Geographic Distribution, Genetic Variability and Biological Properties of Rice Orange Leaf Phytoplasma in Southeast Asia

Rice orange leaf phytoplasma (ROLP) causes clear orange to yellowish leaf discoloration and severe stunting in rice seedlings. The ecological and biological characteristics of ROLP are largely unknown because the disease has not widely caused serious problems in rice cultivated areas, thereby leading to the low accumulation of research data. However, in the past decade, the disease became a threat to rice production, particularly in South China and India; it has also been recognised in other Asian countries, such as Vietnam, Thailand and the Philippines. Here, we observed the occurrence of ROLP in paddies of the Southeast Asian counties (Cambodia, Vietnam and the Philippines) and found that the isolates in the Philippines and Vietnam were monophyletic, while those in India, Thailand and Cambodia were more diverse, suggesting their potential origins. In Cambodia, it was revealed that following polymerase chain reaction (PCR) detection, the known ROLP-insect vectors, N. virescens Distant and Recilia dorsalis Motchulsky, were ROLP-positive, indicating their roles in pathogen dispersal. Moreover, fluorescent and scanning electron microscopy revealed the intensive accumulation of the phytoplasma in phloem tissues and massive accumulation of storage starch in vascular bundle sheath and parenchyma. Altogether, this study illustrated the genetic variability of global ROLP isolates and the pathogen’s biological impact on rice tissue.


Introduction
Phytoplasma is a group of bacteria that are unculturable in vitro due to them lacking a cell wall, localised in the phloem of diverse host plants (more than 1000 species), vectored by phloem-feeding insects and taxonomically classified in the genus "Candidatus Phytoplasma", which accommodates 44 species separated based on 16S ribosomal DNA (rDNA) sequences [1,2]. Rice (Oryza sativa) is known to be affected by two phytoplasmas:

Field Survey of ROLP in the Philippines, Vietnam and Cambodia
From 2015 to 2019, we conducted field surveys in the Philippines (four provinces) [18], Cambodia (five provinces) and Vietnam (two provinces). Some paddies were seen to be affected by ROLP (Figure 1a), as rice plants exhibiting yellowish (Figure 1b) or golden leaf discoloration were sporadically present (Figure 1c,d). Causal correlation between these symptoms and ROLP was roughly assessed by nested PCR or loop-mediated isothermal amplification (LAMP) or both (Figure 2a,b), and the following sequence analyses. ROLP was detected from all provinces where field surveys were conducted, and the phytoplasma detection rate of analysed samples had varying percentages in each country (48.7-88.2%), suggesting the wide occurrence of ROLD in Southeast Asian countries ( Table 1). In most of the sites where ROLD-suspected rice samples were collected, ROLP was present and detected at a rate of over 50%, except for Angkorchey district, Takeo province in Cambodia (3/11, 27.3%) and Tan Hong district, Dong Thap province in Vietnam (5/15, 33.3%) ( Table 1). Although severe damages have not been observed in the paddies in general, rice plants in a paddy of Svay Rieng province in Cambodia showed lethal symptoms (Figure 1e). rDNA in three countries-Vietnam, the Philippines and Cambodia-to understand the genetic variations of global ROLP isolates and their geographic distributions. Furthermore, the population density of insect vectors in Cambodian paddies was monitored, and the rate of ROLP acquisition by insects was assessed. In addition, altered plant condition of ROLD-affected rice was analysed microscopically. Obtained results provide an insight into the nature of the pathogen, Candidatus P. asteris, in rice.

