研究成果紹介


・Introduction
Virological characteristics
Epidemiological studies
Neurovirulence in model animals
Host factors and cytopathogenicity
BDV replication and persistent infection


Introduction

Borna disease, usually fatal, acute nonsuppurative encephalitis in horses, was first described in the late 18th century in South-eastern Germany. Experimental transmission of the disease to experimental animals using brain homogenates from diseased horses together with filtration studies demonstrated a viral etiology for Borna disease. The main natural hosts of Borna disease virus (BDV) are horses and sheep. Natural Borna disease was also diagnosed in donkeys, goats, cattle, rabbits and a dog. BDV is now gaining much of the research attention, because the disturbances seen in animals resemble those of neuropsychiatric disorders in humans.  These observations raise the possibility that BDV infection may be associated with certain human disorders.

We want to clarify the possible relation of human neuropsychiatric disorders with BDV infection in the central nervous system by the approaches to examine how BDV can kill neuronal cells or destroy the neuronal functions at the molecular level. Following projects are ongoing in our Virology Department:

1) Epidemiological studies in patients with neurodegenerative disorders
2) Characterization of BDV pathogenesis using model animals
3) Studies on the mechanism to kill host cells through interaction with host factors.    
4) Studies on the mechanism for the BDV replication and the functions of individual viral proteins


Virological characteristics

Borna disease virus (BDV) has a nonsegmented, negative-, single-stranded (NNS) RNA genome that is similar to other viruses within the order Mononegavirales.  BDV is highly neurotropic RNA virus with noncytolytic replication in the central nervous system.  The name refers to the city of  Borna in  Saxony,  Germany, where many horses died during an epidemic in 1894 and 1896.  BDV is a neurotropic enveloped virus.  Replication and transcription of the BDV genome take place in the nuclei of infected cells.  Such unique features of this virus allowed it to be classified as a new family: Bornaviridae.

    The BDV genomic RNA is approximately 8.9 kb in length. As shown in Fig. 1, the BDV genome encodes six open reading frames (ORF’s): p40 nucleoprotein (N), p24 phosphoprotein (P), p16 matrix protein (M), gp94/gp84 membrane glycoprotein (G), p190 polymerase (L), and p10 protein with unknown function (X). These ORF’s are transcribed in three units: the first transcription unit with a 1.2 kb mRNA for the N protein; the second transcription unit with a 0.8 kb mRNA for the X and P in overlapping ORF’s; and, the third transcription unit with several mRNA’s generated by alternative termination of transcription and splicing of one to three introns for the M, G, and L proteins.

  Culture cell lines are infected with BDV noncytolytically and therefore easily establish persistent infection.  The production rate of viral particles is extremely low.  These are believed to reflect the situation of BDV in the brains of injected model animals, such as rat in which persistent BDV infection is also detectable for a long time.

Fig. 1. Genomic structure and transcription map of BDV. BDV ORF’s are shown by boxes. The location of transcription initiation and transcription termination sites are marked as S and T, respectively. Positions of introns I to III are indicated.

Epidemiological studies


1) BDV infection in animals
     Clinically, pathohistologically and virologically confirmed cases of Borna disease in horses and sheep were mostly restricted to Germany, Switzerland, Austria and the Principality of Liechtenstein. It was also reported that Borna disease also in Swedish horses, associated with an atypical disease pattern. In addition, we recently confirmed two cases of horses with classical Borna disease in Japan. In addition, we also found that BDV was demonstrated at a high rate in restricted regions of the brain from horses with locomotor disease with unknown etiology. Taken together with the reports of subclinical cases of BDV infection in horses and sheep in many countries in the world, the geographical distribution of BDV seems to be worldwide. The reason why Borna disease had mainly been described in the endemic areas in central Europe might be due to the fact that this disease is rather rare and outbreaks are usually sporadic and therefore might have not been always diagnosed in non-endemic areas; however, one could also think certain co-factors, which are present in the endemic areas and contribute to the development of clinical disease.

