Genetic Characterization of the Hemagglutinin Genes of Wild-Type Measles Virus Circulating in China, 1993–2009

Background China experienced several large measles outbreaks in the past two decades, and a series of enhanced control measures were implemented to achieve the goal of measles elimination. Molecular epidemiologic surveillance of wild-type measles viruses (MeV) provides valuable information about the viral transmission patterns. Since 1993, virologic surveillnace has confirmed that a single endemic genotype H1 viruses have been predominantly circulating in China. A component of molecular surveillance is to monitor the genetic characteristics of the hemagglutinin (H) gene of MeV, the major target for virus neutralizing antibodies. Principal Findings Analysis of the sequences of the complete H gene from 56 representative wild-type MeV strains circulating in China during 1993–2009 showed that the H gene sequences were clustered into 2 groups, cluster 1 and cluster 2. Cluster1 strains were the most frequently detected cluster and had a widespread distribution in China after 2000. The predicted amino acid sequences of the H protein were relatively conserved at most of the functionally significant amino acid positions. However, most of the genotype H1 cluster1 viruses had an amino acid substitution (Ser240Asn), which removed a predicted N-linked glycosylation site. In addition, the substitution of Pro397Leu in the hemagglutinin noose epitope (HNE) was identified in 23 of 56 strains. The evolutionary rate of the H gene of the genotype H1 viruses was estimated to be approximately 0.76×10−3 substitutions per site per year, and the ratio of dN to dS (dN/dS) was <1 indicating the absence of selective pressure. Conclusions Although H genes of the genotype H1 strains were conserved and not subjected to selective pressure, several amino acid substitutions were observed in functionally important positions. Therefore the antigenic and genetic properties of H genes of wild-type MeVs should be monitored as part of routine molecular surveillance for measles in China.

Principal Findings: Analysis of the sequences of the complete H gene from 56 representative wild-type MeV strains circulating in China during [1993][1994][1995][1996][1997][1998][1999][2000][2001][2002][2003][2004][2005][2006][2007][2008][2009] showed that the H gene sequences were clustered into 2 groups, cluster 1 and cluster 2. Cluster1 strains were the most frequently detected cluster and had a widespread distribution in China after 2000. The predicted amino acid sequences of the H protein were relatively conserved at most of the functionally significant amino acid positions. However, most of the genotype H1 cluster1 viruses had an amino acid substitution (Ser240Asn), which removed a predicted N-linked glycosylation site. In addition, the substitution of Pro397Leu in the hemagglutinin noose epitope (HNE) was identified in 23 of 56 strains. The evolutionary rate of the H gene of the genotype H1 viruses was estimated to be approximately 0.76610 23 substitutions per site per year, and the ratio of dN to dS (dN/dS) was ,1 indicating the absence of selective pressure.
Conclusions: Although H genes of the genotype H1 strains were conserved and not subjected to selective pressure, several amino acid substitutions were observed in functionally important positions. Therefore the antigenic and genetic properties of H genes of wild-type MeVs should be monitored as part of routine molecular surveillance for measles in China.

