Rev. Biol. Trop.,45(2): 753-771,1997

SPECIAL PAPER

In search of a bacterial spedes definition

EdgardoMoreno Programa de Investigación en Enfermedades Tropicales (PIET), Escuela de Medicina Veterinaria,Universidad Nacional, Heredia,Costa Rica. Fax: (506) 2381298,E-mail: [email protected]

(Received14-IV-1996. Corrected 9-X-1996. Accepted30-X-1996.)

Abstraet: The bacterial concept was ¡;¡xamined within the framework of plant and animal associated a-2 ,taking into consideration the phylogenetic, taxonomic and biological approaches as well as the microbiologists' perception. The virtue of the phylogenetic approach is that it gi.ves an evolutionary perspective of the bacterial lineage; however the methods used possess low resolution for defining species located at the terminal branches of lhe phylogenetic trees. The merit of the taxonomic approach.is that species are defined on the basis of multiple characteristicS allowing high resolution at the terminal branches of dendograms; its disadvantage is the inaccuracy in the earliet nodes. Onan individual level, the qualitative biological characteristics used for the definition of species frequently reveal shortcomings because many of these properties are the result of coevolution, parallel evolution or the horizontal transfer of genes. Nevertheless, when considered together with the phylogenetic and taxonomic approaches,important uncertainties are discovered: these must be weighed if a practical definition of bacterial species is conceived. The microbiologi.sts' perception is thecriterion expressed by a group of sponsors who, basedon scientific and practical grounds, propose a new bacterial species. The success of this new .proposal is measured by its widespread acceptance and itspermanence. A difficult pr(jblem con cerned with definingbacterial species is how to distinguishif tbey are independent evolutionary units or if they are reticulateevolutionary units. In the first case the inherenceis verticallytransmilted as a result of binary fission andclonal expansiono Thismay be the case of sorne animal ceIl associated bacteriain which recombinationappears to be precIuded or exceptional. In the second case adaptive changesoccurring within anindividual canbe horizontally transferredto many or a11 group members. This seems to be the conditión of many intestinal and plant associated . Genetic drift and speciationin c10nal bacteria will depend a1most .exclusively (jn mutation and intemal genetic rearrangement processes, whereas speciation in reticulate bacteria will depend not only on the�e processes bút in their genetic rre o to interactions wíth other bacterial strains. This uncertainty, wruch co sp nds . the evolutionary process,is at the same time one of the key factors in defininga bacterial species.

Key words: Bacterial speci¡;:s,Proteobacteria, review, Brucella, , Ochrobactrum, , Agrobacterium, Rhizobium, Phyllobacterium, . .

"They, and every beast after his kind, and calculate how to accommoqate all species of all the cattle after theirkind, and every animals in afinite ark size (300 cubits long x creeping thingthat creepeth upon theearth . 50 cubits wíde x 30 cubits high) but also due to after his kind, and everyfowl after his kind, the enonnous and exhausting duty of sorting every bird of every sort,GENESIS 7: 14 And and identifying in a brief death line (two theywent in unto N04h into the ark, two and months and seventeen days) thedifferent two ·of all flesh, wherein is the breath of species oí animals, their sexes, their feeding life.GENESIS 7: 15 And they that went in, went habits and aboye aH, to estimate theirchances in male and female of al! flesh, as God had for survival. Conveniently for Noah, his work commanded him: and the Lord shut him in systematics concluded after the ark's door in.GENESIS 7: 16". was closed. However, Noah's task is far from The job assigned to the mythical Noah was complete and the question concerning the monumental, not only because he had to 754 REVISTA DE BIOLOGIA TROPICAL correct identificatíon of species and where to Since the publicatíon of Darwin's locate them is still a major puzzle. masterpiece "On the Origin of Species" in The proximity of an ecologicaJ catac1ysm, 1859, the definitíon of species concerning where human beings actively participate in the animals and plants has been extensively extínction of a considerable number of species, debated by a great number of scientists has alerted many instances in the relevance of (O'Hara 1993 and 1994, Avise 1994,). In sampling and c1assification of living creatures. contrast, the bacterial world was not subjected Due to the fact that plants and animals are to the same detailed analysis for more than one more conspicuous and in consequence more century. This unfortunate fact was mainly due unveiled to our prejudices regarding beauty, to the absence of data sustaining the ugliness and usefulness, most of the new evolutíonary history of bacteria (Woese 1987). "Noah's arks" have been devoted to the census The advent, during the last two decades, of of these biodiversities with little or marginal concise phylogenetic and taxonomic studies attentíon to the microbial world. If we consider based on molecular sequences and structures that the primordial living systems (and very has definitívely contributed to incorporating probably the ultimate) were assembled almost the bacterial world within the constraints of exc1usively by bacteria or bacteria-like evolutionary studies (Woese 1994). As with organisms,and that practically al! the genetics other organisms,the distinctiveness of bacteria as well as the synthetic, catabolic and requires an idiosyncratic analysis for the regulatory biochemical reactions were concept of species. However, we must developed by these organisms and are still distinguish between the various bacterial concentrated in bacteria,then it is obvious that species concepts and the different character we must give to this world its place in our sets used to define species in that the former "modern arks". Furthermore, by considering refers to a phílosophical position and the latter the simple but relevant fact that practicalIy al! refers to the characteristics used by the the eukaryotic systems studied possess or cIassification methods. Furthermore, the interact with at least one unique eubacteria different systems of classification force us to a (including endosymbíonts), it is reasonable to particular model for representing the propose that the bacterial world is not only the dispersion of species which in most cases most abundant but the most diverse of al!. overIap with the different definitions (O' Hara Shortcomings on the isolation and culture 1993). In order to obtain a correct perspective of a great number of bacteria (mainly the on this, 1 will fírst of all discuss the bacterial intracellular and the "extremist") have species concept within the framework of the restricted the correct quantification and different species views which at the same time characterisation of these organisms. Even are related to the different character sets used though this problems have been partially to distinguish the various species. Up to the solved by the advent of recent molecular present, three different views on that which techniques devoted to the amplification and constitutes a species and two perceptions sequencing of genes from bacteria which are dominate our present knowledge: the difficult to isolate or to cultivate under phylogenetic species concept championed by controlled conditíons, the truth is that Cracraft (1983),the taxonomic specíes concept classification of bacteria into species is a (Staley and Krieg 1984), the biological species constant drawback (Krieg, and Holt. 1984). concept (Dobzhansky1937), the Today it is clear (at least for sorne of us) that microbiologists' perception (Pitt 1994, Sneath the success of all conservation programs, 1984 b) and the fixed specíes' perception biodiversíty projects, bacterial collections, reviewed by Schadewald (1986). The three microbial ecologists, medical microbiologists, concepts have strengths and weaknesses, each plant pathologists, food microbiologists, emphasises different aspects of the epidemiologists and geneticists depends on the evolutionary process and are within the correct ídentífication of those extant units boundaries of science. Because the taxonomic known as bacterial species for which and phylogenetic concepts are based on protection, compilation, conservation, control, numerical analysis (Sneath 1984a, Woese destructionor selection is being sought. 1994,Williams et al. 1995), they are the most Moreno: In search of a bacterial species defmition 755 distinctly understood. The biological species THEPHYLOGENETIC APPROACH concept may be considered the ultimate definition, but due to the complexity involved A phylogenetic species may be understood in understanding the biology of a bacteria as 'the smallest diagnosable cluster of regarded as qualitative characters, it has been individual organisms within which there is a relegated to last among the three scientific parental pattem of ancestry and descendent" views (Krieg and Holt 1984). In tbis essay I (Cracraft 1983). This concept has the attempt to rescue the value of the biological advantage that species are defined as the species concept in the framework of the terminal organism in a lineage which has an bacterial population structure. Other ancestor common to other terminal organisms, definitions such as the concordance (Avise and represented by the node joining of their two Ball 1990), recognition (Paterson 1985), branches; thus avoiding the meaningless cohesion (Templeton 1989) and evolutionary elassical ranked classification into kingdoms, (Simpson 1951) species concepts, wbich have phyla, classes,and families. In the light of tbis been used for the definitionof animal and plant definition,debates conceming bacterial ranks species, are either ineluded within the aboye the genus level are trivial and phylogenetic, taxonomic and biological unimportant (O'Hara 1994). A common concepts or are not relevant within the context problem with this concept involves of a practical bacterial species definition. The determining how monophyly is to be two perceptions are anthropocentric and non recognised and how to distinguish gene scientific in nature. Nevertheless, the phylogenies frompedigrees (Avise 1994). microbiologists' perception must be taken into The phylogenetic species concept is account since sooner or later Ua elassification commonly based on two different types of data that is of little use to the microbiologist, no generated by different techniques: the matter how fine a scheme or who devised it, comparison of characters between fossils and will be ignored or significantly modified" the extant species and the comparison of (Staley and Krieg 1984). The fixed species' molecular sequences of the different perception may be the most popular view organisms. Since the availability of bacterial among laymen. Since this perception is not fossils is scarce (Ochman and Wilson 1987), amenable to scientific treatment or open to the evolutionary history of these discussion outside the limits of creationism microorganisms has been mainly reconstructed (Schadewald 1986), I will not consider it in on the basis of molecular sequences which are furtherdiscussions. best represented in the form of evolutionary On attempting to define a concept of a trees or bushes (Olsen et al. 1994). Questions, bacterial species, different and sometimes such as if a gene can represent an organism contrasting views must be considered. and which bacterial genes best represent the However,the concepts should be limited to the evolutionary history of bacteria, have been a constraints of what Ís believed to be a natural matter of dispute among evolutionists (Forterre system of elassification. In short,the bacterial et al. 1993). The weight of evidence generated species must be catalogued in a system during the last few years strongly suggests that commensurate with the currently accepted some genes can in fact represent the phylogeny theory of evolution. This ineludes a concept of bacteria (Woese 1987, Olsen and Woese that takes into account the important role of the 1993, Olsen et al. 1994). The search has been reproductive isolation of organisms and the focused on genes that perform central horizontal· transfer of genes as different and functions, that is, "the entities most tightly complementary modes of bacterial speciation interwoven into the fabric of the cell itself' (Syvanen 1994). Understanding of this is (Olsen and Woese 1993). In this sense, expected to have a profound influence on the bacteria phylogenies based on the sequences of futureof bacterialsystematics. conserved molecules such as ribosomal RNAs, chaperone proteins,cytochrome c,nitrogenase, elongation factor, ATPase synthase and glutamine synthetase, among others (OchInan and Wilson 1987, Kumada et al. 1993, 756 REVISTADE BIOLOGIA TROPICAL

