Molecular epidemiology of dengue viruses in southern China from 1978 to 2006

Dengue fever (DF) and two more severe syndromes, dengue shock syndrome (DSS) and dengue hemorrhagic fever (DHF) are important mosquito-borne diseases in tropical and subtropical regions [1, 2]. Since the first documented outbreak in 1779 in Jakarta, Indonesia, outbreaks have been documented in tropical and subtropical regions. It has been the maximum public health burden in South-East Asia countries [3]. Dengue epidemiology changes fast among epidemic countries, and keeps on expanding to the non-epidemic area [1]. Since the pathogens were first discovered by Japanese scientists in 1943, dengue viruses (DENV) were isolated from almost all South-East Asia countries including Thailand, Cambodia, Vietnam, Laos, Myanmar, Malaysia, Philippines, and Indonesia [3]. The dengue virus belongs to Flaviviridae family, and has four closely related but different serotypes (DENV-1 through -4) in nature that are circulating or co-circulating [1, 2]. Their genetic diversity remains wide, due to 1) absence of a proof-reading capacity in RNA-dependent RNA polymerases [4], 2) emergence of different lineages or clades during epidemic [5, 6], 3) increasing natural recombination [7‐10], and 4) co-circulation of more than one serotypes in a locality [11‐14].

Since imported DF epidemic had been reported in Hankou, Hangzhou, Shanghai and Guangzhou in 1920s and 1940s [15], there was no DF case reported in China till the outbreak occurred in 1978 in Foshan, Guangdong Province, with DENV-4 infection [16]. And then DF was prevalent in Guangdong, Guangxi and Hainan province. A DENV-1 epidemic occurred in 1979 in Zhongshan, Guangdong Province. This serotype of virus continued causing outbreaks over 2-3 years intervals. It become the dominating serotype and caused the latest outbreak in 2006 [17, 18]. DENV-3 epidemic was only recorded once in 1980 at Zhan County, Hainan Island, where in 1985 DENV-2 caused an epidemic during which the first DHF case was reported [19]. After this DENV-2 epidemic, DENV-2 continued to be transmitted into Guangdong, Guangxi and Hainan until 2001, including three outbreaks in Foshan in 1993, 1998 and in Jiangmen in 2001 [20]. The second DENV-4 outbreak occurred in Guangzhou city in 1990 [20]. It was estimated more than 700,540 hospitalized cases with 513 deaths from 1978 to 2007 [15]. Although the first isolate was sampled 30 years ago, dengue epidemic in southern China keeps increasing [15]. Lack of longitudinal research on dengue epidemics has hampered our understanding of dengue molecular origin and evolution in China

In this study, we determined the complete genome sequences of 14 dengue isolates collected in southern China during each epidemic year between 1978 and 2006 and E gene sequences from six patients of Guangdong, 2006. In combination with those published sequences in GenBank, we conducted an extensive molecular epidemiological analysis, aiming to determine where the DENV isolates in China originally came from, and what shaped their evolution.

Since the first well documented DF outbreak in Foshan of Guangdong province in 1978, dengue has been reported periodically in southern China, with the latest epidemic at Guangdong province in 2006. In this study, 14 Dengue isolates from 1978 to 2006 were sequenced and analyzed for molecular evolution.

Dengue virus exists as four antigenically distinct viruses designated as serotypes (DENV-1 through DENV-4), belonging to genus Flavivirus of family Flaviviridae. It has a positive-sense RNA genome that is translated as a single polyprotein and posttranslationally cleaved into three structural proteins and seven nonstructural proteins [28]. The envelope protein (E) is considered to be the immunodominant protein [29]. Dengue viruses also could be divided into different genotypes by the E gene [30, 31]. So there is correlation between serotypes and genotypes of Dengue virus. Phylogenetic trees of DENV-1, -2 and -4 generated by ORF and E gene sequences showed that there was a wide genetic diversity of DENVs isolated in southern China (Figure 1, Figure 2). DENV-1 strains exist in almost all of the clades of genotype I and IV but Asia 1 clade of genotype I (Figure 1B, Figure 2). DENV-2 strains are grouped into four genotypes except American genotype (Figure 1C, Figure 2). DENV-4 strains are grouped into 2 genotypes (I and II) (Figure 1D, Figure 2). It also showed that different serotypes and genotypes epidemic prevalence exist in the same city and even during the same epidemic (Figure 1, Figure 2). It indicated that these dengue viruses maybe have different origination.

