- Email: [email protected]
A Possible Role for Plasmids in Mediating the Cell–cell Proximity Required for Gene Flux
J. theor. Biol. (1996) 181, 237–243
A Possible Role for Plasmids in Mediating the Cell–cell Proximity Required for Gene Flux C F. A´-C† M. E. C‡ † Departamento de Microbiologı´ a, LUSARA. Apartado Postal 102-006, 08930, Me´xico, D.F., MEXICO ‡ Department of Surgical Research, Enders Building, Rm 1012, Children’s Hospital, 300 Longwood Ave., Boston, MA 02115, U.S.A. (Received on 10 July 1995, Accepted in revised form on 17 April 1996)
One of the major requirements for successful gene flux is a close proximity between participating organisms. In previous articles, we have proposed that plasmids act as powerful vehicles transporting genes collected by integration and transposition, mainly via the process of conjugation. However, in addition to conjugation, there are other processes, also mediated by plasmids, in which different cells come into very close contact with each other, such as symbiosis and the formation of multi-specific cellular communities. There is evidence that suggests that such intimate associations between cells may facilitate gene transfer events, even between distantly related organisms. Examples of symbiotic endosymbiotic, and parasitic associations provide evidence in support of the role of plasmids in bridging the genetic gap between species. In this purely theoretical article we attempt to conceptualize existing data on this subject, provide new insights and present testable predictions on how plasmids may facilitate gene flux by bringing cells together. 7 1996 Academic Press Limited
from bacteria to plant cells, seem to constitute the most frequent, versatile, and sophisticated mechanisms of horizontal gene transfer (Ama´bile-Cuevas & Chicurel, 1992, 1993). In addition to conjugation, plasmids enhance or increase the likelihood of other lateral transfer mechanisms. The transfer of a gene by transformation is more successful and has more profound consequences if it involves plasmids as carriers. Likewise, in cases where plasmids are enclosed in phage particles, genes contained in the plasmid are generally more efficiently transferred than genes contained in non-plasmid molecules. A striking example of the effects of plasmids and plasmid-mediated gene transfer can be found in the origin, evolution and spread of antibiotic resistance genes (for review see Ama´bile-Cuevas, 1993; Ama´bile-Cuevas & Chicurel, 1992, 1993). But plasmids can act to facilitate gene flux in additional ways. In this article we discuss how plasmids may have contributed to gene transfer by bringing different cells together.
Introduction Horizontal gene transfer is the transmission of genetic information from one independent and fully developed organism to another. In horizontal gene transfer organisms can acquire novel characteristics, through the incorporation of foreign genetic material, that are directly inherited by future generations. For many years horizontal gene transfer was regarded as a peculiarity of bacteria. However, a large amount of evidence has recently accumulated indicating that the extent of such gene flux is very large, including gene transfers between distantly related species, as well as, in some cases, even eukaryotes. Plasmids and Gene Flux Conjugation between bacteria, between bacteria and yeast, and the conjugative-like T-DNA transfer † Author to whom correspondence should be addressed. 0022–5193/96/150237 + 07 $18.00/0
237
7 1996 Academic Press Limited
238
. . ´ - . .
One of the major requirements for successful gene transfer is a close proximity between the participating organisms. Cell–cell contact is required for all conjugative or conjugative-like processes. However, even for transformation, where cell–cell contact is not necessary, proximity is essential. The lifetime of DNA outside the cell is relatively short, and thus cells to be transformed must be close to the DNA source at the time of its release. Transduction, wherein DNA molecules to be transferred are enclosed in viral particles, may occur between cells widely separated in space. However, the inactivation of phages often happens shortly after their release, especially in bacterially populated media (Saye et al., 1987), and transduction is further limited by its usually narrow spectrum of specific hosts and recipients. Conjugative plasmids encode for the machinery that bring the donor and recipient cells together. In Gram-negative bacteria this is by means of pili, or through aggregation substances induced by ‘‘sex’’ pheromones in some Gram-positive cocci, plasmidencoded mechanisms are responsible for the attachment of the participating bacteria (Clewell, 1993; Ippen-Ihler, 1989). The primary role of such adhesion mechanisms could have been the attachment of invading bacteria to the cells of infected organisms, but this attachment may also have facilitated attachment to other bacterial cells. The microfibrillar surface structures mediating adherence induced by pheromones in Enterococcus faecalis, for example, also seem to be capable of mediating bacterial adherence to pig renal tubular cells (Clewell, 1993). The relationship between conjugation and pathogenicity is not restricted to adherence determinants: pheromones themselves may play a role in virulence (Clewell, 1990), as does the TraT lipoprotein, a Gram-negative plasmid exclusion factor (Sukupolvi & O’Connor, 1990). It is very likely that genes encoding for adhesive mechanisms were ‘‘recruited’’, through transposition or recombination, by plasmids, thus increasing their transfer efficiency and becoming ‘‘conjugative’’ plasmids. Pathogenic invasion, as conjugation, is preceded by the attachment of the involved cells. And this adherence may also enhance the possibilities of horizontal gene transfer. Plasmid-mediated adherence often mediates the aggregation of different microbial cells. These consortia, some of them called biofilms (Costerton et al., 1987), most probably constitute excellent environments for gene exchange. Biofilms can be composed of a variety of microorganisms, including bacteria, algae, protozoa and fungi (Characklis & Marshall, 1990), and it has been shown that, at least
in streams, gene transfer by conjugation occurs at high rates. In biofilms in the West River Taff, for example, transfer frequencies ranged from 1 to 10−6 transconjugants per member of recipient populations (Fry & Day, 1990). It is important to recall that conjugation is possible not only between distantly related bacteria (Mazodier & Davies, 1991; Courvalin, 1994), but also between bacteria and eukaryotic cells (Ama´bile-Cuevas & Chicurel, 1992). Figure 1 summarizes the known instances of bacterial cell–cell adherence that may result in increased gene flux. Bacteria produce adhesins or more complex structures, such as pili and curli, in order to attach to different surfaces, including larger cells. The production of such molecules and structures is often plasmid-mediated (Gulig, 1990; Willshaw et al.,
Adhesive bacteria (a)
(c) Biofilms Predation (b)
Conjugation
Endosymbiosis? F. 1. Bacteria attach to each other in a number of ways. (a) Multi-cellular, multi-specific bacterial associations, called biofilms, adhere to a variety of surfaces. Within these associations, the likelihood of gene transfer is increased. Biofilm formation is sometimes mediated by plasmid-encoded adhesins. (b) The cell-to-cell contact required by conjugation, must start with the attachment of the participating bacteria. This is achieved by very diverse mechanisms, such as the production of pili or the secretion of adhesins. Almost all conjugative processes are plasmid-encoded, and conjugation is the most powerful mechanism of horizontal gene transfer. (c) Bacterial predation must also start with the attachment of the predatory bacterium to its prey. Whether this process is plasmid-encoded, and whether it represents an early evolutionary step in the development of endosymbiosis and the origin of organelles, remains unknown.
