina) are obligately anaerobic spirochetes inha- head-tail-lysis gene cassettes and are transcribed as a biting the intestinal tracts of animals and humans. Some polycistronic message (Anaerobe, this issue). There is no species, such as B. innocens, B. murdochii, B. aalborgii,are evidence that viral genome replication takes place upon commensals. Other species, such as B. hyodysenteriae, VSH-1 induction (Anaerobe, this issue), consistent with the B. pilosicoli, B. intermedia, B. alvinipulli, are pathogens of inability of VSH-1, and other GTAs, to self-propagate. swine, chickens, and perhaps humans. B. hyodysenteriae is Environmental signals and internal regulatory elements the agent of swine dysentery, a severe diarrheal disease of controlling VSH-1 induction have not been well-studied, postweaning swine [33]. although H2O2, a product of swine neutrophils, is an While searching for genetic tools to investigate inducer (Anaerobe, this issue). B. hyodysenteriae, Humphrey et al. [34] purified and VSH-1-like genes and inducible VSH-1-like virions or characterized a GTA (at first, erroneously considered a both have been detected throughout the genus Brachyspira generalized transducing prophage) from mitomycin (27 strains of six species) but have not been found in other C-treated cultures of the spirochete and called it VSH-1. spirochete genera—Treponema, Borrelia, Leptospira [38–42]. VSH-1 particles resemble in morphology small versions of As the only known gene transfer mechanism of B. the coliphage l and contain 7.5 kb dsDNA molecules [35] hyodysenteriae, VSH-1 is likely responsible for the recombi- (Table 1). Restriction digest profiles of VSH-1 purified nant population structure of this enteropathogen [43]. DNA are identical to those of B. hyodysenteriae genomic DNA [34], indicatingVSH-1 virions contain random 4. GTAs of M. voltae and D. desulfuricans fragments of the B. hyodysenteriae genome. Mitomycin C-induced VSH-1 virions transfer, between M. voltae is an obligate anaerobe, an archaebacterium, B. hyodysenteriae cells, bacterial chromosomal genes (gyrB; that can be isolated from marine and freshwater environ- rrs; flaA1; nox) containing antibiotic resistance inserts ments. M. voltae converts H2 and CO2 to CH4. D. desul- [34,36] (and Stanton et al., unpublished). Spontaneously furicans is an obligately anaerobic inhabitant of soil and released VSH-1 particle also mediate horizontal gene aquatic ecosystems. It makes its living by coupling the transfer when B. hyodysenteriae strains are co-cultured oxidation of organic substrates (or H2) to the chemical ARTICLE IN PRESS 46 T.B. Stanton / Anaerobe 13 (2007) 43–49 reduction of sulfate to hydrogen sulfide. GTAs have been 6. GTAs—evolutionary and ecological questions described for both of these free-living anaerobes. The GTA of M. voltae, designated VTA, was discovered 6.1. What are GTAs? What are their evolutionary origins? and characterized by the Bertani lab group following the observation that co-culturing two auxotrophic strains Marrs and colleagues considered two possible origins for (HisÀ and PurÀ) of the methanogen produced prototrophic RcGTA [27]. RcGTA may have evolved from an ancestral offspring [44]. This GTA is filterable, resistant to DNase, bacteriophage that became defective for self-propagation and is produced spontaneously in cultures. Its ability to be and evolved into an effective mechanism for transfer of induced by mitomycin C or other prophage inducing host bacterium genes. Alternatively, RcGTA could have treatments has not been reported. The M. voltae VTA arisen independently of a bacteriophage ancestor, as a initially proved difficult to purify due to its sensitivity to product of bacterial genes involved in gene exchange. Lang centrifugation and to suspension in CsCl [44,45]. VTA and Beatty [20] have noted that genes encoding functional characteristics are given in Table 1. tailed bacteriophages seem to have arisen before a-proteo- Rapp and Wall described a GTA, known as Dd-1, which bacterial divergence (before RcGTA appearance) and appears spontaneously in D. desulfuricans cultures in the suggest ‘‘that RcGTA is a remnant of a prophage ancestor mid- to