Field Survey of ROLP in the Philippines, Vietnam and Cambodia
From 2015 to 2019, we conducted field surveys in the Philippines (four provinces) [18], Cambodia (five provinces) and Vietnam (two provinces). Some paddies were seen to be affected by ROLP (Figure 1a), as rice plants exhibiting yellowish (Figure 1b) or golden leaf discoloration were sporadically present (Figure 1c,d). Causal correlation between these symptoms and ROLP was roughly assessed by nested PCR or loop-mediated isothermal amplification (LAMP) or both ( Figure  2a,b), and the following sequence analyses. ROLP was detected from all provinces where field surveys were conducted, and the phytoplasma detection rate of analysed samples had varying percentages in each country (48.7-88.2%), suggesting the wide occurrence of ROLD in Southeast Asian countries (Table 1). In most of the sites where ROLD-suspected rice samples were collected, ROLP was present and detected at a rate of over 50%, except for Angkorchey district, Takeo province in Cambodia (3/11, 27.3%) and Tan Hong district, Dong Thap province in Vietnam (5/15, 33.3%) ( Table 1). Although severe damages have not been observed in the paddies in general, rice plants in a paddy of Svay Rieng province in Cambodia showed lethal symptoms ( Figure 1e).      Table S1. d Partial sequence (>1.0 kbp) was obtained but not included in phylogenetic analysis ( Figure 3). e Averaged detection rates in a country or in three countries are shown. f Sequence analysis was performed in this study. g Re-sequenced in this study.

Phylogenetic Analysis of ROLP Based on 16S rDNA Sequences
Representatives of nested PCR products, as shown in Figure 2a, were subjected to Sanger sequencing, and obtained sequences were analysed accordingly by basic local alignment search tool (BLAST) searching and were found to be of the ROLP 16S rDNA. Selected ROLP isolates for sequence and phylogenetic analyses are shown in Table 1 and  Supplementary Table S1. Multiple alignments of these DNA sequences were created exclusively from partially determined sequences and subjected to a maximum-likelihood phylogenetic analysis. The result is visualised as a mid-point rooted tree with the J strain of rice yellow dwarf phytoplasma (RYD-J) as an out group in Figure 3. Publicly available sequences of Thai isolates (green) were the most divergent, but formed a single large cluster with all ROLP isolates; however, they were separated into three subgroups, (1) a probable ancestor-like group (Thai-A type), (2) a potential eastwardly transferred group (Mekong-Ph type) and (3) a possible progeniture-like group with genetic variability (Thai-B type). The Thai-A type consists of only four isolates from Thailand. Interestingly, all 11 Philippines isolates and all four Vietnam isolates are grouped with a singly reported Chinese isolate sequence. These sequences are identical to each other and other partial sequences of isolates from Dong Thap province, Vietnam, and the rest of those in the Philippines (Table 1). These are most closely related to a Thai isolate RPKB2-5 from Kanchanaburi and form the Mekong-Ph type subgroup. This subgroup includes 10 out of 16 Cambodian isolates as well. In the Thai-B type subgroup, six Cambodian and four Indian isolates are scattered in the cluster, formed by 25 out of 30 Thai isolates.  Avg. detection rate: 80% e a District or Municipality in the provinces are shown. b Number of ROLP-positive plants per number of total plants, which were examined using PCR or LAMP. c 1.2 kbp PCR fragments of 16S rRNA gene were sequenced and their accession numbers are listed in Table S1. d Partial sequence (>1.0 kbp) was obtained but not included in phylogenetic analysis ( Figure 3). e Averaged detection rates in a country or in three countries are shown. f Sequence analysis was performed in this study. g Re-sequenced in this study.

Figure 3.
Maximum-likelihood phylogenetic analysis of 16S ribosomal DNA (rDNA) sequences of global ROLP isolates. The ROLP 16S rDNA sequences obtained by the nested PCR products with the 2nd PCR primer set (R16F2n and R16R2) were phylogenetically analysed with ROLP sequences deposited in GenBank. The origin of isolates is colour-differentiated as follows: Thailand (green), the Philippines (violet), Cambodia (orange), China (red), India (pink) and Vietnam (blue). Stars indicate the first isolates with 16S rDNA sequence from the Philippines (ROL) and with draft genome sequence from China (LD1). Accession numbers of used sequences are listed in Supplementary Table S1. The ROLP 16S rDNA sequences obtained by the nested PCR products with the 2nd PCR primer set (R16F2n and R16R2) were phylogenetically analysed with ROLP sequences deposited in GenBank. The origin of isolates is colour-differentiated as follows: Thailand (green), the Philippines (violet), Cambodia (orange), China (red), India (pink) and Vietnam (blue). Stars indicate the first isolates with 16S rDNA sequence from the Philippines (ROL) and with draft genome sequence from China (LD1). Accession numbers of used sequences are listed in Supplementary Table S1.