Ludwing et al. 1993, Olsen et al. 1994, catastrophe in the sense that members of the Wallington and Lund 1994, Woese 1994) have Rhizobiaceae (Agrobacterium spp. Rhizobium greatly contributed to discerning the spp. Phyllobacterium spp. and Bradyrhizobium evolutionary relatíonshíps between bacteria, spp.) are not only intermingled with bacteria and therefore have substantially helped to originally included within the Rickettsiaseae identify different bacterial groups. ( and B. visonii, former Unfortunately, the same properties that make Rochalimae species), Bartonellaceae (B. these molecules suítable for phylogenetic bacilliformis) and members of "unknown" classification at a certain level, make them affiliation (Brucella spp. and Afipia spp.), but poor tools for defining bacterial species within are also themselves mixed up (e.g. R. me/flot; close sister dades. For instance, 1 agree with is closer to A. tumefasciens than to other the concept that there are genes (such as those members of the same genus). Furthermore, cited aboye) which can approach the Bradyrhizobium spp. belongs to a different phylogeny of an organism; however, 1 also branch of association, together with understand that at the species level they do not Rhodobacter sphaeroides, the latter bacteria have enough resolution for defining bacterial being considered a member of the species as separate entities (Stackebrandt and Rhodospirillaceae family (Imhoff and Trüper Goebel 1994). Furthermore, sorne of these 1984). In this respect, classification based on bacterial genes are not single units but rather sequence analysis has c1early contributed to are duplicated in the same chromosome discerning sister c1ades and appropriately (Dryden and Kaplan 1990, Viale ei al. 1994) or defining groups among the (J(-2 proteobacteria, repeated in a second (different) chromosome of particularly at the genus level (Stackebrandt sorne bacteria (Allardet-Servent et al. 1988, and Goebel 1994). In addition, it has made us Dryden and Kaplan 1990) and are therefore look more carefully at the common factors of subjected to different positional constraints bacteria belonging to the same lineage, that may promote diversity within the same sometimes in spite of the diverse biology group. It can be argued that, given a gene (or di",:":ayed by sister organisms (Moreno 1992). the entire genome), one can always discern However, at the species level the tree shown in between close bacterial relatives and resolve Fig. 1 has poor resolution for sorne strains: this them into species. The main problem with this is clearly illustrated by the compact clades is the absence of limits in the prediction of a demonstrated in the Brucella, Bartonella and terminal taxa: if one tries hard enough, Bradyrhizobium terminal taxons, whose sequence differences can be found to díagnose similarities are aboye 97%. As a result of this virtually any bacterial populatíon. The problem a combination of analysis of "pendular movement machíne" operated by the idiosyncratic sequences and numerical analysis "lumpers" and the "splitters" of bacterial of phenotypic characteristics has been depends on the data provided by undertaken by several researchers in order to them, generating from oneto an ample number discern between closely related táxones of species. (Wayne el al. 1987, Ficht et al. 1996). For the sake of our discussion, let us consider the phylogenetic tree in Fig. 1 constructed by the neighbor-joining method THE TAXONOMIC APPROACH (Saitou and Nei 1987) from Knuc values derived from 16SrRNA sequences. AH the The taxonomic species concept is a terminal taxa represented in this tree definition clearly linked to its method. For correspond to phylogenetic close relatives of many years, the aim of the microbial Gram negative bacteria assigned to the (J(-2 taxonomist has been to devise a consistent set subdivision of Proteobacteria (Olsen et al. of techniques for the classification of bacteria 1994). On consulting the topology of the tree into different ranks. The most common we can at fírst glance resolve that in this case procedures are based on the analysis of the traditional concept of Family established phenotypic features, biochemical by ranked classification is a complete characteristics, genotypic similarities and immunochemical reactions. Modern bacterial Moreno: In search of a bacterial species definition 757

taxonomists sort the available figures using fonu of clades or dendogrames (Gao et al. �al analysis and synthesise them in the 1994, Williams �t al. 1995). The basic

Agrobacteriumrhizogenes Rhizobiumleguminosarum Soil and plant '------Agrobacteriumtumefaciens '------Rhizobium meliloti associated Rhizobium loti Nunc 0.01 '----- Phyllobacterium�, � - - m!vr�'m(.rce(lrum " �, , {,o - �_ ' '- -

Animal associated } Opportunistic animal associated Afipiafelis Bradyrhizobium e/kanii } Rhodopseudomonas palustris PhotO syntetic, Bradyrhizobiumjaponicum soil and BradyrhizobiumBTAi plant . associated Fig.l PhylC)puetictrae mpresentatiof ve plant-associated, freeliving, opportunistic and animal-associated I"roteobacterla of ¡be a subdivision. Tbe tree was construcled by tbe neigbbour-joining method (Saitou and Nei 1987) from Knuc values derived from 16SrRNA sequences available in EMBL, GeneBank and DDBJNucleotide Sequence Database. The shadowed area ¡ncludes bacteria in which plasmids or Iysogenic phages have no! been found, and in which natural acquisition of genes from foreign sources have nol been demonstrated. Therefore these animal-associated bacteria may be considered as c10nal evolutionary unils.