late exponential phase of growth [46]. It is not that arose before the divergence of the major phylogenetic mitomycin-C inducible. Dd-1 virions have a capsid (head) lines of prokaryotes and that replication, regulatory, and diameter similar to other GTAs but a much shorter tail lysis genes were lost in an a-proteobacterial ancestor’’. (Table 1). DNA extracted from Dd-1 particles migrates as There is no direct evidence to decide the evolutionary a discrete 13.6 kb band during gel electrophoresis [46]. origin of any GTA. For now, the simpler explanation DNA digested by restriction enzymes, however, appears as seems to be a prophage ancestry, based on the over- a smear on gels, indicating the 13.6 kb fragments are whelming predominance of bacteriophages over GTAs, the heterogeneous. Dd-1 particles in cell-free culture filtrates presence of numerous cryptic prophages in bacterial will transfer various bacterial chromosomal genes confer- genomes, the existence of non-GTA host genome packa- ring antibiotic resistance (rifampicin, neomycin, nalidixic ging agents in Bartonella and Bacillus, and the finding of acid, and novobiocin) to D. desulfuricans antibiotic phage-origin bacteriocins. The latter three examples sensitive strains. The heterogeneous nature of Dd-1 DNA indicate that the decay, disruption, and re-direction of and this gene transferring ability indicate Dd-1 virions prophage genes for alternative functions occur in bacteria contain random fragments of the D. desulfuricans chromo- and could feasibly produce a GTA. some. If additional VSH-1 or RcGTA genes are discovered Unfortunately, GTAs of M. voltae and D. desulfuricans within the Brachyspira or Rhodobacter genomes and they have not been characterized further and their respective are homologs of true prophage (early function) genes, that distributions among other species of Methanococcus and discovery would constitute evidence of a prophage origin Desulfovibrio are unknown. for these GTAs. Evolutionary insight might also be gained from the discovery of functional bacteriophage relatives of 5. Defective, host–genome packaging agents (not GTAs) GTAs or from further characterization of other GTAs.

Several bacterial species produce small bacteriophage- 6.2. As a bacterial sexual practice, does GTA-mediated gene like particles that cannot self-propagate. These agents carry transfer offer advantages over other genetic exchange host DNA. Unlike GTAs, however, they do not transfer mechanisms—transformation, conjugation, and genes. transduction? Various strains of Bartonella henselae and B. bacilliformis [47,48] produce small bacteriophage-like particles Unlike transformation by naked DNA, bacterial genes carrying 14 kb, random fragments of the bacterial genome. packaged in GTA capsids are protected from nucleases and Gene transducing ability of cell lysates could not be other DNA-damaging chemicals in the environment. demonstrated [49]. Additionally, the host recipient speciﬁcity of the GTA PBSX and PBS-like agents are common in Bacillus agent makes its DNA content inaccessible as a nutritional species [16–18,50–52]. PBSX particles carry 13 kb random, or genetic source for unrelated, competing bacterial species biased, fragments of the host bacterial genome [17,53].It in the same habitat as the host species. has been difﬁcult to cure Bacillus strains of PBSX Bacterial conjugation (involving plasmids or transpo- suggesting this element has a critical role in bacterial sons) requires the participation of donor and recipient cells growth [17]. Further, PBSX acts like a bacteriocin which that are living, physiologically active, and within close could provide a competitive advantage for B. subtilis host proximity. GTAs described in this review have been shown, strains in their natural environment [16–18]. While PBSX or are presumed to be, oxygen-tolerant and physiologically elements seem important in Bacillus physiology and inert. They maintain their gene transfer ability after ecology, their roles as gene transfer mechanisms have not exposure to environmental conditions in which their been reported. anaerobic bacterial hosts would certainly and rapidly die. ARTICLE IN PRESS T.B. Stanton / Anaerobe 13 (2007) 43–49 47

