- Volume 89, Issue 9, 2008

Acute enteroviral infections ranging from meningitis, pancreatitis to myocarditis are common and normally well controlled by the host immune system comprising virus-specific CD8+ cytotoxic T lymphocytes (CTL). However, in some patients enteroviruses and especially coxsackieviruses of group B are capable of inducing severe chronic forms of diseases such as chronic myocarditis. Currently, it is not known whether divergences in the CTL-related immune response may contribute to the different outcome and course of enterovirus myocarditis. A pre-requisite for the study of CTL reactions in patients with acute and chronic myocarditis is the identification of CTL epitopes. In order to define dominant enterovirus CTL epitopes, we have screened, by using gamma interferon (IFN-γ) ELISPOT, 62 HLA-A*01- and 59 HLA-A*02-positive healthy blood donors for pre-existing CTL reactions against 12 HLA-A*01 and 20 HLA-A*02 predicted CTL epitopes derived from coxsackieviruses of group B. Positive CTL reactions were verified by FACS analysis in a combined major histocompatibility complex-tetramer IFN-γ staining. A total of 14.8 % of all donors reacted against one of the three identified epitopes MLDGHLIAFDY, YGDDVIASY or GIIYIIYKL. The HLA-A*02-restricted epitope ILMNDQEVGV was recognized by 25 % of all tested blood donors. For this peptide, we could demonstrate specific granzyme B secretion, a strong cytolytic potential and endogenous processing. All four epitopes were homologous in 36–92 % of group B enteroviruses, providing a strong basis for monitoring the divergence of T-cell-based immune responses in enterovirus-induced acute and chronic diseases.

Classification of viral sequences should be fast, objective, accurate and reproducible. Most methods that classify sequences use either pair-wise distances or phylogenetic relations, but cannot discern when a sequence is unclassifiable. The branching index (BI) combines distance and phylogeny methods to compute a ratio that quantifies how closely a query sequence clusters with a subtype clade. In the hypothesis-testing framework of statistical inference, the BI is compared with a threshold to test whether sufficient evidence exists for the query sequence to be classified among known sequences. If above the threshold, the null hypothesis of no support for the subtype relation is rejected and the sequence is taken as belonging to the subtype clade with which it clusters on the tree. This study evaluates statistical properties of the BI for subtype classification in hepatitis C virus (HCV) and human immunodeficiency virus-1 (HIV-1). Pairs of BI values with known positive- and negative-test results were computed from 10  000 random fragments of reference alignments. Sampled fragments were of sufficient length to contain phylogenetic signals that grouped reference sequences together properly into subtype clades. For HCV, a threshold BI of 0.71 yields 95.1 % agreement with reference subtypes, with equal false-positive and false-negative rates. For HIV-1, a threshold of 0.66 yields 93.5 % agreement. Higher thresholds can be used where lower false-positive rates are required. In synthetic recombinants, regions without breakpoints are recognized accurately; regions with breakpoints do not represent any known subtype uniquely. Web-based services for viral subtype classification with the BI are available online.

The establishment of clonal infection of hepatitis C virus (HCV) in a small-animal model is important for the analysis of HCV virology. A previous study developed models of molecularly cloned genotype 1a and 2a HCV infection using human hepatocyte-transplanted chimeric mice. This study developed a new model of molecularly cloned genotype 1b HCV infection. A full-length genotype 1b HCV genome, HCV-KT9, was cloned from a serum sample from a patient with severe acute hepatitis. The chimeric mice were inoculated intrahepatically with in vitro-transcribed HCV-KT9 RNA. Inoculated mice developed viraemia at 2 weeks post-infection, and this persisted for more than 6 weeks. Passage experiments indicated that the sera of these mice contained infectious HCV. Interestingly, a similar clone, HCV-KT1, in which the poly(U/UC) tract was 29 nt shorter than in HCV-KT9, showed poorer in vivo infectivity and replication ability. An in vitro study showed that no virus was produced in the culture medium from HCV-KT9-transfected cells. In conclusion, this study developed a genetically engineered genotype 1b HCV-infected mouse. This mouse model will be useful for the study of HCV virology, particularly the mechanism underlying the variable resistance of HCV genotypes to interferon therapy.