2) BDV infection in humans

Epidemiological studies to assess the association of BDV infection with human diseases have been performed serologically and by use of molecular techniques.  Initially, indirect immunofluorescence (IFA) was used in these studies.  Later, several serological methods were developed, including the use of RT-PCR in 1995 in molecular epidemiologic studies to detect BDV footprints in the PBMC’s of patients.  The results reported for patients with psychiatric disorders, chronic fatigue syndrome (CFS), and immunosuppressive state are summarized, according to the year of the report, in Table 1 for anti-BDV antibodies, in Table 2 for BDV antigens, and in Table 3 for BDV RNA.

Epidemiological studies in humans mostly reported on the association between BDV infection and neuropsychiatric disorders including unipolar depression, bipolar disorder and schizophrenia.  Also, BDV linkage has been focused on CFS, AIDS encephalopathy, multiple sclerosis, motor neuron disease and brain tumors. 

In 1985, IFA by use of BDV-infected cells as antigens showed first by German group that sera from a significant proportion of psychiatric patients contained antibodies to BDV.  In contrast, the serum antibodies to BDV were significantly lower in healthy controls.  Thus, it was proposed that BDV infection might be associated with human psychiatric disorders.  Table 1 shows data accumulated, by use of various serologic techniques, by many groups in support of this hypothesis.  However, others have questioned the association of BDV with human diseases because of the low titers of antibodies to BDV seen in humans as compared with those detected in naturally and experimentally infected animals.  One group suggests that the low titer is due to the lower avidity of the human antibodies to BDV, and proposes that such antibodies may not be induced by the virus, but by an antigenically related unknown agent, or by a cellular immunogen that is up-regulated during psychiatric disorders.

Molecular epidemiological analyses by use of the highly sensitive nested RT-PCR were first performed in 1995 to detect BDV nucleic acids in the peripheral blood samples from psychiatric patients.  Data on BDV prevalence in patients with schizophrenia, mood disorders, and CFS has been collected from Austria, Germany, Japan, Korea, Sweden, Taiwan and the USA (Table 3).  The results yielded very divergent percentages in patients and in controls as well.  Several groups also reported no positive case in the human samples studied.  Autopsied brain samples from humans also were examined for BDV antigens and RNA.  The brain samples from patients with a history of various mental disorders and even from apparently normal controls gave positive signals (Table 3).  Again, the positive percentages also were variable among the different studies.  Taken in total, the accumulated results support a possible association between BDV and human diseases.

There might be several possible explanations for the varying BDV prevalence detected in the human populations studied. With the RT-PCR amplification of the BDV RNA’s from the peripheral blood samples, the cell populations examined were different among groups, i.e., whole blood, PBMC’s, and granulocytes. Also, the starting blood sample volumes were variable. This might be critical, because the number of BDV-positive cells in blood can be extraordinarily low. In addition, the methods of nucleic acid extraction from the blood samples were different among groups, and this variation can impact the results critically. Of course, there is always a possibility of accidental laboratory BDV contamination of the samples. This seems unlikely, because most groups reported that blood samples from patients have higher BDV prevalence than controls. In the studies with autopsied brains, the post mortem time for sampling might impact the detection of the BDV signals, especially the viral RNA signals. Also, one of the possible effects on the data is the stage of BDV infection: at acute phase or later stages of the disease. 

Generally, BDV replicates slowly and establishes persistent infection.  BDV transcription and replication have been characterized in several types of actively dividing cell lines, while the major target cells for BDV in vivo are predominantly non-dividing neuronal cells.  Therefore, we need more information on the status of BDV in such non-dividing cells, especially in those of human origin.  If only P is expressed in the non-dividing normal neuronal cells, as a kind of latency, especially in human brains, it will overcome the major critique that argues against the association of BDV with human diseases, i.e., only P protein, its RNA or antibody are detected predominantly as BDV signals in human samples.  If BDV latency in brains can be demonstrated, it would explain why BDV infection is difficult to detect before disease onset in the patients, and that periodic reactivation from latency may give positive BDV signals.

