International Journal of Virology

Year: 2017 | Volume: 13 | Issue: 2 | Page No.: 80-89
DOI: 10.3923/ijv.2017.80.89

Genetic Diversity and Molecular Epidemiology of Rice Yellow Mottle Virus in Central African Republic

R.D.S. Longue , A. Galzi-Pinel, I. Zinga, S. Semballa, D. Fargette, N. Barro and O. Traore

Abstract: Background: Rice yellow mottle virus (RYMV) of the genus Sobemovirus is the most important viral disease of rice in Africa, mainly occurring in the lowland and irrigated ecologies. Materials and Methods: The coat protein gene of forty six isolates of RYMV collected between 2011 and 2014 in Central African Republic was amplified and sequenced. Results: Analysis of sequences revealed that the average nucleotide diversity among isolates was low, 2.0% in nucleotide and 0.9% in amino acid. The ratio of non-synonymous over synonymous nucleotide substitutions per site was 0.07, indicating a virus diversification under a high conservative selective pressure. All isolates shared the amino acids specific of the serotype Ser1, a Val₁₁₅ involved in the response with monoclonal A and a Thr₁₁₅ which accounts for the lack of reaction with monoclonal D. Phylogenetic analyses showed that isolates of the south of Central African Republic belong to two sister monophyletic groups related to the S1ca strain, a strain which gathers all isolates from the East of West Africa to the West of Central Africa. Molecular clock dating of the age of each of the two groups and of their common ancestor suggests that RYMV was introduced a few decades ago in Central African Republic from the West of Central Africa. Conclusion: RYMV remained at a low level and undetected in wild hosts and in traditional rice fields until recently. RYMV emergence occurred less than ten years ago, likely favored by the on-going rice intensification and resulted in the current epidemics. This is the first study of the molecular diversity of RYMV in Central African Republic. It indicates that the isolates of the South of the country belonged to the S1 strain.

Keywords: coat protein, Rice yellow mottle virus, rice cultivation, diversity index and molecular epidemiology

INTRODUCTION

First detected in Kenya¹ in 1966, Rice yellow mottle virus (RYMV) of the genus Sobemovirus is the most important virus of rice in Africa. Only found in Africa and in Madagascar, RYMV is widespread in most countries where rice is grown². The RYMV was reported in Cameroon and Chad³ in 2001, countries bordering the Central African Republic at the West and at the North, respectively. In 2012, RYMV was detected in the south of the Central African Republic⁴.

The diversity among RYMV isolates in Africa was first assessed using polyclonal and monoclonal antibodies raised against an isolate of Madagascar^5-8. A set of six Mabs (A, B, C, D, E and G) was used. Five serotypes were delineated through the pattern of detection by this set of monoclonal antibodies and named Ser1 to Ser5. The RYMV diversity was further characterized by sequencing the highly variable 720 nucteotides-long Coat Protein (CP) gene. Comparatively, CP sequencing is more discriminating than immunological typing since several sequence variants belong to the same serotype⁹. Seven major strains-named S1-S6 and Sa were determined. They had contrasting geographical distribution. Four strains were found in West and Central Africa (S1-S3 and Sa) and three strains in East Africa (S4-S6). Molecular epidemiology of RYMV indicated a center of origin in East Africa and a secondary center of diversification in West Africa¹⁰. Furthermore, genetic differences were found between isolates from Savanna and forest regions of West Africa¹¹.

In this study, we collected and sequenced the coat protein gene of 46 isolates from the South of the Central African Republic, the part of the country where most of the rice is grown. The genetic diversity of the isolates was assessed, their phylogenetic relationships with representative isolates of the main strains in Africa were determined and the serospecific amino-acids were identified. Clock-like diversification of RYMV allows molecular dating¹². Accordingly, the Time of the Most Recent Common Ancestor (TMRCA) of the isolates of the Central African Republic was estimated, providing information on virus introduction and disease emergence in the country. Then, a scenario for the epidemiology of RYMV in the Central African Republic is proposed.

