Polyomaviruses of Birds: Etiologic Agents of Inflammatory Diseases in a Tumor Virus Family

Since their first detection in 1981 (7, 9), the polyomaviruses of birds have been known to cause acute and chronic diseases in several species of birds. Polyomavirus disease of parrots caused by avian polyomavirus (APV) (50, 70) and hemorrhagic nephritis and enteritis of geese (HNEG) caused by goose hemorrhagic polyomavirus (GHPV) (27, 57) are inflammatory diseases characterized by high mortality rates in young birds. The high pathogenicity of polyomaviruses of birds is in contrast to the innocuous persistent infections that the intensively studied mammalian polyomaviruses, such as simian virus 40 (SV40), usually cause in their natural nonimmunocompromised hosts (30, 48). The mammalian polyomaviruses are well known to induce tumors after inoculation into nonpermissive laboratory rodents (4, 14, 23); such events have never been observed for birds with APV or other bird polyomaviruses. Here, we review the available literature on polyomaviruses of birds, focusing on the characterization of the known viruses and the diseases caused by them and on approaches to investigate the mechanisms for their pathogenicity.

POLYOMAVIRUSES OF BIRDS AND DISEASES CAUSED BY THEM

Polyomaviruses are widely known as tumor-inducing agents. However, this property is observed only after experimental infection of newborn laboratory rodents with some of the mammalian polyomaviruses. It is well documented for SV40 (14, 23) and murine polyomavirus (MPyV) (4, 12). In contrast, natural and experimental SV40 infection of macaques, which are permissive for this virus, leads to a persistent infection without any clinical symptoms (48). The human polyomaviruses, JC polyomavirus (JCPyV) and BK polyomavirus (BKPyV), in general also establish subclinical persistent infections in their human host (81). Mammalian polyomaviruses have not been etiologically linked to a severe acute disease in their permissive nonimmunocompromised hosts after natural infection. By contrast, an acute respiratory disease could be induced by experimental infection of newborn mice with the murine pneumotropic virus (25). The MPyV laboratory strain LID, which was derived from the wild-type isolate by repeated passaging in mouse embryo cells and newborn mice, has been shown to cause a rapidly lethal infection of newborn mice (3). Distinct clinical symptoms have been observed after infection of severely immunosuppressed hosts, leading to progressive multifocal leukoencephalopathy in the case of JCPyV infection (61) and to polyomavirus-associated nephropathy in the case of BKPyV infection (29). The etiologic role of the recently discovered human KI and WU polyomaviruses, which have been detected in patients with respiratory diseases, remains to be elucidated (1, 21).

Within the family Polyomaviridae and its sole genus, Polyomavirus (Table 1), at present four polyomaviruses of birds are known, most of them having been discovered only recently. APV was first described in 1981 as the etiologic agent of budgerigar fledgling disease, a devastating disease of young budgerigars, and was subsequently designated budgerigar fledgling disease virus (7, 9). The disease is characterized by hepatitis, ascites, and hydropericardium as the main clinical symptoms and a mortality rate of up to 100% in the fledglings (41, 44). Young budgerigars which survive the infection develop a chronic course of disease with feather disorders as the main clinical symptom. Both types of disease have been reproduced by experimental infection of fledgling budgerigars, by using APV propagated in tissue culture (44). APV infections have also been detected in other parrot species, causing clinical signs similar to those observed for budgerigars; however, the degree of susceptibility for and severity of the disease seems to be dependent on the species infected (15, 70). A distinct disease course found in older parrots is characterized by sudden death accompanied by a membranous glomerulopathy (22, 65). Birds of other families also seem to be susceptible to infection with APV (46, 73, 78). There are several reports on APV infection in diseased finches (18, 71, 77); APV was also detected in buzzards and falcons (34). APV has been found in many countries, suggesting a worldwide distribution (8, 28, 31). All viruses analyzed to date show a very close genomic relationship, and only one genotype (34, 66) and one serotype (42) exist. Although the virus has been designated budgerigar fledgling disease polyomavirus by the International Committee on Taxonomy of Viruses (30), the broad host range of the virus led to the designation APV (34), which is now widely used.

TABLE 1.