Introduction
Measles virus (MeV), a member of the genus Morbillivirus of the family Paramyxoviridae, is an enveloped virus with a nonsegmented negative-sense RNA genome. Measles virus is highly contagious and causes a disease characterized by high fever, cough, coryza, conjunctivitis, and appearance of a maculopapular rash [1]. Despite the existence of an effective vaccine, many developing countries are still experiencing endemic measles. In China, measles remains a major public health concern because of frequent outbreaks although the two effective vaccines, Shang-hai191 (S191) and Changchun47 (C47) have been widely used to prevent illness. In 2005, China set a goal of measles elimination by 2012, and the goverment has been making great efforts to achieve this goal including strengthening routine immunization, expanding epidemiologic and virologic surveillance, increasing vaccination of internal mobile poopulations, and initiation of large-scale supplemental immunization acitivities (SIA). For example, the SIA of September 2010 targeted 100 million children and teenagers.
The genome of MeV contains six genes that encode the nucleoprotein (N), phosphoprotein (P), matrix (M), fusion (F), hemagglutinin (H), and polymerase (L) proteins. The H protein, an 80 kilodalton (kD) glycoprotein, is responsible for receptorbinding and is the major target for neutralizing antibodies [11]. The H protein usually contains five potential N-linked glycosylation sites which are clustered at amino acid positions 168, 187, 200, 215, 238. Some genotypes have an additional N-linked glycosylation site at position 416 [12,13]. MeV uses up to three cellular receptors, including Signaling Lymphocyte Activation Molecule (SLAM, CD150), CD46 and nectin-4 [14,15,16,17] . The H protein contains 13 cysteine residues, seven of which are located at amino acid positions 287, 300, 381, 394, 494, 579, 583. These amino acids play a critical role in maintaining the antigenic structure of the H protein [18]. A few linear neutralizing epitopes have been identified using monoclonal antibodies [19,20].
We carried out this study to describe the genetic variability of the H genes of MeVs circulating throughout China over a 17-year period. This work provides the first detailed report of the sequences of the H genes of Chinese MeVs as well as a unique opportunity to analyze genetic changes in the H genes from a single genotype of MeV over an extended period of time.

The epidemiologic profile of measles in China
The number of reported measles cases and deaths and the incidence of measles were obtained from the National Notifiable Disease Report System (NNDRS) (Figure 1). Nationwide outbreaks of measles occurred every 3-4 years because of the accumulation of susceptible children, especially in areas with lower routine immunization coverage. Viral isolates were obtained throughout the period shown in Figure 1.

Genetic characterization of the H genes of Chinese MeVs
The H glycoprotein is the major target for neutralizing antibodies directed against the virus. For this report, 46 new H gene sequences were obtained from viral isolates and compared to 10 previously published H gene sequences. Analysis of the predicted amino acid sequences of the H proteins showed that genotype H1 cluster1 isolates circulating in 2000-2009 had a conserved substitution of Ser240Asn, which removed the predicted N-linked glycosylation site at amino acid 238 ( Figure 2). In addition, 23 of 56 genotype H1 strains showed an exchange of Pro397Leu, this 397 amino acid is a part of the linear hemagglutinin noose epitope (HNE) which is located at amino acids 379-410. However, all of the putative binding sites for SLAM and CD46 were conserved ( Figure 2). The seven cysteine residues presumably responsible for the antigenic structure of the H molecules (amino acids 287, 300, 381, 394, 494, 579, and 583) were highly conserved [18]. The amino acids 236-250 predicting B Cell epitopes (BCE) were relatively conserved [21], excluding the previously described substitutions at amino acids 240 and 243 in the Chinese vaccine strains (Arg243, Gly243).
The complete H gene sequences of 56 H1 MeVs, the Chinese vaccine strains, and other genotype reference strains were used for phylogenetic analysis (Figure 3). The sequences of genotype H1 isolates could be divided into at least 2 clusters without obvious chronological and geographical distributions. Fifty-one isolates belonged to cluster1 which also contained a few more diverse strains isolated in 1993-1994. Five isolates belonged to cluster2.
The topology of the tree suggested that multiple chains of virus transmission were present and the sequences showed no temporal or geographic distribution ( Figure 3).
The evolutionary rates of the H gene for the genotype H1 and H1 cluster1 viruses were 0.76610 23 and 0.65610 23 substitutions per site per year, respectively. The dN and dS were calculated for selection pressure analysis, the results indicated the dN/dS ratios for the genotype H1 and H1 cluster1 were 0.21 and 0.20, respectively.