MATERIALS AND METHODS

Virus isolates: Symptomatic leaf samples were collected between 2011 and 2014 in large irrigated rice production schemes around 30 km West of Bangui, the capital of the Central African Republic and in lowland subsistence fields in Bangui outskirts. Isolates were propagated by mechanical inoculation on the susceptible cultivar IR64 to increase the viral concentration and to get sufficient material for RNA extraction. Symptomatic leaves were collected 2 weeks after inoculation and used for molecular typing.

Extraction of total RNA from rice leaves: Total RNA was extracted from infected rice leaves using TRIzol reagent (Invitrogen, USA) according to the manufacturer’s instruction. The mains steps were as follows: the mortars were first sterilized 2 h at 120°C and stored overnight at -80°C. The frozen leaves were ground to a fine powder and transferred into sterile 2 mL Eppendorf tubes. The powder was mixed with TRIzol reagent (about 1 mL) and the mixture was vortexed. Then, chloroform was added and the tubes were placed on ice for 5 min to separate RNA into cellular proteins and DNA. After centrifugation 12,000 rpm for 15 min at 4°C, the aqueous phase was transferred into sterile 1.5 mL Eppendorf tubes. To precipitate total RNAs, 550 μL of isopropanol were added and left for 30 min at -20°C. After centrifugation at 12,000 rpm for 10 min, total RNAs (pellet) were then washed with 500 μL of alcohol 75%. The resulting total RNAs were air dried, resuspended in 30 μL of sterile water and stored at -20°C.

RT-PCR of the coat protein gene: The first strand of the CP gene cDNA was synthesized using M-MLV reverse transcriptase (Promega corp, USA). A volume of 5 μL of total RNAs was denatured in presence of 10 μM of reverse primer RYMV II or P052 (5’CTCCCCCACCCATCCCGAGAATT3’) for 5 min at 70°C. The RT reaction mixture contained 5 x RT buffer, dNTP (10 mM), M-MLV-RT (200 U), RNase inhibitor (40 U) and sterile water (8 μL) to a total volume of 25 μL and followed by incubation at 42°C for 1 h.

The PCR amplification was performed as described earlier¹³. Specific primers P013 5’CAAAGATGGCCAGGAA3’ (forward) and P052 5’CTCCCCCACCCATCCCGAGAATT3’ (reverse) were used to cover the full length of the CP gene. The thermal conditions were as follows: Initial denaturation at 94°C for 5 min and 30 cycles of denaturation at 94°C for 1 min, hybridization at 55°C for 30 sec, extension at 72°C for 1 min, followed by final extension 10 min at 72°C and then held at 15°C. All RT-PCR reactions were conducted in Applied Biosystems 2720 Thermal Cycler (Life Technologies, France). The amplified fragments were subjected to electrophoresis in 1% (p/v) agarose gels containing ethidium bromide and were visualized using a UV transilluminator in order to check the size of the amplified fragments according to the covering region. Positives results were sent to Beckman Coulter Genomics (France) for sequencing. Forty sequences were obtained.

Sequence analysis: The sequences were edited and aligned using EditSeq and MegAlign of DNASTAR Lasergene software version 7 followed by similarity assessment of pairwise sequences. Multiple nucleotide sequences alignment was made using Clustal W with default parameters¹⁴. The aligned sequences were checked and corrected manually where necessary. Molecular diversity among the isolates was estimated using the Kimura’s two parameter method¹⁵ for nucleotide and Jones-Taylor-Thornton (JTT) model for amino-acid, as implemented in MEGA-6 software¹⁶. The ratio between non synonymous (π_s) and synonymous (π_a) substitutions per site (ω = π_a/π_s) was calculated to assess the selection pressure acting on the CP gene. Then, a Z-test was computed to validate the type of selection pressure. A p-value less than 0.05 was considered significant at the 5% level. All tests of natural selection were computed using the Nei-Gojobori method as implemented in MEGA-6 software¹⁶.