Members of the family Polyomaviridae and its genus Polyomavirus

Host species	Virus (reference for first description)	Abbreviation	Disease after natural infection
Human	BK polyomavirus (19)	BKPyV	Posttransplantation polyomavirus- associated nephropathy
	JC polyomavirus (61)	JCPyV	Progressive multifocal leukoencephalopathy in AIDS patients
	KI polyomavirus (1)	KIPyV	Not known; may be associated with respiratory disease
	WU virus (21)	WU	Not known; may be associated with respiratory disease
Monkey	Simian virus 40 (80)	SV40	Not known
	Simian virus 12 (55)	SV12	Not known
	Baboon polyomavirus 2 (20)	BPyV-2	Not known
	B-lymphotropic polyomavirus (87)	LPyV	Not known
	Chimpanzee polyomavirus (38)	ChPyV	Not known
Cattle	Bovine polyomavirus (75)	BPyV	Not known
Rabbit	Rabbit kidney vacuolating virus (32)	RKV	Not known
Hamster	Hamster polyomavirus (24)	HaPyV	Skin tumors
Rat	Athymic rat polyomavirus (84)	Rat-PyV	Sialoadenitis in athymic nude rat
Mouse	Murine polyomavirus (26)	MPyV	Not known
	Murine pneumotropic virus (43)	MPtV	Respiratory disease in suckling mice
Parrot and other bird species	Budgerigar fledgling disease polyomavirus (7, 9) (avian polyomavirusa [34])	BFPyV (APVa)	Budgerigar fledgling disease; polyomavirus disease in other bird species
Goose	Goose hemorrhagic polyomavirus (27)	GHPV	Hemorrhagic nephritis and enteritis of geese
Finch	Finch polyomavirus (40)	FPyV	Polyomavirus disease
Crow	Crow polyomavirus (40)	CPyV	Not known

^aSuggested designation according to reference 34.

GHPV, the second polyomavirus of birds, was identified in 2000 in geese suffering from HNEG (27). This disease was described first in 1970 in Hungary (6) and subsequently in Germany (74) and France (83). It is characterized mainly by sudden death of 2- to 10-week-old goslings, resulting in cumulative mortality rates of between 4% and 64% (57, 62). The major pathological sign is a hemorrhagic nephritis with necrosis of the tubular epithelium as well as subcutaneous edemas, whereas enteritis is seen only rarely (45). A chronic course of disease is characterized by visceral gout (57). HNEG has been experimentally reproduced using GHPV purified from organ extracts of diseased geese as well as using GHPV passaged in primary goose kidney cells (27, 45). The GHPV genome has been detected in geese flocks with HNEG situated in France (45), Germany (57), and Hungary (62). There is a high seroprevalence in Germany, even in some flocks in which symptoms of HNEG never had been recorded (86). All GHPV isolates characterized so far are genetically similar; however, a full genome sequence has been obtained for only one strain (36).

Two additional polyomaviruses of birds, the finch polyomavirus (FPyV) and the crow polyomavirus (CPyV), have been detected in 2006 by a screening approach for novel polyomaviruses using a broad-spectrum PCR (40). Both viruses have so far been characterized only by genome sequencing (one representative of each). FPyV was first detected in a young bullfinch in a German aviary that died suddenly, along with his siblings (85). Clinical symptoms like high mortality in young birds and feather disorders in older birds are reminiscent of APV infections and occurred also in other FPyV-infected bird species (finches, grosbeaks, and tits) of the aviary. However, experimental infections to support an etiologic connection between virus infection and disease have so far not been done. The genome of CPyV was detected in a sample of a crow found dead during an epidemic of sudden death in the wild bird population of Spain (40), but an etiologic role of the virus for the disease has not been proven, either.

GENOME ANALYSES AND PHYLOGENETIC RELATIONSHIPS OF POLYOMAVIRUSES

Comparative analysis of the genome sequences of mammalian and bird polyomaviruses has been applied to identify possible factors of pathogenicity. The genome of all members of the family Polyomaviridae consists of one circular double-stranded DNA molecule, approximately 5,000 bp in size (30). It is transcribed bidirectionally for the expression of early and late genes, which are separated by a short noncoding regulatory region (Fig. 1A). The early genes encode the tumor (T) antigens which participate in viral genome replication and transformation of cells. All known polyomaviruses encode a large T antigen and a small T antigen, which are synthesized from differentially spliced mRNAs. In addition, MPyV and the hamster polyomavirus encode a middle T antigen by alternative splicing of mRNA, leading to a frameshift compared to the large T antigen open reading frame (ORF). The late genes code for virus protein 1 (VP1), VP2, and VP3; the latter two of these structural proteins are encoded by the same ORF but use different initiation codons. The primate and human polyomaviruses SV40, JCPyV, and BKPyV encode the so-called agnoprotein by using an ORF located upstream of the VP2-encoding region. This nonstructural protein is dispensable for propagation in cell culture (76). Its perinuclear accumulation (59) and interactions with cellular proteins suggest functions in virus release (60, 79), regulation of the cell cycle, and DNA repair (11). Very recently, an additional protein, consisting of the 125 C-terminal amino acid residues of VP3, has been identified in SV40-infected cells (10). This protein is synthesized very late in the replicative cycle but is not incorporated into the viral capsid and is involved in host cell lysis. Although this protein has no similarities regarding sequence, structure, or biological properties with VP4 of APV mentioned below, it has also been designated VP4.

FIG. 1.

