Expanding and exploring the diversity of phytoplasmas from lucerne (Medicago sativa)

Phytoplasmas are a group of insect-vectored bacteria responsible for disease in many plant species worldwide. Among the crop species affected is the economically valuable forage species lucerne. Here we provide comprehensive molecular evidence for infection in multiple lucerne plants by a phytoplasma not previously known from this plant species. This phytoplasma had a >99% genetic similarity to an unclassified 16S rRNA subgroup previously reported as Stylosanthes little leaf from Stylosanthes spp. and was genetically and symptomatically distinct from a co-occurring but less common 16SrIIA group phytoplasma. Neighbour-joining analyses with publicly available sequence data confirmed the presence of two distinct phytoplasma lineages in the plant population. No PCR detections were made among 38 individuals of 12 co-occurring weed species. Sequence analysis revealed that all nine PCR detections from among 106 individuals of five Hemiptera insect species from the site, three of which had previously been reported as likely vectors, were false positives. This study demonstrates the importance of sequencing to complement PCR detection and avoid potentially inaccurate conclusions regarding vectors, highlights that sampling over a wide spatio-temporal scale is important for vector and alternative host studies, and extends to eight the number of phytoplasma 16 Sr groups known from lucerne.

Scientific RepoRts | 6:37746 | DOI: 10.1038/srep37746 known by common disease symptom names, and 16S rRNA subgroups are a more recent advance in nomenclature, there is no clear cut correspondence between these names and subgroups. Based on publicly available sequence accessions at GenBank (ncbi.nlm.nih.gov), seven 16 Sr phytoplasma groups (I, II, III, V, VI, VII, XII) are known to affect lucerne (Fig. 1). In Australia, the first report of a phytoplasma associated with lucerne was detection of sweet potato little leaf strainV4 (SPLL-V4), a 16SrII group phytoplasma, in a single specimen exhibiting little leaf symptoms from the Northern Territory 16 . A study of Australian lucerne yellows (AluY) disease, which has symoptoms distinct from little leaf 17 , used electron microscopy to visualise phytoplasma bodies in symptomatic plants and initial analysis of 16S rRNA revealed this strain to be closely related phylogenetically to Australian tomato big bud phytoplasma 18 , a 16SrII group phytoplasma 8 . More comprehensive study identified the phytoplasma associated with ALuY disease to be within the faba bean phyllody phytoplasma 16SrII group 19 . Later work on Australian lucerne yellows disease reported a second phytoplasma, 16SrXII-B subgroup strain from lucerne in New South Wales 20 . Over the intervening decade there has been no further report of phytoplasmas from lucerne in Australia, meaning that only two of the seven internationally reported 16Sr groups are known. It is important to establish a more complete understanding of this to inform biosecurity, most particularly the possibility of incursions by additional exotic stains. More generally, there is a lack of integrative studies of the complex of phytoplasma types that affect lucerne globally.
Knowledge of a pathogen, including its genetic diversity, is fundamental to epidemiological understanding and rational management practices. In the case of phytoplasmas, information on the insects capable of vectoring the pathogen, and on any weed species that may constitute alternative hosts, is also key. Accordingly, this study aimed to study the genetic diversity of phytoplasmas from a population of lucerne plants in a single field and to place this into a global context. To maximise the chances of detection, tissue samples were taken from plants exhibiting chlorosis or witches' broom symptoms, whilst corresponding samples were also taken from the nearest healthy neighbouring lucerne plant, as well as from 38 weeds of 12 species growing within the lucerne field. A total of 106 individual Hemiptera insects that were potential vectors were also sampled from the field. All samples were processed in the laboratory using relevant molecular techniques to identify the insects and to detect and characterise phytoplasmas. Scale bar equals 1% equal weighted sequence difference. Cluster node supports > 70% (10,000 bootstrap replicates) as indicated. Terminal 16S rRNA subgroups (refer Methods) collapsed as clusters containing multiple accessions. Tip labels indicate 16S rRNA subgroups, provisional Candidatus Phytoplasma species, and (in parentheses) associated phytoplasma strain or disease acronym. Multiple provisional species in clusters indicated as "spp. ", unknown species as "sp?". Shaded red squares, blue squares, and blue triangles indicate phytoplasma detected in lucerne overseas, Australia and Forbes respectively. Refer Fig. 2  Results PCR detection of putative phytoplasmas. Serial PCR using 2nd stage primers fU5 and m23sr tested positive for putative phytoplasma presence in nine lucerne plants. These nine plants displayed outward symptoms of phytoplasma infection and also tested positive in serial PCR using an alternative 2nd stage primer set 16r758F and m23sr (Table 1). Two additional lucerne plants (one of these symptomless) tested positive using this alternative primer set. Nine of the insects tested PCR positive (Table 1), two using primer set fU5 & m23sr and nine using 16r758F & m23sr. None of the weeds tested PCR positive using either primer set.