Phylogenetic relationships between the Central African Republic isolates (named CF isolates) and between CF isolates and isolates of the main strains in Africa were determined using the Maximum-Likelihood (ML) method implemented in MEGA-6 (Table 1). The Tamura 3-parameter model with rate variation across sites estimated from a gamma distribution (T92+G) was selected as the best fitting model. The significance of the internal branches was evaluated using 1000 bootstrap replications. The phylogenetic tree was visualized and edited using the MEGA-6 software.

Molecular clock dating: The reference dataset used for molecular clock dating consists of 300 sequences that were sampled between 1966 and 2012 in 20 countries across East, West and Central Africa¹⁷ to which were added the CF sequences. The TMRCA of the nodes of interest and their associated 95% highest density probability (HPDs) intervals were estimated within a Bayesian coalescent framework by a Markov Chain Monte Carlo (MCMC) implemented in BEAST v 1.8.2¹⁸. A relaxed (uncorrelated lognormal) clock model, a HKY2+G4 model and the flexible Bayesian skygrid model as prior were selected. In all cases, the BEAST analyses were run until all relevant parameters converged (ESS>200) with 20% of the MCMC chains discarded as burn-in. Statistical confidence in the parameter estimates was represented by values for 95% HPDs intervals around the marginal posterior parameter means.

RESULTS

Agarose gel visualization of RT-PCR products: Forty isolates were analyzed by RT-PCR using specific primers that amplified a zone covering the CP gene. The amplified fragments were about 1000 bp long, corresponding to the expected size according to the genetic marker (Fig. 1). Then, 40 sequences were obtained.

Molecular variability among Bangui (CAR) isolates: The 40 CP sequences of isolates of Central African Republic together with 6 sequences of the same country previously published⁴ were analyzed. The total number of nucleotide sites was 720 nucleotides, encoding 239 amino acid residue. Analyses of the polymorphic sites among sequences revealed 55 and 12 variable sites in nucleotide and amino acid, respectively. Indeed, 75% of the mutations occurred at the 3rd nucleotide position of the codons and none of these mutations result in an amino acid change. Two hundred and thirty nine amino acid residues were mostly hydrophobic amino acids (Table 2). Similarity among the isolates ranged from 94.7-100% in nucleotide whereas the average divergence was 2.0% in nucleotide and 1.1% in amino acid. Divergence among sequences ranged from 0.0-4.5 in nucleotide and from 0.0-2.5% in amino acid. The average number of nucleotide synonymous substitutions per site was low (π_s = 0.06), yet 15 times higher than the number of non-synonymous substitutions per site (π_a = 0.004), i.e., a ω ratio (π_a/π_s) of 0.07 (Table 3).

The maximum of non-synonymous and of synonymous divergence between two sequences was 1.2 and 13%, respectively. As ω<1 is indicative of a purifying selection pressure and as the p-value of the Z-test was highly significant (p<0.001), the CP gene of the CF isolates diversified under a strong purifying selection. The transition-to-transversion ratio was estimated to 7.3.

Phylogenetic relationships: The phylogenetic reconstruction from the nucleotides sequences based on the maximum likelihood method indicates that all isolates from the Central African Republic clustered into two monophyletic sister groups with high bootstrap support which were subsequently designated group 1 and group 2 (Fig. 2). Group 1 contained 17 isolates whereas group 2 included 29 isolates. In some instances, isolates collected in the same rice field were quite similar, sometimes with identical sequences. However, isolates collected in the same rice field may differ. In the Ngola locality in particular, isolates of both groups were found in the same field. Within each group, there were variants with a high bootstrap support. The diversity within and between group 1 and group 2 was assessed. Nucleotide substitution per site between isolates of the two groups ranged from 2.8-4.5% with an average of 3.5%. However, within each group, the average nucleotide diversity was low: 0.8% in group 1 and 0.5% in group 2, with a maximum divergence between two isolates of 1.6 and 1.5%, respectively.