Some of isolates, especially the first emerging of certain dengue serotype strains in southern China were closely related to those strains which were isolated in prior epidemics from neighboring Southeast Asia countries. It indicated that these dengue epidemics may be imported into China from the neighbor countries. For DENV-2, there were two outbreaks in Foshan in 1993 and 1998. Isolates from 1998 epidemic was closely related to the isolates from 1993 epidemic in Thailand (Figure 1C). The same situation was showed in DENV-4 isolates from 1978 and 1990 epidemics. Isolates from 1978 epidemic was closely related to the isolates from 1973 epidemic in Indonesia (Figure 1D). Isolates form 1990 epidemic was closely related to the isolates from 1984 epidemic in Philippines (Figure 1D). It indicated that these epidemics maybe imported into China from neighbor countries.

Although some data showed the dengue fever were imported from neighbor countries, those epidemics with the same serotype of DENV during continues years showed that endemic infection of dengue circulating locally may be also the important cause of Dengue epidemic in southern China. In the study, two isolates (CN.GD01/97 and CN.GD99/99) clustered together as a new Asia-4 clade (Figure 1B), suggesting that the isolates of 1999 maybe originated from the isolates of 1997. Sequence analysis showed that the DENV-1 strains isolated during the 2003 epidemic were closely related to the strains isolated during the 2002 epidemic of Guangzhou, suggesting that the 2002 strains had evolved in local and eventually cause the epidemic of 2003 (Figure 1B). Similar case was also showed in Guangzhou outbreak of DF in 2006. The DENV-1 strain isolated during the 2006 epidemic is closely related to the strains isolated during the 2001 epidemic (Figure 1B). It suggested that DENV probably circulated in local and caused the epidemics of 2001 and 2006. Moreover, there were several clear divisions within the Chinese 2006 isolates in the tip of the tree, even the eight strains collected in Guangzhou city by us during the epidemic of DENV-1 in 2006 also clustered into 2 subclades in Asia 3 clade, showing viral evolution or increase of genetic complexity even during a single epidemic (Figure 1B).

Recombination plays a role in dengue virus evolution and adaptation [7]. We analyzed the potential recombination events of all the Chinese isolates (43 DENV-1,20 DENV-2,2 DENV-3,5 DENV-4) on E gene. Three recombinant events were detected on the E gene of DENV-1 with low P-value (P < 0.01) (Table 2). The incongruence of phylogenetic trees constructed separately by recombinant regions and non-recombinant regions confirmed the recombinant events (Figure 3). The recombination event 1 showed recombination between present GD14/97 isolate [GenBank: AY376737] as the major parent and the GD95/95 isolate [GenBank: FJ196846] as the minor parent, which led to a recombinant isolate of GD23/95 [GenBank: AY373427](Figure 3A). The recombination event 2 showed recombination between present NHD98039/98 isolate [GenBank: EF508198] as the major parent and the GZ02/03 isolate [GenBank: DQ211349] as the minor parent, which led to two recombinant isolates of GD01/97 [GenBank: FJ196847] and GD99/99 [GenBank: FJ196848](Figure 3B). The recombination event 3 showed recombination between present STD06098/06 isolate [GenBank: EF508206] as the major parent and the 71/02 isolate [GenBank: EF025110] as the minor parent, which led to a recombinant isolate of GZ01/04 [GenBank: EF032589](Figure 3C). These studies of recombination provided an important context to understand the evolution of these viruses and sequence diversity generated in viruses in China.

We also analyzed the potential recombination across the 32 complete ORF of all the Chinese isolates, and found no inter-serotype recombination occurred. The results are coincided with previous data [8, 10, 32]. However, we found 22 intra-serotype potential recombination events (17 for DENV-1, 5 for DENV-2) (Table 3). Interestingly, for DENV-1, 13 of 17 potential recombination events happened in the strains isolated in 1995, 1997 and 1999 epidemics (Table 3), prior to which, in 1995, the severest dengue outbreak ever reported, indicating a possibility of viruses escaping human immunity by recombination following epidemic. However, whether the presumption makes sense or just a coincidence remains an open question.

Our study suggested that Dengue fever epidemic in Southern China over the past 30 years presents two important modes, 1) imported-cases-induced endemic prevalence; 2) endogenous epidemic outbreak with natural epidemic focuses. Recombination analysis showed that recombination plays an important role in dengue virus evolution and adaptation.

The authors declare that they have no competing interests.