|Home | About | Journals | Submit | Contact Us | Français|
There have been no long-term systematic analyses of the molecular epidemiology of human parvovirus B19 (B19V). We investigated the variations of nucleotide sequences of B19V strains collected in Sapporo, Japan, from 1980 to 2008. In that period, six outbreaks of erythema infectiosum occurred regularly at 5-year intervals. The B19V strains collected successively, regardless of the outbreak, were analyzed for nucleotide variation in the subgenomic NS1-VP1u junction. The isolated strains can be classified into 10 subgroups. Two patterns of change of endemic strains were observed. One was a dynamic replacement of strains that occurred almost every 10 years, and the other was a gradual change consisting of an accumulation of point mutations.
Human parvovirus B19 (B19V), a small, nonenveloped, single-stranded DNA virus, belongs to the genus Erythrovirus and the family Parvoviridae and is the only erythrovirus known to be pathogenic in humans. B19V usually transmits via a respiratory route and can also be infectious in blood products or blood components (4). B19V DNA replicates restrictively within erythrocyte precursors in bone marrow (23). The most common presentation of B19V infection is erythema infectiosum (fifth disease), which is generally benign and self-limiting (1). B19V causes aplastic crisis in patients with congenital hemolytic anemia (25). Many reports suggest an association between B19V infection and various clinical manifestations, such as hydrops fetalis, arthritis, hepatitis, and meningoencephalitis (2, 5, 26, 33).
The function of NS1, which is composed of 671 amino acids (aa), is related to the transactivation of protein for viral DNA replication, and NS1 is known to have cytotoxic properties (9, 18). Another open reading frame encodes both VP1 (781 aa) and VP2 (554 aa) for capsids. VP1 and VP2 are identical except for the additional 227-aa N terminus called the VP1 unique region (VP1u), which is exposed to the surface of infected cells and the target sites of neutralizing antibodies (17, 27).
Genetic diversity among B19V strains has been considered to be low, and nucleotide variation is less than 1 to 2% for the whole genome (7, 10). However, recent reports suggest that the genetic diversity of B19V is greater than previously believed. B19V strains are now subdivided into three distinct genotypes: 1, 2, and 3 (11, 20, 21, 29). Genotype 1 is thought to be the most common worldwide.
Almost all reports about polygenetic analyses of B19V sequences have shown that genetic diversity is correlated with the respective places and years of isolation (7, 8, 10, 13). However, these studies were based on unsystematic data with an uncoordinated distribution of places and years because sampling occurred at random. We have screened serum samples collected over the last 3 decades for B19V DNA using a nested PCR assay and analyzed these sequences to determine the epidemiological distribution and transmission patterns of B19V. Our study tracks the genetic evolution of the virus more precisely because it is based on serum samples collected from a restricted region over a long period. We report that genotype 1 isolates from Japan can be classified into several subgroups by analyzing the nucleotide variation of the subgenomic NS1-VP1u junction and that sudden substitutions occurred at regular intervals.
Approximately 18,000 serum samples obtained from inpatients and outpatients in Sapporo Medical University and the related hospitals in Sapporo, Japan, from 1980 to 2008 were assayed for the presence of B19V DNA. Samples positive for B19V DNA were then examined for anti-B19V immunoglobulin M (IgM) and IgG by an enzyme-linked immunosorbent assay. A kit with recombinant antigens synthesized in a baculovirus expression system (Denka Seiken Co., Tokyo, Japan) was used. The antigens in this kit were self-assembled recombinants containing both VP1 and VP2 in the same proportions found in native B19V capsids. Anti-B19V IgM was detected by antibody capture methods, and IgG was detected by indirect methods.