Comparison of genome sequences of mammalian polyomaviruses and polyomaviruses of birds. (A) Schematic view of the genomes of SV40 and APV. Regions of low sequence similarity between these viruses are shown as crosshatched boxes. Indicated are the origin of replication (ori) and the regions encoding the small tumor antigen (sT-Ag), large tumor antigen (lT-Ag), agnoprotein (agno), and VP1 through 4. Despite having the same designation, VP4 of SV40 and VP4 of APV have no structural or functional similarities. (B) Phylogenetic relationship of 15 polyomaviruses based on the complete amino acid sequences of large tumor antigen. Abbreviations of the virus names are as in Table 1. The alignment was performed using the Clustal W algorithm of the MegAlign module of DNASTAR software package (LASERGENE).

In principle, the polyomaviruses of birds show a genome organization similar to that of the mammalian polyomaviruses (Fig. 1A), with genomes sizes ranging from 4,981 bp (APV) and 5,278 bp (FPyV). They do not encode a middle T antigen. No ORF encoding a protein with homologies to the agnoprotein is evident. Instead, all polyomaviruses of birds have an additional (multiply) spliced ORF in the 5′ region of the late mRNA (36, 40, 54) encoding VP4 in the case of APV or a homologous protein (designated a product of ORF-X) in the other bird polyomaviruses. A phylogenetic analysis of the genome sequences of all polyomaviruses shows a separate branching of the polyomaviruses of birds, indicating a separate grouping and phylogenetic development of these viruses (40, 63).

A detailed analysis of the genome sequences shows that the main differences between mammalian and bird polyomavirus genome sequences are found in the large-T-antigen-encoding region, the noncoding regulatory region, and in the above-mentioned VP4-encoding (ORF-X) region (Fig. 1A) (40). Minor differences are also present in the sequences of the other proteins, e.g., the lack of the N-terminal nuclear localization signal in VP1 and the lack of a CXCX₂C consensus binding sequence for PP2A in the small T antigen of the bird polyomaviruses.

The analysis of the amino acid sequences of the large T antigens indicates an independent evolution of these proteins in avian and mammalian polyomaviruses over a longer period of time, as demonstrated by separate branching in a phylogenetic tree (Fig. 1B). Most of the functional domains required for enzymatic activity of the large T antigen are conserved in all polyomaviruses; however, differences in DNA-binding activity and sequence differences in binding domains for cellular proteins have been observed (40, 53, 67). The consensus sequence LXCXE, serving as binding sequence for the retinoblastoma protein pRB, is found in all polyomavirus large-T-antigen sequences, including the bird polyomaviruses GHPV, FPyV, and CPyV; however, it is mutated to LXAXE in APV (72). No homologies to the p53 binding sequences of the SV40 large T antigen are evident in the genomes of bird polyomaviruses (40). In contrast to the mammalian polyomaviruses, the bird polyomaviruses show no accumulation of basic amino acid residues within the region of the nuclear localization signal of the large T antigen. In the region known to be responsible for DNA binding of large T antigen in mammalian polyomaviruses, the bird polyomavirus large T antigens have a different consensus sequence, suggesting a different manner of DNA binding for these proteins (40). This difference is also reflected by different consensus sequences of the target DNA in the noncoding regulatory region. Whereas SV40, MPyV, and probably all of the mammalian polyomavirus T antigens use the pentanucleotide GAGGC as binding sequence (82), the bird polyomaviruses carry the palindromic motif CC(A/T₆)GG in the cognate region (36, 40), which has been identified to function as a large-T-antigen binding sequence in APV (53).

ADDITIONAL STRUCTURAL PROTEIN VP4 OF AVIAN POLYOMAVIRUSES

The most obvious difference between the genome sequences of mammalian and bird polyomaviruses is the presence of the additional ORF located upstream of the VP2-encoding region, coding for VP4 of APV or a homologous protein in GHPV, FPyV, and CPyV. As this ORF is located at a position corresponding to that of the agnoprotein ORF in the primate polyomaviruses, the designations agnoprotein 1a and agnoprotein 1b were originally used for the two splicing variants of the protein of APV (54). Because of significant differences in sequences and biological properties, the designations VP4 and VP4Δ are now applied (35). All of the mRNAs encoding these proteins of the bird polyomaviruses are spliced at least once; in addition, in the case of APV, the two variants VP4 and VP4Δ are created by alternative splicing, leading to an internally deleted protein in the case of VP4Δ (54). With sizes of 150 (CPyV) to 205 (FPyV) amino acid residues, the proteins are relatively small; they contain a high percentage of proline residues (14.2% in APV and 16.0% in CPyV) (36, 40).

