Genome wide sequence analysis grants unbiased definition of species boundaries in “Candidatus Phytoplasma”

Abstract

The phytoplasmas are currently named using the Candidatus category, as the inability to grow them in vitro prevented (i) the performance of tests, such as DNA-DNA hybridization, that are regarded as necessary to establish species boundaries, and (ii) the deposition of type strains in culture collections. The recent accession to complete or nearly complete genome sequence information disclosed the opportunity to apply to the uncultivable phytoplasmas the same taxonomic approaches used for other bacteria. In this work, the genomes of 14 strains, belonging to the 16SrI, 16SrIII, 16SrV and 16SrX groups, including the species “Ca. P. asteris”, “Ca. P. mali”, “Ca. P. pyri”, “Ca. P. pruni”, and “Ca. P. australiense” were analyzed along with Acholeplasma laidlawi, to determine their taxonomic relatedness. Average nucleotide index (ANIm), tetranucleotide signature frequency correlation index (Tetra), and multilocus sequence analysis of 107 shared genes using both phylogenetic inference of concatenated (DNA and amino acid) sequences and consensus networks, were carried out. The results were in large agreement with the previously established 16S rDNA based classification schemes. Moreover, the taxonomic relationships within the 16SrI, 16SrIII and 16SrX groups, that represent clusters of strains whose relatedness could not be determined by 16SrDNA analysis, could be comparatively evaluated with non-subjective criteria. “Ca. P. mali” and “Ca. P. pyri” were found to meet the genome characteristics for the retention into two different, yet strictly related species; representatives of subgroups 16SrI-A and 16SrI-B were also found to meet the standards used in other bacteria to distinguish separate species; the genomes of the strains belonging to 16SrIII were found more closely related, suggesting that their subdivision into Candidatus species should be approached with caution.

Introduction

Phytoplasmas are a large and diverse group of bacterial plant pathogens that have not yet been cultivated in vitro. Since their discovery [11] the research on these bacteria has been dominated by the efforts toward their classification. With the introduction of PCR, a taxonomic approach based on the analysis of selectively amplified 16S rRNA genes permitted the definition of a phylogenetically sound classification. Currently, phytoplasmas can be subdivided according of their 16S rDNA sequence into 20 clades [55], or, in a more comprehensive scheme based on observed or virtual RFLP pattern of 16S rDNA, into more than 30 groups, several of which can be further subdivided into subgroups, thus comprising more than 100 so called subgroups lineages [69]. This 16S rDNA based classification scheme is readily applicable in identification and classification of novel strains, and has found very wide application in epidemiological studies. However, it cannot be translated into a binomial nomenclatural system as commonly used in bacterial systematic. Despite its enormous power in resolving phylogenies at the higher hierarchic levels, the 16S rDNA sequence alone is not sufficient to define and circumscribe species [14], [58], [66]. As phytoplasmas have not yet beeen cultivated in vitro they cannot be deposited in culture collections and cannot be investigated for some properties that are regarded as necessary for their classification at the species level [3], [57]. Standing the current inability to comply with the minimal standard for their formal naming [3], the category of Candidatus, implemented out of the Bacteriological Code to record the properties of putative taxa of procaryotes [42], has been adopted for their nomenclature [60]. To date, several “Candidatus Phytoplasma” species have been described, in most cases congruently with the 16S rDNA based schemes. In several cases, however, phytoplasmas sharing high similarity in their 16S rDNA sequence have been described as different candidatae species due to their distinct biological properties. Although the subdivision into candidatae species has been agreed by a large panel of a dedicated working team [60], it implies a large degree of subjectivity due to the paucity and uncertain taxonomic significance of the properties that could be scored for the non cultivable phytoplasmas in contrast to other cultivable mollicutes.

In the 15 years lasted since the meeting when the basis of the classification scheme for phytoplasma were discussed, thanks to the large amount of studies on phytoplasmas their chromosomal and non chromosomal characteristics, their interaction with the plant and insect host, their epidemiology and spread, their biology had become much more clearly understood [12], [21], [30], [50], [59]. The Candidatus species had become more a solid reference than simply a method to record the properties of an organism known only for its ribosomal sequence. According to the paper concerning its implementation, the category of Candidatus apply to organisms that “can be recognized by their molecular structures but cannot be assigned to a known genus because of the lack of enough distinguishing characteristics. Formal recognition will come when new observations allow;” [42]. New observations actually accumulated for phytoplasmas. In view of the formal recognition of species of phytoplasmas, although not yet allowed due to the persisting inability to deposit a live type strain in a culture collection, the scientific community needs to be concerned that the criteria used for the definition of “Candidatus Phytoplasma” species meet the standards used for other bacteria. In recent times, the increased knowledge at the genome level of microorganisms opened unprecedented opportunities. As genome wide comparative sequence analysis is becoming the elective method to unveil the natural classification of prokaryotes, its application to phytoplasmas would prevent the use of arbitrarily chosen boundaries when delimitating Candidatus species and would help defining the Candidatus as a provisional status for formal species. If the very recent claim of axenic cultivation of phytoplasmas [8] will be confirmed and the deposition of strains in culture collections will become possible, then a congruence in classification with widely accepted taxonomic criteria would greatly facilitate the shift from the Candidatus to formal nomenclature.

In this study we analyzed the genomes of 14 different phytoplasmas and showed that the information extracted from their genome sequence can be used to critically evaluate the robustness of Candidatus species definition and contribute to the delineation of species boundaries with methods congruent with those used for formal species definition.