Asian Lineage of Peste des Petits Ruminants Virus, Africa

Interest in peste des petits ruminants virus (PPRV) has been stimulated by recent changes in its host and geographic distribution. For this study, biological specimens were collected from camels, sheep, and goats clinically suspected of having PPRV infection in Sudan during 2000–2009 and from sheep soon after the first reported outbreaks in Morocco in 2008. Reverse transcription PCR analysis confirmed the wide distribution of PPRV throughout Sudan and spread of the virus in Morocco. Molecular typing of 32 samples positive for PPRV provided strong evidence of the introduction and broad spread of Asian lineage IV. This lineage was defined further by 2 subclusters; one consisted of camel and goat isolates and some of the sheep isolates, while the other contained only sheep isolates, a finding with suggests a genetic bias according to the host. This study provides evidence of the recent spread of PPRV lineage IV in Africa.

of PPRV in Sudan affected both sheep and camels (15,16). In camels, a respiratory syndrome was the prominent disease characteristic observed, resembling a previous case reported in Ethiopia during 1995Ethiopia during -1996.
More recently, in summer 2008, Morocco reported outbreaks of PPRV for the fi rst time (18). Because PPRV needs close contact for transmission, this new epizootic was likely the result of introduction of the virus into North Africa through the movement of live infected animals. To characterize PPRV strains identifi ed in Sudan during [2000][2001][2002][2003][2004][2005][2006][2007][2008][2009] and in Morocco during the extensive 2008 outbreaks, a phylogenetic analysis was carried out on wild-type PPRV sequences obtained from biological samples collected from sheep, goats, and camels in these 2 countries.

Biological Materials
Biological specimens were collected in different parts of Sudan during 2000-2009 from camels, sheep, and goats that showed clinical signs of PPRV. Additional camel specimens were collected after outbreaks associated with high death rates were reported from the Kassala Region in 2004. A total of 80 fi eld samples, including lung, liver, and spleen, were obtained from 49 camels, 26 sheep, and 5 goats with PPRV-like clinical signs. For most of the animals, virus detection and identifi cation were performed directly on the tissue samples collected; however, virus isolation in tissue culture was attempted for some samples. A historical PPRV strain isolated in Sudan in 1971, the Gedarif PPRV strain, was included in this study.
In Morocco, a total of 36 samples were collected from sheep displaying PPRV signs during the 2008 outbreaks. These included oral and ocular swabs and mesenteric lymph nodes, spleen, lung, and whole blood samples.

Laboratory Investigations
Sample aliquots fi rst were screened for PPRV antigen detection by immunocapture ELISA (ICE test; Biological Diagnostics Supplies, Ltd., Dreghorn, Scotland). A second aliquot was used for viral RNA extraction and virus isolation in tissue culture. Reverse transcription PCR (RT-PCR) amplifi ed the N protein gene directly from tissue samples by using a set of pan-morbillivirus primers as described by Kwiatek et al. (10).
For the sequencing of the N gene, the 3′ 351 nt were obtained by employing primers NP3 and NP4 (19) and through a modifi cation of the initial protocol by using a 1-step method (OneStep RT-PCR mix; QIAGEN, Courtaboeuf, France). For the F gene, 11 samples were selected from samples that previously tested positive by N amplifi cation. For the specifi c and sensitive detection of PPRV from these clinical fi eld samples, the initial method developed by Forsyth et al. in 1995 (8) was modifi ed by using an additional pair of primer F1/FPPRrev in combination with F1/F2 in a nested PCR. Sequences positions of these primers were 5′-ATCACAGTGTTAAAGCCTGTAGAGG-3′ (PPRV-F1: 777 ± 801), 5′-GAGACTGAGTTTGTGACCTACAAGC-3′ (PPRV-F2: 1124 ± 1148) and 5′-ATATTAATGTGACAA GCCCTAGGGA-3′ (FPPRrev: 2055 ± 2079). The fi nal amplicon was obtained at the expected length of 372 nt after 35 cycles. We analyzed 10 μL of the amplifi ed products by electrophoresis on 1.5% agarose gel. For all positive results, 40 μL of the fi nal product was used directly for sequencing (GATC Biotech, Constance, Germany). All sequences were deposited in GenBank under accession nos. HQ131917-HQ131958. FPPRrev and the study of the target sites of primers F1 and F2 (8) were designed by a comparative sequence analysis of various morbillivirus F sequences from the full genomes available in GenBank (accession nos. X74443, EU267273, EU267274, AY560591, FJ905304, and AJ849636).