Our recent report described the presence of BDV in one of four autopsied brains from schizophrenic patients with a very recent (two years) onset of disease, but no BDV signal was detected in two autopsied brains from healthy individuals.  In this study, the autopsies were performed within eight to 12 hr of death.  Examination to detect BDV RNA by RT-PCR and in situ hybridization (ISH) in a total of 12 different brain regions from the above autopsied brains revealed the presence of BDV in four brain regions [the hippocampus, the cerebellum, the pons, and the temporal lobe of the cerebral cortex] of a BDV-seropositive schizophrenic patient (P2 in Fig. 2), but not in any brain regions from the other three patients and three normal controls.  Several neurons in the hippocampus formation were stained by IHC with a polyclonal serum from a BDV-infected mouse.




Table 1. Summarized data on serepidemiological studies.

文献

疾患

方法

陽性率 (患者 / 対象群) (%)

Rott et al., 1985  

Psychiatric

Affective disorders 

IFA 

IFA 

0.6 / 0

4 / 0 

Amsterdam et al., 1985  

Affective disorders 

IFA 

4.5 / 0  

Bode et al., 1988  

Psychiatric

HIV-positive 

IFA

IFA 

2 / 2

7.8 / 2  

Rott et al., 1991  

Psychiatric/neurological diseases 

IFA/WB 

4-7 / 1 

Bode et al., 1992  

Chronic diseases

EB infection: children 

IFA/IP

IFA/IP 

13.2 / 2

5.6  

Bode et al., 1992  

CFS 

IFA 

0  

Bechter et al., 1992  

Psychiatric 

IFA 

Significantly higher than controls  

Bode et al., 1993  

Psychiatric/Affective disorders 

IFA 

20  

Fu et al., 1993 

Affective disorders 

WB 

38 / 16 (in anti-N)

12 / 4 (in anti-P) 

Bode et al., 1995 

Psychiatric 

IFA

50  

Kishi et al., 1995  

Psychiatric 

WB 

30  

Kishi et al., 1995  

Blood donors 

WB 

1.0  

Waltrip et al., 1995  

Schizophrenia 

WB/IFA 

8.9 / 0 (in anti-N)

27.8 / 20 (in anti-P) 

Bechter et al., 1995  

Psychiatric 

 

Increased CSF/serum index for anti-BDV 

Sauder et al., 1996  

Psychiatric 

WB 

9.6 / 1.4 

Igata-Yi et al., 1996 

Psychiatric 

IFA 

24 / 11 

Nakaya et al., 1996  

CFS 

WB 

24  

Auwanit et al., 1996  

HIV-positive 

ELISA 

0〜48  

Kitze et al., 1996  

Multiple sclerosis 

IFA 

0  

Waltrip et al., 1997 

Deficit schizophrenia 

WB 

33.3 

Richt et al., 1997  

Schizophrenia 

WB 

20  

Iwahashi et al., 1997  

Schizophrenia 

WB 

45  

Takahashi et al., 1997  

Blood donors 

ELISA/WB 

2.6〜14.8  

Kubo et al., 1997  

Psychiatric 

IFA/WB 

Horimoto et al., 1997  

Schizophrenia 

RS-ELISA 

0 / 0  

Deuschule et al., 1998  

Major depression

Multiple sclerosis 

 

6.3 in CSF

0 in CSF 

Yamaguchi et al., 1999  

Schizophrenia 

ECLIA 

3.08 / 1.09  

Backmann et al., 1999  

HIV-positive 

IFA 

12.5〜8.0  

Chen et al., 1999       

Schizophrenia 

WB 

12.1  

Nakaya et al., 1999  

CFS 

WB 

100 (2 family clusters) 

