Characterization of worldwide and regional 'Ca. L. asiaticus' populations from HLB-affected plants can facilitate identification of introduction patterns and predict the possible relationships of HLB-associated Liberibacters found in different citrus growing regions. Multilocus microsatellite marker analysis can provide sufficient resolution for differentiating closely-related isolates and can be useful for tracking genotypes of interest; additionally, these markers may help identify the source of invasive strains.
In this study, seven microsatellite markers successfully genotyped 'Ca
. L. asiaticus' from global populations. Sequence analysis indicated that three of the microsatellites appear to overlap with microsatellites recently developed by Katoh et al. [20
]. Various microsatellite length variations were found in 'Ca
. L. asiaticus' from worldwide collections, with some loci having as many as 30 alleles.
Historical evidence reviewed by da Graça [25
] suggested that HLB was observed in Guangdong province, China in the late 19th century [26
], and later spread to other parts of the country. It is assumed that HLB may have been introduced into China from India along sea trade routes [27
]. The first record of HLB-like symptoms, referred to as 'dieback', was reported from India in the 18th century [28
]; this was later suggested to be HLB [29
]. As 'Ca
. L. asiaticus' has been in Asian countries over a century, the genetic diversity in Asian populations was expected to be high, due to a longer period of mutation accumulation, population differentiation and natural selection. As hypothesized, a higher degree of genetic diversity for 'Ca
. L. asiaticus' was observed in both China and India within the present study (Table ). In contrast, the lower level of allelic and haploid genetic diversity of 'Ca
. L. asiaticus' in Florida and Brazil populations are consistent with the hypothesis that 'Ca
. L. asiaticus' populations in these regions have been derived from recent introductions [30
Human movement of infected plant materials is probably the main cause of long distance dissemination of both 'Ca
. L. asiaticus'-positive psyllids and HLB-affected plant material. The distributions of haplotypes observed in 'Ca
. L. asiaticus' in this study did not detect any identical haplotypes from different continents or even from different countries within the same continent (Additional file 1
). This result does not exclude the possibility of contemporary migration of 'Ca
. L. asiaticus' among different countries through the movement of infected plant materials or by the migration of vector psyllids as rapid mutation and selection could lead to deviation of populations from their original sources. The vector, D. citri
, has been in Brazil for over 60 years without any sign of HLB until its discovery on 2004 [4
]. D. citri
was discovered in Florida in Palm Beach, Broward and Martin counties in 1998 and has spread throughout the state since that time [7
]. However, it is not clear when 'Ca
. L. asiaticus' was introduced into Brazil and Florida. The rapid spread of HLB in Brazil and Florida after the first reports in 2004 and 2005, respectively, suggests that HLB was most likely introduced around that time, rather than evolving from an indigenous source.
As suggested by the historical evidence and review of the early literature related to HLB, the most ancient population of 'Ca
. L. asiaticus' perhaps originated in India. From the 20th
century onward, HLB spread through much of the citrus-growing regions of south and southeast Asia [2
], the Arabian peninsula [31
], East Timor and Papua New Guinea [32
], and the western hemisphere (Brazil and the United States) [1
]. It is difficult to precisely know when the disease entered each country and from where it was introduced. Frequent shipment of plant materials and unlawfully importation of plants has increased the risk of disseminating exotic plant pathogens around the world. The exact pathways responsible for introducing HLB and the Asian citrus psyllid into the United States and Brazil have not yet been determined.
The genetic relationships of the isolates in this study, as determined from the UPGMA based on Nei's genetic distance [22
] and individual based clustering analysis by the STRUCTURE analyses, consistently identified three major genetic groups of 'Ca
. L. asiaticus', with isolates from India included in a distinct genetic group (Figure and Figure ). The similar genetic makeup amongst most isolates from east-southeast Asia and South America (São Paulo, Brazil) support the hypothesis of the introduction of 'Ca
. L. asiaticus' into South America from East Asia or Southeast Asia. While most isolates from Florida were clustered within a separate group, both UPGMA and STRUCTURE analyses showed that some isolates from central Florida overlapped with east-southeast Asian and Brazilian groups. The presence of two genetic groups in Florida suggests at least two introduction events are associated with the recent outbreak of HLB in Florida.