The envelope of bovine viral diarrhea virus (BVDV) contains the glycoproteins Erns, E1 and E2. Complementation of a recombinant vesicular stomatitis virus (VSV) with BVDV glycoproteins resulted in infectious pseudotyped viruses. To elucidate the specific role of each of the single envelope glycoproteins during viral entry, pseudotypes were generated bearing the BVDV envelope proteins in different combinations. Pseudoviruses that contained E1 and E2 but not Erns were infectious, indicating that Erns is dispensable for virus entry. VSV/BVDV pseudotypes with chimeric proteins (the ectodomain of the BVDV glycoprotein and the transmembrane domain of the VSV-G protein) were not infectious. The fact that E1–E2 heterodimers were not detected if one of the proteins was chimeric indicated that the heterodimers are crucial for BVDV entry. It was shown by site-directed mutagenesis that the charged amino acids in the transmembrane domains of BVDV E1 (lysine and arginine) and the charged amino acid in the transmembrane domain of E2 (arginine) play a key role in heterodimer formation. Pseudoviruses bearing the mutation E2-R/A, where the charged amino acid was substituted by alanine, were not infectious, supporting the hypothesis that E1–E2 heterodimers are essential for BVDV entry.

Buggy Creek virus (BCRV) is an unusual arbovirus within the western equine encephalitis complex of alphaviruses. Associated with cimicid swallow bugs (Oeciacus vicarius) as its vector and the cliff swallow (Petrochelidon pyrrhonota) and house sparrow (Passer domesticus) as its amplifying hosts, this virus is found primarily in the western Great Plains of North America at spatially discrete swallow nesting colonies. For 342 isolates collected in Oklahoma, Nebraska, Colorado and North Dakota, from 1974 to 2007, we sequenced a 2076 bp region of the 26S subgenomic RNA structural glycoprotein coding region, and analysed phylogenetic relationships, rates of evolution, demographical histories and temporal genetic structure of the two BCRV lineages found in the Great Plains. The two lineages showed distinct phylogeographical structure: one lineage was found in the southern Great Plains and the other in the northern Great Plains, and both occurred in Nebraska and Colorado. Within each lineage, there was additional latitudinal division into three distinct sublineages. One lineage is showing a long-term population decline. In comparing sequences taken from the same sites 8–30 years apart, in one case one lineage had been replaced by the other, and in the other cases there was little evidence of the same haplotypes persisting over time. The evolutionary rate of BCRV is in the order of 1.6–3.6×10−4 substitutions per site per year, similar to that estimated for other temperate-latitude alphaviruses. The phylogeography and evolution of BCRV could be better understood once we determine the nature of the ecological differences between the lineages.

Me Tri virus (MTV) is a member of the Semliki Forest virus (SFV) complex in the genus Alphavirus, first isolated from Culex tritaeniorhynchus mosquitoes in Vietnam in 1971 and described as a newly recognized alphavirus, based on antigenic characterization. However, based on a partial nucleotide sequence of the E1 envelope glycoprotein gene, it has recently been argued that MTV may represent a variant of SFV rather than a separate species. To enable definitive classification, we determined the complete genome sequence of MTV from original virus stock. Nucleotide homology, as well as phylogenetic analyses based on whole and partial genome sequences confirmed that MTV is an isolate of SFV. Notable differences to other reported SFV sequences included a 122 nt insertion at the 5′ non-translated region (NTR), likely resulting from homologous recombination of part of the nsP2 gene, and differences in the sequence length of the 3′ NTR. To our knowledge, this is the first and only documentation of SFV isolation outside Africa. Further research is needed to clarify whether SFV continues to circulate in Vietnam.

Although the 2003 severe acute respiratory syndrome (SARS) outbreak was controlled, repeated transmission of SARS coronavirus (CoV) over several years makes the development of a SARS vaccine desirable. We performed a comparative evaluation of two SARS vaccines for their ability to protect against live SARS-CoV intranasal challenge in ferrets. Both the whole killed SARS-CoV vaccine (with and without alum) and adenovirus-based vectors encoding the nucleocapsid (N) and spike (S) protein induced neutralizing antibody responses and reduced viral replication and shedding in the upper respiratory tract and progression of virus to the lower respiratory tract. The vaccines also diminished haemorrhage in the thymus and reduced the severity and extent of pneumonia and damage to lung epithelium. However, despite high neutralizing antibody titres, protection was incomplete for all vaccine preparations and administration routes. Our data suggest that a combination of vaccine strategies may be required for effective protection from this pathogen. The ferret may be a good model for SARS-CoV infection because it is the only model that replicates the fever seen in human patients, as well as replicating other SARS disease features including infection by the respiratory route, clinical signs, viral replication in upper and lower respiratory tract and lung damage.