1990). Pathogenic colonization, which starts with this attachment, brings into extremely close proximity bacterial and larger eukaryotic cells. As a result, gene transfer may occur. Genes that may have traveled this way include the gene encoding PapD, a chaperone protein that mediates the assembly of pili in uropathogenic E. coli, whose folding and amino acid sequence are quite similar to the human lymphocyte differentiation antigen Leu-1/CD5 (Holmgren & Bra¨nden, 1989). In addition, there is strong evidence indicating that E. coli acquired a copy of the glyceraldehyde-3-phosphate dehydrogenase gene from a eukaryotic organism. It is possible that this transfer occurred in the gut of an ancestral eukaryotic donor (Smith et al., 1992). The presence of sialidases among bacteria that exist as animal commensals or pathogens may be another case of acquisition of a eukaryotic gene by prokaryotes that live in close association with animals (Roggentin et al., 1993). Again, the laterally transferred gene is itself involved in bacterial virulence. A final, recently reviewed example (Bliska et al., 1993), is provided by the enteropathogen Yersinia. Adherence of these bacteria to mammalian cells is mediated by the YadA adhesin, encoded by the plasmid that confers virulence to Yersinia pseudotuberculosis and Y. enterocolitica. This plasmid also bears other pathogenic determinants, whose products are the Yersinia outer membrane proteins, or Yops. Several of these proteins share homology with eukaryotic signal transduction proteins, such as a tyrosine phosphatase, the GPIba receptor, and a ser/thr kinase. Yops, on the other hand, are not homologous to other known bacterial proteins (Bliska et al., 1993). This suggests that the pathogenic potential of Yersinia, originally provided by a plasmid-encoded adhesin, has been enhanced by the acquisition of eukaryotic genes, whose products alter signal transduction mechanisms within host cells. Gene transfer itself can constitute a form of plasmid-mediated virulence. The infection of plant cells by Agrobacterium for example, involves direct transfer of bacterial DNA to the host plant cell. This parasitic relationship is mediated by Ti plasmids. Agrobacteria respond to compounds released by wounded plant cells, activating vir genes that are responsible for cell adhesion and DNA transfer (Hess et al., 1991; Winans, 1991). This response is encoded by Ti plasmid genes. Furthermore, Ti plasmids also encode an autoinducible system for conjugal Ti plasmid transfer between Agrobacteria (Piper et al., 1993; Zhang & Kerr, 1991). Thus, two distinct systems for increasing cell proximity and DNA transfer are encoded within single Ti plasmids.
239
So far, there are very few examples of gene transfer from bacteria to mammalian cells. This may be a consequence of the more limited ability of eukaryotic cells in assimilating naked DNA and incorporating it into their nuclei for functional expression, in contrast to the relative ease with which bacterial transformation can occur. However, the acquisition and expression of DNA sequences by mammalian cells in vivo (Wolff et al., 1990) suggests that bacterialmammalian gene transfer might not be altogether impossible. Experimental models require very high amounts of intramuscularly administered DNA (Wolff et al., 1991). An injection of 100 mg of Bacillus thuringiensis plasmid DNA was required in order to induce the production of antibodies against plasmidencoded d-endotoxin (Pang, 1994), for example. However, high DNA concentrations might be achieved naturally if DNA is released by bacteria within the microenvironment created by microorganisms firmly attached to a eukaryotic host, as occurs in pathogenic or symbiotic relationships. In an experimental setting, the plasmid-mediated invasivity of Shigella flexneri was used to design a system for delivering DNA into the cytoplasm of mammalian cells (Sizemore et al., 1995). An attenuated strain of Shigella carrying the large plasmid necessary for bacterial invasion, and a eukaryotic expression vector, was used to infect cultured BHK cells. Infected cells overproduced the protein encoded by the vector (b-galactosidase); this expression was only possible in the cytoplasm of the eukaryotic cell but not within the bacterial cell. The system was designed to achieve DNA-mediated immunization; it proved to be capable of invading other kinds of cultured mammalian cells, and also of delivering active DNA to cells in live animals. The authors suggest a potential contribution of naturally-occurring bacterial DNA delivery to evolution. Plasmids and Symbiosis In addition to pathogenic invasion, colonization with plasmid-mediated attachment as the first step, may occur in a variety of other scenarios (Fig. 2). A mixed group of microorganisms inhabiting rumens of cows and other ruminants, for example, enables these animals to digest several of the plant components which they normally feed on. Important members of these heterogeneous populations are anaerobic bacteria and fungi that are capable of degrading cellulose and other polysaccharides of plant cell walls. The amino acid sequence of the fungus Neocallimastix patriciarum xylanase A shares significant homology with the bacterial G family of xylanases, one of which
. . ´ - . .
240
(a) Adhesive bacteria
Symbiosis
(d)
(b)
Bacteria-plant conjugation (Agrobacterium T-DNA transfer)
Endosymbiosis (c)
Pathogenesis F. 2. Attachment and association of bacteria with eukaryotic cells are often plasmid-mediated, and can result in lateral gene transfer. (a) The mere adherence of bacteria to larger organisms for symbiosis or comensalism, generates multi-cellular communities where gene transfer has been reported. More intimate associations such as endosymbiosis (b), are frequently encoded by plasmids and have resulted in gene transfer between participating organisms. Pathogenic invasion (c) is always preceded by bacterial adherence, which is plasmid-encoded in a wide variety of instances. Bacterial pathogenesis can result in, and has been enhanced by the lateral acquisition of genes. T-DNA transfer from Agrobacteria to plant cells (d) is a specialized, plasmid-mediated association, devoted to the trans-kingdom transfer of DNA, and bears an extraordinary similarity to bacterial conjugation.