To assess the phylogenetic relationship between isolates of the Central African Republic and isolates representative of the main strains in Africa¹¹, phylogenetic tree was reconstructed using the ML method as described earlier (Fig. 3). The CF isolates belong to the strain S1 (bootstrap support 92%) made of the S1wa and the S1ca sister strains. The S1wa strain included isolates from the West of West Africa only. The CF isolates clustered within the S1ca strain (node support 95%) referred to as the West-Central African strain, a strain which gathers all isolates from the east of West Africa (Benin, Togo, Niger and Nigeria) to the West of Central Africa (Chad and Cameroun). Consistently, genetic distance between CF isolates and S1ca isolates was up to 4.4% in nucleotide, lower than 5.9% between CF and S1wa isolates (Table 4). Interestingly, one isolate from Chad (Tc3) was close to group 2 and formed a monophyletic group not including group 1 with a high bootstrap support (78%).

Comparison of molecular and serological typing: Although the phylogenetic analyses reconstructed from the nucleotide and amino acid sequences displayed an overall similar pattern, there was yet a noticeable difference. Whereas isolates CF63, CF175, CF177, CF181 CF183 and CF4a belonged to group 1 at the nucleotides level, they did not cluster with the other isolates of this group at the aa level (Fig. 4). This difference is due to specific amino acid changes (which distinguish groups) at positions 68, 80 and 217 corresponding to Lys/Thr, Asp/Asn and Ala/Thr residue changes, respectively (Fig. 5).

These positions also allowed to distinguish CF isolates from Ser1 isolates i.e., when group 1 consensus sequence was compared to that of Ser1 (Fig. 5). In addition, isolates CF28b, CF64 and CF176 from group 2 had an hydrophobic amino acid Leu (TTA) at position 164 instead of a polar residue Ser (TCA). Also, seven other isolates from group 2 had an Arg residue versus a Lys, both basic amino acids, at position 18.

Isolates of the serotype Ser1 are characterized by a detection by Mab A and a lack of recognition by Mab D⁹. Moreover, a Val at position 191 characterizes the reaction with Mab A, whereas a Thr at position 115 accounts for the lack of reaction with Mab D. Sero-specific amino acids of the coat protein of the two groups of CF isolates were identified. Any isolate from the Central African Republic; be it of group 1 or of group 2 had a Val at position 191 and a Thr at position 115 (Fig. 5). This result suggests that, despite their differences in amino acid sequences in the coat protein, isolates of group 1 and 2 shared the serotype Ser1.

Molecular dating: Molecular clock dating provides an estimate of the TMRCA of group 1 and of group 2 and of their common ancestor (Table 5). The diversification of each of the two groups was recent, less than 10 years ago, i.e., after year 2000 (Table 5). The TMRCA was estimated to 8 years (i.e., year 2004) and 6 years (i.e., year 2006) for group 1 and 2, respectively.

The overlap of their 95% HPDs intervals indicates a non-significant difference between the two TMRCAs. The TMRCA of the common ancestor of the two sister group 1 and 2 was substantially older, 18 years old (i.e., year 1994).

The TMRCA of all isolates from strain S1ca, including the CF isolates was 65 years (i.e., year 1947).

