‘Candidatus Liberibacter solanacearum’, associated with plants in the family Solanaceae

The candidate genus ‘Candidatus Liberobacter’ was first proposed by Jagoueix et al. (1994) with two members, ‘Candidatus Liberobacter asiaticum’ and ‘Candidatus Liberobacter africanum’. Later, they were renamed ‘Candidatus Liberibacter asiaticus’ and ‘Candidatus Liberibacter africanus’ in order to conform with the International Code of Nomenclature of Bacteria (Garnier et al., 2000). In 2004, a third candidate species, ‘Candidatus Liberibacter americanus’, was described, from Brazil (Teixeira et al., 2005). Collectively, these three liberibacter species are commonly known as huanglongbing (formerly citrus greening). They have only been found naturally to infect citrus and other members of the family Rutaceae.

The liberibacters are phloem-limited, Gram-negative, unculturable bacteria that are spread from infected to healthy plants by grafting and psyllid insect vectors (reviewed by Bové, 2006). Scanning electron microscopy images of liberibacters reveal that they have a rod-shaped morphology (Tanaka et al., 2007). ‘Candidatus Liberibacter’ species belong to the Alphaproteobacteria, with their closest relatives belonging to the genera Bradyrhizobium, Bartonella, Agrobacterium, Brucella and Afipia (Jagoueix et al., 1994).

An investigation to determine the aetiology of a new disease of glasshouse-grown tomato (Solanum lycopersicum) and capsicum (Capsicum annuum) in New Zealand was done by Liefting et al. (2009). Extensive testing ruled out the presence of pathogenic fungi, culturable bacteria, viruses, viroids and phytoplasmas. Transmission electron microscopy of thin sections of leaf tissue from symptomatic plants revealed the presence of phloem-limited bacterium-like organisms. A range of universal and specific 16S rRNA PCR primers were used in different combinations on DNA extracted from healthy and symptomatic plants. One of the primer combinations produced a unique product from symptomatic plants only. Sequence analysis of this fragment revealed that it shared high identity with ‘Candidatus Liberibacter’ species. The remainder of the 16S rRNA gene was sequenced and phylogenetic analysis showed that it is distinct from the three citrus-infecting liberibacter species described previously (Liefting et al., 2009). Subsequently, with the development of a specific PCR diagnostic method, this new liberibacter was also detected in potato (Solanum tuberosum), tamarillo (Solanum betaceum), cape gooseberry (Physalis peruviana) (Liefting et al., 2008a, b; Abad et al., 2009) and chilli (Capsicum sp.).

A 1.7 kb fragment of the β operon containing the ribosomal proteins L10 (rplJ) and L12 (rplL) and the RNA polymerase beta subunit (rpoB) was amplified using the primer pair fp 1898/rp 1897 (Planet et al., 1995). From the resulting sequence, the primers rp01F (5′-CTCTAAGATTTCGGTTGGTT-3′) and rp01R (5′-TATATCTATCGTTGCACCAG-3′) were designed in order to obtain the full-length sequence of the 1.7 kb PCR product. PCR products were sequenced directly and overlapping fragments were assembled using the SeqMan software of the LaserGene package. Multiple sequence alignments were performed using Prankster (Löytynoja & Goldman, 2008) including sequences from other liberibacters and the genome sequences of related taxa. Multiple methods of phylogenetic tree building were used; all were generally congruent. Bayesian inference trees are presented here as representative. Trees were constructed with MrBayes 3.12 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003) using the GTR+I+Γ model for DNA and the mixed model for amino acid sequences. Analyses were run for 5×10⁶ generations using a flat Dirichlet distribution (assuming no prior knowledge on taxa relations). The first 2×10⁶ generations before convergence of likelihoods were discarded from the analysis.

Phylogenetic analyses based on sequences of the 16S rRNA gene (Liefting et al., 2009), the deduced RplJ protein (Fig. 1⇓) and the β operon (Fig. 2⇓) show that isolate NZ082226 is unambiguously a member of ‘Candidatus Liberibacter’ and is distinct from the currently described candidate species and strains. The exact phylogenetic relationship to the other liberibacters is uncertain. In the tree based on β operon nucleotide sequences (Fig. 2⇓), isolate NZ082226 was located at the base with ‘Ca. L. americanus’; however, in the rplJ deduced amino acid sequence tree (Fig. 1⇓), it was located between ‘Ca. L. asiaticus’ and ‘Ca. L. africanus’. The posterior probability for this node is also low (0.55), indicating uncertainty. The sequence divergence of isolate NZ082226 may be the cause of this uncertainty, as a result of long-branch attraction (Bergsten, 2005). Pairwise nucleotide identity matrices of the rpoB sequence were performed using MatGat (Campanella et al., 2003). Isolate NZ082226 had 88.1 % identity to ‘Ca. L. asiaticus’, 89.1 % to ‘Ca. L. africanus’, 87.4 % to ‘Ca. L. africanus subsp. capensis’ and 82.5 % to ‘Ca. L. americanus’. Under the proposal of Adékambi et al. (2008), members of the same species have an rpoB gene sequence identity of >97.7 % and members of different genera show identity of <85.5 %. With these criteria, isolate NZ082226 represents a novel species in the same genus as ‘Ca. L. asiaticus’, ‘Ca. L. africanus’ and ‘Ca. L. africanus subsp. capensis’. However, this proposal would determine ‘Ca. L. americanus’ to be in a separate genus and, with an identity score of 92.1 %, ‘Ca. L. africanus’ and ‘Ca. L. africanus subsp. capensis’ would be considered different species.

The 16S rRNA gene and approximately 580 bp of the β operon including partial rplJ and rplL genes were sequenced from the liberibacters isolated from capsicum, potato, tamarillo, cape gooseberry and chilli. All sequences were identical to that of isolate NZ082226 from tomato.

Naming of the citrus-infecting liberibacter species is based on the continent where they were first detected and occurred. We propose that the novel liberibacter species identified in solanaceous plants is named after its plant host family rather than the geographical region. Although this liberibacter species was first identified in New Zealand, it has subsequently been detected in potato plants exhibiting zebra chip symptoms in the USA (Abad et al., 2009). Zebra chip disease of potato was first reported in Mexico in 1994 and then in the USA in 2000 (Secor & Rivera-Varas, 2004). Transmission studies by Munyaneza et al. (2007) strongly suggest that the tomato/potato psyllid Bactericera cockerelli is the vector of the zebra chip pathogen. B. cockerelli has a very wide host range, including species in 20 plant families, with a strong preference for solanaceous species (Wallis, 1955). Therefore, isolate NZ082226 could potentially infect a wider host range than the six solanaceous hosts identified to date.

Evidence suggests that the solanaceae-infecting liberibacter did not originate in New Zealand but came into the country with the psyllid vector. B. cockerelli was first discovered in an Auckland glasshouse tomato crop in May 2006, after which the plant disease symptoms caused by the liberibacter were first observed. Furthermore, the liberibacter has been detected by PCR in individuals from a colony of B. cockerelli started from one of the initial incursion sites [Peter Workman (New Zealand Institute for Crop & Food Research Ltd), personal communication].

The results of phylogenetic analyses obtained in the present and previous studies, along with the novel host-range, show that isolate NZ082226 represents a novel candidate species within the genus ‘Candidatus Liberibacter’, for which the name ‘Candidatus Liberibacter solanacearum’ is proposed. This is the first liberibacter species described from a non-rutaceous host.