Evengard et al., 1999  

CFS 

ELISA/WB 

0 / 0  

Valenkamp et al., 2000  

psychiatric 

IFA 

17.4 / 0  

Tsuji et al., 2000  

Psychiatric/Schizophrenia 

WB 

0  

Fukuda et al., 2001  

Schizophrenia 

TCPR 

4  

Prudlo et al., 2002

Amyotrophic lateral sclerosis

WB

3 / 1.5 (in anti-N)

Rybakowski et al., 2002

Psychiatric  

ECLIA

2.4 / 1.0 (in anti-P)

10.2 (recent onset of disease)

Lebain et al., 2002

Psychiatric

IFA

12.6 / 15.5

CFS: chronic fatigue disease; IFA: indirect immunofluorescence assay; WB: Western blotting; IP: immunoprecipitation assay; ELISA: enzyme-linked immunosorbent assay; RS-ELISA: reverse-type ELISA; ECLIA: electrochemiluminescence immunoassay; TCPR: T-cell proliferative response; N: nucleoprotein; P: phosphoprotein


Data on BDV infection from autopsied brain samples might be greatly affected by a long post mortem interval before processing for examination.  In addition, the duration of disease, frequency of relapses, the clinical course of disease prior to death and autopsy could be very important factors as well.  Most of the examinations for BDV were performed on autopsied brains from old patients.

Fig. 2. Detection by immunohistochemistry of BDV RNA in autopsied brain tissue from a patient with schizophrenia.  Sections from hippocampus, cerebellum and pons from P1 and P2 cases, which were negative and positive for BDV ‘P’ RNA by RT-PCR, respectively, were subjected to immunohistochemistry using a BDV ‘P’ antisense riboprobe.


We are now trying to detect BDV signals in the autopsied brains from neurodegenerative disorders.


Neurovirulence in model animals

Pathological findings in naturally infected horses and sheep are similar to those in experimentally infected animals such as rats.  Immunocompetent adult rats infected with BDV develop severe encephalitis and neural dysfunction because of anti-BDV immunopathogenesis.  Initial immune cell infiltrates in the perivascular spaces are CD8-positive and CD4-positive T cells, NK cells, and macrophages.  The encephalitic reactions in the CNS correlate to strong immune responses, especially CD8-positive T-cell-mediated immune responses against the N proteins.  Also, experimentally infected rats have a hyperactive movement disorder because of abnormal dopamine activity.

Neonatal rats infected with BDV develop persistent infection and show developmental disturbances affecting specific areas in the brain.  Neonatal BDV infection also results in a variety of behavioral abnormalities and neuroanatomical disturbances without generalized meningitis or encephalitis.  A chronic upregulation of proinflammatory cytokines such as IL-1b has been observed in the hippocampus and cerebellum of the neonatally infected rats.  Recent studies have demonstrated that neonatal BDV infection directly alters the concentrations of neurotransmitters in the brain, including norepinephrine and serotonin.  Furthermore, BDV infection displays a progressive decrease in synaptic density and plasticity, especially in the cortex and hippocampus, which precedes a significant dropout of the cortical neurons in the infected rats.  Reduced mRNA expression levels of neurotrophin-3, brain-derived neurotrophic factor (BDNF) and nerve growth factor (NGF) are found in the hippocampus of newborn infected rats.  These observations indicate that BDV infection has direct effects on the microenvironment of neuronal cells in the infected brain.


Host factors and cytopathogenicity

A wide range of host cell types and species can be infected with BDV that usually replicates in vitro without cell lysis, and establishes persistent infection chronically producing the virus.  BDV is tightly cell-associated.  Although BDV infects cells easily through cell-to-cell interaction, the infectivity titers of cell-free virus in the culture medium or even in the cell lysate are usually very low.  Interestingly, BDV replication rates are different among cell types.  As previously shown, more BDV is produced by neuronal cell lines than by astrocyte cell lines.  In addition, NGF treatment can enhance virus production.