Sequence Data, Alignment, and Phylogenetic Analysis
Sequencing reactions were performed by GATC Biotech. Each nucleic acid segment on the N and F genes of the Sudan and Morocco strains was aligned with other sequences from PPRV maintained in the database or retrieved from GenBank (Table 1) (10,(20)(21)(22). Multiple alignments of 255 and 322 nt of the N and F genes, respectively, were made by using the ClustalW program (www.ebi.ac.uk/clustalw). Phylogenetic analysis was carried out by using the criterion of neighborhood based on the principle of parsimony (23). Dissimilarities and distances between the sequences fi rst were determined with Darwin software (24). Tree construction was based on the unweighted neighbor-joining method proposed by Gascuel (25). Trees were generated with the TreeConMATRIXW program of Darwin (26). Bootstrap confi dence intervals were calculated on 1,000 replicates.

Virus Isolation
For the samples from Sudan, isolation was successful on MDBK cells with lung samples from Cam_8, Cam_169, Cam_318, Ov_Soba, and Ov_Al Azaza, collected from 3 camels and 2 sheep, respectively. For the samples from Morocco, PPRV Morocco_08_02 virus was isolated successfully from a sheep lung sample after infection of Vero.DogSLAMtag cells (27). These isolates were sequenced and compared as described above.

Detection of PPRV by N gene RT-PCR
A total of 80 animals from Sudan that initially tested positive in the ICE test were analyzed further by using RT-  (Table  2). Of the 36 samples tested from Morocco, 16 yielded a positive RT-PCR.

Characterization of Strains Involved in the Infection of Sheep, Goats, and Camels in Sudan, and in Sheep in Morocco
Of the tissue samples tested, lung and lymph nodes were the most suitable for sequencing because they had the highest viral load, thereby yielding a suffi cient amount of PCR product. Sequence also was obtained from several virus isolates, including the historical Gedarif isolate from 1971. The partial N gene sequences were obtained for 26 of the 64 samples from Sudan and 6 of the 16 samples from Morocco. Sequences were aligned with an extended set of PPRV isolate sequences that either were in the database or were retrieved from GenBank where they were described by Kwiatek et al. (10) and Banyard et al. (20).
In contrast to what would have been expected for isolates from eastern Africa, most of the PPRV strains collected in Sudan during 2000-2009 were clustered in lineage IV (24); only a few remained in lineage III (2). Molecular typing also showed for the fi rst time the presence of PPRV lineage IV in Morocco ( Figure 1).
Lineage IV isolates from Sudan could be further distinguished into 2 clusters. The grouping of the sequences within the clusters was supported by bootstrap percentages >70%. In the fi rst group, 17 sequences from all camel and goat isolates and some sheep isolates (camel Sudan Cam_1, Cam_3, Cam_8, Cam_169, Cam_223, Cam_264, Cam_268, Cam_304, Cam_330, Cam_352, Cam_318; sheep Ov_1, Ov_140, Ov_25, Ov_39300; goat Cap_9, Cap_1) showed 100% identity and matched with Saudi Arabia_1999_7 strain. Sheep Sudan Ov_41 differed from the previous sequences by 2 nt at position 68 (C to T) and 142 (A to G). The divergence of this cluster with consensus IV ranged from 1.2% (Cam_318) to 2%  (20). This cluster was less homogeneous than the cluster previously described; nucleotide variations from the lineage IV consensus sequence ranged from 1.6% (Ov_Al Azaza) to 3.9% (Ov_23).
Only 2 isolates collected at the start of the study (in mid-2000) from sheep in western and eastern Sudan fell into lineage III. Sequences of these 2 fi eld isolates remained similar to the historical strains of Sinnar 72 that was isolated 40 years earlier from the Blue Nile region (2), and with the Gedarif PPRV strain that was isolated from sheep in 1971 (13). In addition, these isolate sequences were close to an Omani strain, Ibri_1983 (28) and a United Arab Emirates strain, Dorcas_1986 (29), but they differed from these strains with a nucleotide variation of 4.7%.
Because the dataset within the Saudi Arabia cluster was highly homogeneous, we compared an N phylogenetic tree (Figure 2, panel A) with an F phylogenetic tree ( Figure  2, panel B) on a panel of viruses. These viruses were selected from the 2 lineage IV clusters defi ned in Figure 1 and from lineage III with additional PPRV from Asia, for which both gene sequences were available (21,22). The F phylogenetic tree confi rmed the high homology within the Saudi Arabia cluster, and the cluster's closeness to strains from Morocco. It also allowed for assignment of strains from Sudan to the 2 described clusters represented by the Arabia_1999_7 and the Central Africa strains.
Unlike the N gene amplifi cations, amplications with F-specifi c primers (8) resulted in a number of negative samples by RT-PCR because of mismatches occurring mainly in the F2 reverse primer (Table 3). To circumvent this problem, another reverse primer (FPPRrev) was designed that matched with all lineages.
The sequences and the corresponding lineage classifi cation generated in this study were collated to fi eld information. Occurrence among species of the different lineages show that most of the isolates belong to lineage IV; all camel isolates grouped in the cluster Saudi Arabia, whereas only sheep isolates were found within the other cluster, Central Africa. Only sheep isolates were found in lineage III (Table 4). Geographic and times distribution according to animal species and lineage in Sudan during [2000][2001][2002][2003][2004][2005][2006][2007][2008][2009] show that lineage IV was circulating as early as mid-2000 and that progressive substitution of lineage III in a large zone encompassing the eastern, northern, Blue Nile, and Khartoum regions has taken place since this date ( Figure 3).  Figure 1. Other designated strains were as published (10,(20)(21)(22). The phylogram was generated by analyzing 1,000 bootstrap replicates; clusters were supported by bootstrap percentages >70%. Strains from Sudan are represented by prefi xes: cam, camel; cap, caprine; ov, ovine. The Kabete 0 strain of rinderpest (RBOK) vaccine strain of rinderpest virus retrieved in GenBank (accession no. Z30697) was used as an outgroup. Scale bar indicates nucleotide substitutions per site. -