taxonomic unit generated by this analysis is phylogenetic grouping based on sequence defined as a taxospecies (Sneath 1984a). The comparisons with taxonomic classification is a same procedures used to determine a species very powerful method for approaching a are employed to further characterise the practical definition of the bacterial species terminal taxons into subspecies, biotypes, (Stackebrandt and Goebel 1994). In spite of al! serotypes, chemotypes, phagotypes, this advantages, this proposed polyphasic phenotypes and so on. The virtue of the approach possesses sorne drawbacks as will be taxonomic proposition is that it is defined in a dernonstrated in the following paragraphs. polyphasic manner allowing high resolution at For instance, Fig. 1 shows that the plant the terminal branches of dendograrns; its endosymbiont B. japonicum is drawback is tbe inaccuracy in the earlier nodes phylogenetically c10sely related to the of the b.,anches (e.g. compare phylogenetic Bradyrhizobium strain BTAl; however, the trees from Yanagi and Yamasato [1993] with latter is photosynthetic while the former is noto phenic dendograms generated by Gao et al. Since this characteristic is cmomosornal coded [1994] ) This is rnainly due to the faet that and unlikely to be horizontally transferred, many eharacteristics displayed by the different these two bacteria could be tentatively bacteria (e.g. serological cross reactivity, classified as different species. SimilarIy, the staining properties, resistance to antibiotics, phylogeneticalIy closely related Bartonella etc.) may be the produet of eoevolution, species can be differentiated into ten individual parallel evolution or the result of horizontal species according not only to different transfer of genes. However, thecombination of phenotypic, chernotypic and immunochemical 758 REVISTA DE BIOLOGIA TROPICAL criteria, but more importantly to DNA-DNA interacting different hábitat. In spite of sorne reassociation values (Birtles et al. 1995). The shortcomings (Ayala 1976), the biological latter procedure has been pragmatically concept still has a profound impact on the established as a "superior method" for the recognition of animal and plant species (Avise classification of species (Stackebrandt and 1994). One important difference between this Goebel 1994). Nevertheless, our experience on consideration and the previous definitions is its studying the microbiological world has taught predictive nature, where time projects ¡nto the us that reality challenges the imagination. Not future from the moment in which a particular wíthout lengthy debate and a considerable species is recognised. Such a claim about the input of energy (and sorne bile), the genus future may of course turn out to be false since Brucella is believed to comprise six species the separation between two or more species (Corbel 1989,Meyer 1990),in spite of the fact may be temporary or permanent depending that phylogenetic analysis based on conserved upon the way the world goes (Danto 1985). molecules sueh as 16SrRNA símilaríty and With respect to bacteria,the biological species DNA relatedness values indicate figures aboye concept has been defined on the basís of 99% and 98% respectively (Fig. 1 and Verger certain distinctive biological characterístícs et al. [1985]). The reason is not only because such as the basic comprehension of the life the different Brucella species may be resolved cycle (preferred hábitat, potentí al hábitat, on the basis of DNA restriction patterns preferred host, secondary hosts, vectors, (Allardet-Servent et al. 1988, Fekete et al. reservoirs, ínteraction with other bacteria, 1992), phylogenetic analysis of idiosyncratic symbíosis, commensalism, parasitism, sequences (Ficht et al. 1996) as well as saprophytic Iife and modes of transmissíon), phenotypic, chemotypic and antigenic bacterial physiology (metabolism, generation characteristics (Moreno 1992), but more time, optimal growth temperature, defence important because of the distinct biological mechanisms,offensive mechanisms,reparation behaviour of the species that distinguishes one systems and internal rearrangement of genes), from the other (Corbel 1989, Meyer 1990). In and reproductive behaviour (vertical vrs. this particular case the DNA-DNA horizontal inherence, conjugation, reassociation technique has been considered an transduction,retrotranspositíon, transformation "inferior method" compared to the biological and capsductíon). The bacterial specie defined characteristics of these bacteria and to other on the basis of qualitative biological assays (Meyer 1990). characteristics is sometimes referred to as genospecies (Sneath 1984a). However, on finding that similar biological traÍts may occur THE BIOLOGICAL APPROACH due to convergent or parallel evolution or may be acquired by the horizontal transfer of genes In contrast to the numerical analyses used (Lambert et al. 1990),the traditional biological in the phylogenetic and taxonomic studies for species concept must be modified in favour of the definition of bacterial species, the a more versatile notion. biological concept is based upon the Experience has demonstrated that identification of fundamental qualitative biological characteristics themselves are of characteristics. In modern biological usage, httle use in defining bacterial species species are defined as a particular group of (Weisburg et al. 1989). However, when organisms which retaín their distinctness from phylogeny and taxonomy are considered, the other kinds in nature over a period of biological properties shown by a bacterium successive generations. When referring to contribute to resolving important uncertainties. sexually reproducing organisms, "distínctness" For instance, B. bacilliformis, the etiological commonly designates the abilíty of members agent of Oroya Fever, can be recognised from of the same species to interbreed freely, other Bartonella species (Fig. 1), not only by whereas interbreeding between members of its clear taxonomic features (Brenner et al. different species is prevented or restricted by 1993, Birtles et al. 1995) but because this natural causes. In another context, bacterium, which is restricted to a small "distinctness" means isolation in a non geographical area in Peru,is a human pathogen Moreno: In search of a bacteria! species definition 759 exclusively transmitted by one distinct species THE MICROBIOLOGISTS' PERCEPTION of sand flies (Gray et al. 1990). Since humans are the only vertebrate host for B. CaroIus Linnaeus himself admitted that his bacilliformis, and our species reached the classification did not always categorise the highlands of Oroya no more than 12,000 years most similarthings together (including humans ago, it is reasonable to propose that this and apes who were recorded as separate particular parasitic bacterium emerged in its taxons), but he felt that his system was very present form around that time. Likewise, in well adapted to the interests of plant growers spite of the indistinguishable DNA-DNA and botanists (as well as clergies) of his time. reassociation similarity index shown by the The virtue of the binomial system of differentBrucella species (Verger et al. 1985), nomenclature established by this Swedish they can be recognised by their biological naturalist in 1753 has been its clarity and attributes,such as host range and pathogenicity simplicity. These attributes constitute the main (CorbeI 1989,Meyer 1990). Sheep are the only reason why the linnaen system is still used for animal host from which B. ovis has been naming all the organisms within the three recovered, whereas the wood-rat Neotoma domains of living things: Bacteria, Archea and lepida is the only known host for B. neotomae. Eucarya (Woese 1987). Prior to Linnaeus' Furthermore, B. neotomae is non pathogenic Species Plantarum publication, the names of for its host; in contrast B. ovis primarily affects the species were in reality short Latín males (causing ram epididymitis) and has descriptions, usually including several words. seldom been encountered among ewes. This complicated designation attempted to arethe natural host for B. canis; this bacterium present the main characters by which the seldom infects other species, and when ít does species could be recognised. The wisdom of infection occurs without any epidemiological Linnaeus was not only to create a importance since the bacteria cannot be comprehensíble classificationsystem, but more transmitted. Similarly,the host animals for B. importantly,a useful one. abortus, B. melitensis and B. suis are bovines, The development of a natural system, in caprines and suines respectively (with the only wruch living things are classifiedaccording to exception of B. suis biotype 4, which seems to the totality of their similarities and differences be specific for reindeer, and therefore defined and evolutionary traits,has occupied the minds by sorne taxonomist as a distinct species). of post-linnaen investigators to this day,and it These three Brucella species are very will very probably never be satisfactorily aggressive pathogens whích may infect other completed. The definition of a bacterial species hosts; however, when this occurs infection is furnishes the basis of our knowledge of it, but often terminal and of minor epidemiological that same knowledge also helps us to use it importance. What makes the biological more effectively. The art of the systematic attributes of these microorganisms important in bacteriologist lies not only in the correct use of the context of the species definition is the fact methods and in the construction of elaborated that under natural conditions the distinct theories to reach the "final" goal of biochemical,phenotypic, immunochemical and successfully defining an entíty that he or she genomic attributes that identify each of the thinks is a species, but also in convincing the different Brucella species have remained (scientific) community that this definition is constant in space and time (Corbel 1989, more appropriate to the reality of their Meyer 1990). These clear-cut examples force disciplines. Although, sorne general rules for us to consider the qualitative biological devising a valid classification may be characteristics as a valid alternative in the established by many, eventually only a few scheme of a species definition. Nevertheless, will be responsible for actualIy naming a with respect to other bacteria we must be species. In this respect the microbiologist cautious in our judgement since distinct species perception is sustained upon the biological behaviour may depend on the criteria expressed by a sponsor or a group of presence of accessory genetic elements. sponsors (experts, scientists, professionals, organisations) who based on scientific grounds (phylogenic,taxonomic and biological studies) 760 REVISTA DE BIOLOGIA TROPICAL and on practical parameters (epidemiology, many ways reveals the success of thechanges, production, potential use, risks, ranking orders, because the new classification is in accordance and tradition) propose a new bacterial species. with the attributes of the bacterium. As in a In considering the phylogenetic trees, fairytale happy ending, phylogenetic, taxonomic dendograms or biological taxonomic and biological studies have greatIy separation based on what is believed to be the contributed to joining this bacterium with its most natural system of classification for sister species (Fig. 1). However, other bacterial designation of species, the microbiologists groups are not as fortunate and despite must be aware that at higher systematic levels extensive studies they still resist classification these representations are retrospective, but capable of convincing a wide audience. For close to the species level they are prospective, instance, sorne species which seem to arrange that is, dependent upon the interpretation of the themselves into networks such as those future (O'Hara 1993). This is commonly members of the Rhizobiaceae, or referred as the "problem of prospective , and others which may narrative predication" (Danto 1985) since the have clonal structures such as the Brucella separation between two bacterial populations organisms, represent difficult challenges for as distinct species depend upon future taxonomists. contingencies which may be estimated but not The majority of plant pathologists and predicted. As a consequence, it is logically other professionals working with plants, impossible for aIl the species concepts to be consider a Rhizobium species one that is applied with certainty to the present because capable of nodulating root 1'lants and fixing they all depend upon the future (O'Hara 1993). nitrogen, whereas an Agrobacterium species is However, taking into consideration the capable of producing tumors in plants (after relatively short life span of humans, acquisition of S Y m and Ti plasmids, microbiologists are confronted with the respectively). During the last decade, dilemma of reaching a consensus for defining a prominent research has demonstratedthat these bacterial species that is useful for all purposes. bacteria are phylogenetica11y intermixed In this respect the data obtained by the various (Yanagi and Yamasato 1993), "promiscuous" approaches must be weighed in the light of the with respect to the horizontal transfer of genes different bacterial species concepts. The (Amábile-Cuevas and Chicurel 1992) and success of this enterprise will be measured by notably diverse (Laguerre et al. 1994, So et al. its widespread acceptance and its permanence. 1994), a11 findings that have considerably Therefore, the closest approximation to a complicated the classification of these bacterial species definition will be one that is organisms. Similar problems have been widely accepted by scientists and practising observed in the Salmonella and Pseudomonas professionals and which resists the inexorable species, whose diversity defeats the number of vicissitudes of time. This problem is similar to bacteriologists devoted to their study (Le­ the disadvantage oí the relativistic theory when Minor 1984,Palleroni 1984). It may be that the considering space and time in our daily lives definition of species concerning these with respect to the Newtonian approach which, endosymbionts, plant pathogens, commensals, in spite of its drawbacks, is more practical for saprophytes and sorne animal pathogens must thispurpose. be viewed in the context of a reticular structure From its initial "baptism" as Rickettsiae rather than as a unique terminal taxon (O'Hara quintana, followed by its "re-baptism" as 1994). Unfortunately, the undertaking of this Rochalimae quintana, succeeded by its "super­ laborious and difficult task does not guarantee baptism" as Bartonella quintana, this that plant pathologists, soil microbiologists and bacterium has changed none of its attributes medical doctors will accept the new (although new ones have been discovered), nomenclature if it is of liule practical use. apart from its name (Brenner et al. 1993). The One has the tendency to think that the fast and widespread acceptance of the new opposite would facilitate the classification of a nomenclature--bf the scientific community, bacterial species; however, this is not true. For veterinarians and physicians (Breitschwerdt et instance, it is clear that the Brucella species are al. 1994, Kordick and Breitschwerdt 1995),in a close related phylogenetic and taxonomic Moreno: In searcn of a bacterial species definition 761 cluster of organisms (Fig; 1). As stated before, shortened to "B. abortus 1". In this respect it on the basis of DNA-DNA reassociation values would be wise to weigh up the advantage of it has beeo proposed that aH Brucella spp. are adopting a restrictive definition of species biovars of a single species, B. melitensis against well-established bacterial groups. (Verger et al. 1985). Thus, this monospecies With respect to animals and plants it is would include six biovars with their respective accepted practice to define a new species on biotypes, e. g. B. abortus biotype 1 would the basis of one or a few members or on the become B. melitensis, biovar abortus, biotype grounds of the discovery of a few fossiI bones l. Even though this proposition may approach or seeds (Anderson and Jones-Jr 1984). the suggestion of a "superior rank" of Although sorne microbiologists have followed classification based on DNA-DNA the same practice, in rnodern times very few hybridisation values (Stackebrandt and Goebel scientists attempt to define a species on the 1994), the truth is that this molecular technique basis of a single bacterial ¡solate. In general, has only revealed genome similarities and no the characterisation of repeated distinct isolates discreet, consisten! and substantial differences are required before a seriol.ls committee in that are known to exist among these organisms bacterial 110menclature would consider it (AlIardet-Servent et al. 1988, Meyer 1990, appropriate to definea new species (Graham et Fekete et al. 1992, Ficht et al. 1996). Although al. 1984). Nevertheless, this is no! a this may be one of the main reasons why this straightforward subject. No one really knows proposition has had little acceptance within the how many isolates are necessary,and if these scientific community devoted to bruceHae isolates should be obtained duril1g certain time research (Meyer 1990), the ultimate judgement periods from different niches, hábitats, host8, has come from veterinarians, microbiologists, localities and so 011. In many cases the celerity physicians, epidemiologists and many other for naming a new bacteria is based on the professionals who do not consider the new grounds of pure anthropocentric arguments. recommendation relevant for their purposes For instance in the case of sorne "unimportant" (Nielsen and Duncan 1990, Wong et al. 1992). so11 bacteria, endosymbionts or the pathogens The motive for this undoubtedly Hes in the fact of sorne marine animals there is no rush to that for many years it has been known that B. provide a species name (Ewalt et al. 1994, Gao melitensis is a "vicious bug" (once upon a time et al. 1994, So et al. 1994). However, when considered as one of the candidates lor certain "emerging" microbial warfare), while B. neotomae is a suddenly invade humans or animals and plants "nice bug" confined to desert wood rats relevant for human use, the names of the new (Corbel 1989). This is not a matter of organisms are fiercely pursued (Regnery et al. semantics; in every single country people 1992). Our human self-centred view of tile responsible for control and eradication world frequently makes us forget that bacteria programs know the different pathological themselves are not susceptible to systematics behaviour and epidemiological circumstances as we are. that distinguish the various Brucella species: talk of B. abortus is one thing, but the mention of B. melitensis is a very different matter. The ARE BACTERIAL CLONAL OR name of the latter species would dispatch a RETICULATE EVOLUTIONARY UNITS? general alarm among aH public health units (Flores-Castro and Baer 1979). Furthermore, Accessory genetic elements of bacterial tradition in many western countries indícates genomes are plasmids, temperate phages, that polynomial names given to persons or to transposons, insertion sequences and retrons living things have a tendency to disappear in (Amábile-Cuevas and Chicurel 1992, Rice et favour of (linnaean) binomial nomenclature. al. 1993). These elements, besides being Therefore, just as one of the precursors of transmitted verticaHy during bacterial evolutionism, the "Conde Georges-Louis replication are also capable of horizontal Leclerc de Buffon" is simply known as transfer by means of conjugation, "Buffon", sooner or later "Brucella melitensis, transforrnation,transduction, capsduction or biovar abortus, biotype 1", will probably be retrotransposition. Under natural conditions the 762 REVISTA DE BIOLOGIA TROPICAL