is present in Ruminococcus flavefaciens, a rumendwelling bacterium (Gilbert et al., 1992). Horizontal gene transfer, within the shared habitat, is a likely explanation. The symbiotic association of microbes and higher organisms, where different cells come into very close contact with each other is, in some cases, plasmid-mediated. One example is the endosymbiosis
observed between Rhizobacteria and the root cells of leguminous plants. Bacteria enter the host cell to begin their endosymbiotic existence, losing their cell wall and becoming bacteroids dedicated to the reduction of molecular nitrogen to ammonia. Both the nitrogen fixation and endosymbiotic capabilities of these bacteria are encoded by plasmids (Long,
1989). Although definitive data is still lacking, it appears that one of the Rhizobacteria genes may have originated from a plant host. The gene in question encodes hemoglobin, a protein present in plants (called leghemoglobin) and animals, but mostly absent in bacteria. The intimate association of Rhizobacteria with plant cells could easily explain the eukaryotic to prokaryotic transfer of this gene (Doolittle et al., 1990). An interesting speculation arises when one considers the endosymbiotic origin of intracellular organelles as proposed by Lynn Margulis (Sagan (Margulis), 1967). It seems possible that just as several present day endosymbiotic interactions are mediated by plasmids, those that led to the origin of the first eukaryotic cell might also have been. It has even been proposed that the ‘‘sequestered’’ genomes of cellular organelles are remnants of ancestral plasmids, instead of chromosomes (Jacobs, 1991). Similarities have been reported between the sequences of ribosomal RNAs of plant mitochondria and the ribosomal RNAs of Agrobacteria and Rhizobacteria, two kinds of bacteria that are plasmid-mediated symbionts or parasites of present day plants (Singer & Berg, 1991). It is therefore conceivable that early endosymbiotic associations may have been mediated by plasmids. If that is the case, plasmid genes that encoded symbiotic relationships may have indirectly promoted the incorporation of bacterial sequences into the eukaryotic genome. It has been shown that movement of genetic information from mitochondria to chloroplasts and vice versa (Nakazono & Hirai, 1993; Stern & Palmer, 1984), and from the nucleus to mitochondria and vice versa, has occurred and, in fact, continues to occur in present day eukaryotic cells (Thorsness & Fox, 1990). Invasivity may constitute an evolutionary prelude to endosymbiosis. It has been proposed that predatory bacteria such as Vampirococcus, Bdellovibrio and Daptobacter may represent early stages of organelle evolution (Guerrero, 1991). Daptobacter, for example, is able to penetrate directly into the cytoplasm of the purple sulfur bacterium Chromatium. Several initial steps prior to the predatory attack, are markedly similar to plasmidencoded conjugation or conjugative-like transfer: the chemical detection of the prey, the chemotactic attraction of the predator by the prey, the attachment of the predator to the surface of the prey. Is bacterial predation plasmid-encoded? The acquisition of bacterial endosymbionts by Amoeba proteus constitutes an example of how endosymbiosis may start as a pathogenic invasion, and become interdependency after some time of
241
coexistence. The symbionts, called x-bacteria, are Gram-negative rods, bearing two kinds of plasmids on which bacterial infectivity seems to depend (Jeon, 1991). Other related members of the kingdom Protoctista are also known to interact symbiotically with bacteria. Interestingly, it has been observed that when exposed to stressful environments, several bacterial species survive by living within amoebae and ciliate protozoans (King et al., 1988; Leff et al., 1992). Entamoeba histolytica, which causes amoebiasis, may derive its pathogenic ability from its association with bacteria (Mirelman, 1987). And E. histolytica seems to have acquired the gene for a Fe-containing superoxide dismutase (one of only two known cases of Fe-SOD in eukaryotes) from bacteria (Smith et al., 1992). There are several other cases of symbiotic bacteria that can affect gene expression in their hosts: the bacterium Vibrio fischeri seems to influence the development of the squid Euprymna scolopes (McFall-Ngai & Ruby, 1991) and different bacterial endosymbionts induce parthenogenesis in a wide variety of insects (Stouthamer et al., 1993). Are these associations plasmid-mediated? Do these regulatory effects involve gene transfer? In this regard it is interesting to note that genes that affect plant development have been transferred from Agrobacterium to Nicotiana species by T-DNA conjugative-like transfer (Furner et al., 1986). Conclusions At this time, we can only speculate on the general impact plasmid-mediated intercellular interactions have had on gene flux and the evolution of living organisms; more experimental data are required to obtain a reasonable appreciation of the extent and limits of these effects. However, a large body of evidence indicates that microbial consortia are not uncommon and may be the rule rather than the exception (Finlay, 1990). If, as proposed in this paper, bacterial plasmids have played an important evolutionary role in bringing cells together, one might speculate that plasmids may have been involved in the formation of early multi-cellular colonies and, perhaps, even in the early origins of multi-cellular organisms. Bacterial cells are known to express multi-cellular behaviors in a number of instances (Budrene & Berg, 1991; Shapiro, 1991). Natural selection may foster the development of intimate microbial associations under a wide variety of conditions, as supported by the previously described observations indicating that many bacterial species are capable of surviving environmental stress by being ingested by protozoans (Smith et al., 1992; Winans,
242
. . ´ - . .
1991), and the widespread tendency of bacteria to form aggregates such as biofilms. In the last half of this century, with the introduction of antibiotics and, more recently, with the advent of recombinant DNA technology, humans have directly affected the evolution and the dispersal of the mechanisms underlying gene flux. In this way, and although indirectly, every ecosystem has been or will be modified by the new evolutionary dynamics of horizontal gene transfer. Many phylogenetic studies that use sequence homology analysis have provided abundant and accurate data, even though they generally assume only vertical transfer of genetic information. It is becoming increasingly evident, however, that even in eukaryotic organisms, instances of horizontal transfer do occur and it is thus essential to gain further knowledge on the nature and extent of such events. When Joshua Lederberg proposed the term plasmids to describe extrachromosomal genetic elements, he also highlighted their ‘‘endosymbiotic’’ nature (Lederberg, 1952), defining them as subcellular organisms. The evolution of these entities is clearly independent from that of their host organisms nevertheless, they can promote dramatic changes in the latter. Any way we choose to look at them, either as Lederberg, or as the most sophisticated manifestation of the ‘‘selfish gene’’, the participation of plasmids in the present and future evolution of bacteria and, apparently, of all living things, should be a source of continuous research and concern. We wish to thank Dr. Abigail Salyers for her very valuable comments on this manuscript.
REFERENCES A´-C, C. F. (1993). Origin, Evolution and Spread of Antibiotic Resistance Genes. Austin: R. G. Landes Co. A´-C, C. F. & C, M. E. (1992). Bacterial plasmids and gene flux. Cell 70, 189–199. A´-C, C. F. & C, M. E. (1993). Horizontal gene transfer. Am. Sci. 81, 332–341. B, J. B., G´, J. E. & F, S. (1993). Signal transduction in the mammalian cell during bacterial attachment and entry. Cell 73, 903–920. B, E. O. & B, H. C. (1991). Complex patterns formed by motile cells of Escherichia coli. Nature 349, 630–633. C, D. B. (1990). Movable genetic elements and antibiotic resistance in enterococci. Eur. J. Clin. Microbiol. Infect. Dis. 9, 90–102. C, D. B. (1993). Bacterial sex pheromone-induced plasmid transfer. Cell 73, 9–12. C, J. W., C, K.-J., G, G. G., L, T. I., N, J. C., D, M. & M, T. J. (1987). Bacterial biofilms in nature and disease. Annu. Rev. Microbiol. 41, 435–464.