    We recently demonstrated that the BDV P protein binds specifically with the host cellular protein.  The P protein is a nucleus-associated phosphoprotein and is putatively a cofactor for the BDV polymerase during replication and transcription.  Being one of the abundantly expressed viral proteins in infected cells, it is possible that the binding of P to host factor(s) would induce functional alterations in the infected neural cell environment.  Amphoterin, also named high-mobility group box 1 (HMGB1) protein, is a neurite outgrowth factor of 30-kDa that is present in abundance in developing brains (Fig. 3).  BDV infection, as well as the purified P protein in culture medium, can significantly inhibit cell process outgrowth of cells maintained on laminin (Fig. 4).  Furthermore, migration activity of the cells to laminin was also decreased by BDV infection.  These results suggest that BDV infection causes a functional disturbance of HMGB1 in cells by the interaction of the P protein.  This functional disturbance may result in neurodevelopmental damage in the developing brain, as reported in BDV-infected neonatal rats.  Thus, the interaction of HMGB1 with BDV P may similarly occur in the developing brains of the persistently BDV-infected humans.  However, there has been no report on the neonatal infection or the possible vertical transmission of BDV in humans.

Fig. 3. Blocking of HMGB1 function by its interaction with BDV P (phosphoprotein). HMGB1 synthesized is extracellularly secreted and interacts with its receptor molecule RAGE. Based on the interaction, RAGE induces the signals necessary for not only ofr neurite outgrowth but for the cellular survival. In addition, a part of the HMGB1 is located in the nuclei and involved in the transactivation. BDV P phosphoprotein can interact with HMGB1 and could block these functions of HMGB1.
Fig. 4. BDV infection inhibits neurite outgrowth of neural cells. Persistently BDV-infected (panels b, e, f) and uninfected (panels a, c, d) cells (a and b, human oligodendrocyte-derived cell line OL; and c-f, rat glia-derived cell line C6).

   HMGB1 also has intranuclear functions, which have been extensively studied. HMGB1 can interact via the HMG boxes with a broad range of proteins from nuclear proteins to viral components. Interaction with HMGB1 has been described in several transcription factors (p53, Hox, Pou, Oct, steroid hormone receptors and TATA-binding protein), viral proteins (adeno-associated virus Rep) and the recombination activation gene proteins (RAG1 and 2). It has been reported that HMGB1 has a function in the nucleus as a specific enhancer of p53 activity (Fig. 3). In general, HMGB1 increases the DNA binding affinity of those factors and shows either negative or positive effect on transcription.  These observations raised the possibility that BDV P can influence the transcriptional activity of p53 by interference of HMBG1 in the nucleus of infected cells.

We have demonstrated that P protein expression was shown to suppress the transcription activation of cyclin G1 and p21Waf1 via interference with HMGB1 binding to p53 (Fig. 5).

Fig. 5. Reduction of p53-mediated transcriptional activation of cycline G promoter by BDV P protein. The p53-deficienct NCI-H1299 cells were transiently transfected with pGL-cycline G-Leu and tested plasmids (pcD-HMGB1, pcD-p53, pcD-P, and pcD-PDM1. pcD-PDM1 contains a small deletion corresponding to the HMGB1-binding domain of BDV P amino acid 78-86. The cell extracts 24 hr after transfection were assayed for luciferase activity.

Then, we prepared transgenic mice expressing BDV P protein in the brain (Fig. 6).  The transgenic mouse line (GFP20) expressing high level of P protein showed aggressiveness and reduced spatial reference memory, while no apparent such signs in the GFP4 line expression of low level of P protein.  Neurobiological analysis revealed that the expression levels of BDNF and specific serotonin receptors were significantly reduced in GFP20 mouse brains.  Also, synaptic density in the brain was decreased in GFP20 mouse brains (Fig. 6).  Thus, these studies may indicate the P expression can induce the impairment of glial cell function, a common cause for the induction of neurobehavioral disorders.  Although the role of BDV infection in the induction of psychiatric disorders remains controversial, above studies using transgenic mouse promote further investigation regarding this question.

