Analysis of whole genome sequences of 16 strains of rubella virus from the United States, 1961–2009
- Emily Abernathy†1,
- Min-hsin Chen†1,
- Jayati Bera2,
- Susmita Shrivastava2,
- Ewen Kirkness2,
- Qi Zheng1,
- William Bellini1 and
- Joseph Icenogle1Email author
© Abernathy et al.; licensee BioMed Central Ltd. 2013
Received: 12 September 2012
Accepted: 16 January 2013
Published: 25 January 2013
Rubella virus is the causative agent of rubella, a mild rash illness, and a potent teratogenic agent when contracted by a pregnant woman. Global rubella control programs target the reduction and elimination of congenital rubella syndrome. Phylogenetic analysis of partial sequences of rubella viruses has contributed to virus surveillance efforts and played an important role in demonstrating that indigenous rubella viruses have been eliminated in the United States. Sixteen wild-type rubella viruses were chosen for whole genome sequencing. All 16 viruses were collected in the United States from 1961 to 2009 and are from 8 of the 13 known rubella genotypes. Phylogenetic analysis of 30 whole genome sequences produced a maximum likelihood tree giving high bootstrap values for all genotypes except provisional genotype 1a. Comparison of the 16 new complete sequences and 14 previously sequenced wild-type viruses found regions with clusters of variable amino acids. The 5′ 250 nucleotides of the genome are more conserved than any other part of the genome. Genotype specific deletions in the untranslated region between the non-structural and structural open reading frames were observed for genotypes 2B and genotype 1G. No evidence was seen for recombination events among the 30 viruses. The analysis presented here is consistent with previous reports on the genetic characterization of rubella virus genomes. Conserved and variable regions were identified and additional evidence for genotype specific nucleotide deletions in the intergenic region was found. Phylogenetic analysis confirmed genotype groupings originally based on structural protein coding region sequences, which provides support for the WHO nomenclature for genetic characterization of wild-type rubella viruses.
KeywordsRubella virus Whole genome
Rubella virus (RV) is a positive-polarity, single-stranded RNA virus and the sole member of the Rubivirus genus of the Togaviridae family. The virus causes a mild childhood disease, but is also a potent teratogenic agent when contracted by a pregnant woman. Therefore, the goal of rubella control and elimination programs is the reduction or elimination of the congenital rubella syndrome (CRS) that occurs in 90% of infants whose mothers were infected with rubella in their first trimester . The genome contains two open reading frames (ORFs), both of which encode precursor proteins that are proteolytically cleaved into functional proteins. The 5′ proximal ORF encodes the non-structural proteins (NSP) P150 and P90. The 3′ proximal ORF encodes the structural proteins (SP), the capsid (C) and two glycoproteins, E2 and E1. There are 3 untranslated regions (UTRs) in the rubella virus genome: a 40-nucleotide (nt) sequence at the 5′ terminus, an approximately 120-nt intergenic region (IR) between the two ORFs, and a 59-nt region at the 3′ terminus.
There are currently whole genome sequences for 21 RVs in Genbank (there are multiple sequences for some viruses, e.g. the vaccine strain RA27/3). Almost half of the sequences (11/21) are from wild-type and vaccine viruses from the 1960′s, a decade when a world-wide epidemic of rubella occurred, resulting in 20,000 cases of CRS in the United States alone . Vaccination against rubella was initiated in the 1970′s, mainly in industrialized countries. No further global epidemics were recorded and interest in collecting and characterizing rubella viruses waned. As a result, very few virus isolates exist from the 1970′s, 1980′s and the early 1990′s. In the 1990′s efforts to improve rubella control were strengthened globally as evidenced by an increase in the number of countries including rubella-containing vaccine in their national immunization programs (from 43% of WHO member states in 1996 to 67% of member states in 2009) .