basic trend is that genes carried by accessory species definitíon, sorne paragraphs will be elements constitute a particular sample of the devoted to their analysis. bacterial genome that is required occasionally For sorne time it was believed that rather than continually. By contrast, genes recombination among bacteria occurred at such needed for normal functions located in low frequency. relative to mutation, that it was chromosomes are the "housekeeping" stock considered evolutionary insignificant (Guttman that is indispensable (Krawiec and Riley and Dykhuizen 1994). However, recentIy it has 1990). However, there are a number of been demonstrated that recombination is the exceptions to these rules. Transposition and dominant force driving the clonal divergence other mobilisation actions can shuttle specific of sorne enterobacterial species (Guttman and genes and sequen ces between chromosomes Dykhuizen 1994), thus it must be considered a and plasmids as well as group different genes significant factor in the structuring of bacterial into a single replicon (Amábíle-Cuevas and populations capable of horizontal gene Chicurel 1992). During this process sorne transfer. Among enterobacterial species, additional genes may be transferred by a recombination seems to take place more often particular phenomenon known as "DNA between closely related bacteria than across hitchhiking". The generalisation of this event "species". In this respect the genetic has led to the suggestion that all bacteria may relationships between closely related share a pool of genetic information that is individuals could be regarded as tokogenetic accessible to virtually every other bacterial (Henning 1966) in contrast to the phylogenetic cel!. Therefore, the bacterial community could links that ínterconnect the different species. be envisioned as a single, heterogeneous Tokogenetic recombination are expected to multicellular organism, with elements homogenise the gene pool among the continually moving from one group of cells to interacting organisms, restricting the network another. In this respect, the bacterial universe structure of the species to a limited level (Fig. would function as a superorganism with a 2). In the case of sorne strains, network structure rather than as single entities the divergence from a comrnon ancestor was with a branching family organisation (Sonea traced as recently as 2400 years ago (Guttman 1991). If this nOlÍon were assumed, the and Dykhuizen 1994). In cornparison, bacterial species concept would remain an recombination between different "species" illusion. Although thismay sound exaggerated (Amábile-Cuevas and Chicurel 1992) may in the light of many different arguments given result in genetic diversification promoting not by population geneticists and evolutionary only the expansion of the network structure of biologists who have convinced us that genetic the species. but also favouring a fast and identity is a reality in the microbial world severe genetic drift which may eventually (Ochman and Wilson 1987, Woese 1987, cause speciation (Syvanen 1994), in the Olsen et al. 1994,), sorne bacterial species such manner proposed by the punctuated as those belonging to the Rhizobiaceae and equilibrium hypothesis (Eldredge and Gould Enterobacteriaceae may assemble in a network 1972). A similar recombinational phenomenon fashion. In other cases, certain animal may explain the difficulties of defining pathogenic bacterial species such as Bartonella bacterial species among certain groups of soil spp. Brucella spp. Rickettsiae spp. and sorne and plant associated bacteria (Laguerre et al. endosymbionts seem to have a clonal structure 1994, So et al. 1994). because under certain local circumstances The constant changes of the plant­ horizontal gene flux between bacteria, associated bacteria from soil to plants would although possible, is an improbable natural exert selective forces on the versatility of the evento From this perspective, the bacterial genome, better achíeved by non-essential superorganism hypothesis, which is based on genes carried in plasmids than in chromosomes horizontal gene transfer between different (Eberhard 1990). However, the fact that species of bacteria, surrenders. Since these two chromosomal genes could move from one standpoints (network vs. clonal structures) are bacteria to another (Amábile-Cuevas and important for a practical concept of bacterial Chicurel 1992, Syvanen 1994), forces us to consider that sorne species of bacteria may Moreno: In search of a bacterial species definition 763 arrange in a network structure. For instance, induce plant tumors or root nodulation, members of the genera Agrobacterium and respectively (Jordan 1984, Lambert et al. Rhizobium have beendefined by their ability to 1990).

Extant species

Clonal Reticulate Clonal-Reticulate ABCD E F GH I ABCD E F G H ABCDEF GHI

(J en ..... a. (J)..c. c: en 0>(J) e 0.0- - ..... >oro .c - a..�

Time Ancestor Tokogenetic relationships

Fig. 2.Hypothetical phylogenetic trees and tokogenetic relationships among bacteria. 'Three different altematives for generating bacterial species or strains (A-I) during evolution are presented. Speciation will depend on the probabilistic transmission and expansion of exogenous and endogenous genetic events as well as natural selection. Endogenous events such as mutation, internal recombination of genes and duplication or deJetion of sequences are vertícally transmitted and clonally expanded by the ancestor (1). Exogenous events inelude !he horizontal acquisition or recombination of foreign bacteria sequences, plasmids and Iysogenic phages ( A ). Examples of clonal species are the Brucella, Bartonella and Rickettsiae spp.; examples of reticulate or clonal-retículate bacteria are members of the Rhizobiaceae and Enterobacteriaceae families. Tokogenetic relationships have been documented in enterobacterial species (Guttm¡¡.n and Dykhuizen 1994).

However, manynon tumorigenic agrobacteriae genera are intertwined (Fig. 1). Since Sym, Ti and non nodulating rhizobiae have been or Ri sequences are also intermingled with the isolated from soil (Bouzar et al. 1993, different bacterial species,it is possible that the Laguerre et al. 1993, Kalogeraki and Winans horizontal flux of these and other genes among 1995). In these bacteria the genes responsible the plant-associated bacteria has occurred for neoplastic transformation (onc and vir), several times during evolution (Deng et al. root nodules (nod) and nitrogen fixation (fix 1995). It is also probable that many of these and ni!) are present in large accessory elements genes located in plasmids or chromosomes such as the Ti and Ri megaplasmids for gall have a common origino This is further and hairy root transformation respectively and supported by the fact that Sym, Ti and R i the Sym megaplasmid for nodulation and plasmids sharesequences arnongthemselves as nitrogen fixation (Jordan 1984, Lambert et al. well as with genes located in fue chromosomes 1990, Deng et al. 1995). The phylogeny of (Prakosh and Schilperoort 1982, Kündig et al. these plant-associated bacteria, based on 1993, Rivilla and Downie 1994, Schwedock ribosomal RNA sequences and supported by and Long 1994, Kalogeraki and Winans 1995). other genotypic characteristics (Dooley et al. Another important example that may sustain 1993, Yanagi and Yamasato 1993, Wallington the network structure hypothesis of sorne and Lund 1994), show that the species of these species is the horizontal transfer of cytochrome 764 REVISTA DE BIOLOGIA TROPICAL