Fig. 6. Transgenic mouse expressing BDV P protein in the brain. A, P rpteon expressionin cerebellum of 8 months old transgenic mouse (arrows; Bergmann’s glia cells expressing BDV P); B, Reduction of brain-derived neurotrophic factor (BDNF), Non-Tg, control non-transgenic mouse; C, reduction of synaptic density in the brains by synaptophysin staining (a, cerebellum of BDV P transgenic mouse; b, control mouse); D, Offensive intermale aggression of BDV P transgenic mouse(high expression of BDV P in GFP20 transgenic mouse line; low expression of BDV P in GFP4 transgenic mouse line; and control non-transgenic mouse). 


BDV replication and persistent infection

1)     BDV replication

BDV has unique properties in replication and transcription of its genome; BDV is the only known animal NNS RNA virus that replicates and transcribes in the nucleus of infected cells, whereas the other animal viruses of this group undergo their life cycle in the cell cytoplasm.  BDV produces several polycistronic transcripts. Among six different proteins of BDV, only N is translated from monocistronic mRNA, while transcripts of other viral proteins are polycistronic (Fig. 1). .

The N, P and X are the major proteins expressed in infected cells.  The N protein exists in two forms: 40- and 38-kDa (p40N and p38N, respectively).  The latter lacks 13 N-terminal amino acids of the 40-kDa protein, because it is translated from the second AUG codon of the same mRNA.  In cells independently transfected with plasmids encoding each of the two N isoforms, p40N and p38N are accumulated in the nucleus and cytoplasm, respectively.  This difference in the subcellular distribution of the two forms is derived from a nuclear localization signal (NLS) present in the N-terminal of p40N.

The smallest transcript of BDV, as mRNA of 0.8 kb, is bicistronic and encodes ORFs for 10-kDa, X and P proteins. The ORF of X starts 49 nucleotides upstream from P ORF and overlaps with the 71 N-terminal amino acid of P protein in a different frame.  As X protein has been shown to bind directly or indirectly to P and N, these viral proteins would promote nuclear targeting of X in infected cells.  The amino terminus of X was recently shown to be critical for interaction with P and to contain a consensus leucine-rich sequence, which is very similar to nuclear export signal (NES)-like motif. 

Although there is no direct evidence for the role of the P protein in the life cycle of BDV, P seems to have a function similar to that proposed for phosphoproteins of other known NNS RNA viruses, i.e., it serves as an essential cofactor in virus transcription and replication.  The P protein is phosphorylated predominantly by protein kinase-e and, to a lesser extent, by casein kinase II. It contains two strong NLSs.  The P protein interacts with itself, with X and N, and co-localizes with N and X in distinct areas within the nuclei of the infected cells.  In addition to the P protein, the P ORF also produces a 16-kDa protein (P’) by translation from the second in-frame AUG codon.  This 16-kDa P’ has been detected in BDV-infected cultured cells and in brains of experimentally infected animals.  The smallest mRNA codes X, P, and P’ (Fig. 7).

BDV p16, the putative BDV M protein, is a nonglycosylated matrix protein associated at the inner surface of the viral membrane.  M forms stable tetramers and contains hydrophobic sequences characteristics of membrane spanning proteins.  It has been demonstrated that anti-sera against M protein have neutralizing activity and that BDV infection into cultured cells is prevented in the presence of M protein.  In addition, G protein, N-glycosylated with high-mannose and/or hybrid oligosaccharides, yields a full length type I membrane protein that is cleaved and the C-terminal cleavage product is identified as gp43.  Subcellular localization studies demonstrated that gp94/gp84 accumulates in the endoplasmic reticulum, whereas gp43 reaches the cell surface. Furthermore, both products were found to be associated with infectious virions. The presence of neutralization epitopes on the G protein and its capacity to interfere with infectivity suggested that the G protein is important for viral entry, as well as triggering the fusion events.  A recent pseudotype approach based on a recombinant vesicular stomatitis virus revealed that the N-terminal domain of the gp94/gp84 is sufficient for receptor recognition and virus entry. 