The use of molecular epidemiology was recognized as a powerful tool for virus surveillance and isolations of RVs to be used for this purpose greatly increased in the late 1990′s and 2000′s. In 2005, a systematic nomenclature was adopted by the WHO [4, 5]. Briefly, genetic characterization has identified 2 clades which differ by 8-10% at the nt level. Clade 1 is divided into 10 genotypes (1a, 1B, 1C, 1D, 1E, 1F, 1G, 1h, 1i, and 1j), of which 6 are recognized and 4 are provisional (designated by lower case letters). Clade 2 contains 3 genotypes (2A, 2B, and 2C). Rubella virus was declared eliminated from the United States in 2005  and molecular epidemiological analysis of rubella viruses was used to demonstrate that indigenous rubella virus had been eliminated from the country . Imported cases of rubella into the United States continue to occur and efforts are made to obtain viruses for molecular surveillance. Although sequence analysis of a 739-nt region in the E1 coding region has been determined sufficient for phylogenetic analysis [4, 5], obtaining the sequences of the whole viral genome can contribute to a greater understanding of the virus. This work adds 16 complete wild-type rubella virus sequences with an emphasis on recent isolates from currently circulating genotypes. An initial analysis of these virus sequences is provided here.
Phylogenic analysis of 30 viruses
List of rubella viruses used in this study
Isolation site, year
New Jersey, USA, 1961
New York, USA, 1998
California, USA, 1991
Texas, USA, 1998
California, USA, 1988
AGMK 2, BSC 2, Vero 2
Maine, USA, 2008
Massachusetts, USA, 1998
Washington, USA, 2008
Massachusetts, USA, 2007
New Hampshire, USA, 2005
California, USA, 1997
RK13 1, BSC 1, Vero 3
Michigan, USA, 2007
Washington, USA, 2000
California, USA, 2008
Minnesota, USA, 2009
North Dakota, USA, 2008
Connecticut, USA, 1964
Leipzig, Germany, 1984
Toyama, Japan, 1967
Matsue, Japan, 1968
Cordoba, Argentina, 1988
Baja California, Mexico, 1997
Auckland, New Zealand, 1991
Pavia, Italy, 1997
Stuttgart, Germany, 1992
Beijing, China, 1979
Seoul, South Korea, 1996
Tel Aviv, Israel, 1968
Moscow, Russia, 1967
Moscow, Russia, 1997
Clusters of variable amino acids and hyper-variable regions (HVR)
Clade-specific amino acid variation
(n = 20)
(n = 10)
P150 (Q domain)
The clusters of variable amino acids in SP were shorter than those in NSP, with lengths about 20 amino acids, and they did not exhibit variability as high as those found in NSP. Cluster IV (aa 26 to 47) overlapped with the major RNA binding domain of the C protein (aa 28–56) . Clusters V and VI were located at the N-terminal half of E2 and overlapped with the putative antigenic domain of E2 (aa 301–416) . Two out of the 10 changes in cluster V were clade-specific, including Met308 (to Ile or Val in clade 2) and Gln323 (to Lys in clade 2) (Table 2). A clade specific change was also found in cluster VI, where Ala420 (or Val) was preserved among all clade 1 viruses, while it was substituted by Thr in clade 2 viruses (Table 2).
Conserved regions in RV genomes
In addition to clusters of variable amino acids, regions with little amino acid variation were also identified among the 30 viruses. In general, the putative methyltransferase (MT), helicase (Hel) and RNA-dependent RNA polymerase (RdRp) were less variable and E1 was the least variable among the SP. Among these 30 sequences at the nucleotide level, the percentage of variable nt positions (at least one virus differed from the consensus sequence) across the whole genome was about 26%; approximately 23% of the nucleic acids were variable in the Hel, RdRp and E1. The MT, containing 12% variable nucleic acids, was the most conserved region (Figure 3).
When examining the invariant amino acids among the 30 viruses, 26 regions of more than 30 consecutive identical amino acids were found: 5 of them were within the N-terminal 400 amino acids and 7 were in the putative RdRp domain. The majority of the amino acid conserved regions in the SP gene were within the E1 coding region which overlapped with the putative motifs encoding fusion/antigenicity peptides [18–20]. One of these regions in C also encompassed the antigenic/transmembrane domain of C (nt 7196–7318). Interestingly, while these regions possessed identical amino acid sequences, the nucleic acid identity varied from 10% (the N-terminal 41 amino acids of NSP) to 27% (the C-terminal half of C) (Figure 3).
An examination of the nucleic acid sequences of all 26 regions of more than 30 consecutive identical amino acids as well as the 3 UTRs in the 30 viruses revealed several regions with more than 20 consecutive identical nucleotides. Most of them clustered within the 5′ 250 nts of the genome. Among these stretches of identical nucleic acid sequences, two clusters were identified within the IR. One of these stretches, nt 6429–6448, encompassed the subgenomic RNA start site, suggesting that this sequence may contain critical regulatory elements for subgenomic RNA synthesis.