Sequence analysis of putative phytoplasma positives. Sequence queries at GenBank using BLAST confirmed phytoplasma identities for the nine lucerne samples shown to be PCR positive using 2nd stage primer set fU5 & m23sr, and alternative primer set 16r758F and m23sr. Sequences of the two additional PCR positive lucerne samples detected using the alternative primer set were matched to bacteria other than phytoplasma (Supplementary Table S1). PCR positives detected for all nine insects were false positives, matching a variety of non-phytoplasma bacteria (Supplementary Table S2 of two distinct genetic lineages of phytoplasma differing by ~9.76% among the lucerne positives. Two specimens (ww18841, ww18842) were nested in the "Ca. P. aurantifolia" species cluster containing 16SrII-A subgroup phytoplasma strains associated with a variety of diseases and plant hosts (Fig. 2). The two specimens more closely matched sequences in 16SrII-A subgroup (associated with multiple strains of witches' broom and virescence in several host plants), than to other 16SrII subgroup strains accessioned to lucerne, including LYSP-E2 (accession JX861231), ALuY (accession AJ315965 19 ) strain in Australia, and a variety of other lucerne disease strains in the Middle East and Europe.
The 2nd phytoplasma lineage detected was from seven lucerne plants and clustered with an undescribed 16S rRNA subgroup previously identified to Stylosanthes little leaf phytoplasma (StLL) previously reported from Stylosanthes legumes [21][22][23] . StLL is unusual among phytoplasmas in containing two independent 16S-23S rRNA operons, one lacking the internal tRNA Ile gene that is normally present and characteristic in all other phytoplasmas 21 . Absence of this tRNA Ile gene from a phytoplasma 16S-23S rRNA-encoding operon has not been reported elsewhere. Sequences here identified both presence and absence of the tRNA Ile gene among samples identified to the StLL lineage, in agreement with prior reports for this phytoplasma.
Phytoplasma subgroup 16SrXII-B strain previously reported infecting NSW lucerne 20 and seven other 16 S rRNA subgroups infecting lucerne outside of Australia (Fig. 1), were not evident among the samples tested in this study.

Discussion
Several genetically distant 16Sr groups of phytoplasma have been reported as the potential etiological agents leading to disease symptoms in lucerne described as "yellows" and "witches' broom". For example, in Australia, Lucerne yellows disease (ALuY), has been associated with separate phytoplasma 16S rRNA groups II 19 and XII 20 , respectively. Worldwide, seven distinct 16S rRNA subgroups have been reported in lucerne symptomatic for yellows and witches broom related diseases. Cataloguing the diversity of phytoplasma sequence strains and 16Sr groups present in lucerne, their associations with disease symptoms and their host and vector arthropod specificity is a necessary primary step for evidence-based biosecurity measures and management.
Here we report molecular genetic evidence of two distinct phytoplasmas detected among lucerne plants. Lucerne symptomatic in the field for witches' broom had > 99.85% sequence similarity to various near identical 16S rRNA subgroup II-A strains associated with diseases in other, botanically unrelated crops including sweetpotato little leaf (SPLL-V4). 16SrII-A strains such as SPLL-V4 have been previously reported in lucerne sampled in Australia 16,24 . Other 16SrII subgroups ascribed to the Ca. P. aurantifolia species, contain a broad variety of sequence strains associated with lucerne diseases including Alfalfa witches broom (in the Middle East and Europe) 25 and Australian lucerne yellows (ALuY & LYSP-E2) 26,27 ; each of which is marginally less similar in sequence identity to the present study's specimens than are the aforementioned 16SrII-A accessions.