Over distances of both space and time, GTAs would enable producing higher levels of carbohydrate permeases or an anaerobe to transmit its genes to a different host oxidative response proteins such as NADH oxidase, bacterium. It would not be surprising, for example, if VSH- peroxidase, and catalase may have survival advantages 1 particles survived fecal–oral transmission and transferred over B. hyodysenteriae haploid cells. genes between B. hyodysenteriae cells in different swine hosts (although there is not yet evidence for this event). GTAs are the ultimate ‘‘lean, mean, generalized transdu- 7. Is there a gene transfer agent in your anaerobe? cing machines’’. In some respects, a greater efficiency is associated with GTA-mediated gene transfer than with A direct approach towards determining whether or not bacteriophage-mediated generalized transduction. Tradi- an anaerobic bacterium carries a GTA is to demonstrate tional generalized transducing bacteriophages, such as P1 GTA-mediated gene transfer. Two strains, each with a or P22, package a host DNA fragment into 0.1% or fewer different selectable genetic marker (antibiotic resistance, of the virions [54]. By contrast, every GTA particle for example) are co-cultured, and, after several generations, contains a fragment of the bacterial genome. From another are plated onto culture medium selective for growth of perspective, if a single, dying cell of B. hyodysenteriae recombinant offspring. Growth of greater numbers of produces 200 VSH-1 particles, each carrying a unique colonies from the co-culture than are obtained from a 7.5 kb piece of the spirochete genome, then 1.5 Mb of monoculture of either strain would be an indication that bacterial DNA or 50% of the B. hyodysenteriae 3.2 Mb gene transfer has taken place. Subsequent demonstrations genome would survive death of the host. that the gene transfer mechanism is in cell-free, filter- sterilized culture supernates and is resistant to DNAse 6.3. Might there be additional benefits from GTA production treatment, would be presumptive evidence that the transfer other than horizontal gene exchange? Is it only about sex? mechanism is a prophage or GTA. Confirming evidence would be phage-like particles visible by electron micro- A GTA-introduced fragment of DNA carrying either a scopy. GTAs can be differentiated from traditional mutated gene or an extra new gene may confer on the prophage based on their small size and their small recipient bacterium a new, advantageous phenotype for (o15 kb) DNA content (Table 1). Purified GTA DNAs survival in the environment. The spread of genetic diversity give restriction digest patterns that are indiscrete (smears) by GTAs (or any exchange system) undoubtedly contri- or resemble the electrophoretic patterns of host DNA. The butes to species evolution, and the involvement of VSH-1 final conclusive evidence that an anaerobe produces a GTA in B. hyodysenteriae evolution has been proposed [43]. would be to demonstrate that purified particles transfer Nevertheless, the most frequent gene transfer by VSH-1 bacterial genes. (used as a GTA example) within its natural environment, is For some anaerobes, genetically markered strains may likely to be the transfer of identical alleles between be unavailable or spontaneous production of a GTA might neighboring spirochetes. These extra, linear, 7.5 kb DNA be undetectable by gene transfer in co-cultures. VSH-1 was fragments would temporarily make the recipient cell initially induced and detected directly in B. hyodysenteriae diploid with respect to genes on the fragment. Given the cultures [35]. Following methods used for VSH-1, cultures slow doubling times (12–24 h) estimated for anaerobes in of an anaerobic species in early to mid-exponential phase of growth are treated with a range of two-fold concentra- the animal digestive tract [55,56], the duplicate genes could 3 serve as templates for transcription for hours before tions of a chemical inducer such as mitomycin C. The disappearing through homologous recombination into the culture OD’s are monitored to detect reductions in culture replicating bacterial chromosome. (The