Insect Population Dynamics and Detection of ROLP from Field-Collected Insects
Zig-zag leafhopper (ZLH, R. dorsalis) and green leafhoppers (GLH, N. cincticeps and N. virescens) are known ROLP vectors [6,14,18]. In the previous study, N. virescens was found abundant in the paddies of the Philippines; however, N. cincticeps was not observed [18]. Similarly, N. virescens but not N. cincticeps was found in Cambodian paddies where rice samples were collected. A low population of ZLH was found in the same paddies, regardless of seasons and rice crop stages ( Figure 4). Alternatively, in all seasons,

Phloematic Accumulation of ROLP
The behaviour of ROLP in its lifecycle in rice and insects is largely unknown. 4′,6-Diamidino-2-phenylindole (DAPI) staining of phytoplasma genomic DNA and observation under a fluorescent microscope was applied to develop an instant observation protocol for ROLP [10]. Using the rice plants that were ROLP-positive following nested PCR, thin sections of rice leaves were prepared and stained with DAPI. By this conventional method, ROLP was presumably detected in rice phloem as bright-blue fluorescence (Figure 5a) that was absent in the control specimen (ROLP-free healthy rice plants) (Figure 5b). Additionally, this fluorescence was observed in root phloem (Figure 5c). Notably, most of the phloem in the same root By nested PCR, the ecological roles of ZLH and GLH (N. virescens) for ROLP transmission in Cambodia were assessed. Although ZLH was less present in the paddies, insects carried ROLP at a high rate (62.5-100%) (Figure 2c, Table 2). In contrast, while the GLH (N. virescens) population was high in the paddies, the ROLP detection rate was not high as that of ZLH (16.7-42.9%) (Figure 2c, Table 2). These results suggest that both insects have vital roles in the spread of ROLP in the rice ecosystem of Cambodia.

Phloematic Accumulation of ROLP
The behaviour of ROLP in its lifecycle in rice and insects is largely unknown. 4 ,6-Diamidino-2-phenylindole (DAPI) staining of phytoplasma genomic DNA and observation under a fluorescent microscope was applied to develop an instant observation protocol for ROLP [10]. Using the rice plants that were ROLP-positive following nested PCR, thin sections of rice leaves were prepared and stained with DAPI. By this conventional method, ROLP was presumably detected in rice phloem as bright-blue fluorescence (Figure 5a) that was absent in the control specimen (ROLP-free healthy rice plants) (Figure 5b). Additionally, this fluorescence was observed in root phloem (Figure 5c). Notably, most of the phloem in the same root tissue exhibited a fluorescent signal, suggesting that ROLP could be propagated and spread throughout the plant (Figure 5d).

Histological Observation of ROLP-Infected Rice
In the ROLP-infected rice plant, drastic physiological changes occur. Thick sections of rice leaf blade were prepared and observed under a scanning electron microscope (SEM) to assess the changes in plants. The major difference between healthy and diseased rice plants is the massive accumulation of starch-like granules in the ROLP-infected rice leaf ( Figure 6). In the vascular bundle sheath and parenchyma of the ROLP-infected rice, compacted, amorphous blobs of possible storage starch were observed (Figure 6c), while a very low amount was seen in healthy plants (Figure 6a). Similarly, in the phloem's enclosed view, particle compaction was observed only in the ROLP-infected rice (Figure 6b,d). The sizes of these particles appeared to be very small (100-200 nm in diameter), and this range agrees with the general size of the phytoplasma (80-800 nm in diameter) (Figure 6e,f). These obvious changes are associated with ROLD symptoms, such as leaf discoloration, growth inhibition and fast senescence.