Nucleotide deletions in the junction region
Nine of the 16 new sequences were 9762 nucleotides in length, while the other 7 viruses had either one or two nucleotide deletions in the IR. These deletions confirm data reported in 2007  and provide additional evidence that these deletions are genotype specific. All five of the genotype 2B viruses sequenced in this study (years of collection ranging from 2000 to 2009; Table 1) had a single deletion at nt 6422, as did the 2 older (1968 and 1996) 2B viruses sequenced previously . Thus, all genotype 2B viruses sequenced thus far contain a one nucleotide deletion which appears to be specific for this genotype. In addition, Zhou et al. reported that the two nucleotide deletion (nts 6480 and 6481) was found in five of seven genotype 1B viruses. Based on the updated rubella nomenclature , analyses of the phylogenetic groupings of these 5 viruses with other rubella virus sequences showed that two of the five would currently be classified as 1G viruses (Guz_GER92 and I76 ISR 92), two would belong to genotype 1i (Tom_UK_86 and 0754_GER_92), and the last would be an outlier to genotype 1h (4008_ITA_91). The remaining two 1B viruses, Cba_ARG_88 and Fan_UK_78, which did not contain the two nucleotide deletion, remain classified as genotype 1B in the current standard nomenclature. In this report a genotype 1B virus, Brooklyn.NY.USA/98, did not have the two nucleotide deletion while two genotype 1G viruses, Boston.MA.USA/10.07 and Lebanon.NH.USA/3.05, did have the two nucleotide deletion. The 2 nucleotide deletion in the IR sequences was consistently seen in 4 additional 1G and 1 each additional 1i and 1h viruses, while no deletion was observed in 3 additional 1B viruses (author’s unpublished result). In summary, using the current nomenclature system, the 1B viruses do not contain the two nucleotide deletion, but all of the 1G, 1h, and 1i viruses do have the two deleted nucleotides. Therefore, this deletion appears to be specific for genotypes 1G, 1h, and 1i.
To investigate possible recombination events among the 30 rubella whole genome sequences, six different algorithms implemented in the RDP3 program , RDP, GENECONV, MaxChi, Chimaera, SiScan and 3Seq, were used for analysis. Since no single program provided optimal performance under all conditions, we looked for consistency of the results from different methods. Only MaxChi and Chimaera showed evidence of possible recombination events among CA.USA/88, JC2_NZL91 and Springfield.MA.USA/98 from nt 715 to 2786 with P-values of 7.764×10-3 and 4.989×10-3, respectively. Positive data from only two out of six methods were not significant enough to support the conclusion of a recombination event in this region.
In this study, the whole genome sequences of 30 rubella viruses were analyzed, including the 16 viruses that were sequenced for this report. Although all 16 viruses were isolated from cases within the United States, 10 out of 16 were determined by standard epidemiological means to be importations from other countries. The analysis is consistent with the previous report on the genetic characterization of rubella virus using 19 sequences . The data in this study provides more confidence in the previous analysis by the use of a larger dataset. The phylogenetic analysis also provided high confidence for the current genotype groupings (originally based on structural protein coding region sequences), as evidenced by the high bootstrap values for all genotypes represented except for genotype 1a. Use of the whole genome sequences, thus, provides robust support for the WHO standardized nomenclature for genetic characterization of wild-type rubella viruses.
It was found that 74% of the nucleotides in the 30 genomic sequences were invariant (78% were found previously with a smaller data set ) while the amino acid identity was higher at about 89%. Six clusters, located in the coding sequences in the middle of the p150 and the N-terminal half of the SP, were found to contain a higher percentage of variable positions than the rest of the genome. Previously the HVR was identified as the region with the highest local variability; in this study, we identified another region containing even more nucleic and amino acid variability (Cluster III). The nucleotide variabilities found in the coding regions of the E1 and E2 proteins are 23% and 30% respectively, while the amino acid variability is notably different between these two surface proteins, at 7% and 17%, respectively. E1 is believed to be the major immunodominant protein while the role of E2 in immunogenicity has not been fully characterized.