Feline infectious peritonitis virus (FIPV), a coronavirus that causes a lethal chronic disease in cats, enters feline monocytes via endocytosis. In this study, the pathway of internalization is characterized by evaluating the effect of chemical inhibitors and/or expression of dominant-negative (DN) proteins on the percentage of internalized virions per cell and infection. Further, co-localization studies were performed to determine the involvement of certain cellular internalization proteins. FIPV is not internalized through a clathrin-mediated pathway, as chlorpromazine, amantadine and DN eps15 did not influence virus uptake and FIPV did not co-localize with clathrin. The caveolae-mediated pathway could be excluded based on the inability of genistein and DN caveolin-1 to inhibit virus uptake and lack of co-localization between FIPV and caveolin-1. Dynamin inhibitory peptide and DN dynamin effectively inhibited virus internalization. The inhibitor strongly reduced uptake to 20.3±1.1 % of uptake in untreated cells. In the presence of DN dynamin, uptake was 58.7±3.9 % relative to uptake in untransduced cells. Internalization of FIPV was slightly reduced to 85.0±1.4 and 87.4±6.1 % of internalization in control cells by the sterol-binding drugs nystatin and methyl-β-cyclodextrin, respectively. Rho GTPases were inhibited by Clostridium difficile toxin B, but no effect was observed. These results were confirmed with infection studies showing that infection was not influenced by chlorpromazine, amantadine and genistein, but was significantly reduced by dynamin inhibition and nystatin. In conclusion, these results indicate that FIPV enters monocytes through a clathrin- and caveolae-independent pathway that strongly depends on dynamin and is slightly sensitive to cholesterol depletion.

Rift Valley fever virus (RVFV) is responsible for large and recurrent outbreaks of acute febrile illness among humans and domesticated animals in Africa. It belongs to the family Bunyaviridae, genus Phlebovirus, and its negative-stranded RNA genome consists of three segments. Here, we report the establishment and characterization of two different systems to rescue the RVFV wild-type strain ZH548. The first system is based on the BHK-21 cell clone BSR-T7/5, which stably expresses T7 RNA polymerase (T7 pol). Rescue of wild-type RVFV was achieved with three T7 pol-driven cDNA plasmids representing the viral RNA segments in the antigenomic sense. The second system involves 293T cells transfected with three RNA pol I-driven plasmids for the viral segments and two RNA pol II-driven support plasmids to express the viral polymerase components L and N. It is known that the 5′ triphosphate group of T7 pol transcripts strongly activates the antiviral interferon system via the intracellular RNA receptor RIG-I. Nonetheless, both the T7 pol and the pol I/II system were of similar efficiency. This was even true for the rescue of a RVFV mutant lacking the interferon antagonist nonstructural proteins. Further experiments demonstrated that the unresponsiveness of BHK-21 and BSR-T7/5 cells to T7 pol transcripts is most probably due to a deficiency in the RIG-I pathway. Our reverse genetics systems now enable us to manipulate the genome of RVFV and study its virulence mechanisms. Moreover, the finding that BHK-derived cell lines have a compromised RIG-I pathway may explain their suitability for propagating and rescuing a wide variety of viruses.

The nucleocapsid (N) protein of hantaviruses (family Bunyaviridae) is the most abundant component of the virion; it encapsidates genomic RNA segments and participates in viral genome transcription and replication, as well as in virus assembly. During RNA encapsidation, the N protein forms intermediate trimers and then oligomers via ‘head-to-head, tail-to-tail’ interactions. In previous work, using Tula hantavirus (TULV) N protein as a model, it was demonstrated that an intact coiled-coil structure of the N terminus is crucial for the oligomerization capacity of the N protein and that the hydrophobic ‘a’ residues from the second α-helix are especially important. Here, the importance of charged amino acid residues located within the coiled-coil for trimer formation and oligomerization was analysed. To predict the interacting surfaces of the monomers, the previous in silico model of TULV coiled-coils was first upgraded, taking advantage of the recently published crystal structure of the N-terminal coiled-coil of the Sin Nombre virus N protein. The results obtained using a mammalian two-hybrid assay suggested that conserved, charged amino acid residues within the coiled-coil make a substantial contribution to N protein oligomerization. This contribution probably involves (i) the formation of interacting surfaces of the N monomers (residues D35 and D38, located at the tip of the coiled-coil loop, and R63 appear particularly important) and (ii) stabilization of the coiled-coil via intramolecular ionic bridging (with E55 as a key player). It is hypothesized that the tips of the coiled-coils are the first to come into direct contact and thus to initiate tight packing of the three structures.