DISCUSSION

Isolates from the Central African Republic were specific as they belong to two phylogenetic groups not found in other countries, even in the neighboring countries Cameroun and Chad. Yet, they are genetically related to isolates of the West-Central strain S1ca as described by Traore et al.³. This result jointly indicates that RYMV in the Central African Republic belongs to the S1ca strain and that the S1ca strain had a geographical distribution more extended at the East than earlier reported, by at least 500 km¹⁹. Interestingly, isolates of the S4-S6 strains of East Africa and at the East of Central Africa (Burundi, Rwanda) and at the east of the Democratic Republic of the Congo, a country bordering the Central African Republic at the South were not found in the Central African Republic. Conversely, isolates of the S1ca strain were neither found at the East of Central Africa nor in East Africa. This clear-cut split between the geographical distribution of the S1ca strain and that of the S4-S6 strains suggests that the large forests of the Democratic Republic of the Congo with scarce rice cultivation make an efficient and long-lasting barrier to RYMV propagation across them¹².

The isolates of the Central African Republic belonged to two distinct genetic groups. There was a small genetic distance between the two groups and a low genetic diversity within each group. Similar diversity (π = 1.0%) was found within isolates from Sierra Leone at the far West of Africa and originating from mangrove habitats, according to Abubakar et al.²⁰. Based on these low variabilities between

equences, we suppose that diversification of the two groups in the South of the Central African Republic is relatively recent. This genetic diversity is also characterized by a low ratio of non-synonymous over synonymous substitution per site, indicative of a strong purifying selective pressure acting on the CP gene and a high transition to transversion ratio, two characteristics shared with isolates of the main strains of RYMV¹¹. The results confirm those showed by Pinel et al.¹³ suggesting a conservative evolution pressure over the CP gene and explains that variation is less at amino acid than at nucleotide levels.

According to N’Guessan et al.⁸ and Fargette et al.⁹, amino acid sequences of the coat protein gene suggested that isolates from the Central African Republic shared the Ser1 serotype. This is consistent with serological studies which showed that isolates from the Central African Republic were all recognized by Mab A but were undetected by Mab D, indicative of the Ser1 serotype^4,21.

Molecular clock dating indicates a recent diversification of each group (<10 years). This recent and simultaneous emergence of RYMV in the South of the Central African Republic is consistent with farmer’s opinion who reported that disease emergence coincided with the recent rice intensification in the country and occurred not only in traditional rice areas grown since the 1960s but also in new rice cropping areas. Detection of isolates of each group in proximate fields, sometimes even within the same fields, indicates a current and rapid spread of isolates of the two strains within the area and probably beyond. The inclusion of each of the two groups within the S1ca strain and the relationship with isolates from Cameroon and Chad, suggest introduction(s) of the virus, a few decades ago, from these countries located at the West and North of the Central African Republic, respectively, as suggested earlier¹⁷. According to the previous studies, the Time of the Most Recent Common Ancestor (TMRCA) of the mains strains of West Africa is estimated around 100 years²². Whereas, the results showed the TMRCA of S1ca strain is around 65 years suggesting that RYMV remained at a low level and undetected in wild hosts and in traditional rice fields until recently. The RYMV emergence occurred less than ten years ago, likely favored by the on-going rice intensification and resulted in the current epidemics. Such a scenario is consistent with what has been observed in other countries in Africa, but need to be documented from denser and larger surveys in the Central African Republic and neighboring countries.

CONCLUSION

Rice yellow mottle virus emergence in the Central African Republic is recent and coincided with rice intensification in the country. Particular attention should be paid to the current epidemics in order to prevent major crop loss as what occurred in West Africa in 1990s.

This study contributes to the understanding of the genetic relationships of RYMV through Africa. Since these data are from isolates from of the South of the country, it would be interesting to extend the study to isolates from other parts of country. This could result in a better understanding of the virus spread within the country and would also contribute to the understanding of its evolutionary history across Africa.

SIGNIFICANT STATEMENT

ACKNOWLEDGMENTS

This study was supported financially by the International Foundation for Science (IFS) through fellowship No. C/5359-1 to R D Longué Sokpé and SCAC from the French Embassy in Bangui. We also gratefully acknowledged the laboratoire mixte international (LMI-PathoBios), Burkina Faso, for providing technical assistance.

REFERENCES