Discussion
This study presents fi ndings from PPRV surveillance in Sudan over a 10-year period and the results of tests conducted in Sudan and in Morocco during Morocco's 2008 outbreaks (18). Analysis of 80 samples confi rmed the wide distribution of PPRV throughout Sudan, which has been known since 1971 (2,13,15). Genetic characterization of 26 samples positive for PPRV also provided strong evidence of the introduction and spread of Asia PPRV lineage IV in the country. Surprisingly, the number of samples containing the indigenous lineage III decreased dramatically during the study.
Viruses collected in Morocco also were classifi ed as lineage IV. The origin of the outbreaks in Morocco remains unknown, although Ayari-Fakhfakh et al. (30) recently reported a PPRV seroprevalence in Tunisia of up to 7.45%, indicating that the virus is present across a wider area of northern Africa.
Analysis of the N and F genes showed that 2 clusters were identifi ed in strains from Sudan: 1 related to a strain from Saudi Arabia, the other related to central Africa viruses. The 6 strains from Morocco collected during the 2008 outbreaks were related closely to the Saudi Arabia cluster.
Within the Saudi Arabia cluster, a remarkable genetic stability over a decade was observed; notably, all camel isolates fell into this cluster. The variability and nucleotide sequence was slightly higher for isolates in the central Africa cluster, within which only sheep were found. The inconsistent low genetic variability of isolates from Sudan may be a consequence of a species bias because of the limited contact between camels and sheep. When replicated in a single host that has a limited exposure to new variants, the viral genome thus may remain highly conservative. This hypothesis was verifi ed previously for the measles virus over a shorter period (1997)(1998)(1999)(2000). For the measles virus, it appears that the lower the level of circulation, the higher the sequence conservation (31). Camels were not regarded as possible hosts for PPRV until 1992, when a number of authors reported PPRV seroconversion in these animals (32)(33)(34)(35)(36). The fi rst documented outbreak of PPRV in camels, reported from Ethiopia in 1996, consisted of highly contagious respiratory syndromes with high illness rates but low death rates (17,37). The causative agent was confi rmed to be a lineage III PPRV (A. Diallo, unpub. data). This present study confi rmed the etiology of the disease in camels through the virologic and epidemiologic investigations and the isolation of PPRV. Surveillance of camels furthermore allowed the virus to be detected in consecutive outbreaks in Kassala, eastern Sudan (2004); Atbara, northern Sudan (2005); and Tambool, Blue Nile region, Sudan (2007) (16). Viruses from the same cluster also were recovered from sick sheep and goats in a large zone encompassing the eastern, northern, Blue Nile, and Khartoum regions as early as mid-2000. Under extensive pastoral farming conditions in these regions, camels may have served as a bridge with areas of northern Africa and contributed to the spread of a camel-derived strain of lineage IV, as seen in Morocco. In contrast, none of the camel and goat isolates were distributed in the central Africa cluster, although viruses of this cluster were cocirculating with camel viruses during the same period and in the same areas of northern Sudan, Khartoum, and Blue Nile.
The reason that isolates of lineage IV became predominant in Sudan, progressively replacing lineage III viruses, during the last decade remains unclear. Lineage IV has been present in Asia and in part of the Middle East for a long time, probably as low virulent strains (5). However, a constant rise of disease incidence recently has been associated with this lineage and suggests increased virulence (3). A virulent lineage IV strain may have been introduced in Africa during the 1990s, resulting in outbreaks in both camels and small ruminants. In parallel, rinderpest virus control and eradication may have favored the decline of cross-immunity in small ruminants and their increased risk for PPRV as predicted earlier by Taylor (38).