In addition to variable regions, regions containing conserved peptides of more than 30 consecutive identical amino acids were identified among the 30 viruses. The p90 is more conserved than p150 and SP. This includes a region of 114 consecutive identical amino acids in the putative RNA-dependent RNA polymerase domain. The nucleotide sequences coding for these conserved peptides were 80% identical overall; however the nucleotide identity was greater than 90% near the 5′ terminus of the genome, while the nucleotide sequence was more variable in the C and E1 coding regions, with only 73% identity. The presence of highly conserved nucleotide sequences suggests their importance in virus replication and may potentially be used as targets for molecular detection of RV. It is not known whether the highly conserved 5′ 250-nt region contains critical secondary structures. Two conserved regions in the IR were also noticed which could be critical for the regulation of subgenomic RNA synthesis. Nevertheless, analysis of the 30 sequences also revealed genotype-specific deletions in this region: 6 genotype 2B viruses in this report had a single deletion at 6422 and two 1G viruses had nucleotides 6480 and 6481 deleted, indicating the flexibility of the IR in length. This dinucleotide deletion has also been found in 1h and 1i viruses, providing strong evidence that genotypes G, h, and i diverged from a common ancestor. Unlike two previous reports [9, 22], we did not find significant evidence of recombination using the 30 whole genomes and six different programs. Weak possible recombination was predicted close to the 5′ termini for three viruses; however, the evidence was not considered significant.
Sixteen wild-type rubella virus isolates from the Centers for Disease Control and Prevention (CDC) were chosen for whole genome sequencing at the J. Craig Venter Institute (JCVI) (Table 1). All viruses in this study were collected in the United States, most as a result of routine surveillance, and dates of collection range from 1961 to 2009, with the majority coming from the late 1990′s and 2000′s. One of the virus isolates is a laboratory strain (NJ.USA/61/1a), commonly known as M33 (ATCC, Manassas, VA) which was the first rubella virus isolate . There are currently 4 entries in GenBank for the M33 strain: X05259 (3382 nts), X72393 (6600 nts), J02620 (1822 nts), and AJ438491 (948 nts). Although the combined X05259 and X72393 sequences cover most of the genome (9749 nts), they contain errors making it difficult to align and use the sequences. Therefore, a resequencing of this historic virus was deemed beneficial. Ten of the isolates were known to have been imported into the United States from other countries, either as acute cases acquired abroad or as CRS cases whose mothers spent time in other countries during the early stages of their pregnancies. Two of the isolates, LA.CA.USA/91/1C and Seattle.WA.USA/16.00/2B, serve as WHO reference viruses (Table 1).
Stocks of the 16 viruses were inoculated into T75 flasks of confluent Vero cells. Five to seven days post-infection, the culture medium was removed and total cellular RNA was extracted using Tri-Reagent (Molecular Research Center, Cincinnati, OH) according to the manufacturer’s protocol. The RNA was resuspended in nuclease-free water and stored at −80°C.
Oligonucleotide primers were designed using an automated primer design tool [24, 25]. Clade 1 primers were designed from an alignment of the following reference sequences: 1a_AF435865, 1a_AB222609.1, 1B_DQ085339.1, 1C_DQ085341.1, 1D_DQ388281 and Clade 2 primers were designed from 2A_AY258322.1, 2B_DQ085338.1, and 2C_DQ085340.1. Primers, with M13 tags added, were designed at intervals along both the sense and antisense strands, and provided amplicon coverage of at least 4-fold (Additional file 1: Table S1). RT-PCRs were performed with 1 ng of RNA using OneStep RT-PCR kits (Qiagen, Valencia, CA) according to manufacturer’s instructions with minor modifications. Reactions were scaled down to 1/5 the recommended volumes, the RNA templates were denatured at 95°C for 5 min, and 1.6 U RNase Out (Invitrogen, Carlsbad, CA) was used. The RT-PCR products were sequenced with an ABI Prism BigDye v3.1 terminator cycle sequencing kit (Applied Biosystems, Carlsbad, CA). Raw sequence traces were trimmed to remove any primer-derived sequence as well as low quality sequence, and gene sequences were assembled using Minimus, part of the open-source AMOS project . The gene sequences were then manually edited using ClOE (Closure Editor; JCVI) and ambiguous regions were resolved by additional sequencing when possible. Finally, the Viral Genome ORF Reader  was used to check segment lengths, perform alignments, ensure the fidelity of open-reading frames, correlate nt polymorphisms with amino acid changes, and detect any potential sequence errors. The 5′ and 3′ termini were determined using the 5′/3′ RACE kit as directed by the manufacturer (Roche Diagnostics, Mannheim, Germany) and oligo dT priming, respectively. The termini sequences were incorporated into the whole genome sequences as described above.