oxidase genes between soiVplant bacteria. It seems an improbable natural evento In general has been found that a phylogenetic tree terms, these intra�ellular organisms are constructed on the basis of different distinguished by the following characteristics cytochrome oxidase genes has unexpected which are in accordance with the absence of assemblage with respect to phylogenetic trees foreign genetic material transference: 1) the constructed on the basis of ribosomal genes metabolic and antibiotic susceptibility patterns (Castresana et al. 1994). The incongruency are maintained constant throughout space and seen between the cytochrome oxidase gene tree time; 2) structures such as piti, which favour and the 16S rRNA tree suggests a case for the horizontal transference of genes between horizontal gene transfer between these parties. bacteria are absent in these pathogens ; 3) as a Finally, several studies have demonstrated rule the animal pathogens are isolated as pure discrepancies between phylogenetic trees cultures, indicating that these bacteria are in based on 16S rRNA and dendograms less intimate contact with other species of constructed on the basis of numerical bacteria; 4) infections by the animal pathogens taxonomy (Ladha and So 1994, Leung et al. are initiated by a few bacteria that arec10nally 1994, So et al. 1994, Eardly et al. 1995). These expanded in the host; 5) multiple infections by findings support the notion that within this different biovars seldom occur in the same group of bacteria, the horizontal transfer of host; 7) striking similarity among isolates of genes, rather than convergence or parallel the same species obtained from different evolution, has influenced the evolution of localities during different periods of time has species (Syvanen 1994). Interestingly, the high been demonstrated; 8) bacterial diversity (e. g. genetic diversity of plant-associated bacteria number of species or biotypes) is found worldwide, which overlaps with the commensurate with host and vector diversity clonal population structure encountered in (Ristic and Kreier 1984a, 1984b, Moulder specific ecosystems, is one of most striking 1985, lsenberg 1988, Corbe1 1989, Halling and characteristics of these organisms (Bouzar et Zeher 1990, Meyer 1990, Regnery et al. 1991, al.1993, Dooley et al. 1993, Leung et al. 1994, Koehler et al. 1992, Schwartzman 1992, Wong et al. 1994). Brenner et al. 1993, Ewald 1993, Weiss and Al¡ mentioned previously, there are a Moulder 1894, Birtles et al. 1995). considerable number of intracellular bacteria With respect to these intracellular animal such as Brucella, Bartonella, Rickettsiae, and pathogenic cx-proteobacteria, and some Anaplasma species (Fig. 1) in which natural bacterial endosymbionts (Yubin et al. 1994), plasmids or temperate bacteriophages have not the concept of species must be considered been detected in spite of numerous attempts to within the context of c10nality rather than that findthem during chromosomal studies (Meyer of a network or reticulate structure. Since 1990, Reschke et al. 1990, Regnery et al. c10nality is maintained by the vertical, asexual 1991, Elzer et al. 1993, Roux and Raoult 1993, transmission of genetic material from parent to Grasseschi and Minnick 1994, Krueger et al. offspring, all evolutionary change through time 1994). Furthermore, some of these bacteria, must be considered the result of a mutational which have been experimentally transformed process or the autogenous rearrangement of with conjugative plasmids, are incapable of genes mediated by insertion sequences, retrons transferring the extrachromosomal DNA to a or transposons. The relative contribution of second host of the same or a different species; mutation and autogenous recombination to and no evidence indicating that the plasmid divergence is not known. This is mainly due to had integrated into the chromosome was problems associated with the collection of attained during these studies (Elzer et al. accurate figures during internal 1993). Gene transfer by other mechanisms recombinational processes. However, in some such as translocation of naked DNA, plasmid-free bacteria such as Brucella spp. conjugation mediated by specialised with an upper limit DNA-DNA reassociation transposons, or reception of foreign DNA as values, the relative contribution between these result of heterologous conjugation cannot be . two processes could be estimated by ruled out. However, due to the characteristics comparing the gene sequences at homologous of this bacterial group, horizontal gene transfer sites in two different species, with the position Moreno: In search of a bacteria! species definition 765 of the genes in the chromosomes (Brucella dimension which has greatly contributed to spp. possesses two different chromosomes correcting sorne of the distorted perceptions [Michaux-Charachon et al. 1993]). Differences "not in tune with the position of in the number and location of insertion microorganisms in the natural order of things" sequences between different Brucella species (Woese 1987). The methods used for the (Ouahrani et al. 1993), may suggest that definition of biological and taxonomic species autogenous recombination in addition to concepts lack sufficient evolutionary mutation could be a driving force in the perspective and, hence provide an speciation of bacteria devoid of heterologous inappropriate guide to the origins and products recombination promoted by plasmids. It is of evolutionary diversification (O'Hara 1993, possible that the speciation of intracellular Woese 1994). This has led to a call for the bacteria and endosymbionts is commensurate replacement of the biological-taxonomic with the evolutionary history of their hosts concept by the phylogenetic species concepto (Ochman and Wilson 1987). In this case the However, as stated before, the methods used biological approach takes on an important for constructing bacterial phylogenies do not in dimension, since most of the environmental themselves possess sufficient resolution to constraints that direct natural selection seem to distinguish those terminal taxons we call be present in the host (Brunham et al. 1993, species; therefore a polyphasic phylogenetic­ Laguerre et al. 1993). These interactions could taxonomic approach was suggested attain extreme proportions such as (Stackebrandt and GoebeI 1994). Although tbis mitochondria and chloroplasts in which the latter proposition is closer to a practical eubacterial identity is forever lost within the definition of bacterial species (or host environment, maintaining however, their nomenspecies), it is vague with respect to the phylogenetic relationship with the respective biological properties of the bacterial parental group (Yang et al. 1985). An organisms. Furthermore, it does not take into important exception may be the modern and consideration the microbiologist point of view, indiscriminate use of vaccines, antibiotics and which I consider essential for a practical other antimicrobial agents which may redirect definition of bacterial species. As a scientist I sorne of the natural selective processes in these am obligated to observe, measure, analyse and and other bacteria. describe phenomena; however as a human being I may have preconceived notions of the object I am devoted to studying; tbis I can not WHAT IS THE BEST BACTERIAL neglect but if I recognise its existence, I may SPECIES DEFINITION? be able lo control it. The members of the (X-2 proteobacteria For over a century, the biological species shown in Fig. 1 wbich I have used to illustrate concept in the first instan ce, and following the virtues and drawbacks of the phylogenetic, this, the taxonomic species concept were the taxonomic and biological approaches as well major theoretical postulates orienting as the microbiologists' perception, constitute a bacteriological research. In the 19th century good exercise. For instance, it is clear that and the first half of the 20th century molecular sequencing based on 16SrRNA had bacteriology was primarily confined within the resolved the phylogenetic position of limits of medical research (Smith and Martin facultative cell associated bacteriawhich were 1949). During this time many different located haphazardly in different families (e.g. pathogenic bacteria were described, and sets of Rochalimae and Bartonella species), in the key phenotype characters were established to same family (e.g. Bradyrhizobium and identify members of the same. It was mainly in Rhizobium species) or in groups with uncertain the second half of this century that research in affiliation (e.g. Brucella and Ochrobactrum bacteria devoid of medical importance began species). In addition, this approach has to participate in the microbiologist game. revealed that genera and the corresponding Owing to the incursion of molecular. biology species of plant associated bacteria (e.g. during the last decade, the bacterial world Agrobacterium, R h izob i u m a n d began to develop an evolutionary historical Phyllobacterium species) and sorne stem- 766 REVISTA DE BIOLOGlA TROPICAL nodulating and photosynthetic bacteria clustered as a single species; although the (Bradyrhizobium spp. and Rhodopseudomonas qualitative biological attributes (e.g. different palustris) are intermixed in the phylogenetic hosts and virulence) supported by fine analysis tree, a fact that has caBed the attention of (Allardet-Servent et al. 1988, Meyer 1990, several investigators who have suggested Fekete et al. 1992, Ficht et al. 1996), clearly reclassification of the group (Sawada et al. demonstrates that these bacteria behave as 1993, Willems and Collins 1993, Yanagi and independent urtits deserving a distinct species Yamasato 1993). From a different perspective, classification. In this regard, it would be the phylogenetic analysis did not have enough prudent that microbiologists base their resolution in defining sorne of the animal (e.g. perception on the clonal structure exhibited by Brucella, Bartonella and Afipia species) and the Brucella organisms, rather than on the plant (Bradyrhizobium spp) associated species. narrow classification displayed by the In this regard numerical taxonomy has clearly phylogenetic analysis and DNA-DNA resolved the position of sorne of the animal hybridisation studies of these organisms associated bacteria (Bartonellaspp. and Afipia (Meyer 1990). spp.) as demonstrated by the polyphasic approach undertaken to classify these organisms (Brenner et al. 1991, Brenner et al. A NEVER ENDING STORY 1993). However, while the data provided by 16SrRNA analysis gave a more reliable In considering a definition of bacterial indication of the phylogeny, the degree of species there is at least one major uncertainty phenotypic divergence observed among sorne left which probably would never be plant and soil bacteria (e.g. photosynthetic and satisfactorily resolved. For instance, if bacteria non photosynthetic Bradyrhizobium spp.) and behave as independent evolutionary · units, the degree of genotypic convergence estimated adaptive changes may expand clonally in Brucella spp. (Verger et al. 1985) raises resulting in a distinctive variety or in a new questions conceming the polyphasic taxonomic species. On the contrary, if they behave as approach to bacteria! systematics (Yanagi and reticulate evolutionary units, adaptive changes Yamasato 1993, So et al. 1994). Should these occurring within an individual are horizontally examples "persuade us that numerical transferred to many or aH members of the taxonomy based on phenotypic characteristics species which will "homogenise the species does not pro vide reliable evidence concerning gene pool, even as it creates dilj:erent evolutionary relationships? Or should it instead genotypes and diversifies clones" (Guttman remind us that sequence data alone do not tell and Dykhuizen 1980). Recombination between us everything (or even anything) of interest different species may result in unbounded about the biology of organisms" (So et al. divergence and speciation. Depending on local 1994). At the present time both scenes may be circumstances bacteria may alternate from one correct. With respect to the soil and plant form to another, that is, behaving as associated a-2 proteobacteria, the qualitative evolutionary units propagating clonaHy for biological characteristics seem to be of little long periods of time, followed by periods of use in discerning different bacterial species; "promiscuity" where clonal divergence and a nevertheless it gives an excellent perception of network structure may . result from the network structure of the group. Therefore it recombination. Mutatión without (foreign) would be wise that microbiologists base their recombination followed by clonal expansion in perception on this understanding, rather than an isolated environment (e.g. the intracellular on trying to dissect the soil and plant milieu) may ptomote fast speciation, whereas associated bacteria species on the grounds of recombination between clones of the same 16SrRNA sequence comparisons and numeral species (tokogenetic relationships) may delay taxonomy data alone. On the contrary, the speciation. These ideas are schematical1y phylogenetic data represented as trees as well proposed in Fig. 2. as the DNA-DNA hybridisation analysis Without doubt, modern bacteriologists will suggest to us that Brucella organisms are a devote much of their time to resolving these monophyletic group of bacteria that may be and other questions. Researchers will attempt Moreno: In seíirChoh Iiacterial species defihition . 767