Punta Toro virus (PTV; genus Phlebovirus, family Bunyaviridae) causes apoptosis of hepatocytes in vivo in experimentally infected hamsters and in vitro in cultured HepG2 cells. Screening for expression of apoptosis-related genes has shown alterations in the genes for tumour necrosis factor-α (TNF-α) and the TNF receptor family. This study examined the roles of the TNF receptor-related extrinsic pathway and the Bcl-2 family-associated mitochondrial pathway in PTV-induced cell death. The effects of caspase inhibitors (caspIs) and TNF on cellular viability, virus replication, and morphological and biochemical changes in apoptosis were examined in HepG2 cells at different time points after infection with PTV (Adames strain). The results showed that caspIs dampened the virus-induced reduction in cellular viability, partially suppressed and delayed viral titres and antigen expression, and partially decreased the expression of apoptotic genes, caspase activities and DNA fragmentation. TNF treatment further decreased cellular viability after PTV infection and increased the level of apoptosis, whilst caspIs partially inhibited these effects. These findings indicate that TNF, caspase-8 and caspase-9 contribute to PTV-induced hepatocytic apoptosis and that additional mediators are probably also involved in this process. These mediators from different pathways correlated with one another and may be interlinked.

Besides enormous economic losses to the poultry industry, recent H5N1 highly pathogenic avian influenza viruses (HPAIVs) originating in eastern Asia have posed serious threats to public health. Up to April 17, 2008, 381 human cases had been confirmed with a mortality of more than 60 %. Here, we attempt to identify potential progenitor genes for H5N1 HPAIVs since their first recognition in 1996; most were detected in the Eurasian landmass before 1996. Combinations among these progenitor genes generated at least 21 reassortants (named H5N1 progenitor reassortant, H5N1-PR1–21). H5N1-PR1 includes A/Goose/Guangdong/1/1996(H5N1). Only reassortants H5N1-PR2 and H5N1-PR7 were associated with confirmed human cases: H5N1-PR2 in the Hong Kong H5N1 outbreak in 1997 and H5N1-PR7 in laboratory confirmed human cases since 2003. H5N1-PR7 also contains a majority of the H5N1 viruses causing avian influenza outbreaks in birds, including the first wave of genotype Z, Qinghai-like and Fujian-like virus lineages. Among the 21 reassortants identified, 13 are first reported here. This study illustrates evolutionary patterns of H5N1 HPAIVs, which may be useful toward pandemic preparedness as well as avian influenza prevention and control.

Antigen processing of respiratory syncytial virus (RSV) fusion (F) protein epitopes F85–93 and F249–258 presented to cytotoxic T-lymphocytes (CTLs) by the murine major histocompatibility complex (MHC) class I molecule Kd was studied in different viral contexts. Epitope F85–93 was presented through a classical endogenous pathway dependent on the transporters associated with antigen processing (TAP) when the F protein was expressed from either RSV or recombinant vaccinia virus (rVACV). At least in cells infected with rVACV encoding either natural or cytosolic F protein, the proteasome was required for epitope processing. In cells infected with rVACV encoding the natural F protein, an additional endogenous TAP-independent presentation pathway was found for F85–93. In contrast, epitope F249–258 was presented only through TAP-independent pathways, but presentation was brefeldin A sensitive when the F protein was expressed from RSV, or mostly resistant when expressed from rVACV. Therefore, antigen-processing pathways with different mechanisms and subcellular localizations are accessible to individual epitopes presented by the same MHC class I molecule and processed from the same protein but in different viral contexts. This underscores both the diversity of pathways available and the influence of virus infection on presentation of epitopes to CTLs.