The synthesized oligonucleotide primers used for PCR are located in the VP2 region: E1 (sense, 5′-GCT GTT AAG GAT GTT ACA GA-3′, nucleotides [nt] 3502 to 3521) and E2 (antisense, 5′-GGA TCC GTA TAA GGG ATT GT-3′, nt 3882 to 3901) for the first PCR and E3 (sense, 5′-CAG GTT ACT GAC AGC ACT AC-3′, nt 3541 to 3560) and E4 (antisense, 5′-TGT TGA CTG CAG CCC TCT AA-3′, nt 3848 to 3867) for the second PCR. Ten serum samples (10 μl each) were pooled into a single specimen. Five microliters of this 100-μ1 specimen was used for the template for PCR. Template dissolved in 30 μl of distilled water was added to the PCR mixture. The total mixture was adjusted to 50 μl. Thirty-five cycles of 94°C for 1 min, 60°C for 2 min, and 72°C for 2 min were carried out. The products were analyzed on a 1.5% agarose gel by electrophoresis in a Tris-acetate-EDTA buffer containing ethidium bromide. The expected size of the PCR products was 307 bp. When a specimen was positive, the 10 original serum samples were retested separately to identify the positive sample. This assay is sensitive enough to detect a few copies of the B19V genome.
DNA was extracted from 200 μl of positive serum samples by using the QIAmp UltraSens virus kit (Qiagen, Japan) according to the manufacturer's instructions. Nucleic acids were eluted from the filter column with 30 μl of nuclease-free double-distilled water and were stored at −20°C.
Viral DNA for sequencing was obtained by nested PCR amplification of a 1,027-bp region (nt 1738 to 2765) spanning the NS1-VP1u junction for B19V DNA-positive cases. Nested PCR was performed using the following primer pairs: Pfo (sense, 5′-GGC ATG GTT AAC TGG AAT AA-3′, nt 1481 to 1499) and Pro (antisense, 5′-TGC AGC ACT GTC AAC AGC ACT-3′, nt 2873 to 2894) for the first PCR and Pfi (sense, 5′-AAG TCT ACA ATT GTA GAA GC-3′, nt 1564 to 1593) and Pri (antisense, 5′-CAG GCT TGT GTA AGT CTT CAC-3′, nt 2778 to 2798) for the second PCR. Both the first and the second PCRs were carried out with 35 cycles (at 94°C for 45 seconds, 60°C for 15 seconds, and 72°C for 1 min), and 1.25 units of PrimeSTAR HS DNA polymerase (Takara Bio Co. Ltd., Japan.) was used to ensure more-precise sequence results. Sequence analysis of DNA fragments amplified using PrimeSTAR HS demonstrated a mutation frequency of only 0.005%. PrimeSTAR HS DNA polymerase has a higher fidelity than alternative high-fidelity enzymes and 10 times higher fidelity than Taq DNA polymerase. The PCR products were visualized by electrophoresis in a 1.5% agarose gel and subsequently purified by using Microcon centrifugal filter devices. The sequencing reaction was carried out with the BigDye Terminator cycle sequencing kit, version 3.0 (PE-ABI). The products were cleansed of any free dye dideoxynucleotides using the BigDye XTerminator purification kit and finally were analyzed with an ABI PRISM 3100 genetic analyzer (Applied Biosystems) according to the manufacturer's instructions.
B19V sequences were aligned by using CLUSTAL_W and BLAST (DNA Data Bank of Japan; http://www.ddbj.nig.ac.jp/Welcome-j.html). The reliability of the alignment was additionally checked, phylogenetic trees were constructed by the neighbor-joining method (28), and genetic distances were calculated by the Kimura two-parameter method (14) using MEGA4 software (http://www.megasoftware.net).
One hundred four cases were positive for B19V DNA (54 males, 50 females; mean age, 6.9 years; range, 0.5 to 20 years) with the nested PCR assay. These cases had developed various clinical manifestations, such as erythema infectiosum, papular purpuric gloves and socks syndrome, aplastic crisis, encephalopathy, acute transverse myelitis, and hepatitis. They were examined for anti-B19V IgM and IgG. Results for both anti-B19V IgM and IgG were positive in 95 samples. Six samples were negative for anti-B19V IgM and positive for anti-B19V IgG, suggesting that they were obtained in the late phase of infection. One sample was positive for anti-B19V IgM and negative for anti-B19V IgG. The remaining two samples were both anti-B19V IgM and anti-B19V IgG negative. These last three samples had very high B19V DNA titers, up to 1 × 1012 genomes/ml, and were probably obtained in the early acute phase of infection.