Discussion
The results of the sequence analysis reported here provide another example of the genetic stability of MeV. The H genes of 56 MeVs isolated over a 17 year time span showed remarkable conservation of functionally important amino acids. In particular, the stability of the amino acids comprising the receptor-binding sites likely contributes to the monotypic nature of MeV [22][23][24]. The cysteine residues responsible for the tertiary structure of the H molecule were also highly conserved [18] . The amino acid substitutions detected appeared to be the result of a gradual accumulation of genetic changes and the degree to which these amino acid mutations contribute to antigenic changes of the H protein is being investigated.
The H protein of MeV usually contains five or six potential N-linked glycosylation sites. The fifth site, located at amino acid 238, was absent in all the H1 cluster1 strains circulating in 2000-2009. Hu et al. showed that the fifth glycosylation site had minimal impacted on the processing and antigenicity of the H molecule, though the other glycosylation sites played important functional roles [25,26]. Shi et al. demonstrated that the neutralization titers of serum samples from human vaccinees were lower against some of the wild-type viruses than against vaccines strains and antigenic differences have been detected by monoclonal antibodies directed against the H protein [27,28]. The linear hemagglutinin noose epitope (HNE; amino acids 379-410) was highly stable in both vaccine and wild-type strains [29][30][31]. We found that nearly half of the genotype H1 strains showed an exchange of Pro397Leu, a mutation that results in loss of recognition of two monoclonal antibodies directed against HNE [32]. Though the H protein of a few genotype H1 MeVs would presumably not be recognized by antibodies directed against the HNE, serum from human vaccines neutralized H1 MeVs with either Pro or Leu at position 397 [36]. This conservation of neutralizing epitopes was also noted by Bouche et al. [33]. The strong conservation of the seven cysteine residues also clarified the antigenic structure and processing of the H molecules [18]. The Chinese vaccines have been very effective at dramatically reducing the incidence of measles in China. However, since amino acid substitutions were observed on the H protein between genotype H1 strains and the Chinese vaccine strains, it will be important to continually monitor the antigenicity and genetic characteristics of the H protein of circulating MeVs.
Molecular surveillance of wild-type MeVs from 1993 to 2008 based on analysis of the 450 nucleotides coding for the carboxylterminus of the N gene indicated that genotype H1 cluster1 viruses were continuously circulating in China [6,9,10]. The phylogenetic tree constructed based on the H gene sequences, confirmed the analysis based on the N-450 sequences that genotype H1 cluster1 viruses were the predominant strains in China during 1993-2009. Cluster1 was divided into 3 lineages and lineages 1 and 2 were considered to be an intermediate group that appeared to be the predominant strains in 1993-1994 and have not been found since 1995 [6][7][8]. Lineage 3 was further divided into lineage 3a and 3b and all of the strains isolated from 2000-2009 belonged to lineage 3b. Genotype H1 cluster2 viruses have not been detected since 2002. Although there was a temporal distribution among the MeV strains of the 3 lineages of cluster1, multiple transmission chains of were apparent in lineage 3b without an obvious temporal distribution ( Figure 3). The observation that the phylogenetic analysis based on the H gene sequences gave results that were consistent with results obtained after analysis of the much larger set of N-450 sequences suggests that the H gene sequences provide a reasonable representation of MeVs circulating in China. The caveat is that virologic surveillance is incomplete and we cannot be sure that all of the lineages were represented by the H gene or N-450 data sets. Viral RNA polymerases have high error rates due to the lack of proofreading ability, which leads to the high mutation rate of the viral genome and rapid evolutionary rates. In some cases, these mutations allow the virus to adapt to a new host or escape the host's anti-viral responses. This study estimated that the mutation rate of the predominant group (Cluster1) of genotype H1 MeVs was 0.65610 23 8610 23 ), and equivalent to the mutation rates reported for dengue virus and human rotavirus [34][35][36]. The dN to dS ratios for both H1 and H1 cluster1 were ,1 that demonstrating that the H gene of the MeVs analyzed was not subject to antigenic selection. Rather, the data suggest that the amino acid substitutions in the H gene were the result of random genetic drift, rather than accumulated mutations.
In summary, the H gene of the MeVs endemic in China should be monitored continuously for genetic variations that could affect antigenic properties. Our group is currently using the bioinformatics methods to understand the disappearance of cluster2 and map the antigenic domains on the 3-dimentional crystal structure Figure 2. Alignment of the predicted amino acid sequences of the partial H gene. The variation of amino acid Ser240Asn (highlighted in blue) leads to the absence of a glycosylation site; the exchange of Pro397Leu (highlighted in pink) that results in loss of recognition of two monoclonal antibodies directed against HNE; the amino acid residues of putative binding sites for CD46, SLAM and seven cysteine residues were highlighted in red, yellow and purple, respectively. doi:10.1371/journal.pone.0073374.g002 of the H protein. This information will be helpful to evaluate the efficiency of the current vaccines.