Cross-neutralization between rabies virus (RABV) and two European bat lyssaviruses (EBLV-1 and -2) was analysed using lentiviral pseudotypes as antigen vectors. Glycoprotein (G-protein) cDNA from RABV challenge virus standard-11 (CVS-11) and EBLV-1 and -2 were cloned and co-expressed with human immunodeficiency virus (HIV) or murine leukemia virus (MLV) gag–pol and packageable green fluorescent protein (GFP) or luciferase reporter genes in human cells. The harvested lentiviral (HIV) vector infected over 40 % of baby hamster kidney (BHK) target cells, providing high-titre pseudotype stocks. Tests on blinded antibody-positive (n=15) and -negative (n=45) sera, predetermined by the fluorescent antibody virus neutralization (FAVN) test approved by the World Health Organization (WHO) and Office International des Epizooties (OIE), revealed that the CVS-11 pseudotype assay had 100 % concordance with FAVN and strongly correlated with neutralization titres (r 2=0.89). Cross-neutralization tests using sera from RABV-vaccinated humans and animals on pseudotypes with CVS-11, EBLV-1 and EBLV-2 envelopes showed that the relative neutralization titres correlated broadly with the degree of G-protein diversity. Pseudotypes have three major advantages over live-virus neutralization tests: (i) they can be handled in low-biohazard-level laboratories; (ii) the use of reporter genes such as GFP or β-galactosidase will allow the assay to be undertaken at low cost in laboratories worldwide; (iii) each assay requires <10 μl serum. This robust microassay will improve our understanding of the protective humoral immunity that current rabies vaccines confer against emerging lyssaviruses, and will be applicable to surveillance studies, thus helping to control the spread of rabies.

A DNA vaccine expressing human immunodeficiency virus type 1 (HIV-1) southern African subtype C Gag (pTHGag) and a recombinant baculovirus Pr55gag virus-like particle prepared using a subtype C Pr55gag protein (Gag VLP) was tested in a prime–boost inoculation regimen in Chacma baboons. The response of five baboons to Gag peptides in a gamma interferon (IFN-γ) enzyme-linked immunospot (ELISPOT) assay after three pTHGag immunizations ranged from 100 to 515 spot-forming units (s.f.u.) per 106 peripheral blood mononuclear cells (PBMCs), whilst the response of two baboons to the Gag VLP vaccine ranged from 415 to 465 s.f.u. per 106 PBMCs. An increase in the Gag-specific response to a range of 775–3583 s.f.u. per 106 PBMCs was achieved by boosting with Gag VLPs the five baboons that were primed with pTHGag. No improvement in Gag responses was achieved in this prime–boost inoculation regimen by increasing the number of pTHGag inoculations to six. IFN-γ responses were mapped to several peptides, some of which have been reported to be targeted by PBMCs from HIV-1 subtype C-infected individuals. Gag VLPs, given as a single-modality regimen, induced a predominantly CD8+ T-cell IFN-γ response and interleukin-2 was a major cytokine within a mix of predominantly Th1 cytokines produced by a DNA–VLP prime–boost modality. The prime–boost inoculation regimen induced high serum p24 antibody titres in all baboons, which were several fold above that induced by the individual vaccines. Overall, this study demonstrated that these DNA prime/VLP boost vaccine regimens are highly immunogenic in baboons, inducing high-magnitude and broad multifunctional responses, providing support for the development of these products for clinical trials.

Many aspects of the complex interaction between human immunodeficiency virus type 1 (HIV-1) and the human immune system remain elusive. Our objective was to study these interactions, focusing on the specific roles of dendritic cells (DCs). DCs enhance HIV-1 infection processes as well as promote an antiviral immune response. We explored the implications of these dual roles. A mathematical model describing the dynamics of HIV-1, CD4+ and CD8+ T-cells, and DCs interacting in a human lymph node was analysed and is presented here. We have validated the behaviour of our model against non-human primate simian immunodeficiency virus experimental data and published human HIV-1 data. Our model qualitatively and quantitatively recapitulates clinical HIV-1 infection dynamics. We have performed sensitivity analyses on the model to determine which mechanisms strongly affect infection dynamics. Sensitivity analysis identifies system interactions that contribute to infection progression, including DC-related mechanisms. We have compared DC-dependent and -independent routes of CD4+ T-cell infection. The model predicted that simultaneous priming and infection of T cells by DCs drives early infection dynamics when activated T-helper cell numbers are low. Further, our model predicted that, while direct failure of DC function and an indirect failure due to loss of CD4+ T-helper cells are both significant contributors to infection dynamics, the former has a more significant impact on HIV-1 immunopathogenesis.