DNA of bacter- turbidity. (There can be a difference of only two- to four- iophage T7 is pulled into E. coli by bacterial RNA fold between an inducing concentration and a bactericidal polymerase transcribing its genes [57].) Could there be a concentration of mitomycin C !!) Generally, the ideal benefit for a B. hyodysenteriae cell to have, even inducer concentration will inhibit growth temporarily or temporarily, duplicate genes? will cause a small decrease in OD from which the culture Duplicate genes are beneficial for bacteria growing in will recover after further incubation. Cells from growth environments under nutrient limiting conditions or under inhibited cultures are harvested by centrifugation, nega- inhibitor stress. For example, starvation for arabinose or tively stained, and examined by electron microscopy to other carbohydrates results in the selection of Salmonella detect cell free, prophage-like particles. The appearance of enterica Typhimurium strains carrying duplicate permease additional migrating bands of DNA in DNA preparations genes for those carbohydrates. The strains outcompete from treated cultures (compared to untreated cultures) and strains haploid for those genes [58]. separated by gel electrophoresis is an indication that GTAs At their highest population levels in the intestinal tracts or prophages, or both, have been induced in the cultures. of dysenteric swine, B. hyodysenteriae cells are presumably 3We recently found the antibiotic Carbadox (Sigma) is a highly effective nutrient limited and undoubtedly exposed to oxidative inducer of VSH-1. On a molar basis, a VSH-1-inducing concentration of stress [59]. In such an environment, B. hyodysenterie carbadox is approximately 10À6 the cost of an equivalent mitomycin C ‘‘phenotypic mutant’’ cells with temporary, duplicate genes concentration (Sam Humphrey, personal communication). ARTICLE IN PRESS 48 T.B. Stanton / Anaerobe 13 (2007) 43–49

Prophages and GTAs can be distinguished by DNA [11] Lawrence JG. The dynamic bacterial genome. In: Higgins NP, editor. characteristics, as described above. This approach provides The bacterial chromosome. Washington DC: ASM Press; 2005. preliminary evidence of a GTA. As above, conclusive p. 19–37. [12] Lichens-Park A, Smith CL, Syvanen M. Integration of bacteriophage evidence that an anaerobe produces a GTA requires a lambda into the cryptic lambdoid prophages of E. coli. J Bacteriol demonstration that purified particles transfer bacterial 1990;172:2201–8. genes. [13] Canchaya C, Proux C, Fournous G, Bruttin A, Bru¨ssow H. Prophage genomics. Microbiol Mol Biol Rev 2003;67:238–76. [14] Figueroa-Bossi N, Bossi L. Resuscitation of a defective prophage in 8. Concluding comments Salmonella cocultures. J Bacteriol 2004;186:4038–41. [15] Nakayama K, Takashima K, Ishihara H, Shinomiya T, Kageyama Pseudomonas aeruginosa Brachyspira M, Kanaya S, et al. The R-type pyocin of is GTAs have been reported in species of , related to P2 phage, and the F-type is related to lambda phage. Mol Methanococcus, Desulfovibrio, and Rhodobacter (Table 1). Microbiol 2000;38:213–31. The existence of GTAs in these ecologially diverse, [16] Okamoto K, Mudd JA, Mangan J, Huang WM, Subbaiah TV, phylogenetically distinct bacteria implies that GTAs may Marmur J. Properties of the defective phage of Bacillus subtilis. J Mol be more common in nature than is now appreciated. It is Biol 1968;34:413–28. hoped that this review will serve not only as a source of [17] Wood HE, Dawson MT, Devine KM, McConnell DJ. Characteriza- tion of PBSX, a defective prophage of Bacillus subtilis. J Bacteriol information on GTAs but also an incentive for research 1990;172:2667–74. both to further characterize the GTAs of Methanococcus [18] Krogh S, Oreilly M, Nolan N, Devine KM. The phage-like element and Desulfovibrio species and to search more broadly for PBSX and part of the skin element, which are resident at different GTAs among other anaerobic species. locations on the Bacillus subtilis chromosome, are highly homologous. Microbiology 1996;142:2031–40. [19] Juhala RJ, Ford ME, Duda RL, Youlton A, Hatfull GF, Hendrix Acknowledgments RW. Genomic sequences of