Based on the history of HLB, it could be predicted that populations of 'Ca. L. asiaticus' in Florida were most likely established following the introduction of HLB-affected plant materials or 'Ca. L. asiaticus'-carrying psyllid from Asia or other countries through human-mediated transport. The analyses in this study do not support the hypothesis of introduction of HLB into the Americas through biological materials sourced from India. Only a single isolate from India (Prakasam District, Andhra Pradesh) overlapped with the east-southeast Asian and Brazilian group (Figure , red). STRUCTURE analysis revealed that less-dominant clusters (Figure , red) in central Florida (Polk, Pasco, and Lake Counties) were observed in the same lineage (q ≥ 0.90) with east-southeast Asian and Brazilian clusters suggesting that the origin of members of this cluster in Florida might be derived from Asia or via Brazil. Moreover, some admixed (q < 0. 90) isolates between Florida and east-southeast Asia also support the hypothesis of introducing 'Ca. L. asiaticus' into Florida from Asia.
eBURST analysis provided further insights into the origin of 'Ca
. L. asiaticus'. Founder haplotypes were identified from China, Brazil, and India. Based on their position within the eBURST network, these founders are predicted to have given rise to the three global genetic groups, consistent with prevailing theories of the geographic origins of HLB [1
]. While one founder type was predicted in Brazil, the similar genetic makeup of Brazilian and east-southeast Asian isolates suggest that this founder could have been introduced into Brazil from any of these Asian countries. Consistent with the STRUCTURE analysis, the eBURST diagram also predicted the introduction of 'Ca
. L. asiaticus' into Florida citrus groves through at least two separate introduction events. While a primary network was detected between a founder haplotype from China and two unique haplotypes in Florida, clear differentiation was observed between most isolates from China and Florida by Bayesian clustering and UPGMA analyses. Differences between the dominant groups found in Florida and China were also reported in a recent study using a single VNTR locus [21
]. It is uncertain whether the dominant group of Florida isolates were introduced en masse
or if a small population of nearly-identical 'Ca
. L. asiaticus' haplotypes from China were introduced, evolved quickly, and established a large population. The recent discovery and rapid spread of HLB in Florida, along with wide distribution of dominant 'Ca
. L. asiaticus' group observed in the present study suggests that isolates of this group have been directly introduced from an unknown location. Another recent study also indicated that some isolates of 'Ca
. L. asiaticus' from Florida may have been introduced through two different events, and sources were unknown [21
]. The analyses of microsatellites in the present study, however, suggest that the introduction of the less-dominant cluster was likely from a single source either Asia or Brazil. The low occurrence of less dominant group in some central counties in Florida suggests that the members of this group were perhaps introduced more recently (Figure ). However, it is certainly plausible that these two haplotypes were introduced into Florida at nearly the same time. Isolates from one of the sources may have spread quickly due to selective advantage under a favorable set of biological or environmental conditions.
Figure 4 Sample distribution of 'Candidatus Liberibacter asiaticus' from 15 citrus-growing counties (gray highlighted) in Florida, USA. Green circles indicate the counties where only the dominant 'Ca. L. asiaticus' group were observed based on STRUCTURE analysis (more ...)
Our analysis showed that a dominant group of 'Ca
. L. asiaticus' genotypes are widely distributed in south-central Florida (Figure ). HLB is highly invasive and spatial spread of the disease is rapid compared to other arboreal pathosystems [4
]. Increased diversity of host, pathogen, vector, and environmental conditions likely influence the rates of HLB distribution. Moreover, the rates of HLB increase are directly related to increase and spread of the psyllid vector population: in June 1998, the Asian citrus psyllid was first detected in Palm Beach County; within two years of this discovery the disease occurred to 31 counties in Florida [33
]. The vector is now present in nearly all citrus-growing areas of Florida [34
]. In Florida, HLB was first discovered in Miami-Dade County in August 2005, seven years after detection of the vector in the same region [30
]. By mid-October, the disease was found in many residential properties stretching northwards more than 250 km from Miami-Dade County to St. Lucie County and several commercial citrus groves were also affected in Palm Beach and Hendry Counties [1
]. However, no epidemiological survey has clearly demonstrated HLB or 'Ca
. L. asiaticus'-carrying psyllids being introduced into the southern part of Florida and then spreading northward through the continuous movement of psyllid vectors. Since 2005, HLB has spread to most citrus-producing counties in Florida [34
]. The rapid and widespread distribution of this disease among citrus growing counties in Florida is most likely due to the result of the multiple secondary introductions of HLB-associated 'Ca
. L. asiaticus'. Based on the present analyses, it appears that there were at least two 'Ca
. L. asiaticus' introduction events in Florida. Moreover, the rapid distribution of HLB within Florida after 2005 is concomitant with the discovery of a dominant genetic cluster within south-central Florida. Taken together, this suggests that dominant 'Ca
. L. asiaticus' haplotypes, possibly from different countries may have established a population within Florida through multiple introduction events.