Sequence and phylogenetic analyses
The sequences were aligned by ClustalW in the MEGA 4.0 software . Phylogenetic analysis was performed using the Mega 5.05 program . The first 21 nts of the 30 sequences were deleted from the sequence alignment due to a gap in one sequence (AF435865). Analysis of variable and conserved positions relative to a consensus sequence was done using Microsoft Excel by comparing alignments of the nt and aa sequences of the 30 viruses. The comparison does not necessarily reflect the evolutionary distance between the viruses. Recombination analysis was performed using the RDP3 program .
The findings and conclusions in this report are those of authors and do not necessarily represent the views of the U.S. Department of Health and Human Services.
Congenital rubella syndrome
Open reading frame
RNA dependent RNA polymerase
J.Craig Venter Institute
Centers for Disease Control and Prevention.
The authors would like to thank public health officials in the state health departments that contributed viruses and the MMR epidemiology team at CDC. This project has been funded in part with federal funds from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, Department of Health and Human Services under contract number HHSN272200900007C. We thank Ludmila Perelygina and Pierre Rivailler for critical reading of the manuscript.
- Peckham C, Tookey P, Hardelid P: Rubella epidemiology: surveillance to monitor and evaluate congenital rubella prevention strategies. In Rubella Viruses. Edited by: Banatvala J, Peckham C. Amsterdam: Elsevier; 2007:95-114.Google Scholar
- Reef S, Plotkin SA: Rubella vaccine. In Rubella Viruses. Edited by: Banatvala J, Peckham C. Amsterdam: Elsevier; 2007:79-93.Google Scholar
- World Health Organization: Rubella vaccines: WHO position paper. Wkly Epidemiol Rec 2011, 86: 301-316.Google Scholar
- World Health Organization: Standardization of the nomenclature for wild-type rubella viruses. Wkly Epidemiol Rec 2005, 80: 125-132.Google Scholar
- World Health Organization: Update of standard nomenclature for wild-type rubella viruses. Wkly Epidemiol Rec 2007, 82: 209-224.Google Scholar
- Centers for Disease Control and Prevention: Elimination of rubella and congenital rubella syndrome-United States, 1969–2004. Morb Mortal Wkly Rep 2005, 54: 279-282.Google Scholar
- Icenogle JP, Frey TK, Abernathy E, Reef SE, Schnurr D, Stewart JA: Genetic analysis of rubella viruses found in the United States between 1966 and 2004: evidence that indigenous rubella viruses have been eliminated. Clin Infect Dis 2006,43(Suppl 3):127-132.Google Scholar
- Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S: MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 2011, 8: 2731-28.View ArticleGoogle Scholar
- Zhou Y, Ushijima H, Frey TK: Genomic analysis of diverse rubella virus genotypes. J Gen Virol 2007, 88: 932-941. 10.1099/vir.0.82495-0PubMedView ArticleGoogle Scholar
- Hofmann J, Renz M, Meyer S, von Haeseler A, Liebert UG: Phylogenetic analysis of rubella virus including new genotype 1 isolates. Virus Res 2003, 96: 123-128. 10.1016/S0168-1702(03)00180-1PubMedView ArticleGoogle Scholar
- Kakizawa J, Nitta Y, Yamashita T, Ushijima H, Katow S: Mutations of rubella virus vaccine TO-336 strain occurred in the attenuation process of wild progenitor virus. Vaccine 2001,19(20–22):2793-2802.PubMedView ArticleGoogle Scholar
- Sakata M, Komase K, Nakayama T: Histidine at position 1042 of the p150 region of a KRT live attenuated rubella vaccine strain is responsible for the temperature sensitivity. Vaccine 2009,27(2):234-242. 10.1016/j.vaccine.2008.10.049PubMedView ArticleGoogle Scholar