Vaccination with live attenuated simian immunodeficiency virus (SIVmacC8) confers potent, reproducible protection against homologous wild-type virus challenge (SIVmacJ5). The ability of SIVmacC8 to confer resistance to superinfection with an uncloned ex vivo derivative of SIVmac251 (SIVmac32H/L28) was investigated. In naïve, Mauritian-derived cynomolgus macaques (Macaca fascicularis), SIVmac32H/L28 replicated to high peak titres (>108 SIV RNA copies ml−1), persisted at high levels and induced distinctive pathology in lymphoid tissues. In cynomolgus macaques vaccinated with SIVmacC8, no evidence of detectable superinfection was observed in 3/8 vaccinates following challenge 3 or 20 weeks later with SIVmac32H/L28. Analyses after SIVmac32H/L28 challenge revealed a significant reduction in viral RNA (P<0.001) and DNA levels between 20 week vaccinates and challenge controls. Amongst 3 week vaccinates, less potent protection was observed. However, analysis of env from breakthrough virus indicated >99 % sequence similarity with the vaccine virus. Highly sensitive PCR assays that distinguish vaccine and challenge virus stocks demonstrated restimulation of replication of the vaccine virus SIVmacC8 in the face of potent protection against a vigorous, homologous challenge virus. Vaccine-induced antiviral neutralizing antibodies and anti-Nef CD8+ cytotoxic T cell responses did not correlate with the outcome of the challenge. Defining the mechanism of vaccine protection will need to account for the effective control of a genetically closely related challenge virus whilst remaining unable to suppress replication of the pre-existing vaccine virus. The role of innate and intrinsic anti-retroviral immunity in the protection conferred by live attenuated SIV vaccines warrants careful study.

Rz1–7 is a multimeric hammerhead ribozyme targeting seven unique sites within the human CCR5 mRNA that is active in vitro. Mouse stem cell virus-based MGIN and human immunodeficiency virus (HIV)-1-based HEG1 vectors were used to express Rz1–7 in a human CD4+ T lymphoid cell line. Stable transductants expressed Rz1–7, which was further shown to be active, since CCR5 mRNA and surface CCR5 protein expression levels decreased. High levels of progeny virus were produced when the transduced cells were challenged with an X4-tropic HIV-1 (NL4-3) strain, suggesting that Rz1–7 expression does not affect X4-tropic virus replication. When the transduced cells expressing Rz1–7 were challenged with the R5-tropic HIV-1 (BaL) strain, 99–100 % inhibition of progeny virus production was observed for the duration of the experiment (∼2 months). When the cells were precultured for 2–3 months prior to HIV-1 infection, inhibition was more prominent in cells transduced with MGIN-Rz1–7 than with HEG1-Rz1–7. Inhibition occurred at the level of viral entry, as no HIV-1 DNA could be detected. These results demonstrate that Rz1–7 confers excellent inhibition of R5-tropic HIV-1 replication at the level of entry. Therefore, we anticipate that this multimeric ribozyme will be beneficial for HIV-1 gene therapy.

It was recently shown that the transcription factor T-bet is crucial for adequate innate and acquired immune responses to genital herpes simplex virus type 2 (HSV-2) infection in mice. To test the possible genetic influence of variations in the TBX21 gene encoding T-bet on susceptibility to infection, this study evaluated the frequencies of five different single-nucleotide polymorphisms (SNPs) in the human TBX21 gene in 159 HSV-2-infected individuals and compared them with those in 186 healthy HSV-2-seronegative controls. The data showed that one variation (rs17244587) in the 3′-untranslated region of TBX21 was strongly associated with the incidence of genital HSV-2 infection. The frequency of the A allele at this position was 0.19 in the group of HSV-2-infected individuals compared with 0.05 in the group of uninfected controls (P=9.3×10−8). Furthermore, a homozygous AA genotype was found only among HSV-2-infected individuals and not in seronegative controls. These results indicate that the host genetic background may affect susceptibility to HSV-2 infection in humans, with TBX21 as a strong candidate gene.