The last BDV protein is L, predicted to encode the viral RNA-dependent RNA polymerase.  Expression of L protein from the third transcription units is dependent on a splicing event that fuses a small upstream ORF that overlaps with the 5’ end of the G ORF. The L protein is predominantly present in the nuclei of cells transfected with an L expression plasmid in the absence of other viral proteins.  BDV L was also shown to be phosphorylated by cellular kinases and to interact with the P protein.  Recently, over-expression of recombinant L showed interaction with P and further indicated that L is phosphorylated by cellular kinases.


Fig. 7. Mechanism for translation of BDV 0.8 kb mRNA.  A, Detection of P’ protein expression in COS cells transfected with cDNA homologous with 0.8 kb mRNA and persistently BDV-infected OL cells.  As a control, COS and OL cells were used.  B, Schematic presentation of start translation point of individual X, P, and P’ proteins.

BDV belonging to the order Mononegavirales has similarity in its genome organization to other members of this order.  However, BDV has several unique features.  One of the most striking characteristics of BDV is its location for transcription.  Therefore, BDV employs the RNA splicing machinery for gene expression and the genome contains at least three introns: Intron I, nucleotide (nt) 1932-2025; intron II, nt 2410-3703; intron III, nt 2410-4559 (Fig. 1).  Transcripts that retain intron I serve as messages for expression of the M protein of BDV, and those that retain intron II serve as messages for expression of the G.  Transcripts that lack both introns serve as messages for expression of the L protein of BDV.  The intron III-spliced RNAs could form predicted ORF’s in all three frames. Frame 1 generated the largest ORF, which encodes the first 58 amino acids of G protein fused to L protein lacking one quarter of the N-terminus.  Frames 2 and 3 generated ORF’s encoding putative products (Fig. 1).  We have demonstrated that the alternative splicing of introns II and III in BDV is regulated by an alternative polyadenylation and cis-acting exon splicing suppressor (ESS) element within the L gene that contains similar motifs with other viral and cellular ESS (Fig. 1). The pre-mRNAs that are terminated at the E5 signal might splice the intron III sequence efficiently (SA3 splicing), because the pre-mRNAs could not carry the ESS region in their sequences (Fig. 8).

Fig. 8. Alternative splicing of BDV.The SA3 splicing could be inefficient in the E5 readthrough pre-mRNAs, because the pre-mRNAs carry the exon splicing suppressor (ESS) in the sequences. On the other hand, pre-mRNAs that are terminated at the E5 could employ SA3 signal efficiently.


2) BDV persistent infection
    BDV easily establishes persistent infection and this property could be derived from special regulation mechanisms for viral protein expression levels and their localization in the cells. Neuronal cell lines infected with BDV express the major two viral proteins, N and P. In the early stage of BDV infection, these antigens are found in the nucleus as dot-likestaining by immunofluorescence (Fig. 9). The BDV antigen is gradually translocated to the cytoplasm from the nucleus. In the persistently infected cells, the BDV antigen is abundantly expressed in the cytoplasm (Fig. 9). These dot structures are believed to be the center for viral transcription and replication. Therefore, it seems that viral protein distribution in the cells may be related with the mechanism for persistent infection. 

Fig. 9. Intracellular localization of BDV antigen.a, In the early stage of BDV infection, BDV antigen is found in the nucleus as dot-like structure. b, The BDV antigen is gradually translocated to the cytoplasm from the nucleus. c and d, In the persistently infected cells, the BDV antigen is abundantly expressed in the cytoplasm. Green: anti-Ki67 monoclonal antibody, Red: anti-BDV P monoclonal antibody. e,MDCK cells persistently infected with BDV.


