Proposal that Agrobacterium radiobacter has priority over Agrobacterium tumefaciens. Request for an Opinion

The specific epithet ‘radiobacter’ was first proposed in the combination Bacillus radiobacter by Beijerinck & van Delden (1902) in a study of saprobic soil bacteria associated with nitrogen utilization. Subsequently, the species was reclassified successively in the genera Bacterium, Rhizobium, Achromobacter and Alcaligenes (Buchanan et al., 1966) before being transferred by Conn (1942) to the new genus Agrobacterium. The specific epithet ‘tumefaciens’ was first proposed in the combination Bacterium tumefaciens by Smith & Townsend (1907) for the pathogen of crown gall and was successively transferred to the genera Pseudomonas and Phytomonas (Buchanan et al., 1966) before Conn (1942) proposed transfer to the genus Agrobacterium, designating Agrobacterium tumefaciens as the type species.

The similarity and probable synonymy of Agrobacterium radiobacter and A. tumefaciens has been widely acknowledged (reviewed by Kersters & De Ley, 1984). Holmes & Roberts (1981) proposed synonymy of the two species and this is now unquestioned (Sawada et al., 1993; Bouzar, 1994; Young et al., 2001, 2005). Although the epithet ‘radiobacter’ has priority over ‘tumefaciens’ according to the International Code of Nomenclature of Prokaryotes [‘the Code’ – hitherto the International Code of Nomenclature of Bacteria; Lapage et al. (1992)] by reason of its earlier publication (the Code; Rule 38), Holmes & Roberts (1981) noted that A. tumefaciens was the type species of the genus and concluded that the union of the species would require that this name be retained over A. radiobacter.

Kersters & De Ley (1984) strongly supported the classification of Holmes & Roberts (1981), accepting a concept of phenotypic species and differentiating pathogen populations as infrasubspecies, but they nevertheless adopted an alternative species nomenclature based on pathogenic characters. This was contrary to the taxonomic concept guiding the inclusion of names in the Approved Lists of Bacterial Names (Skerman et al., 1980), and endorsed in subsequent considerations of classification, that species should be supported by circumscriptions as natural taxa (Wayne et al., 1987; Murray et al., 1990; Stackebrandt et al., 2002). Pathogenicity is not an essential species characteristic of a species named with the epithet ‘tumefaciens’, nor, of itself, the basis for species discrimination, especially where pathogenic characters are borne on transmissible plasmids with acquisition, exchange or loss leading to a change in species identity (Kersters & De Ley, 1984). Kersters & De Ley (1984) could have adopted a natural classification, even involving the epithet ‘tumefaciens’ (Holmes & Roberts, 1981; Bradbury, 1986; Moore et al., 2001, Young et al., 2001, 2005), but they claimed that the Judicial Commission of the International Committee on Systematic Bacteriology (1970) had designated A. tumefaciens as the type species and therefore this name must be used. They appear to have assumed, mistakenly (Sneath, 1984), that names are descriptive and that the epithet ‘tumefaciens’ was reserved for a species based on pathogenic characters.

Sawada et al. (1993) reviewed the nomenclatural relationship of A. radiobacter and A. tumefaciens and, in a Request for an Opinion (Sawada et al., 1993), proposed rejection of A. tumefaciens and acceptance of A. radiobacter (the Code; Rule 38). Their proposal makes it clear that they did not intend rejection according to Rule 56a (Rejection of names and placement in a list of nomina rejicienda) but that, believing that only one name could be applied to a species, the correct name should be A. radiobacter. Sawada et al. (1993) did not take account of the earlier discussion (Holmes & Roberts, 1981; Kersters & De Ley, 1984).

Subsequently, Bouzar (1994) reviewed the nomenclatural case and argued, contra Sawada et al. (1993) and reiterating the conclusion of Kersters & De Ley (1984), that, because A. tumefaciens was the type species, A. tumefaciens took priority over A. radiobacter. However, Bouzar's specific Request for an Opinion was confined to the proposal that A. tumefaciens be confirmed as the type species of Agrobacterium and a response (by Oyaizu and Sawada; Authors' reply in Bouzar, 1994) conceded the validity of this claim. With this accord, the Judicial Commission decided that there was no need for a formal response to either Request (Frederiksen, 1995) and therefore the question of the relative priority of the naming of the species, either as A. radiobacter or A. tumefaciens, was not considered by the Commission. Subsequently, Young et al. (2001, 2005) followed the interpretation that A. tumefaciens was the correct name for the united species.