Present evidence of 16SrII-A in lucerne, confirms broad host use by this subgroup of phytoplasma, reported previously from tomato and eggplant (Solanaceae), sweetpotato (Convolvulaceae), and four taxonomically diverse weeds (Alysicarpus sp. (Fabaceae), Amaranthus sp. (Amaranthaceae), Passiflora foetida (Passifloraceae) and Evolvulus sp. (Convolvulaceae) in northern Australia 28 . Amaranthus was noted by Gibb, et al. 28 as potentially of epidemiological significance for spread of this phytoplasma to crops in northern Australia because it grows in close association with sweetpotato and supports high levels of the supposed vector insect O. argentatus. Fletcher, et al. 29 provide evidence that O. argentatus is a valid species and distinct from O. orientalis with which it had previously been synonymised by Kwon and Lee 30 .
The other phytoplasma strain detected in the present study was associated with symptoms of yellowed/ stunted leaves in lucerne plants, and genetically identified (> 99.9% sequence similarity) to the ungrouped but phylogenetically unique Stylosanthes little leaf [StLL] phytoplasma previously identified in Stylosanthes legumes 23 . The discovery here of the StLL phytoplasma present in lucerne raises the number of 16Sr groups reported from this important crop to three within Australia and to eight globally. Significantly, lucerne and other legumes in the genus Stylosanthes share at least two disparate 16S rRNA subgroups indicating the two legume genera have shared susceptibility to distantly related phytoplasmas. It remains to be determined if the plants are also hosts to a common assemblage of vector insects infective for the two phytoplasma subgroups.
In our survey, phytoplama presence was not detected among any of the phloem feeding insects sampled in the vicinity of the infected lucerne. This is surprising given that our insect sample was dominated by presence of A. torrida, a leafhopper species previously identified as a potential vector of lucerne phytoplasma 31 , and a lesser number of other putative vector species (O. argentatus and O. orientalis). The presence of false positives among the insects suggests that there were no deficiencies in sample preparation and processing; rather that phytoplasmas were absent. This indicates that the collection of Hemiptera species (even those present in large numbers) from host plants that are proven to be infected by phytoplasmas (two diverse strains) is not evidence for vector status. Clearly sampling over a larger spatio-temporal scale will be necessary to establish which insect species may test PCR positive though transmission tests are necessary to provide definitive evidence of vector capacity 32 .
False positives were observed in each of the two independent serial PCR assays using different forward primers in 2nd stage PCR. The effect was more prevalent when primer 16r758F was used as an alternative forward primer to fU5 in 2nd stage PCR. Optimization of PCR annealing temperature above that used here may increase targeted stringency to phytoplasma amplification and eliminate false positive presence, but also potentially result in increased frequency of false PCR negatives in instances where there is nucleotide variation among phytoplasma strains at primer annealing sites. Regardless, the presence of false positives in serial PCR indicates sequencing is a necessary requisite for confirming phytoplasma presence and identity when PCR positives are detected. Alternatives to sequencing such as restriction digest profiling of PCR positives may be expedient for confirming presence of phytoplasma and even subgroups of phytoplasma, but has limited capability for detection of novel phytoplasma varieties.

Methods
Sampling. Leaf samples from 64 plants from an agricultural field site at Forbes New South Wales (NSW), Australia (− 33.381S, 147.976 E) sampled Feburary 2013 were genetically tested for phytoplasma presence and identity (Table 1 and Supplementary Table S1). Samples included lucernes (N = 14) and weeds (N = 6) with witches' broom symptoms of leaf chlorosis, stunting and/or leaf bunching. Neighbouring asymptomatic lucernes (N = 12) and various weeds (N = 32) were also sampled, as were phloem feeding Hemiptera (N = 106) captured using sweep nets. Specimens were individually catalogued with unique specimen ID labels, preserved in > 70% ethanol, and curated at NSW Department of Primary Industries agricultural institutes in Orange (insects) and Wagga Wagga (plants).
Non-destructive DNA extraction. DNA extraction was preceded by a non-destructive tissue digestion.