Histological Observation of ROLP-Infected Rice
In the ROLP-infected rice plant, drastic physiological changes occur. Thick sections of rice leaf blade were prepared and observed under a scanning electron microscope (SEM) to assess the changes in plants. The major difference between healthy and diseased rice plants is the massive accumulation of starch-like granules in the ROLP-infected rice leaf ( Figure 6). In the vascular bundle sheath and parenchyma of the ROLP-infected rice, compacted, amorphous blobs of possible storage starch were observed (Figure 6c), while a very low amount was seen in healthy plants (Figure 6a). Similarly, in the phloem's enclosed view, particle compaction was observed only in the ROLP-infected rice ( Figure  6b,d). The sizes of these particles appeared to be very small (100-200 nm in diameter), and this range agrees with the general size of the phytoplasma (80-800 nm in diameter) (Figure 6e,f). These obvious changes are associated with ROLD symptoms, such as leaf discoloration, growth inhibition and fast senescence.

Global Population Structure of ROLP Based on Currently Available 16S rDNA Sequences
The current study detected a considerable number of ROLP-infected rice plants (236 samples out of 295 plants tested) in the Philippines, Vietnam and Cambodia from 2015 to 2019 ( Table 1) and confirmed that these are associated with insect vectors ZLH and GLH in Cambodia (Table 2). A series of surveys were triggered by the first observation of ROLD by our team in Laguna, the Philippines, in 2015 and that in Svay Rieng, Cambodia in 2016. At this stage, ROLP in Cambodia was expected to have migrated from Vietnam [22]. Together with the reported ROLP 16S rRNA gene sequences from Thailand, India, Vietnam, the Philippines and South China   Table 1) and confirmed that these are associated with insect vectors ZLH and GLH in Cambodia (Table 2). A series of surveys were triggered by the first observation of ROLD by our team in Laguna, the Philippines, in 2015 and that in Svay Rieng, Cambodia in 2016. At this stage, ROLP in Cambodia was expected to have migrated from Vietnam [22]. Together with the reported ROLP 16S rRNA gene sequences from Thailand, India, Vietnam, the Philippines and South China [3,18,[21][22][23]28], those of the Philippines, Vietnam and Cambodia analysed in this study were phylogenetically assessed for the first time ( Figure 3).
It was found that isolates in India, Cambodia and Thailand were more diverse than those in other countries, and thus these may be considered as autochthonous genotypes of ROLP in these areas or international translocation-resultant isolates. Although further analysis is required, it is anticipated that the potential origin of ROLP is Thailand because of its rich genetic diversity (Figure 3). Additionally, probable spread of a single population in Cambodia, Vietnam, China and the Philippines was assumed based on the distribution of the Mekong-Ph type ROLP in these countries as shown in Figure 7 (red stars). Indeed, the evolutional relationship between the Chinese and Indian isolates is far apart; the disease outbreaks in these countries may not be attributed to the genetic commonality of ROLP. To our knowledge, ROLP can be found commonly in the paddies of Asian countries; thus, the pathogen is expected to be more broadly present in rice cropping countries today. To further understand the genetic variations of ROLP, disease surveys must be conducted more widely together with genotyping of the pathogen by sequencing of 16S rDNA and multilocus house-keeping genes [26,29,30] or even by genome-wide comparisons.
Pathogens 2021, 10, x FOR PEER REVIEW 10 of 16 [3,18,[21][22][23]28], those of the Philippines, Vietnam and Cambodia analysed in this study were phylogenetically assessed for the first time ( Figure 3). It was found that isolates in India, Cambodia and Thailand were more diverse than those in other countries, and thus these may be considered as autochthonous genotypes of ROLP in these areas or international translocation-resultant isolates. Although further analysis is required, it is anticipated that the potential origin of ROLP is Thailand because of its rich genetic diversity (Figure 3). Additionally, probable spread of a single population in Cambodia, Vietnam, China and the Philippines was assumed based on the distribution of the Mekong-Ph type ROLP in these countries as shown in Figure 7 (red stars). Indeed, the evolutional relationship between the Chinese and Indian isolates is far apart; the disease outbreaks in these countries may not be attributed to the genetic commonality of ROLP. To our knowledge, ROLP can be found commonly in the paddies of Asian countries; thus, the pathogen is expected to be more broadly present in rice cropping countries today. To further understand the genetic variations of ROLP, disease surveys must be conducted more widely together with genotyping of the pathogen by sequencing of 16S rDNA and multilocus house-keeping genes [26,29,30] or even by genome-wide comparisons.