bacteriophages HK97 and HK022: pervasive genetic mosaicism in the lambdoid bacteriophages. J Mol Biol 2000;299:27–51. The author acknowledges and thanks Sam Humphrey, [20] Lang AS, Beatty JT. The gene transfer agent (GTA) of Rhodobacter Rich Zuerner, Tom Casey, and Vijay Sharma for carefully capsulatus: evolution and regulation. Trends Microbiol 2007;15: reading an early draft of this article and for providing 54–62. critical, insightful suggestions that have been incorporated [21] Solioz M, Marrs B. The gene transfer agent of Rhodopseudomonas capsulata. Purification and characterization of its nucleic acid. Arch into this final version. Biochem Biophys 1977;181:300–7. [22] Marrs B. Genetic recombination in Rhodopseudomonas capsulata. Proc Natl Acad Sci 1974;71:971–3. References [23] Scolnik PA, Marrs BL. Genetic research with photosynthetic bacteria. Annu Rev Microbiol 1987;41:703–26. [1] Perna NT, Plunkett 3rd G, Burland V, Mau B, Glasner JD, Rose DJ, [24] Wall JD, Weaver PF, Gest H. Genetic transfer of nitrogenase- et al. Genome sequence of enterohaemorrhagic Escherichia coli hydrogenase activity in Rhodopseudomonas capsulata. Nature 1975; O157:H7. Nature 2001;409:529–33. 258:630–1. [2] Canchaya C, Fournous G, Bru¨ssow H. The impact of prophages on [25] Solioz M, Yen H-C, Marrs B. Release and uptake of gene transfer bacterial chromosomes. Mol Microbiol 2004;53:9–18. agent by Rhodopseudomonas capsulata. J Bacteriol 1975;123:651–7. [3] Casjens S. Prophages and bacterial genomics: what have we learned [26] Wall JD, Weaver PF, Gest H. Gene transfer agents, bacteriophages, so far? Mol Microbiol 2003;49:277–300. and bacteriocins of Rhodopseudomonas capsulata. Arch Microbiol [4] Casjens S, Hendrix RW. Bacteriophages and the bacterial genome. 1975;105:217–24. In: Higgins NP, editor. The bacterial chromosome. Washington DC: [27] Yen HC, Hu NT, Marrs BL. Characterization of the gene transfer ASM Press; 2005. p. 39–52. agent made by an overproducer mutant of Rhodopseudomonas [5] Cheetham BF, Katz ME. A role for bacteriophages in the evolution capsulata. J Mol Biol 1979;131:157–68. and transfer of bacterial virulence determinants. Mol Microbiol [28] Lang AS, Beatty JT. Genetic analysis of a bacterial genetic exchange 1995;18:201–8. element: the gene transfer agent of Rhodobacter capsulatus. Proc Natl [6] Boyd EF, Bru¨ssow H. Common themes among bacteriophage- Acad Sci USA 2000;97:859–64. encoded virulence factors and diversity among the bacteriophages [29] Lang AS, Beatty JT. The gene transfer agent of Rhodobacter involved. Trends Microbiol 2002;10:521–9. capsulatus and ‘‘constitutive transduction’’ in prokaryotes. Arch [7] Bru¨ssow H, Canchaya C, Hardt W-D. Phages and the evolution of Microbiol 2001;175:241–9. bacterial pathogens: from genomic rearrangements to lysogenic [30] Lang AS, Beatty JT. A bacterial signal transduction system controls conversion. Microbiol Mol Biol Rev 2004;68:560–602. genetic exchange and motility. J Bacteriol 2002;184:913–8. [8] Hendrix RW, Smith MC, Burns RN, Ford ME, Hatfull GF. [31] Schaefer AL, Taylor TA, Beatty JT, Greenberg EP. Long-chain acyl- Evolutionary relationships among diverse bacteriophages and pro- homoserine lactone quorum-sensing regulation of Rhodobacter phages: all the world’s a phage. Proc Natl Acad Sci USA 1999; capsulatus gene transfer agent production. J Bacteriol 2002;184: 96:2192–7. 6515–21. [9] Wei J, Goldberg MB, Burland V, Venkatesan MM, Deng W, [32] Lang AS, Taylor TA, Beatty JT. Evolutionary implications of Fournier G, et al. Complete genome sequence and comparative phylogenetic analyses of the gene transfer agent (GTA) of Rhodo- genomics of Shigella flexneri serotype 2a strain 2457T. Infect Immun bacter capsulatus. J Mol Evol 2002;55:534–43. 2003;71:2775–86. [33] Hampson DJ, Atyeo RF, Combs BG. Swine dysentery. In: Hampson [10] Campbell AM, Cryptic prophages. In: Neidhardt F et al, editors. DJ, Stanton TB, editors. Intestinal spirochaetes in domestic animals E. coli and Salmonella: cellular and molecular biology. 2nd ed., vol. 2, and humans. Wallingford and New York: CAB International; 1997. Washington, DC: ASM Press; 1996. p. 2041–6. p. 175–209. ARTICLE IN PRESS T.B. Stanton / Anaerobe 13 (2007) 43–49 49