The 1,027-nt sequences ranging from nt 1738 to 2764 in the NS1-VP1u regions were determined for all 104 strains in our study, isolated from 1980 to 2008 in Japan. The nucleotide mutations and amino acid substitutions of these strains were compared to the sequence of the B19V strain Au (numbering according to GenBank accession no. M13178), which was isolated in North America (30). All nucleotide differences were attributed to base substitutions. Compared to the standard strain, no insertions or deletions among the isolates were identified. The nucleotide and amino acid diversity of the NS1-VP1u coding region between each strain in our study and strain Au was close to 2%, as previously published (13).
Ninety-nine nucleotide mutation sites were found in the sequences of the 104 strains. Of these 99 sites, 61 single mutation sites were sporadic, not specific for a certain subtype. In all of the 104 strains, there were nucleotide differences from strain Au at nt 2268 and 2453. The other nucleotide mutation sites are shown in Fig. Fig.1.1. Based on these nucleotide mutations, the 104 isolates could be divided into 10 subtypes (A1, A2, AB, B, BC1, BC2, C1, C2, D1, and D2). Subtype A had 9 to 16 nucleotide mutations. There were several nucleotide differences between subtypes A1 and A2, i.e., a change from G to A (to C in strain Au) at nt 1749 and variations at nt 1884, 2412, 2148, and 2427. There was only one characteristic mutation at nt 2352 in subtype B compared to strain Au. E27, isolated in 1991, has the same mutations at nt 2107, 2400, and 2527 as subtype A and therefore was designated subtype AB. Subtype C has mutations at nt 1865, 2011, 2096, 2416, and 2539. Moreover, subtype C was subdivided, with the subtype with an additional mutation at nt 2412 designated C2 and the rest designated C1. The intermediate type between B and C was designated type BC. Subtype BC1 has additional changes at nt 2260, 2439, and 2618, while BC2 does not. Subtype D1 has mutations at nt 2244, 2392, 2531, 2578, and 2736. D2, with additional mutations at nt 2370, 2461, and 2638, could be distinguished from D1.
We experienced six outbreak seasons of erythema infectiosum, one every 5 years, from 1980 to 2008 (Fig. (Fig.2A).2A). All subtype A viruses were isolated before 1987; A1 was dominant in 1981 to 1982, and A2 was dominant in 1986 to 1987. Subtype B was dominant in 1991 to 1992, and subtype C was dominant in the 1996-to-1997 outbreak seasons. Subtype D was detected sporadically in 1988, 1994, and 1996. It became a main subtype in the latter half of 1990.
The nucleotide sequences of the subgenomic NS1-VP1u region of B19V were analyzed by using the phylogenetic tree method (Fig. (Fig.3).3). The two previous reference sequences are those of strain Au, isolated in the United States in 1982 from a child with transient aplastic crisis, and strain Wi (GenBank accession no. M24682), isolated in the United Kingdom in 1973 from a healthy blood donor (3, 30). There were at least four clusters in our analysis. Cluster 1 consisted of subtypes B and BC1 and strain Au; cluster 2 comprised BC2, C1, and C2; cluster 3 comprised D1 and D2; and cluster 4 comprised A1 and A2. Reference strain Wi and E27/1991 had no clusters with our main strains.
B19V transmits via the respiratory tract. Immunocompetent individuals who are infected by B19V usually have coldlike symptoms with fever and headache during a period of high-titer viremia, and then erythema appears 2 to 3 weeks after the onset. The viremia lasts about a week and rapidly resolves following seroconversion for antibody against B19V, but low-level viremia sometimes persists for several months (16). In recent studies, lifelong persistence of viral DNA was detected in various tissues, such as skin, liver, synovium, and myocardium, but not in serum (6, 12, 15, 22, 24, 31). Therefore, serum samples are considered suitable material to investigate the evolution of B19V, compared with the other tissue samples. In this study, the analysis was based on clinical cases with acute infections.