Paddy Leafhopper and Planthopper Populations and the Transmission of ROLP in Southeast Asia
ZLH carried ROLP at high rates in the diseased rice fields in Cambodia, but the population density of ZLH was very low throughout the rice growth stages and in different cropping seasons (Table 2, Figure 4). GLH had relatively low rates of ROLP detection; however, large numbers of these were found in the paddies. Although the insect's population structure has not been assessed in the Philippines, similar trends of less ZLH and many GLH in rice fields were observed previously [18]. This also agrees with the situation in South China where another GLH (N. cincticeps) of the temperate region played an essential role in the pathogen transmission [6]. Moreover, while the acquisition rate of this GLH species (16.7-55.3%) was comparable to that of the GLH in the tropics (N. virescens) (16.7-42.9%) ( Table 2), the germ-carrying rate of ZLH in China (31.7-61.6%) was lower than those in Cambodia (62.5-100%) ( Table 2). This fact may reflect different affinity levels of the ZLH biotypes to ROLP. In fact, ZLH in Cambodia seemed much smaller in size than it is generally known. To this regard, an investigation into the affinity of ROLP surface proteins with each vector is required. VmpA of Flavescence dorée phytoplasma (16SrV-C and-D) was associated with Spiroplasma citri transmission ability [31], and the antigenic membrane protein (Amp) of onion yellows phytoplasma (16SrI) formed Amp-microfilament complexes in insects that defined transmissibility [32]. Altogether, we hypothesized that both ZLH and GLH are important vectors of ROLP in Cambodian rice fields, which is probably the case in Vietnam and the Philippines [18].
In addition, it was observed that BPH was the most abundant and the primary pest in two paddies in Cambodia (Figure 4). The high population density of BPH might be associated with the opportunistic transmission of ROLP; therefore, two additionally monitored insects were subjected to ROLP detection by nested PCR. As a preliminary result, only one of tested BPHs and white backed planthoppers (WBPHs) was ROLPpositive, suggesting the low acquisition rate of ROLP by these insects, although their transmission ability is currently unknown (Supplementary Table S2). These observations raise awareness of the requirement of further investigation on field insects for ROLP acquisition and transmission.
Small holder cultivated Cambodian rice fields and the timing of rice planting among those farmers is not uniform, being situated in the concomitant presence of rice plants at different growing stages in a small area [33,34]. The quick drop in the population density of BPH at the dough stage ( Figure 4) is expected to be a result of BPH migration to other rice fields with a feasible environment [35]. If BPH has ROLP transmission ability, it will contribute to the distant spread of the disease. In contrast, the GLH density at the dough stage was maintained during the humid season (early wet and wet seasons) ( Figure 4). This ecological characteristic of GLH might contribute to the constant maintenance of ROLP in the rice ecosystem, as for the tropics; farmers keep cropping rice three, up to four, times in irrigated areas. Thus, controlling the GLH population before rice planting may be a useful measure of ROLP control.