C, P. (1994). Transfer of antibiotic resistance genes between gram-positive and gram-negative bacteria. Antimicrob. Agents Chemother. 38, 1447–1451. C, W. G. & M, K. C. (1990). Biofilms: a basis for an interdisciplinary approach. In: Biofilms. (Characklis, W. G. & Marshall, K. C. eds) pp. 3–15, New York: Wiley. D, R. F., F, D. F., A, K. L. & A, M. R. (1990). A naturally occurring horizontal transfer from a eukaryote to a prokaryote. J. Mol. Evol. 31, 383–388. F, B. J. (1990). Physiological ecology of free living protozoa. Adv. Microbial. Evol. 11, 1–35. F, J. C. & D, M. J. (1990). Plasmid transfer in the epilithon. In: Bacterial Genetics in Natural Environments. (Fry, J. C. & Day, M. J. eds) pp. 55–80, London: Chapman and Hall. F, I. J., J, G. A., A, R. M., G, D. J., G, M. P. & N, E. W. (1986). An Agrobacterium transformation in the evolution of the genus Nicotiania. Nature 319, 422–427. G, H. J., H, G. P., L, J. I., O, C. G. & X, G. P. (1992). Homologous catalytic domains in a rumen fungal xylanase: evidence for gene duplication and prokaryotic origin. Mol. Microbiol. 6, 2065–2072. G, R. (1991). Predation as prerequisite to organelle origin: Daptobacter as example. In: Symbiosis as a Source of Evolutionary Innovation. (Margulis, L. & Fester, R. eds) pp. 106–117, Cambridge: MIT Press. G, P. A. (1990). Virulence plasmids of Salmonella typhimurium and other salmonellae. Microb. Pathog. 8, 3–11. H, K. M., D, M. W., L, D. G., J, R. D. & B, A. N. (1991). Mechanism of phenolic activation of Agrobacterium virulence genes: development of a specific inhibitor of bacterial sensor/response systems. Proc. Natl. Acad. Sci. U.S.A. 88, 7854–7858. H, A. & B¨, C.-I. (1989). Crystal structure of chaperone protein PapD reveals an immunoglobulin fold. Nature 342, 248–251. I-I, K. (1989). Bacterial conjugation. In Gene Transfer in the Environment. (Levy, S. B. & Miller, R. V. eds) pp. 33–72, New York: McGraw-Hill Publishing Co. J, H. T. (1991). Structural similarities between a mitochondrially encoded polypeptide and a family of prokaryotic respiratory toxins involved in plasmid maintenance suggest a novel mechanism for the evolutionary maintenance of mitochondrial DNA. J. Mol. Evol. 32, 333–339. J, K. W. (1991). Bacterial evolution without speciation. In: Symbiosis as a Source of Evolutionary Innovation. (Margulis, L. & Fester, R. eds) pp. 118–131, Cambridge: MIT Press. K, C. H., S J, E. M., W, R. E. & P, K. G. (1988). Survival of coliforms and bacterial pathogens within protozoa during chlorination. Appl. Environ. Microbiol. 54, 3023–3033. L, J. (1952). Cell genetics and hereditary symbiosis. Physiol. Rev. 32, 403–430. L, L. E., MA, J. V. & S, L. J. (1992). Information spiralling: movement of bacteria and their genes in streams. Microb. Ecol. 24, 11–24. L, S. R. (1989). Rhizobium-legume nodulation: life together in the underground. Cell 56, 203–214. M, P.& D, J. (1991). Gene transfer between distantly related bacteria. Annu. Rev. Genet. 25, 147–171. MF-N, M. J. & R, E. G. (1991). Symbiont recognition and subsequent morphogenesis as early events in an animalbacterial mutualism. Science 254, 1491–1494. M, D. (1987). Ameba-bacterium relationships in amebiasis. Microbiol. Rev. 51, 272–284. N, M. & H, A. (1993). Identification of the entire set of transferred chloroplast DNA sequences in the mitochondrial genome of rice. Mol. Gen. Genet. 236, 341–346. P, A. S. D. (1994). Production of antibodies against Bacillus thuringiensis delta-endotoxin by injecting its plasmids. Biochem. Biophys. Res. Commun. 202, 1227–1234.
P, K. R., B, S. B. & F, S. K. (1993). Conjugation factor of Agrobacterium tumefaciens regulates Ti plasmid transfer by autoinduction. Nature 362, 448–450. R, P., S, R., H, L. L. & V, E. R. (1993). The sialidase superfamily and its spread by horizontal gene transfer. Mol. Microbiol. 9, 915–921. S (M), L. (1967). On the origin of mitosing cells. J. Theoret. Biol. 14, 225–274. S, D. J., O, O., S, G. S. & M, R. V. (1987). Potential for transduction of plasmids in a natural freshwater environment: effect of plasmid donor concentration and a natural microbial community on transduction in Pseudomonas aeruginosa. Appl. Environ. Microbiol. 53, 987–995. S, J. (1991). Multi-cellular behavior of bacteria. ASM News 57, 247–252. S, M. & B, P. (1991). Genes and Genomes. Mill Valley: University Science Books. S, D. R., B, A. A. & S, J. C. (1995). Attenuated Shigella as a DNA delivery vehicle for DNAmediated immunization. Science 270, 299–302. S, M. W., F, D.-F. & D, R. F. (1992). Evolution by acquisition: the case for horizontal gene transfers. Trends Biochem. Sci. 17, 489–493. S, D. B. & P, J. D. (1984). Extensive and widespread homologies between mitochondrial DNA and chloroplast DNA in plants. Proc. Natl. Acad. Sci. U.S.A. 81, 1946–1950.
243
S, R., B, J. A. J., L, R. F. & W, J. H. (1993). Molecular identification of microorganisms associated with parthenogenesis. Nature 361, 66–68. S, S. & O’C, D. (1990). TraT lipoprotein, a plasmid-specified mediator of interactions between gramnegative bacteria and their environment. Microbiol. Rev. 54, 331–341. T, P. E. & F, T. D. (1990). Escape of DNA from mitochondria to the nucleus in Saccharomyces cerevisiae. Nature 346, 376–379. W, G. A., S, H. R., MC, M. M., G, W., T, A., H, M. & R, B. (1990). Plasmid-encoded production of coli surface-associated antigen 1 (CS1) in a strain of Escherichia coli serotype 0139.H28. Microb. Pathog. 9, 1–11. W, S. C. (1991). An Agrobacterium two-component regulatory system for the detection of chemicals released from plant wounds. Mol. Microbiol. 5, 2345–2350. W, J. A., M, R. W., W, P., C, W., A, G., J, A. & F, P. L. (1990). Direct gene transfer into mouse muscle in vivo. Science 247, 1465–1468. W, J. A., W, P., A, G., J, S., J, A. & C, W. (1991). Conditions affecting direct gene transfer into rodent muscle in vivo. Biotechniques 11, 474–485. Z, L. & K, A. (1991). A diffusible compound can enhance conjugal transfer of the Ti plasmid transfer by autoinduction. J. Bacteriol. 173, 1867–1872.