fo organise the knowledge of the diversity and RESUMEN ' variability among bact�ria into systems of cla:ssifi cation which reflecUhe evolution of Se examinó la definición'de especie bacteriana usando como ejemplos protebacterias. del subgruPQ a--2 asociadas these microorganisms as wetl as their con las plantas y con los animales, y tomando en similarities and differences under certa:in consideración estudios filogenéticos, taxonómicos, biológicos y el criterio de los microbiólogos. La virtud que biological limits. Alth()ugh this sounds tienen los estudios filogenéticos es que proPQrcionan una relatively straightforwatd, the truth is that perspectiva evolutiva de los linajes bacterianos; sin regardless of how much we Iearn about a embargo el los métodos usados PQseen poca J:"eSolución para definir aquellas especies localizadas en las ramas o particular bacterial $r up, the ambiguity terminale� de los árboles filogenéticos. El mérito que related to the definition of species . will prevail, tienen los métodos taxonómicos se debe ha que definen las because it is not possible to formulate "any especies con baseen múltiples características, permitiendo así una excelente n;solución.en las ramas terminales de los cleardistinction between irtdividual differences dendogramas; su desventaja radicaen su Poca exactitud en and slight vari.eties; or between more plainly los nodos iniciales de los c1adogramils. A nivel individual los estudios biológicos frecuentemente presenta marked varieties a:nd sub-species, and species" dificultades debido a que muchas de las características son (Darwin 1859 ). No two individuals are exactly el resultado de cQevolución, evolución paralela o bien alike in aH details. "Even if we were presented transmisiónhorizontal de, genes; sin embargo cuando este concepto es C9nsiderado conjtpt�nte con el filogenético with " a complete DNA sequence of every y el taxonómico, se logran responder algunas preguntas individual ofevery species, we would stiH not que ninguna de las estrategias puede resolveráisladamente be able to discern what makes species El juicio expresado por un grupo de expertos con base en . evidencias científicas y en la práctica cotidiana para different. We need to be abl� to modify genes proPQneruna nueva especie de b�teria es lo que se conoce to see if we can produce neW strains and even como el . criterio de los Wic1"()biólogos. �I éxito de este criterio se nUde por su , aceptación 'gen�ral y .por su new species, and see whether the species act as pennanencia 'a través del tiempo. Uno de los problemas if !hey possess. SOrne inertia, resisting mas. difíciles respe!)to.a la definiciónde especie bac�riana modification" (Statltin 1985). Thediversity of es como 4istinguir si , estas son 1l!l1dade.s .evolutivas indepenciieptes o si ellas son unidades .e�olutivas livil\g tllings i§ constantIy émerging . an� es reticulBrl?s;·.En el pripier.ca sol la herenc�a trasmitida .the s 1i"aliti tlJ(e tbe ey ol li,on verticalm,éí:lte como resilltádp de la división binari¡l y la chaOtlt ; hi �� � ' oÍ. u ' pto ceSs '. itit�uirs 'i ,an ip�scapable mwtiplicación clq!l�. .Este podrla ser el . caso. de algunas asociíl¡1a� .!r¡>n los animales en. las que la indeteinrirlateJ:les$ to the notión oí species. . reOO�binacióA ge nética posibleirien� no OCIU'f!:.0 bi�p es This'b ió1ogical prit:l;ciple i� cqmparabl� in cipnal. Jiln el Segu!1do c�p, los cambios ��vos respects ;to of e. s1J9:e,4ell.ell un.. iy 4�yi9qo:'J�ueden �r tr��Mjclos ' m�riY,' the uneertain't)l principl. ,()nzon,t�epte •• a ,vlll1osp a,tpdos t.os �eil;!bros .ae un quan:tlJm m�haíliCs pro¿1ai ined. by HeiseWerg gruPQ.�� p�ie¡:a �relcasa �.mú.i;has delaS l;!ac1'etias ''more thaq 60 years a,go, tbat establisbed the , intestillaitls y ,de �Jls asocjada.$ con las planta&� La divergencla'gen�i:a y laéspecl ación de bacteri� can una ')iirnpossibility of simQlt�t:l;�Pllsly det�rmining estryctí:!ra 'iclonal· depender@ · c¡u;i exelusivl¡lmellte de . , the velocity a.J.Í,dpos itiónóta> particl�(Wheeler ¡putaciones y d� 10s. n;a,iTéglos �enéticos.iJltef1l oS que and Zurek 1983). The speeÍEls problern is not a: puedaíl sú.cei�ri .de l� wte¡:ac�i¡'nes.gené �c� con '. sornetbingi tbafwe need to ove17COJIle, �c.f·the .otras b�enlll!iG�t¡l.mce ,, �, lil.qu .,I,\:duda correspol)de ,al proceso e. . o, es ,al very existel1ce of tbis pncertaihty i$ itselfofie de los éle!DentóseseriHales paracotnpt en. er . a. of lPe ll1;ostsisrtif ic�nt pieces,.�f evidenc.e. oí de especie baCteriana. the historieal process:we call evolution. ,

REFERI3NCES , , AOKNOWL:.EDGMENTS Allardet-Secve\lt; A. G. BoWg� M. ��qi. � •. ,flages M· Bellis, &. .G. Roi:¡res, H� 8�., Dl'-l� po1ymorpbisro in 1 thank Natasha McIver and several strains of the gc;lnl!s Bru cellp . J. BaIfl�tiol, .179: 4603- 4607. at'onymQus reviewersfor tbeir careful review 'oí tbe ll)a:n,�scriPt.. TJ:ii s worlcwas supported by AIl'!ábile,(::uev�, C. F. & M. E.' Chicurel. 1992. Bacterial the International .Atomic Energy Agency plasmids II!ld �ene flux. Ce1l 10: 189-199. (8277/RB). AndersQn, S. & J. K. Jones-Jr.1984. Orders andlamilí�s of recel1t mammals of the word. A Wiley-Interscjence Publications, JOM WileY & Sons, NeW'YorkPR, 1�68(i¡ 768 REVISTA DE BIOLOGIA TROPICAL

Avise, J. C. & R. M. Ball. 1990. PrincipIes of genealogical Danto, A. 1985. Narration and knowledge. Columbia concordance in species concepts and biological Univ. Press, NewYork. taxonomy. Oxford Surv.Evo!. Biol. 7: 45-67. Darwin, C. 1859. On the origen of species, p. 469. Murray, Avise, J. C. 1994. Molecular markers, natural history and London. evolution. Chapman & Hall Inc. , New York, London pp. 1-511. Deng, W. Y. M. P. Gordon & E. W. Nester. 1995. Sequence and distribution of IS 1312: evidence for Ayala, F. J. 1976. Molecular genetics and evolution, p.l- horizontal DNA transfer from Rhizobium melíloti to 20. In F. J. Ayala (ed.). Molecular evolution. Sinauer, Agrobacterium tumefaciens. J. Bacterio!. 177: 2554- Sunderland MA. 2559.

Birtles, R. J. T. G. Harrison, N. A. Saunders & D. H. Dobzhansky, T. 1937. Genetics and the origin of species. Molyneux. 1995. Proposal to unify the genera Columbia UniversityPress, New York. Grahamella and Bartonella with description of combo nov. Dooley, J. J. S. P. Harrison, L. R. Mytton, M. Dye, A. combo nov. and the new species Cresswell, L. Sko!, & J. R. Beeching. 1993. sp. nov. Bartonella ray/orii sp. nov. and Bartonella Phylogenetic grouping and identification of Rhizobium doshiae sp. nov. ¡nt. 1. Syst. Bacteriol. 45: 1-8. isolates on the basis of random amplified po!ymorpbic DNA profiles. Can. J. Microbio!. 39: 665-673. Bouzar, H. D. Ouadah, Z. Krimi, J. B. Jones, M. Trovato, A. Petit & Y. Dessaux. Correlative association between Dryden, S. C. & S. Kaplan. 1990. LocalizatÍon and residentplasmids and the host chromosome in a diverse structural analysis of the ribosomal RNA operons of Agrobacterium soil population. Appl. Environ. Rhodobacter sphaeroides Nucleic. Acids Res 18: 7267- Microbio!. 59: 1310-1317. 7277.

Breitschwerdt, E. B. D. L. Kordick, D. E. Malarkey, B. Eardly, B. D. F. S. Wang, T. S. Whittam & R. K. Selander. Keene, T. L. Hadfield & K. Wilson. 1994. Endocarditis 1995. Speeies limits in Rhizobium populations Ihat in dogs due to infection with a novel Bartonella nodulate ¡he common bean (Phaseolus vulgaris). Appl. subspeciesJ. Clin. Microbio!. 33: 154-160. Environ. Microbio!. 61: 507-512.

Brenner, D. J. D. G. Hollis, C. W. Moss, C. K. English, G. Eberhard, W. G. 1990. Evolution in bacterial pla."Ilids and S. Hall, 1. Vincent, J. Radosevic, K. A. Birkness, W. F. levels of selection. Quarterly Rev. Bio!. 65: 3-22. Bibb, F. D. Quinn, B. Swaminathan, R. E. Weaver, M. W. Reeves, S. P. O'Connor, P. S. Hayes, F. C. Eldredge, N. & S. 1. Gould. 1972. Punctuated equilibria: Tenover, A. G. Steigerwalt, B. A. Perkins, M. 1. an altemative lo phyletic gradualism, p.82-115. ln T. J. Daneshvar, B. C. Hill, J. A. Washington, T. C. Woods, M. Schopf (ed.). Models in paleobiology, Freeman, S. B. Hunter, T. L. Hadfield, G. W. Ajello, A. F. Cooper & CO. San Francisco CA. Kaufmann, D. J. Wear & J. D. Wenger. 1991. Proposal of Afip ia gen. nov. with Afip ia felis sp. nov. (formerly Elzer, P. H. M. E. Kovach, R. W. Phillips, G. T. the cal scratch disease bacíllus), Afip ia cleve/andensis Robertson, K. M. Peterson & R. M. Roop-Il. 1995. In sp. nov. (formerly the c\eveland elinic foundation vivo and in vitro stability of broad-host-range cloning stmin), Afip ia broomeae sp. nov. and Three Unnamed veclor pBBiMCS in six Brucella species. Plasmid 33: Genospecies, J. Clin. Microbio!. 29: 2450-2460. 51-57.