Whole insects, and plant leaf laminar (< 0.2 g), were individually placed in 480 μ L aliquots of DXT tissue digestion buffer (QIAGEN, Doncaster, Australia) incorporating 1% DX digestion buffer additive (QIAGEN) and digested overnight at 55 °C. Specimens were later removed from the digests and stored in 70% ethanol. DNA was extracted from 240 μ L of each specimen digest using a Corbett Research 1820 X-tractor Gene robotic system and associated DNA extraction kit reagents (QIAGEN). DNA was eluted to 150 μ L and stored at − 20 °C. PCR for DNA barcoding 33 of insects targeted a 667 base pair (bp) portion of the 5′ mitochondrial cytochrome c oxidase I (COI) gene using primers described in Fletcher, et al. 29 . PCR products were visualized by UV trans-illumination after electrophoresis through a 1.5% agarose gel in 1% TAE buffer containing SYBR ® Safe DNA gel stain (Invitrogen TM ), and qualitatively checked for expected fragment size against E-Gel size marker (Invitrogen TM ). PCR products were sent to the Australian Genome Research Facility (Brisbane) for purification and bidirectional sequencing using an Applied Biosystems 3730xl DNA Analyzer.
PCR for positive/negative detection of infective phytoplasma genes in host plants and insects targeted amplification of the partial 16S-23S nuclear ribosomal DNA gene region (16S-23 S rRNA) containing the complete tRNA Ile gene and surrounding intergenic spacer regions. A two-stage serial PCR procedure modified from Pilkington, et al. 19 was used to enhance amplification of phytoplasma gene targets, given the expected low titre of bacterial DNA in host specimen extracts 34 . The 1st stage PCR amplified > 1.8-kbp product using primers P1 35 and P7 36 . 1st stage PCR products were diluted (1:100) and used as templates in 2nd stage PCR with primers fU5 37 and m23Sr 38 to amplify an internal > 1.4-kbp product. 2nd stage PCR was repeated substituting primer 16r758F 28 for primer fU5, to amplify a smaller portion (> 1.0-kbp) of the 16S-23S rRNA region. PCR reactions were prepared as described earlier for DNA barcoding with the exceptions of primers used, and an increased PCR annealing temperature set at 55 °C. 2nd stage PCR products were visualized after electrophoresis as described earlier, and positives in expected size ranges were sequenced (as described earlier). In several instances dual PCR products in the expected range, but separated by approximately 100 bp, were observed in individual specimen PCR. In these instances, 2nd stage PCR products were re-run through a 1% agarose gel in 1% TBE buffer to allow size separation of the dual products. Size separated products were excised from the gel and individually purified of agarose using QIAquick Gel Extraction Kit (QIAGEN) prior to their independent sequencing. Sequence analyses. Forward and reverse sequence chromatograms were assembled to specimen ID, primer truncated and checked for signal quality using Lasergene SeqMan Pro ver. 8.1.0(3) (DNASTAR Inc., Maddison, WI, USA). Phytoplasma PCR positive sequences were re-aligned for indel positions using MUSCLE 39 implemented in MEGA version 6 40 .
Insect DNA barcode sequences were queried for species identity (20 June 2016) at the Barcode of Life Data systems 41 online sequence repository. BOLD specimen sequences with > 98% sequence similarity to our query sequences were considered conspecific. Insect specimen records and DNA barcode sequences are available as a BOLD dataset (http://dx.doi.org/10.5883/DS-AUHEMI01).
Sequences of phytoplasma-positive PCRs were queried for identity against GenBank (and EMBL) sequence accessions using the online NCBI BLAST tool. 16S rRNA phytoplasma accessions matched at > 99% similarity and coverage to query sequences were included in an alignment containing accessions (Supplementary Table S3) representative of 16S rRNA subgroups and 37 provisionally identified 'Candidatus Phytoplasma' species 42 , and accessions reported in Medicago sativa (lucerne, alfalfa). The alignment (N = 172) was truncated to 1171nt (including gaps) corresponding to positions 421-1527 of accession FJ943262 (Candidatus Phytoplasma australiense strain NZ09156). Pair-wise % genetic distances (sites equally weighted; missing sites and indels excluded) between sequences were compared as a neighbor-joining (NJ) tree 43 using MEGA version 6 40 . Support values for NJ clusters were estimated by non-parametric bootstrapping (10 000 replicates). Representative phytoplasma sequences from field collected specimens were deposited at GenBank under accession records KX421793-KX421797.