A Possible Role for Plasmids in Mediating the Cell–cell Proximity Required for Gene Flux C F. A´-C† M. E. C‡ † Departamento de Microbiologı´ a, LUSARA. Apartado Postal 102-006, 08930, Me´xico, D.F., MEXICO ‡ Department of Surgical Research, Enders Building, Rm 1012, Children’s Hospital, 300 Longwood Ave., Boston, MA 02115, U.S.A. (Received on 10 July 1995, Accepted in revised form on 17 April 1996)
One of the major requirements for successful gene flux is a close proximity between participating organisms. In previous articles, we have proposed that plasmids act as powerful vehicles transporting genes collected by integration and transposition, mainly via the process of conjugation. However, in addition to conjugation, there are other processes, also mediated by plasmids, in which different cells come into very close contact with each other, such as symbiosis and the formation of multi-specific cellular communities. There is evidence that suggests that such intimate associations between cells may facilitate gene transfer events, even between distantly related organisms. Examples of symbiotic endosymbiotic, and parasitic associations provide evidence in support of the role of plasmids in bridging the genetic gap between species. In this purely theoretical article we attempt to conceptualize existing data on this subject, provide new insights and present testable predictions on how plasmids may facilitate gene flux by bringing cells together. 7 1996 Academic Press Limited
from bacteria to plant cells, seem to constitute the most frequent, versatile, and sophisticated mechanisms of horizontal gene transfer (Ama´bile-Cuevas & Chicurel, 1992, 1993). In addition to conjugation, plasmids enhance or increase the likelihood of other lateral transfer mechanisms. The transfer of a gene by transformation is more successful and has more profound consequences if it involves plasmids as carriers. Likewise, in cases where plasmids are enclosed in phage particles, genes contained in the plasmid are generally more efficiently transferred than genes contained in non-plasmid molecules. A striking example of the effects of plasmids and plasmid-mediated gene transfer can be found in the origin, evolution and spread of antibiotic resistance genes (for review see Ama´bile-Cuevas, 1993; Ama´bile-Cuevas & Chicurel, 1992, 1993). But plasmids can act to facilitate gene flux in additional ways. In this article we discuss how plasmids may have contributed to gene transfer by bringing different cells together.
Introduction Horizontal gene transfer is the transmission of genetic information from one independent and fully developed organism to another. In horizontal gene transfer organisms can acquire novel characteristics, through the incorporation of foreign genetic material, that are directly inherited by future generations. For many years horizontal gene transfer was regarded as a peculiarity of bacteria. However, a large amount of evidence has recently accumulated indicating that the extent of such gene flux is very large, including gene transfers between distantly related species, as well as, in some cases, even eukaryotes. Plasmids and Gene Flux Conjugation between bacteria, between bacteria and yeast, and the conjugative-like T-DNA transfer † Author to whom correspondence should be addressed. 0022–5193/96/150237 + 07 $18.00/0
237
7 1996 Academic Press Limited
238
. . ´ - . .
One of the major requirements for successful gene transfer is a close proximity between the participating organisms. Cell–cell contact is required for all conjugative or conjugative-like processes. However, even for transformation, where cell–cell contact is not necessary, proximity is essential. The lifetime of DNA outside the cell is relatively short, and thus cells to be transformed must be close to the DNA source at the time of its release. Transduction, wherein DNA molecules to be transferred are enclosed in viral particles, may occur between cells widely separated in space. However, the inactivation of phages often happens shortly after their release, especially in bacterially populated media (Saye et al., 1987), and transduction is further limited by its usually narrow spectrum of specific hosts and recipients. Conjugative plasmids encode for the machinery that bring the donor and recipient cells together. In Gram-negative bacteria this is by means of pili, or through aggregation substances induced by ‘‘sex’’ pheromones in some Gram-positive cocci, plasmidencoded mechanisms are responsible for the attachment of the participating bacteria (Clewell, 1993; Ippen-Ihler, 1989). The primary role of such adhesion mechanisms could have been the attachment of invading bacteria to the cells of infected organisms, but this attachment may also have facilitated attachment to other bacterial cells. The microfibrillar surface structures mediating adherence induced by pheromones in Enterococcus faecalis, for example, also seem to be capable of mediating bacterial adherence to pig renal tubular cells (Clewell, 1993). The relationship between conjugation and pathogenicity is not restricted to adherence determinants: pheromones themselves may play a role in virulence (Clewell, 1990), as does the TraT lipoprotein, a Gram-negative plasmid exclusion factor (Sukupolvi & O’Connor, 1990). It is very likely that genes encoding for adhesive mechanisms were ‘‘recruited’’, through transposition or recombination, by plasmids, thus increasing their transfer efficiency and becoming ‘‘conjugative’’ plasmids. Pathogenic invasion, as conjugation, is preceded by the attachment of the involved cells. And this adherence may also enhance the possibilities of horizontal gene transfer. Plasmid-mediated adherence often mediates the aggregation of different microbial cells. These consortia, some of them called biofilms (Costerton et al., 1987), most probably constitute excellent environments for gene exchange. Biofilms can be composed of a variety of microorganisms, including bacteria, algae, protozoa and fungi (Characklis & Marshall, 1990), and it has been shown that, at least
in streams, gene transfer by conjugation occurs at high rates. In biofilms in the West River Taff, for example, transfer frequencies ranged from 1 to 10−6 transconjugants per member of recipient populations (Fry & Day, 1990). It is important to recall that conjugation is possible not only between distantly related bacteria (Mazodier & Davies, 1991; Courvalin, 1994), but also between bacteria and eukaryotic cells (Ama´bile-Cuevas & Chicurel, 1992). Figure 1 summarizes the known instances of bacterial cell–cell adherence that may result in increased gene flux. Bacteria produce adhesins or more complex structures, such as pili and curli, in order to attach to different surfaces, including larger cells. The production of such molecules and structures is often plasmid-mediated (Gulig, 1990; Willshaw et al.,
Adhesive bacteria (a)
(c) Biofilms Predation (b)
Conjugation
Endosymbiosis? F. 1. Bacteria attach to each other in a number of ways. (a) Multi-cellular, multi-specific bacterial associations, called biofilms, adhere to a variety of surfaces. Within these associations, the likelihood of gene transfer is increased. Biofilm formation is sometimes mediated by plasmid-encoded adhesins. (b) The cell-to-cell contact required by conjugation, must start with the attachment of the participating bacteria. This is achieved by very diverse mechanisms, such as the production of pili or the secretion of adhesins. Almost all conjugative processes are plasmid-encoded, and conjugation is the most powerful mechanism of horizontal gene transfer. (c) Bacterial predation must also start with the attachment of the predatory bacterium to its prey. Whether this process is plasmid-encoded, and whether it represents an early evolutionary step in the development of endosymbiosis and the origin of organelles, remains unknown.
