Background High-throughput sequencing (HTS) has been applied successfully for virus and viroid discovery in many agricultural crops leading to the current drive to apply this technology in routine pathogen detection. The validation of HTS-based pathogen detection is therefore paramount. Methods Plant infections were established by graft inoculating a suite of viruses and viroids from established sources for further study. Four plants (one healthy plant and three infected) were sampled in triplicate and total RNA was extracted using two different methods (CTAB extraction protocol and the Zymo Research Quick-RNA Plant Miniprep Kit) and sent for Illumina HTS. One replicate sample of each plant for each RNA extraction method was also sent for HTS on an Ion Torrent platform. The data were evaluated for biological and technical variation focussing on RNA extraction method, platform used and bioinformatic analysis. Results The study evaluated the influence of different HTS protocols on the sensitivity, specificity and repeatability of HTS as a detection tool. Both extraction methods and sequencing platforms resulted in significant differences between the data sets. Using a de novo assembly approach, complemented with read mapping, the Illumina data allowed a greater proportion of the expected pathogen scaffolds to be inferred, and an accurate virome profile was constructed. The complete virome profile was also constructed using the Ion Torrent data but analyses showed that more sequencing depth is required to be comparative to the Illumina protocol and produce consistent results. The CTAB extraction protocol lowered the proportion of viroid sequences recovered with HTS, and the Zymo Research kit resulted in more variation in the read counts obtained per pathogen sequence. The expression profiles of reference genes were also investigated to assess the suitability of these genes as internal controls to allow for the comparison between samples across different protocols. Conclusions This study highlights the need to measure the level of variation that can arise from the different variables of an HTS protocol, from sample preparation to data analysis. HTS is more comprehensive than any assay previously used, but with the necessary validations and standard operating procedures, the implementation of HTS as part of routine pathogen screening practices is possible.

Greening disease of citrus is a serious disease known in South Africa since the late 1920s. In South Africa, it is associated with infection by ‘Candidatus Liberibacter africanus’, a heat sensitive, phloem-limited, noncultured alpha-proteobacterium. Huanglongbing (HLB), a similar, but more devastating disease that was described initially from China but which now occurs in several citrus producing countries, is associated with a different Liberibacter species, ‘Ca. L. asiaticus’. A ‘Ca. L. africanus’ subspecies, ‘Ca. L. africanus subsp. capensis’, has been found only in South Africa infecting an indigenous Rutaceous species, Calodendrum capense (Cape Chestnut), in the Western Cape in 1995. The discovery of a new Liberibacter species in Brazil, ‘Ca. L. americanus’, and the spread of ‘Ca. L. asiaticus’ to a number of additional countries over the last few years prompted us to assess whether only ‘Ca. L. africanus’ is present in commercial citrus orchards in South Africa. Samples displaying greening or similar symptoms were collected from 249 citrus trees from 57 orchards distributed throughout the greening affected citrus production areas of South Africa. Multiplex polymerase chain reaction (PCR) was performed on DNA extracts to detect the known citrus Liberibacters. Amplicons were obtained from 197 samples. None of the samples yielded a 1,027-bp amplicon indicative of ‘Ca. L. americanus’ infection. The amplicons of 84 samples were sequenced, and all were identical to the cognate ‘Ca. L. africanus’ Nelspruit sequence in GenBank. No instance of ‘Ca. L. asiaticus’ or ‘Ca. L. africanus subsp. capensis’ sequence was found. Geographically representative samples that tested negative for Liberibacter also tested negative for phytoplasmas based on real-time PCR results. Based on the results of this survey, it is concluded that to date only ‘Ca. L. africanus’ is associated with citrus greening in commercial citrus in South Africa.

Genotypic characterization of Citrus tristeza virus (CTV) strains has progressed significantly, but their phenotypic expression is poorly established as CTV naturally occurs as mixed‐strain populations. A screening system for the analysis of mixed‐strain populations is required for population studies and the correlation with symptom expression. In this study, a published CTV strain‐specific detection assay was expanded and improved to facilitate detection of currently known CTV strains. Supplementary RT‐PCR assays were developed for two variant groups of the RB strain and the HA16‐5 strain, and assays for the T36 strain and generic CTV detection were improved. The value of the strain‐specific assays was shown by the ability to identify the strain components of two CTV cross‐protecting sources, GFMS35 and LMS6, used in the South African budwood certification scheme and to demonstrate the segregation of strains in budwood source trees.

Citrus tristeza virus (CTV) is endemic to southern Africa and the stem pitting syndrome that it causes was a limiting factor in grapefruit production prior to the introduction of cross-protection in the Citrus Improvement Scheme. This disease mitigation strategy, using various field-derived CTV sources, has significantly extended the productive lifespan of grapefruit orchards in South Africa. CTV commonly occurs as a population of various strains, masking the phenotypic effect of individual strains. Likewise, current South African CTV cross-protection sources are strain mixtures, obscuring an understanding of which strains are influencing cross-protection. The severity of various CTV strains has mostly been assessed on sensitive indicator hosts, but their effect on commercial varieties has seldom been investigated. Single-variant CTV isolates were used to investigate the phenotypic expression of CTV strains in commercial grapefruit varieties as well as CTV indicator hosts. They were biologically characterized for their ability to cause stem pitting and their rate of translocation and titer in the different hosts, monitored by enzyme-linked immunosorbent assay. Complete genome sequences for three CTV strain variants were generated. Isolates of CTV strains VT, T68, RB, and HA16-5 did not induce severe stem pitting in four grapefruit hosts in a glasshouse trial. Viral titers of the strains differed in the grapefruit hosts, but the RB isolate reached a higher titer in the grapefruit hosts compared with the VT, T68, and HA16-5 isolates. Additionally, horticultural assessment of two grapefruit varieties inoculated with the RB isolate in two field trials demonstrated that mild stem pitting did not negatively influence the horticultural performance of the grapefruit trees over an eight-year assessment period. ‘Star Ruby’ trees containing the CTV source GFMS35 showed less stem pitting than trees inoculated with the RB isolate, but had smaller canopy volumes and lower yields than trees containing the RB isolate. This suggests that the influence of CTV sources on tree performance is not limited to the effect of stem pitting.

Two isolates of the T68 genotype of citrus tristeza virus (CTV) were derived from a common source, GFMS12, by single aphid transmission. These isolates, named GFMS12-8 and GFMS12-1.3, induced stem pitting with differing severity in ‘Duncan’ grapefruit (Citrus × paradisi [Macfad.]). Full-genome sequencing of these isolates showed only minor nucleotide sequence differences totaling 45 polymorphisms. Numerous nucleotide changes, in relatively close proximity, were detected in the p33 open reading frame (ORF) and the leader protease domains of ORF1a. This is the first report of full-genome characterization of CTV isolates of a single genotype, derived from the same source, but showing differences in pathogenicity. The results demonstrate the development of intragenotype heterogeneity known to occur with single-stranded RNA viruses. Identification of genetic variability between isolates showing different pathogenicity will enable interrogation of specific genome regions for potential stem pitting determinants.