The Code is not transparent on the attributes of synonymy (the Code; Rule 24a) and, as follows, it is easy to see how the above interpretation was arrived at. A species may have only one validly published name (the Code; Rule 23a). If the two species are united, the species name A. radiobacter, as the earlier heterotypic (subjective) synonym, should be the name of all strains identified as members of the species, including those previously named as A. tumefaciens (the Code; Rule 38). Consequently: (i) all strains of A. tumefaciens would be identified as A. radiobacter; (ii) the type species of the genus would cease to have strain representation and thus (iii) the name A. tumefaciens would cease to have nomenclatural standing. However, this creates a paradox because the type ‘is that element of the taxon with which the name is permanently associated’ (the Code; Rule 15) and loss of identity of the type species would imply loss of the genus name. Therefore, the type strain of A. tumefaciens must continue to be named A. tumefaciens and the name of the species containing the type strain must be A. tumefaciens. Because the epithet ‘tumefaciens’ is closely associated with identification of the species as a pathogen (Kersters & De Ley, 1984), this line of reasoning has gone unquestioned and has been instrumental in inhibiting the development of a phenotypic concept for species in Agrobacterium for more than 20 years.

Neither the International Code of Botanical Nomenclature (Greuter et al., 2000) nor the International Code of Zoological Nomenclature (2000) support this interpretation. Although the Bacteriological Code was established with the intention that the special needs of microbiologists be recognized, an early principle in the development of the Code was that ‘insofar as they are applicable and appropriate, the nomenclatural codes of Botany and Zoology should be followed in the naming of micro-organisms' (Editorial Board, 1958). Furthermore, if the intention to amalgamate all biological codes in a single International Biocode (Greuter et al., 1996) is pursued, then the determinations of these codes are likely to be applied in bacterial nomenclature.

The contention that A. tumefaciens must be the correct name of a united species because it is the name of the type species of the genus Agrobacterium arises from confusion between two independent principles; the principle of nomenclatural priority and the principle of the application of types. The proposition on which the present Request for an Opinion is based is that nomenclature in which A. tumefaciens has priority over A. radiobacter is inconsistent with the Code in respect to these two principles.

1. The correct nomenclature for the genus is

Agrobacterium Conn 1942

Type species: Agrobacterium tumefaciens (Smith & Townsend 1907) Conn 1942 [=a later heterotypic synonym of A. radiobacter (Beijerinck & van Delden 1902) Conn 1942].

2. In the decision of the Judicial Commission (1970) to conserve the genus name Agrobacterium over Polymonas, they did not ‘designate’ (Kersters & De Ley, 1984) the type species, merely noting that the type species was A. tumefaciens by the original designation of Conn (1942). Furthermore, conservation of names was explicitly denied in the Introduction to the Approved Lists. This type species is not conserved (the Code; Rule 56b) and the epithet ‘tumefaciens’ has not been conserved over the epithet ‘radiobacter’.

3. The application of names is determined by means of nomenclatural types, which are ‘that element of the taxon with which the name is permanently associated’ (the Code; Principle 5). Nomenclatural types, usually of genera for taxa at higher taxonomic levels, species for genera and strains for species and subspecies, are the name-bearers of those taxa. A type species of a genus, in this case A. tumefaciens, is always associated with the name of the genus (the Code; Rule 15).

4. When species are amalgamated, there is only one name for the united species (the Code; Rule 23a). The principle of priority, central to ordered nomenclature (the Code; Principle 6), is that the correct name of the united species is chosen by priority of publication, embodied in Rule 38. However, the name of a species that is united as a later heterotypic synonym does not lose standing in nomenclature (the Code; Rule 24a); all such synonyms are maintained with their individual type strains. Although a type strain may be correctly named after the earlier synonym of a united species, its status as the type of a type species is not affected. The point is not clearly expressed in the Code because the relevant text (the Code; Rule 24a, Note 3) is in the context of synonyms published in the Approved Lists. This should be clarified in a future edition of the Code.

5. In the case of a united species amalgamating A. radiobacter and A. tumefaciens, the correct name is A. radiobacter (as earlier heterotypic synonym) and the type strain of the united species is the original type strain of the name A. radiobacter. The name A. tumefaciens is a later heterotypic synonym, but nevertheless retains its original type strain and remains the type species of the genus Agrobacterium.

6. This proposal is in accord with the codes of botanical and zoological nomenclature.

7. This revision will reduce the confusion of nomenclature in Agrobacterium [but not of the genus itself (Young et al., 2001)] by permitting the application of a species name that does not have ambiguous connotations (Holmes & Roberts, 1981; Kersters & De Ley, 1984).

A Request for an Opinion is sought.