1990). Pathogenic colonization, which starts with this attachment, brings into extremely close proximity bacterial and larger eukaryotic cells. As a result, gene transfer may occur. Genes that may have traveled this way include the gene encoding PapD, a chaperone protein that mediates the assembly of pili in uropathogenic E. coli, whose folding and amino acid sequence are quite similar to the human lymphocyte differentiation antigen Leu-1/CD5 (Holmgren & Bra¨nden, 1989). In addition, there is strong evidence indicating that E. coli acquired a copy of the glyceraldehyde-3-phosphate dehydrogenase gene from a eukaryotic organism. It is possible that this transfer occurred in the gut of an ancestral eukaryotic donor (Smith et al., 1992). The presence of sialidases among bacteria that exist as animal commensals or pathogens may be another case of acquisition of a eukaryotic gene by prokaryotes that live in close association with animals (Roggentin et al., 1993). Again, the laterally transferred gene is itself involved in bacterial virulence. A final, recently reviewed example (Bliska et al., 1993), is provided by the enteropathogen Yersinia. Adherence of these bacteria to mammalian cells is mediated by the YadA adhesin, encoded by the plasmid that confers virulence to Yersinia pseudotuberculosis and Y. enterocolitica. This plasmid also bears other pathogenic determinants, whose products are the Yersinia outer membrane proteins, or Yops. Several of these proteins share homology with eukaryotic signal transduction proteins, such as a tyrosine phosphatase, the GPIba receptor, and a ser/thr kinase. Yops, on the other hand, are not homologous to other known bacterial proteins (Bliska et al., 1993). This suggests that the pathogenic potential of Yersinia, originally provided by a plasmid-encoded adhesin, has been enhanced by the acquisition of eukaryotic genes, whose products alter signal transduction mechanisms within host cells. Gene transfer itself can constitute a form of plasmid-mediated virulence. The infection of plant cells by Agrobacterium for example, involves direct transfer of bacterial DNA to the host plant cell. This parasitic relationship is mediated by Ti plasmids. Agrobacteria respond to compounds released by wounded plant cells, activating vir genes that are responsible for cell adhesion and DNA transfer (Hess et al., 1991; Winans, 1991). This response is encoded by Ti plasmid genes. Furthermore, Ti plasmids also encode an autoinducible system for conjugal Ti plasmid transfer between Agrobacteria (Piper et al., 1993; Zhang & Kerr, 1991). Thus, two distinct systems for increasing cell proximity and DNA transfer are encoded within single Ti plasmids.
239
So far, there are very few examples of gene transfer from bacteria to mammalian cells. This may be a consequence of the more limited ability of eukaryotic cells in assimilating naked DNA and incorporating it into their nuclei for functional expression, in contrast to the relative ease with which bacterial transformation can occur. However, the acquisition and expression of DNA sequences by mammalian cells in vivo (Wolff et al., 1990) suggests that bacterialmammalian gene transfer might not be altogether impossible. Experimental models require very high amounts of intramuscularly administered DNA (Wolff et al., 1991). An injection of 100 mg of Bacillus thuringiensis plasmid DNA was required in order to induce the production of antibodies against plasmidencoded d-endotoxin (Pang, 1994), for example. However, high DNA concentrations might be achieved naturally if DNA is released by bacteria within the microenvironment created by microorganisms firmly attached to a eukaryotic host, as occurs in pathogenic or symbiotic relationships. In an experimental setting, the plasmid-mediated invasivity of Shigella flexneri was used to design a system for delivering DNA into the cytoplasm of mammalian cells (Sizemore et al., 1995). An attenuated strain of Shigella carrying the large plasmid necessary for bacterial invasion, and a eukaryotic expression vector, was used to infect cultured BHK cells. Infected cells overproduced the protein encoded by the vector (b-galactosidase); this expression was only possible in the cytoplasm of the eukaryotic cell but not within the bacterial cell. The system was designed to achieve DNA-mediated immunization; it proved to be capable of invading other kinds of cultured mammalian cells, and also of delivering active DNA to cells in live animals. The authors suggest a potential contribution of naturally-occurring bacterial DNA delivery to evolution. Plasmids and Symbiosis In addition to pathogenic invasion, colonization with plasmid-mediated attachment as the first step, may occur in a variety of other scenarios (Fig. 2). A mixed group of microorganisms inhabiting rumens of cows and other ruminants, for example, enables these animals to digest several of the plant components which they normally feed on. Important members of these heterogeneous populations are anaerobic bacteria and fungi that are capable of degrading cellulose and other polysaccharides of plant cell walls. The amino acid sequence of the fungus Neocallimastix patriciarum xylanase A shares significant homology with the bacterial G family of xylanases, one of which
. . ´ - . .
240
(a) Adhesive bacteria
Symbiosis
(d)
(b)
Bacteria-plant conjugation (Agrobacterium T-DNA transfer)
Endosymbiosis (c)
Pathogenesis F. 2. Attachment and association of bacteria with eukaryotic cells are often plasmid-mediated, and can result in lateral gene transfer. (a) The mere adherence of bacteria to larger organisms for symbiosis or comensalism, generates multi-cellular communities where gene transfer has been reported. More intimate associations such as endosymbiosis (b), are frequently encoded by plasmids and have resulted in gene transfer between participating organisms. Pathogenic invasion (c) is always preceded by bacterial adherence, which is plasmid-encoded in a wide variety of instances. Bacterial pathogenesis can result in, and has been enhanced by the lateral acquisition of genes. T-DNA transfer from Agrobacteria to plant cells (d) is a specialized, plasmid-mediated association, devoted to the trans-kingdom transfer of DNA, and bears an extraordinary similarity to bacterial conjugation.