Genetic diversity within B19V genotype 1 has been previously described. Mori et al. (19) first classified B19V strains into four genome types on the basis of analysis using 13 restriction enzymes. Umene and Nunoue (32) also analyzed B19V strain variation by a restriction enzyme fragment length polymorphism assay and demonstrated that B19V isolated in 1981 differed from isolates detected from 1986 to 1987 in Japan. In the present study, the isolates of subtype A1 from 1981 to 1982 and of subtype A2 from 1986 to 1987 were also different. Kerr et al. (13) divided B19V isolates into six types (1a, 1b, and 2 to 5) by a single-stranded conformational polymorphism assay and partial nucleotide sequencing. They collected 50 samples from various countries (United Kingdom, United States, and Japan). Fukada et al. (8) also showed four subtypes (A to D) of B19V isolates using amino acid polymorphism. The serum samples were obtained in 1980 to 1981 in Tokyo and in 1992 to 1993 and 1996 in Kyushu, which is situated in southwest Japan. Both reports pointed out that genotypes differed according to the time and collection location, while there was no correlation between genome type and clinical manifestation. In the present study, we utilized serum samples from patients with various clinical types of B19V infection; however, no specific correlation was found between clinical features and nucleotide/amino acid sequences.
The representative strain of each subtype can be selected as follows: E1/1980 in subtype A1, E7/1986 in A2, E27/1991 in AB, E14/1990 in B, E47/1993 in BC1, E68/1996 in BC2, E44/1992 in C1, E67/1996 in C2, E84/2003 in D1, and E98/2007 in D2. The numbers of nucleotide differences between Au and E1/1980, E7/1986, and E27/1991 are 11, 17, and 7, respectively, and those between Wi and E1/1980, E7/1986, and E27/1991 are 8, 9, and 5, respectively. There have been no strains similar to subtype A1 found anywhere in the world. The one most similar to subtype A2 is the strain with accession number AB126265, isolated in Japan, which differs in two bases. These results suggest that subtype A is endemic in Japan and unique to Japanese isolates. The strains most similar to subtypes BC1, C1, BC2, and C2 were those isolated in northern Europe (accession numbers AY504945 [United Kingdom] and AY028234 [Sweden]). Subtypes D1 and D2 were similar to the German (AJ781038) and Japanese (AB126267) strains.
In this study, nucleotide variations in B19V genotype 1 over 28 years were analyzed in serum samples obtained sequentially. In previous reports on the genetic evolution of B19V, the places where the samples were collected, the kinds of materials (serum and/or tissue), and the times varied widely and unsystematically. Our study was restricted to serum samples all gathered in the same fixed region. We speculate that the evolution of B19V strains seems to have at least two patterns (Fig. (Fig.1).1). Drastic alteration from subtype A to B in the late 1980s might have been caused by a replacement with a new, stronger strain. The change from C to D1 also seems to be significant. Subtype D1 has existed for 10 years and became a main endemic strain in the latter half of the 1990s. The changes from subtype A1 to A2, B to C, and D1 to D2 were gradual; these may be caused by point mutations. It was impossible to guess the genetic source of each subtype in the past reports. Subtype A could be an original strain in Japan, while subtype B might be descended from strain Au, isolated in North America. The other subtypes were similar to strains found in northern Europe. Genotype 1 strains may be circulating in the world.
Six outbreak seasons of B19V infection occurred regularly at 5-year intervals from 1980 to 2008 in Japan. The drastic changes from subtype A2 to subtype B and from subtype C to subtype D1 occurred in nonoutbreak periods in 1989 and 1998, respectively. The dynamic alteration of endemic strains seems to occur every 10 years (Fig. (Fig.2B);2B); however, those alterations were not always associated with outbreaks of B19V infection. The herd immunity to B19V in nonoutbreak periods could permit an invasion of novel B19V; however, it may need some special conditions to cause another outbreak. Our study was intended to determine the distribution and evolution of B19V genotype 1 in Sapporo, Japan. Based on nucleotide polymorphism in the NS1-VP1u region, 104 strains of B19V could be classified into 10 subgroups. There are two patterns of change of the endemic strains. One is a dynamic replacement of strains that occurred about every 10 years, and the other is a gradual change consisting of an accumulation of point mutations.
We thank Peter M. Olley for helpful advice and English revision of our paper.
Published ahead of print on 26 August 2009.