 Then, we characterized the viral protein distribution during the course of BDV infection.  BDV replicates and transcribes in the nucleus of infected cells, and therefore, nuclear import and export of the viral genome are critical for the viral life cycle.  Recent studies have suggested that the nuclear transport activity of BDV is mediated by the major viral antigens, N and P, which contain NLS and/or NES within their sequences.  The NES of BDV N contains a canonical leucine-rich motif, and the nuclear export activity of the protein is mediated through the chromosome region maintenance protein pathway, indicating that BDV N has two contrary activities: nuclear localization and export.  These observations suggested that the nuclear localization of the N and P proteins is critical for nuclear targeting of the BDV RNA-protein complexes (vRNP). BDV must employ a switch mechanism that changes the direction of nuclear transport of the vRNP’s dependent on the viral life cycle in infected cells. The nucleocytoplasmic shuttling proteins such as N could play an essential role in the switch mechanism. The mechanism of N may be mediated by the interaction with P protein and production of an NLS-lacking N isoform, p38N. Interestingly, we found that the NES of N exactly overlaps one of the P-binding sites. Interestingly, we found that the nuclear export activity of BDV p38N is blocked by the expression of P in transfected cells (Fig. 10).

Fig. 10. BDV P prevents nuclear export of p38N.(A) COS-7 cells were transfected with p38N or P plasmid. BDV p38N localizes in the cytoplasm in the solo expression. (B) p38N was retained in the nucleus in the presence of P in the cells
Fig. 11. Nucleocytoplasmic transport of BDV RNP. BDV N and P may actively retain vRNPs in the nucleus and enhance their transcription and replication. An increased amount of X protein in the nucleus may regulate interaction between P and vRNPs, leading to the decreased level of BDV replication. The dissociation of P from vRNPs could enhance nuclear export of vRNPs, because BDV N contains NES (nuclear export signal). Molecular ratio or interaction among these proteins in the nucleus seems to be important for determination of the direction of vRNPs.

Recent studies have also revealed that binding between BDV P and X proteins is mediated via a putative-NES region within the X protein. These observations suggested that P might act as a nuclear retention factor of the N and X proteins by binding directly to the NES.  Indeed, the significance of the interaction between X and P proteins have suggested in a recent study.  We have demonstrated that transient transfection analysis using a cDNA clone corresponding a bicistronic transcript that expresses both the X and P proteins revealed that P efficiently localizes in the cytoplasm only when BDV X is expressed in the cells. Furthermore, it has been revealed that the direct binding between X and P is necessary for the cytoplasmic localization of the P. Interestingly, we showed that X is not detectably expressed in the BDV-infected cells in which P is predominantly found in the nucleus with little or no signal in the cytoplasm. These observations suggested that BDV P can modulate their subcellular localization through the binding to the X and that BDV may regulate the expression ratio of each viral product in infected cells to control the intracellular movement of the viral protein complexes.

     Recently, we proposed a model of nucleocytoplasmic trafficking of BDV RNP’s, in which relative levels of N, X and P in the nucleus may play a key role for determining the direction of BDV RNP movement.  Increased levels of P could mediate retention of the vRNP’s in the nucleus by the masking of the NES of N during the nuclear replication stage (Fig. 11). On the other hand a lower concentration of P in the nucleus can increase free NES, mediating nuclear export of N-containing RNP complexes (Fig. 11).  An increased amount of X protein in the nucleus could promote nuclear export of P protein, resulting in decreased level of P in the nucleus. This could subsequently lead to cytoplasmic localization or maturation of BDV RNP’s (Fig. 11).  In addition, the presence of the NLS-lacking p38N in N multimer can increase the relative number of NES when compared with the NLS. An increased number of NES in N complex would enhance nuclear export of the viral RNPs for maturation or assembly of the progeny virions. At present, however, there is a little evidence supporting significant changes in the ratio between N and P during the viral life cycle. Further experiments will be necessary to understand the nucleocytoplasmic trafficking of BDV RNP’s in detail.