is present in Ruminococcus flavefaciens, a rumendwelling bacterium (Gilbert et al., 1992). Horizontal gene transfer, within the shared habitat, is a likely explanation. The symbiotic association of microbes and higher organisms, where different cells come into very close contact with each other is, in some cases, plasmid-mediated. One example is the endosymbiosis
observed between Rhizobacteria and the root cells of leguminous plants. Bacteria enter the host cell to begin their endosymbiotic existence, losing their cell wall and becoming bacteroids dedicated to the reduction of molecular nitrogen to ammonia. Both the nitrogen fixation and endosymbiotic capabilities of these bacteria are encoded by plasmids (Long,
1989). Although definitive data is still lacking, it appears that one of the Rhizobacteria genes may have originated from a plant host. The gene in question encodes hemoglobin, a protein present in plants (called leghemoglobin) and animals, but mostly absent in bacteria. The intimate association of Rhizobacteria with plant cells could easily explain the eukaryotic to prokaryotic transfer of this gene (Doolittle et al., 1990). An interesting speculation arises when one considers the endosymbiotic origin of intracellular organelles as proposed by Lynn Margulis (Sagan (Margulis), 1967). It seems possible that just as several present day endosymbiotic interactions are mediated by plasmids, those that led to the origin of the first eukaryotic cell might also have been. It has even been proposed that the ‘‘sequestered’’ genomes of cellular organelles are remnants of ancestral plasmids, instead of chromosomes (Jacobs, 1991). Similarities have been reported between the sequences of ribosomal RNAs of plant mitochondria and the ribosomal RNAs of Agrobacteria and Rhizobacteria, two kinds of bacteria that are plasmid-mediated symbionts or parasites of present day plants (Singer & Berg, 1991). It is therefore conceivable that early endosymbiotic associations may have been mediated by plasmids. If that is the case, plasmid genes that encoded symbiotic relationships may have indirectly promoted the incorporation of bacterial sequences into the eukaryotic genome. It has been shown that movement of genetic information from mitochondria to chloroplasts and vice versa (Nakazono & Hirai, 1993; Stern & Palmer, 1984), and from the nucleus to mitochondria and vice versa, has occurred and, in fact, continues to occur in present day eukaryotic cells (Thorsness & Fox, 1990). Invasivity may constitute an evolutionary prelude to endosymbiosis. It has been proposed that predatory bacteria such as Vampirococcus, Bdellovibrio and Daptobacter may represent early stages of organelle evolution (Guerrero, 1991). Daptobacter, for example, is able to penetrate directly into the cytoplasm of the purple sulfur bacterium Chromatium. Several initial steps prior to the predatory attack, are markedly similar to plasmidencoded conjugation or conjugative-like transfer: the chemical detection of the prey, the chemotactic attraction of the predator by the prey, the attachment of the predator to the surface of the prey. Is bacterial predation plasmid-encoded? The acquisition of bacterial endosymbionts by Amoeba proteus constitutes an example of how endosymbiosis may start as a pathogenic invasion, and become interdependency after some time of
241
coexistence. The symbionts, called x-bacteria, are Gram-negative rods, bearing two kinds of plasmids on which bacterial infectivity seems to depend (Jeon, 1991). Other related members of the kingdom Protoctista are also known to interact symbiotically with bacteria. Interestingly, it has been observed that when exposed to stressful environments, several bacterial species survive by living within amoebae and ciliate protozoans (King et al., 1988; Leff et al., 1992). Entamoeba histolytica, which causes amoebiasis, may derive its pathogenic ability from its association with bacteria (Mirelman, 1987). And E. histolytica seems to have acquired the gene for a Fe-containing superoxide dismutase (one of only two known cases of Fe-SOD in eukaryotes) from bacteria (Smith et al., 1992). There are several other cases of symbiotic bacteria that can affect gene expression in their hosts: the bacterium Vibrio fischeri seems to influence the development of the squid Euprymna scolopes (McFall-Ngai & Ruby, 1991) and different bacterial endosymbionts induce parthenogenesis in a wide variety of insects (Stouthamer et al., 1993). Are these associations plasmid-mediated? Do these regulatory effects involve gene transfer? In this regard it is interesting to note that genes that affect plant development have been transferred from Agrobacterium to Nicotiana species by T-DNA conjugative-like transfer (Furner et al., 1986). Conclusions At this time, we can only speculate on the general impact plasmid-mediated intercellular interactions have had on gene flux and the evolution of living organisms; more experimental data are required to obtain a reasonable appreciation of the extent and limits of these effects. However, a large body of evidence indicates that microbial consortia are not uncommon and may be the rule rather than the exception (Finlay, 1990). If, as proposed in this paper, bacterial plasmids have played an important evolutionary role in bringing cells together, one might speculate that plasmids may have been involved in the formation of early multi-cellular colonies and, perhaps, even in the early origins of multi-cellular organisms. Bacterial cells are known to express multi-cellular behaviors in a number of instances (Budrene & Berg, 1991; Shapiro, 1991). Natural selection may foster the development of intimate microbial associations under a wide variety of conditions, as supported by the previously described observations indicating that many bacterial species are capable of surviving environmental stress by being ingested by protozoans (Smith et al., 1992; Winans,
242
. . ´ - . .
1991), and the widespread tendency of bacteria to form aggregates such as biofilms. In the last half of this century, with the introduction of antibiotics and, more recently, with the advent of recombinant DNA technology, humans have directly affected the evolution and the dispersal of the mechanisms underlying gene flux. In this way, and although indirectly, every ecosystem has been or will be modified by the new evolutionary dynamics of horizontal gene transfer. Many phylogenetic studies that use sequence homology analysis have provided abundant and accurate data, even though they generally assume only vertical transfer of genetic information. It is becoming increasingly evident, however, that even in eukaryotic organisms, instances of horizontal transfer do occur and it is thus essential to gain further knowledge on the nature and extent of such events. When Joshua Lederberg proposed the term plasmids to describe extrachromosomal genetic elements, he also highlighted their ‘‘endosymbiotic’’ nature (Lederberg, 1952), defining them as subcellular organisms. The evolution of these entities is clearly independent from that of their host organisms nevertheless, they can promote dramatic changes in the latter. Any way we choose to look at them, either as Lederberg, or as the most sophisticated manifestation of the ‘‘selfish gene’’, the participation of plasmids in the present and future evolution of bacteria and, apparently, of all living things, should be a source of continuous research and concern. We wish to thank Dr. Abigail Salyers for her very valuable comments on this manuscript.
REFERENCES A´-C, C. F. (1993). Origin, Evolution and Spread of Antibiotic Resistance Genes. Austin: R. G. Landes Co. A´-C, C. F. & C, M. E. (1992). Bacterial plasmids and gene flux. Cell 70, 189–199. A´-C, C. F. & C, M. E. (1993). Horizontal gene transfer. Am. Sci. 81, 332–341. B, J. B., G´, J. E. & F, S. (1993). Signal transduction in the mammalian cell during bacterial attachment and entry. Cell 73, 903–920. B, E. O. & B, H. C. (1991). Complex patterns formed by motile cells of Escherichia coli. Nature 349, 630–633. C, D. B. (1990). Movable genetic elements and antibiotic resistance in enterococci. Eur. J. Clin. Microbiol. Infect. Dis. 9, 90–102. C, D. B. (1993). Bacterial sex pheromone-induced plasmid transfer. Cell 73, 9–12. C, J. W., C, K.-J., G, G. G., L, T. I., N, J. C., D, M. & M, T. J. (1987). Bacterial biofilms in nature and disease. Annu. Rev. Microbiol. 41, 435–464.