- Zheng DP, Zhou YM, Zhao K, Han YR, Frey TK: Characterization of genotype II Rubella virus strains. Arch Virol 2003,148(9):1835-1850. 10.1007/s00705-003-0132-7PubMedView ArticleGoogle Scholar
- Tzeng WP, Frey TK: Functional replacement of a domain in the rubella virus p150 replicase protein by the virus capsid protein. J Virol 2009, 83: 3549-3555. 10.1128/JVI.02411-08PubMedPubMed CentralView ArticleGoogle Scholar
- Liang Y, Yao J, Gillam S: Rubella virus nonstructural protein protease domains involved in trans- and cis-cleavage activities. J Virol 2000, 74: 5412-5423. 10.1128/JVI.74.12.5412-5423.2000PubMedPubMed CentralView ArticleGoogle Scholar
- Wolinsky JS, McCarthy M, Allen-Cannady O, Moore WT, Jin R, Cao SN, Lovett A, Simmons D: Monoclonal antibody-defined epitope map of expressed rubella virus protein domains. J Virol 1991, 65: 3986-3994.PubMedPubMed CentralGoogle Scholar
- Liu Z, Yang D, Qiu Z, Lim KT, Chong P, Gillam S: Identification of domains in rubella virus genomic RNA and capsid protein necessary for specific interaction. J Virol 1996, 70: 2184-2190.PubMedPubMed CentralGoogle Scholar
- Mitchell LA, Zhang T, Ho M, Décarie D, Tingle AJ, Zrein M, Lacroix M: Characterization of rubella virus-specific antibody responses by using a new synthetic peptide-based enzyme-linked immunosorbent assay. J Clin Microbiol 1992, 30: 1841-1847.PubMedPubMed CentralGoogle Scholar
- Terry GM, Ho-Terry L, Londesborough P, Rees KR: Localization of the rubella E1 epitopes. Arch Virol 1988, 98: 189-197. 10.1007/BF01322168PubMedView ArticleGoogle Scholar
- Wolinsky JS, Sukholutsky E, Moore WT, Lovett A, McCarthy M, Adame B: An antibody- and synthetic peptide-defined rubella virus E1 glycoprotein neutralization domain. J Virol 1993, 67: 961-968.PubMedPubMed CentralGoogle Scholar
- Martin DP, Lemey P, Lott M, Moulton V, Posada D, Lefeuvre P: RDP3: a flexible and fast computer program for analyzing recombination. Bioinformatics 2010, 26: 2462-2463. 10.1093/bioinformatics/btq467PubMedPubMed CentralView ArticleGoogle Scholar
- Zheng DP, Frey TK, Icenogle J, Katow S, Abernathy ES, Song KJ, Xu WB, Yarulin V, Desjatshove RG, Aboudy Y, Enders G, Croxson M: Global distribution of rubella virus genotypes. Emerg Inf Dis 2003, 9: 1523-1530.View ArticleGoogle Scholar
- Parkman PD, Buescher EL, Artenstein MS: Recovery of rubella virus from army recruits. Proc Soc Exp Biol Med 1962, 111: 225-230.PubMedView ArticleGoogle Scholar
- Li K, Brownley A, Stockwell TB, Beeson K, McIntosh TC, Busam D, Ferriera S, Murphy S, Levy S: Novel computation methods for increasing PCR primer design effectiveness in directed sequencing. BMC Bioinforma 2008, 9: 191. 10.1186/1471-2105-9-191View ArticleGoogle Scholar
- Li K, Shrivastava S, Brownley A, Katzel A, Bera J, Nguyen AT, Thovarai V, Halpin H, Stockwell TB: Automated degenerate PCR primer design for high-throughput sequencing improves efficiency of viral sequencing. Virol J 2012, 9: 261. 10.1186/1743-422X-9-261PubMedPubMed CentralView ArticleGoogle Scholar
- The AMOS project. http://amos.sourceforge.net
- Wang S, Sundaram JP, Spiro D: VIGOR, an annotation program for small viral genomes. BMC Bioinforma 2010, 11: 451. 10.1186/1471-2105-11-451View ArticleGoogle Scholar
- Tamura K, Dudley J, Nei M, Kumar S: MEGA4: molecular evolutionary genetics analysis software version 4.0. Mol Biol Evol 2007, 24: 1596-1599. 10.1093/molbev/msm092PubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.