Epidemiologic information
As a class B reportable disease, suspected measles cases are reported to the National Notifiable Disease Reporting System (NNDRS) in China. The number of measles cases, the annual measles incidence, and measles-associated deaths were retrieved directly from NNDRS reports.

Selection and isolation of viruses
In 2001, the National Measles Laboratory Network was established to carry out measles surveillance in China. Genotyping and genetic analysis of wild-type viruses were included in the laboratory surveillance. These strains were chosen from viral strains bank of the National Measles Laboratory (NML) to achieve representative chronological and geographical distributions (covering 25 provinces). All of the selected strains were cultured and propagated in Vero/hSLAM cells, which were maintained in minimal essential medium supplemented with 2% fetal bovine serum [37]. Viruses were harvested when the classic measles CPE was maximal.

Determination of the complete H gene nucleotide sequences
Total viral RNA was extracted from the infected Vero/hSLAM cells with the QIAamp Viral RNA mini kit (Qiagen, Valencia, CA) according to the manufacturer's instruction. The entire proteincoding region of the H gene was amplified with the sequencespecific primers (Table 1) by using the SuperScript TM III One-Step RT-PCR kit (Invitrogen, Carlsbad, CA). The PCR products were sequenced directly after purification (QIA gel extraction kit, Qiagen, Valencia, CA) by the dye terminator method (BigDye Terminator, version 3.1, cycle sequencing kit; Applied Biosystems) in an ABI PRISM 3100 genetic analyzer (Applied Biosystems, Hitachi, Japan). The primers used for sequencing are listed in Table 1.

Sequence and phylogenetic analysis
Sequence data were assembled and edited using Sequencher software version 4.0.5 (GeneCodes, Ann Arbor, MI); sequence alignments were assembled with BioEdit version 7.0 (Tom Hall, North Carolina State University, Raleigh, NC), and phylogenetic trees were constructed by the neighbor-joining method (1000 bootstraps) using MEGA version 4.0 [38,39]. The 24 WHO reference genotype strains, 56 genotype H1 strains and two Chinese vaccine strains were used to construct the phylogenetic trees.

Estimation of evolutionary rate
Because of the lack of a true ''ancestor'' strain and the apparent presence of multiple lineages, to estimate the evolutionary rate of genotype H1 and genotype H1 cluster1 viruses, we used the oldest strain in each target group as a reference. The evolutionary distances were calculated by overall mean scope according to the Kimura two-parameter method of the MEGA program. The dS (synonymous substitutions per synonymous site), dN (non-synonymous substitutions per non-synonymous site), and the ratios of dS to dN values also were analyzed.

Nucleotide sequence accession numbers
Nucleotide sequences of the complete H gene (1854 nucleotides) of the 46 measles virus strains which were determined have been submitted and deposited in the GenBank database, the accession numbers are JN997482 to JN997527. Accession numbers of Chinese circulating strains in 1993-1994, all the WHO reference strains of each genotype and Chinese vaccine strains are given in Figure 3.