C, P. (1994). Transfer of antibiotic resistance genes between gram-positive and gram-negative bacteria. Antimicrob. Agents Chemother. 38, 1447–1451. C, W. G. & M, K. C. (1990). Biofilms: a basis for an interdisciplinary approach. In: Biofilms. (Characklis, W. G. & Marshall, K. C. eds) pp. 3–15, New York: Wiley. D, R. F., F, D. F., A, K. L. & A, M. R. (1990). A naturally occurring horizontal transfer from a eukaryote to a prokaryote. J. Mol. Evol. 31, 383–388. F, B. J. (1990). Physiological ecology of free living protozoa. Adv. Microbial. Evol. 11, 1–35. F, J. C. & D, M. J. (1990). Plasmid transfer in the epilithon. In: Bacterial Genetics in Natural Environments. (Fry, J. C. & Day, M. J. eds) pp. 55–80, London: Chapman and Hall. F, I. J., J, G. A., A, R. M., G, D. J., G, M. P. & N, E. W. (1986). An Agrobacterium transformation in the evolution of the genus Nicotiania. Nature 319, 422–427. G, H. J., H, G. P., L, J. I., O, C. G. & X, G. P. (1992). Homologous catalytic domains in a rumen fungal xylanase: evidence for gene duplication and prokaryotic origin. Mol. Microbiol. 6, 2065–2072. G, R. (1991). Predation as prerequisite to organelle origin: Daptobacter as example. In: Symbiosis as a Source of Evolutionary Innovation. (Margulis, L. & Fester, R. eds) pp. 106–117, Cambridge: MIT Press. G, P. A. (1990). Virulence plasmids of Salmonella typhimurium and other salmonellae. Microb. Pathog. 8, 3–11. H, K. M., D, M. W., L, D. G., J, R. D. & B, A. N. (1991). Mechanism of phenolic activation of Agrobacterium virulence genes: development of a specific inhibitor of bacterial sensor/response systems. Proc. Natl. Acad. Sci. U.S.A. 88, 7854–7858. H, A. & B¨, C.-I. (1989). Crystal structure of chaperone protein PapD reveals an immunoglobulin fold. Nature 342, 248–251. I-I, K. (1989). Bacterial conjugation. In Gene Transfer in the Environment. (Levy, S. B. & Miller, R. V. eds) pp. 33–72, New York: McGraw-Hill Publishing Co. J, H. T. (1991). Structural similarities between a mitochondrially encoded polypeptide and a family of prokaryotic respiratory toxins involved in plasmid maintenance suggest a novel mechanism for the evolutionary maintenance of mitochondrial DNA. J. Mol. Evol. 32, 333–339. J, K. W. (1991). Bacterial evolution without speciation. In: Symbiosis as a Source of Evolutionary Innovation. (Margulis, L. & Fester, R. eds) pp. 118–131, Cambridge: MIT Press. K, C. H., S J, E. M., W, R. E. & P, K. G. (1988). Survival of coliforms and bacterial pathogens within protozoa during chlorination. Appl. Environ. Microbiol. 54, 3023–3033. L, J. (1952). Cell genetics and hereditary symbiosis. Physiol. Rev. 32, 403–430. L, L. E., MA, J. V. & S, L. J. (1992). Information spiralling: movement of bacteria and their genes in streams. Microb. Ecol. 24, 11–24. L, S. R. (1989). Rhizobium-legume nodulation: life together in the underground. Cell 56, 203–214. M, P.& D, J. (1991). Gene transfer between distantly related bacteria. Annu. Rev. Genet. 25, 147–171. MF-N, M. J. & R, E. G. (1991). Symbiont recognition and subsequent morphogenesis as early events in an animalbacterial mutualism. Science 254, 1491–1494. M, D. (1987). Ameba-bacterium relationships in amebiasis. Microbiol. Rev. 51, 272–284. N, M. & H, A. (1993). Identification of the entire set of transferred chloroplast DNA sequences in the mitochondrial genome of rice. Mol. Gen. Genet. 236, 341–346. P, A. S. D. (1994). Production of antibodies against Bacillus thuringiensis delta-endotoxin by injecting its plasmids. Biochem. Biophys. Res. Commun. 202, 1227–1234.
P, K. R., B, S. B. & F, S. K. (1993). Conjugation factor of Agrobacterium tumefaciens regulates Ti plasmid transfer by autoinduction. Nature 362, 448–450. R, P., S, R., H, L. L. & V, E. R. (1993). The sialidase superfamily and its spread by horizontal gene transfer. Mol. Microbiol. 9, 915–921. S (M), L. (1967). On the origin of mitosing cells. J. Theoret. Biol. 14, 225–274. S, D. J., O, O., S, G. S. & M, R. V. (1987). Potential for transduction of plasmids in a natural freshwater environment: effect of plasmid donor concentration and a natural microbial community on transduction in Pseudomonas aeruginosa. Appl. Environ. Microbiol. 53, 987–995. S, J. (1991). Multi-cellular behavior of bacteria. ASM News 57, 247–252. S, M. & B, P. (1991). Genes and Genomes. Mill Valley: University Science Books. S, D. R., B, A. A. & S, J. C. (1995). Attenuated Shigella as a DNA delivery vehicle for DNAmediated immunization. Science 270, 299–302. S, M. W., F, D.-F. & D, R. F. (1992). Evolution by acquisition: the case for horizontal gene transfers. Trends Biochem. Sci. 17, 489–493. S, D. B. & P, J. D. (1984). Extensive and widespread homologies between mitochondrial DNA and chloroplast DNA in plants. Proc. Natl. Acad. Sci. U.S.A. 81, 1946–1950.
243
S, R., B, J. A. J., L, R. F. & W, J. H. (1993). Molecular identification of microorganisms associated with parthenogenesis. Nature 361, 66–68. S, S. & O’C, D. (1990). TraT lipoprotein, a plasmid-specified mediator of interactions between gramnegative bacteria and their environment. Microbiol. Rev. 54, 331–341. T, P. E. & F, T. D. (1990). Escape of DNA from mitochondria to the nucleus in Saccharomyces cerevisiae. Nature 346, 376–379. W, G. A., S, H. R., MC, M. M., G, W., T, A., H, M. & R, B. (1990). Plasmid-encoded production of coli surface-associated antigen 1 (CS1) in a strain of Escherichia coli serotype 0139.H28. Microb. Pathog. 9, 1–11. W, S. C. (1991). An Agrobacterium two-component regulatory system for the detection of chemicals released from plant wounds. Mol. Microbiol. 5, 2345–2350. W, J. A., M, R. W., W, P., C, W., A, G., J, A. & F, P. L. (1990). Direct gene transfer into mouse muscle in vivo. Science 247, 1465–1468. W, J. A., W, P., A, G., J, S., J, A. & C, W. (1991). Conditions affecting direct gene transfer into rodent muscle in vivo. Biotechniques 11, 474–485. Z, L. & K, A. (1991). A diffusible compound can enhance conjugal transfer of the Ti plasmid transfer by autoinduction. J. Bacteriol. 173, 1867–1872.