Biological Properties of ROLP in the Host Rice Plants and Future Research Prospects
The molecular mechanisms of infection, spread and pathogenicity of ROLP in rice plants are largely unknown. Akin to other phytoplasmas, ROLP was observed in the phloem by transmission electron microscopy [6,14]. In this study, highly accumulated ROLP as a DAPI-stained genome in the phloem of rice leaves and roots was confirmed ( Figure 5). This led to the speculation that the microparticle compaction in the phloem observed under the SEM (Figure 6d-f) was of ROLP bodies, which agrees with the sizes of known phytoplasma cells [1]. PCR detection of ROLP from rice samples is very effective as it is generally detectable in the first round of nested PCR (1.5 kbp product). This, therefore, suggests a high propagation of ROLP in the phloem, which is correlational with microparticle compaction in the phloem. It is interesting to quantitatively compare phytoplasma accumulations between different phytoplasma or host plants. Technically, the DAPI stain-ing/fluorescent observation to detect phytoplasma is a useful approach for understanding the systemic distribution of the pathogen in plant tissues [10,36]. The combination of this and development of artificial inoculation cycle in the laboratory is highly desired.
Phytoplasma infection causes drastic phenotypic changes in host plants-many examples report physiological changes upon infection. For example, mulberry yellow dwarf phytoplasma (16SrI group) infection changed the content of host metabolites, including carbohydrates, amino acids, organic acids and others [37]. It should be noted that the phytoplasma also attenuates starch digestion ability by the downregulation of αand β-amylase genes, whereas photosynthesis was inactivated, resulting in higher starch accumulation levels in plant [37]. In this regard, as observed in Figure 6c, ROLP infection caused high accumulation of starch-like granules in vascular bundle sheath and parenchyma. To further understand the biological impact of ROLP infection in rice, transcriptomic and metabolomic analyses need to be addressed.
Other microbes and viruses may affect the host rice or ROLP once a mixed infection is established. The detection of a DNA virus, rice tungro bacilliform virus (RTBV), was attempted using the available rice DNA samples. As previously observed [18], the virus was not detected in most ROLP-positive rice samples tested in Cambodia. However, those from Bati, Takeo province, contained a few positive and faintly positive samples (Supplementary Figure S1). Interestingly, samples showing a strong signal of RTBV were negative for the ROLP detection, whereas faint signals were found in a few ROLP-positive samples. In tomato, tomato big bud phytoplasma (16SrII-D) and tomato yellow leaf curl virus were found to be antagonistic to each other in terms of their accumulation levels and symptom expressions [38]. As rice tungro viruses rely on the insect vector GLH for their transmission, it is highly interesting to investigate whether these viruses act counteractively or synergistically against ROLP in rice plants and insect vectors.

Field Survey of ROLP-Infested Rice Paddies
A rice field survey was conducted from 2015 to 2019 in Svay Rieng, Prey Veng, Takeo, Kampong Thom, Kampong Seng provinces and Phnom Penh in Cambodia, Tien Giang and Dong Thap provinces in Vietnam and Mindanao and Laguna provinces in the Philippines. A portion of this survey was previously reported [18]. Paddies with ROLD-suspected rice plants were targeted and leaf or whole plant samples as well as insects were harvested or collected upon farmers' agreement. Plant and insect samples were processed for DNA extraction or stored at −68 • C.

DNA Extraction, Nested PCR and LAMP
Rice and insect samples were homogenised in liquid nitrogen, and genomic DNA fractions were extracted using a Cetyl trimethyl ammonium bromide (CTAB)-mediated conventional extraction method [18], DNeasy plant mini kit or DNeasy blood and tissue kit (Qiagen, Hilden, Germany) by following the manufacturer's instructions.
In total, 50 ng of purified DNA was subjected to nested PCR with well-designed universal primer sets: P1 and P7 for the first round PCR and R16F2n and R16R2 for the second round PCR as previously described [6,18]. The first PCR program was 94 • C for 2 min; 30 cycles at 94 • C for 30 s, 52 • C for 30 s, and 72 • C for 2 min; 72 • C for 10 min. The second PCR program was 94 • C for 2 min; 30 cycles at 94 • C for 30 s, 50 • C for 30 s and 72 • C for 2 min; 72 • C for 10 min. The second PCR amplicons with lengths of 1.2 kbp were examined by 1% agarose gel electrophoresis.
According to the provided protocol, a phytoplasma universal detection kit (Nippon Gene, Tokyo, Japan) was used for phytoplasma detection by LAMP, and confirmed by nested PCR later on for ROLP confirmation.