[34] Humphrey SB, Stanton TB, Jensen NS, Zuerner RL. Puriﬁcation and [46] Rapp BJ, Wall JD. Genetic transfer in Desulfovibrio desulfuricans. characterization of VSH-1, a generalized transducing bacteriophage Proc Natl Acad Sci 1987;84:9128–30. of Serpulina hyodysenteriae. J Bacteriol 1997;179:323–9. [47] Umemori E, Sasaki Y, Amano K, Amano Y. A phage in Bartonella [35] Humphrey SB, Stanton TB, Jensen NS. Mitomycin C induction of bacilliformis. Microbiol Immunol 1992;36:731–6. bacteriophages from Serpulina hyodysenteriae and Serpulina innocens. [48] Anderson B, Goldsmith C, Johnson A, Padmalayam I, Baumstark B. FEMS Microbiol Lett 1995;134:97–101. Bacteriophage-like particle of Rochalimaea henselae. Mol Microbiol [36] Stanton TB, Matson EG, Humphrey SB. Brachyspira (Serpulina) 1994;13:67–73. hyodysenteriae gyrB mutants and interstrain transfer of coumermycin [49] Barbian KD, Minnick MF. A bacteriophage-like particle from

A1 resistance. Appl Environ Microbiol 2001;67:2037–43. Bartonella bacilliformis. Microbiology 2000;146:599–609. [37] Li C, Corum L, Morgan D, Rosey EL, Stanton TB, Charon NW. The [50] Zahler SA. Temperate bacteriophages of B. subtilis. In: Calendar R, spirochete FlaA periplasmic flagellar sheath protein impacts flagellar editor. The bacteriophages, vol. 1. New York and London: Plenum helicity. J Bacteriol 2000;182:6698–706. Press; 1988. [38] Matson EG, Thompson MG, Humphrey SB, Zuerner RL, Stanton [51] Seaman E, Tarmy E, Marmur J. Inducible phages of Bacillus subtilis. TB. Identification of genes of VSH-1, a prophage-like gene transfer Biochemistry 1964;3:607–12. agent of Brachyspira hyodysenteriae. J Bacteriol 2005;187:5885–92. [52] Garro AJ, Marmur J. Defective bacteriophages. J Cell Physiol 1970; [39] Zuerner RL, Stanton TB, Minion FC, Li C, Charon NW, Trott DJ, 76:253–64. et al. Genetic variation in Brachyspira: chromosomal rearrangements [53] Anderson LM, Bott KF. DNA packaging by the Bacillus subtilis and sequence drift distinguish B. pilosicoli from B. hyodysenteriae. defective bacteriophage PBSX. J Virol 1985;54:773–80. Anaerobe 2004;10:229–37. [54] Birge E. Bacterial and bacteriophage genetics, 4th ed. New York: [40] Stanton TB, Thompson MG, Humphrey SB, Zuerner RL. Detection Springer; 2000. of bacteriophage VSH-1 svp38 gene in Brachyspira spirochetes. [55] El-Shazly K, Hungate RE. Fermentation capacity as a measure of net FEMS Microbiol Lett 2003;224:225–9. growth of rumen microorganisms. Appl Environ Microbiol 1965; [41] Calderaro A, Dettori G, Collini L, Ragni P, Grillo R, Cattani P, et al. 13:62–9. Bacteriophages induced from weakly beta-haemolytic human intest- [56] Gibbons RJ, Kapsimalis B. Estimates of the overall rate of growth of inal spirochaetes by mitomycin C. J Basic Microbiol 1998;38:323–35. the intestinal microflora of hamsters, guinea pigs, and mice. [42] Calderaro A, Dettori G, Grillo R, Plaisant P, Amalfitano G, Chezzi J Bacteriol 1967;93:510–2. C. Search for bacteriophages spontaneously occurring in cultures of [57] Zavriev SK, Shemyakin MF. RNA polymerase-dependent mechan- haemolytic intestinal spirochaetes of human and animal origin. ism for the stepwise T7 phage DNA transport from the virion into J Basic Microbiol 1998;38:313–22. E. coli. Nucleic Acids Res 1982;10:1635–52. [43] Trott DJ, Oxberry SL, Hampson DJ. Evidence for Serpulina [58] Sonti RV, Roth JR. Role of gene duplications in the adaptation of hyodysenteriae being recombinant, with an epidemic population Salmonella typhimurium to growth on limiting carbon sources. structure. Microbiology 1997;143:3357–65. Genetics 1989;123:19–28. [44] Bertani G. Transduction-like gene transfer in the methanogen [59] Stanton TB, Rosey EL, Kennedy MJ, Jensen NS, Bosworth BT. Methanococcus voltae. J Bacteriol 1999;181:2992–3002. Isolation, oxygen sensitivity, and virulence of NADH oxidase [45] Eiserling F, Pushkin A, Gingery M, Bertani G. Bacteriophage-like mutants of the anaerobic spirochete Brachyspira (Serpulina) hyody- particles associated with the gene transfer agent of Methanococcus senteriae, etiologic agent of swine dysentery. Appl Environ Microbiol voltae PS. J Gen Virol 1999;80:3305–8. 1999;65:5028–34.