Brenner, D. J. S. P. O'Connor, H. H. Wink1er & A. G. Ewald, P.W. 1993. Evolution of Infectious Disease, Steigerwalt. 1993. Proposals to unify the genera Oxford University Press lne. New York. Bartonella and Rochalimaea, with descriptions of Bartonella quintana combo nov. Bartonella vinsonii Ewalt, D. R. J. B. Payeur, B. M. Martin, D. R. Cummins & combo nov. Bartonel/a henselae combo nov. and W. G. Miller. 1994. Characteristics of a Brucella Bartonella elizabethae combo nov. and to remove the species from bottlenose dolphin (Tursiops truncatus). 1. family Bartonellaceae from the order . Int. Vet. Diag. Iovest. 6: 448-452. J. Syst. Bacteriol. 43: 777-786. Fekete, A. J. A. Bant!e, S. M. Halling & R. W. Stich. 1992. Brunham, R. C. F. A. Plummer & R. S. Stephens. 1993. Amplifieation fragment length polymorphism in Bacterial antigenic variation,. hos! Immune response, Brucella strains by use of polymerase chain reaction and pathogen-host coevolution. Infeet. Immun. 61: with arbitrary primers. J. BacterioL 174: 7778-7783. 2273-2276. Ficht, T. A. H. A. Husseinen, J. Derr & W. Bearden. 1992. Castresana, 1. M. Lübben, M. Saraste & D. G. Higgins. Speeies specific sequences al the omp2 locus for 1994. Evolution of cytochrome oxidase, an enzyme Brucella type strains. Int. J. Syst. Bae!. 46: 329-331 oIder than atmospeheric oxigeno EMBO J. 13: 2516- 2525. Flores-Castro, R. & G. M. Baer. 1979. Brueellosis (Brucella melitensis) - Zoonotic implications, p. 195- Corbel, M. J. 1989. Microbiol.ogy of the genus Bruceila, p 216. /n J. H. Steele (ed.). Handbook series i!l zoonoses 53. In E. J. Young & M. J. Cocbel (eds.). : eRC Press, lnc. Boca Raton, FL. chemieal and laboratory aspects. CRC Press, Inc. Boca Raton, F!.... Forterre, P. N. Benachenhou-Lahfa, F. Confalonieri, M. Duguet, C. Elie & B. Labedan, 1993. The nature ofthe Cracraft, J. 1983. Species concept and speciation analysis, universal ancestor and the rool of the Iree of life, still p.159-187. In R. F. Johnson (ed.). Current omithology, open questions. BioSystems 28: 15-32. Plenum Press, NY. Moreno: In search of a bacterial species definition 769

Gao, J. L. J. G. Sun, y. Li, E. T. Wang & W. X. Chen, Kumada, Y. D. R. Benson, D. HiIlemann, T. 1. Hosted, D. 1994. Numerical taxonomy and DNA relatedness of A. Rochefort, C. J. Thompson, W. Wohlleben & Y. tropical rhizobia isolated from Hainan province, China. Tateno. 1993. Evolution of the glutamine synthetase Int. J. Syst. Bacterio!. 44: 151-158. gene, one of the oldest existing and functioning genes. Proc. Na!. Acad. Sci. 90: 3009-3013. Graham, P. H. M. J. Sadowsky, H. H. Keyser, Y. M. Barnet, R. S. Bradley, J. E. Cooper, D. 1. De-Ley, B. D. Kündig, C. H. Hennecke & M. Gottfert. 1993. Correlated W. Jarvis, E. B. Roslycky, B. W. Strijdom & 1. P. W. physical and genetic map of the Bradyrhizobium Young. 1984. Proposed minimal standards for the japonicum 110 Genome. J. Bacterio!. 175: 613-622 . description of new genera and species of root- and steam-nodulating bacteria. Int. J. Syst. Bacterio!. 41: Ladha, J. K. & R. B. So. 1994. Numerical taxonomy of 582-587. photosynthetic nodulating Aeschynomene species. Int. J. Syst. Bacterio!. 44:62- 73. Grasseschi, H. A. & M. F. Minnick. 1994. Transformation of by electroporation. Can. J. Laguerre, G. M. AlIard, F. Revoy & N. Amarger, 1993. Microbiol. 40: 782-786. ¡solation from soil of symbiotic and nonsymbiotic Rhizobium leguminosarum by DNA hybridization. Can. Gray, G. C. A. A. Johnson, S. A. Thornton,W. A. Smith, J. Microbiol. 39: 1142-1149 . J. Knobloch, P. W. Kelly, L. O. Escudero, M. A. Huayd & R. L. Hoover. 1990. An epidemic of Oroya Laguerre, G. M. Bardin & N. Amarger. 1994. Rapid fever in Peruvian Andes. Ame. J. Trop. Med. Hyg. 42: identification of Rhizobia by restriction fragment 215. length polymorphism analysis of PCR-amplified 16S rRNA genes. Appl. Environ. Microbio!. 60: 56-63. Guttman, D. S. & D. E. Dykhuizen. 1994. Clonal divergence in Escherichia coli as a result of Lambert, B. H. Joos, S. Dierickx, R. Vantomme, J. recombination, no! mutatíon. Scíence 266: 1380-1382. Swings, K. Kersters & M. Van-Montagu. 1990. ldentification and plant interaction of Phyllobacterium Halling, S. M. & E. S. Zeher. 1990. Analysis of Brucella sp. a predominan! rhizobacterium of young sugar heet genotypic DNA for polymorphism and plasmids, p. plants. Appl. Environ. Microbio!. 56: 1093-1102. 476. In L. G. Adams (ed.). Advances in brucellosis, Texas A&M University Press, College Station, TX. Le-Minor, L. 1984. Genus m. Salmonella Lingieres 1900, 389AL, p. 427-458. In N. R. Krieg & J. G .Holt (eds.). Imhoff, J. F. & H. G. Trüper. Purple 1984. Purple Bergey's manual of systematic bacteriology, Vol 1, nonsulfur bacteria (Rhodospirillaceae) Pfenning & William & Wilkins, Baltimore. Trüper 171, 17AL,p.l6 58. ln N. R. Krieg & J. G. Holt (eds.). Bergey's manual of systematic bacteriology, Vol Leung, K. S. R. Strain, F. J. DeBruijn & P. J. Bottomley. l. William& Wilkins, Baltimore. 1994. Genotypic and phenotypic comparisons of chromosomal types within an indigenous soil Isenherg,H. D. 1988. Pathogenicity and virulence: another population of Rhizobium leguminosarum bv. trifolii. view. Clin. Microbiol. Rev. 1: 40-53. Appl. Environ. Microbio!. 60: 416-426.

Jordan, D. C. 1984. Family m, Rhizobiaceae Conn 1938, Ludwing, W. J. Neumaier, N. Klugbauer, E. Brockmann, 321AL, p. 234. In N. R. Krieg & J. G. Holl (eds.), C. RolIer, S. Jilg, K. Reetz, I. Schachtner, A. Bergey's manual of systematic bacteriology, Vol 1, Ludvigsen, M. Bachleitner, U. Fischer & K. H. William & Wilkins, Baltimore. Schleifer. 1993. Phylogenetic relationships of bacteria based on comparative sequence analysis of elongation Kalogeraki, V. S. & S. C. Winans. 1995. The octapine­ factor Tu and A TP-synthase beta subunit genes. type Ti plasmid pTiA6 of Agrobacterium tumefaciens Antonie Van Leeuwenhoek Int J. Gen. Mol. Biol. 64: contains a gene homologous to the chromosomal 3-4. virulence gene acvB. J. Bacterio\. 177: 892-897. Meyer, M. E.1990. Current concepts in the taxonomy of Koehler,1. E. Quinn, F. D. Berger, T. G. LeBoít, P. E. & J. the genus Brucella, p. ¡-17. ln K. Nielsen &B. Duncan W. Tappero. 1992. Isolation oC Róchalimaea species (eds.). Animal brucellosis, CRC Press, lnc. Boca from cutaneous and osseous lesions of baciJIary Raton, FL. angiomatosis. New Engl. J. Med. 327: 1625-163 1. Michaux-Charachon, S. J. Paillisson, M. G. Carles-Nurit, Kordick, D. L. & E. B. Breitschwerdl. 1995. G. Bourg, A. Allardet-Servent & M. Ramuz. 1993. Intraerythroytic presence of . J. Presence of two independent chromosomes in the Clin. Microbio!. 33: 1655-1656. Brucella me/itensis 16M genome. J. Bacterio!. 175: 701 -705. Krawiec, S. & M. Riley. 1990. Organization of bacterial chromosome. Microbio!. Rev. 54: 502-539. Moreno, E. 1992. Brucella evolution, p. 198-218. In M. Plommet (ed.). Prevention of bruceIlosis in Krieg, N. R. & J. G. Holt. 1984. Bergey's manual of mediterranean countries. International Centre fo! systematic bacteriology, Vol 1, William & Wilkins, Advanced Mediterranean Agronomic Studies, Pudoc Baltimore. Scientific Publishers,Wageningen, Netherlands.

Krueger, C. M. K. L. Marks & G. M. Ihler. 1994. Mouldet, J. W. 1985. Comparative biology of intracelIular Bartonella bacilliformis genome size estimate and parasitismo Microbiol. Rev. 49: 298-337. preliminary macrorestriction map. Abstracts of th 94th general meeting, Ame. Soco Microbiol. H-187: 233. Nielsen, K. & B. Duncan. 1990. Animal brucellosis. CRC Press,Inc. BocaRaton, FL. 770 REVISTA DE BIOLOGIA TROPICAL

O'Hara, R. J. 1993. Systematic generalization, historical systematie bacteriology, Vol 1, William & Wilkins, fate, and the species problem. Syst. Biol. 42: 231-146. Baltimore.

O'Hara, R.J. 1994. Evolutionary history and the species Rivilla, R. & J. A. Downie. 1994. Identification of problem. AmZoologist 34: 12-22. Rhizobium leguminosarum gene homologous to nodT but located outside the symbiotie plasmid. Gene 144: Ochman, H. & A. Wilson. 1987. Evolution in bacteria: 87-91. evidence for a universal substitution cate in Cellular Genomes. 1. Mol. Evo!. 26: 74-86. Roux, V. & D. Raoult. 1993. Genotypic identification and phylogenetie analysis of the spotted fever group Olsen, G. J. C. Woese & R. Overbeek. 1994. the winds of riekettsiae by pulsedcfield gel e1eetrophoresis. 1. (evolutionary) change: breathing new Jife into Bacteriol. 175: 4895-4904. mirobiology. 1. Bacteriol. 176: 1-6. Saitou, N. & M. Nei. 1987. The neighbour-joining method: Olsen, GJ. & C.R. Woese. 1993. Ribosomal RNA: a key a new method foc reconstructing phylogenetic trees. lo phylogeny. FASEB J. 7: 113-123. Mol. Biol. Evo!. 4: 406-425.

Ouahrani, S. S. Michaux, J. S. Widada, G. Bourg, R. Sawada, H. H. Ieki, H. Oyaizu & S. Matsumoto. 1993. Tournebize, M. Ramuz & J. P. Liaulard. 1993. Proposal . for · rejection of Agrobacterium tumefaciens Idenlification and sequence analysis of IS6S01, an and revised deseription for the · genus Agrobacterium insertion sequence in Brucella spp. Relationship and for Agrobacterium radiobacter andAgrobacterium between genomic structure and number if IS6501 rhizogenes. Int. J. Syst. Bacteriol. 43: 694-702. copies. J. Gen. Microbio!. 139: 3265�3273. Sehadewald, R. 1986. Creationism pseudoseience p.305- Palleroni, N. J. 1984. Genus I. Pseudomonas Migula 1894, 317. In K. Fraizier (ed.). Seienee eonfronts the 237 al (Nom. con. Opino 5, judo Cornm. 1952, 237), p. paranormal, Prometheus books, NY. 141-199. In N. R. Krieg & J. G. Holt (eds.). Bergey's manual of systematic bacteriology, Vol 1, William & Sehwartzman, W. A. 1992. Infeetions due to Rochalimaea: Wilkins, Baltimore. The expanding elinical speetrum. Clin. Infeet. Dis. 15: 893-902. Paterson, H. E. H. 1985. The recognition concept of species, p. 21-29. In E. S. Vrba (ed.). Transvaal Schwedock, J. & S. R. Long. 1994. An openreading frame Museum Monograph N° 4, Pretoria, South Africa. downstream of Rhizobium meliloti nod Ql shows nucleotide sequenee similarity to an Agrobacterium Pitt, T. L. 1994. Bacterial typing: the way ahead. 1. Med. turnejaciens insertion sequenee. Mol. Plant·Microbe Microbio!. 40: 1-2. Interaet. 7: 151-153.

Prakosh, R. K. & R. A. Schilperoort. 1982, The Simpson, G. G. 1951. The species coneept. Evolution 5: relationship between nif plasmids of fast growing 285-298. Rhizobium species and Ti plasmids of Agrobacterium tumefaciens. J. Bacteriol. 149: 1129-1134. Smith, D. T. & D. S. Martin. 1949. Zinsser's texbook of bacteriology. Appl.eton-Century-Crofts, Ine. New Regnery, R. L. C. L Spruill,. & B. D. Plikaytis. 1991. York, NY. Genotypic identification of Rickettsiae and estimation of intraspecies sequence divergence for portions of two Sneath, P. H. A. 1984a. Numerieal taxonomy, p. 5·7. In N. rickettsial genes. J. Bacteriol. 173: 1576-1589. R. Krieg & J. G. Holl (eds.). Bergey's manual of systematic bacteriology, Vol 1, WiIliam & Wilkíns, Regnery, R. L. B. E. Anderson, 1. E. Carridge-III, M. e. Baltimore. Rodriguez-Barradas, D. C. Jones & J. H. Carro 1992. Characterization of a novel Rochalimaea species, R. Sneath, P. H. A. 1984b. Bacterial nomenclature, p.1 9-23. henselae sp. nov. isolated from blood of a febrile, In N. R. Krieg & J. G. Holt (eds.). Bergey's manual of human immunodeficiency virus-positive patient. J. systematie bacteriology, Vol 1, William & Wilkins, Clín. Microbio!. 30: 265-274. Baltimore.

Reschke, D. K. M. E. Fraizer & L. P. Mallavia. 1990. So, R. B. J. K. Ladha & J. P. W. Young. 1994. Transformation of Rhochalimaea quintana, a membec Photosynthetie symbionts of Aeschynomene spp. form a of the family . J. Bacteriol. 172: 5130- cluster with Bradyrhizobia on the basis of fatty acíd 5134. and rRNA analysis. Int. 1. Syst. Bacteriol. 44: 392-403.

Rice, S. J. Bieber, J. Chung,G. Stacey & B. C. Lampson. Sonea, S. 1991. Bacterial evolution without speciation, p. 1993. Diversity of retrons e1ements of rhizobia and 95. In L. Margulis & R. Fester (eds.). Symbiosis as a other gram-negative bacteria. 1. Bacteriol. 175: 4250- souree of evolutionary innovation. MIT Press, 4254. Cambridge, MA.

Ristic, M. & J. P. Kreier. 1984a. Family n. Bartonellaceae Stackebrandt, E. & B. M. GoebeL .1994. A place for DNA­ Gieszcykiewicz 1939, 25 AL p.717. In N. R. Krieg & J. DNA reassociation and 16S rRNAsequence analysis in G. Holt (eds.). Bergey's manual of systematic ¡he present speeies definition in bll.cteriology. Int. J. baeteriology, Vol 1, WilIiam & Wilkins, Baltimore. Syst. Bacteriol. 44:846-849.

Ristie, M. & J. P. Kreier. 1984b. Family III. Staley, J. T. & N. R. Krieg. 1984. Calssifieation of Anaplasmataeeae Philip J957, 980 AL, p.719. In N. R. procaryotic organisms: an overview, p. 1-4. In N. R. Krieg & J. G. Holt (eds.). Bergey's manual of Krieg & J. G. Holt (eds.). Bergey's manual of Moreno: In search of a bacterial species definition 771

systematic bacteriology, Vol 1, WiJliam & Wilkins, Wheeler, 1. A. & W. H. Zurek. 1983. Quantum theory and Baltimore. measurement, Princenton University Press.

Statkin, M. 1985. The descent of genes. Science 85, 6: 80- WilIems, A. & M. D. Collins. 1993. Phylogenetic analysis 83. of rhizobia and agrobacteria based on 16S rRNA gene sequences. In!. 1. Syst. Bacterio!. 43: 305-313. Syvanen, M. 1994. Horizontal gene transfer: evidence and possible consequences. Ann. Rey. Genet. 28: 237-26l. Williams, E. E. H. Rand, S. Rand & R. 1. O'Hara. 1995. A computer approach to the comparison of species in Templeton, A. R. 1989. The meaning of species and difficult taxonomíc groups. Breyiora50 2: 1-47. speciation: a genetic perspectiye, p.3-27. In D. OUe & J. A. Endler (eds.). Speciation and its consequences, Woese, C. R. 1987. Bacterial evolutíon. Microbio\. Rey. Sinauer, Sunderland, MA. 51: 22 1-271.

Verger, J. F. Grimont, P. A. D. Grimont & M. Grayon. Woese, C. R. 1994. There must be a prokaryote 1985. Brucella a monospecific genus as shown by somewhere: microbiology's search for itself. deoxyribonucleic acid hybridization. Int. J. Syst. Microbio!. Rey. 58: 1-9. Bacterio!. 35: 292-295. Wong, F. Y. K. E. Stackebrandt, J. K. Ladha, D. E. Viale, A. M. A. K. Arakaki, F. C. Soncini & R. G. Fleischman, R. A. Date & J. A. Fuerst. 1994. Ferreyra. 1994. Evolutionary relationships among Phylogenetic analysis of Bradyrhizobium japonicum eubacterial groups as inferred from GroEL and photosynthetic stem-nodulating bacteria from (chaperonin) sequence comparisons. Int. J. Syst. Aeschynomene species grown in separated geographical Bacterio!. 44:527 -533. regions Appl. Environ. Microbio\. 60: 940-946.

Wallington, E. J. & P. A. Lund. 1994. Rhizobium Wong, J. D. J. M. Janda & P. S. Duffey, 1992. Preliminary leguminosarum contains multiple chaperonin (cpn60) studies on the use of carbon substrate utílization genes. Microbiology 140: 113-122. patterns for the identification of Brucella species. Diagn. Microbiol.lnfect. Dis. 15: 109-113. Wayne, L. G. D. J. Brenner, R. R. ColweJl, P. A. D. Grimont, O. Kandler, I. Krichevsky, L. H. Moore, W. Yang, D. Y. Oyaizu, H. Oyaizu,G. J. Olsen & C. R. E. C. Moore, R. G. E. Murray, E. Stackebrandt, P. Star Woese, 1985. Mitochondrial origens. Proe. Nat!. Acad. & H. G. Trüper. 1987. Report of the ad hoc committee Sci. 82: 4443-4447. on reconciliatíon of approaches to bacterial systematics Int J. Syst. Bacterio\. 37: 463-464. Yanagi, M. & K. Yamasato. 1993. Phylogenetie analysis of the family Rhizobiaceae and related bacteria by Weisburg, W. G. M. E. Dobson, J. E. Samuel, G. A. sequeneing of 16S rRNA gene using PCR and DNA Dasch, L. P. Mallavia, O. Baca, L. Mandelo, J. E. sequencer. FEMS Microbio\. Lett. 107: 115-120. Sechrest, E. Weiss & C. R. Woese. 1989. Phylogenetic diyersity of rickettsiae. J. Bacterio!. 171: 4202-4206. Yubin, D. A. M. Dmitri & K. Chango 1994. Monophyletíc origen of a-division proteobacterial endosymbionts and Weiss, E. & J. W. Moulder. 1984. Family 1. Rickettsiaceae their eoevolution with inseet trypanosomatid protozoa Pinkerton 1936 186AL,p. 687-717. In N.R. Krieg & J. Blastocrithidia culicis and Crithidia spp. Proc. Nat!. G. Holt (eds.). Bergey's manual of systematic Acad. Sci. USA 91: 8437-8441 bacteriology, Vol 1, WilJiam & Wilkins, Baltimore.