Up to now, functional studies have been performed only with VP4 of APV. This protein is regularly present in purified viral particles, and interactions with the major structural protein VP1 as well as with double-stranded DNA have been identified (35). This has initially led to the assumption that the protein is a structural component that also has functions in the packaging of the viral genome into the virus particle. However, it has been shown that this protein is not necessary for the formation of virus-like particles (37). An additional property of the protein is the induction of apoptosis, which is observed after its expression in insect cells as well as in chicken embryo cells, the latter also supporting APV replication (33). Limited deletion mapping has shown that the central part of VP4 is responsible for DNA binding (35), whereas the N- and C-terminal parts are involved in the induction of apoptosis (33) (Fig. 2). Sequence analysis shows that a typical leucine zipper motif is present in the central part of VP4 and may be responsible for multimerization of this protein (35). Thus, basic amino acid residues are positioned in adjacent positions, enabling a DNA-binding mechanism similar to that of some cellular transcription factors, e.g., c-fos/c-jun (47). There are multiple sites of phosphorylation in VP4 as identified by mutational analysis. The functional significance of phosphorylation, however, is not clear at present (51, 52). Also, the reason for the difference in the apparent molecular mass of this protein in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (32 kDa) and its calculated molecular mass (19.8 kDa) (51), even after expression in Escherichia coli (35), has not been elucidated until now.

FIG. 2.

Schematic map of VP4 of APV. General characteristics of the amino acid composition are shown above the map. Leucine residues (L) involved in the formation of a leucine zipper-like structure and phosphorylation sites (P) are indicated. Functional domains involved in induction of apoptosis (apoptosis), DNA binding, and the proposed multimerization of VP4 are labeled in italics. The numbers beneath the map refer to amino acid positions. Amino acids 70 to 132 are deleted in the splicing variant VP4Δ.

Several deletion mutants of APV have been generated and studied to assess the function of VP4 during viral replication. A mutation of the common initiation codon for VP4 and VP4Δ expression totally abolished virus replication in chicken embryo cells, indicating that expression of at least one of these proteins is essential (33). Further studies using splicing mutants expressing either VP4 or VP4Δ indicated, however, that infectious virus was produced with slower replication kinetics and lower end-point titers than those of the wild-type virus (39). All mutants were attenuated with respect to the induction of apoptosis in chicken embryo cells. Therefore, VP4 and VP4Δ have been suggested to be pathogenicity factors (33, 39). The infectious mutants have also been tested in chickens, and serological investigations indicated viral replication in the birds. The induction of neutralizing antibodies suggests these viruses as candidate vaccines. However, as chickens infected with wild-type APV do not show any symptoms (16, 17), the safety of the mutants has to be assessed in further experiments with susceptible animals in which APV causes disease, e.g., parrots.

PATHOGENESIS OF DISEASES CAUSED BY POLYOMAVIRUSES OF BIRDS

After infection of budgerigars with APV, the virus can be detected in essentially all organs (64). APV, which has the ability to replicate in a broad variety of organs, is in contrast to the mammalian polyomaviruses, which usually show a distinct tissue tropism (30), e.g., that of JCPyV to the central nervous system (68), BKPyV to the kidney (49), or mouse pneumotropic polyomavirus to the lung (25). Remarkably, the highly pathogenic laboratory MPyV strain LID shows also a broad tissue distribution in infected newborn mice, which has been linked to a mutation in the receptor-binding site of VP1, resulting in an increased viral spread throughout the organism (2, 3). Pathohistological studies in geese infected with GHPV also showed a broad tissue tropism for this virus (5, 45); however, a closer look revealed a preferential infection of endothelial cells which are present in all organs (13). The restricted cell type specificity of GHPV is also reflected by the fact that GHPV does not replicate efficiently in cultures of goose embryo fibroblasts or goose kidney cells (5, 36), in contrast to APV, which grows to high titers in cultures of chicken embryo fibroblasts as well as in some other cell culture systems (58). The tissue tropism of FPyV and CPyV has not been investigated so far; neither virus could be propagated after inoculation onto cultures of chicken embryo cells (40).

Generally, the ability of the polyomaviruses of birds to destroy a high number of cells in the infected organism is thought to be the cause of the severe disease. According to this hypothesis, the multisystemic acute disease is caused by destruction of multiple organs of APV-infected parrots (33), and the edemas and hemorrhages in the skin and in the kidney are caused by destruction of endothelial cells in GHPV-infected geese (13). The reason for the marked cellular damage caused by these viruses is not known. However, it has been speculated that the induction of apoptosis by these viruses could contribute to pathogenicity (33). Indeed, APV infection of chicken embryo fibroblasts led to marked cellular damage due to the induction of apoptosis, in contrast to the effect of SV40 infection of Vero cells, which is characterized mainly by necrosis (33). Induction of apoptosis may facilitate an efficient release of virus progeny from the infected cells in the absence of a primary inflammatory response, leading to efficient virus spread throughout the whole organism.

As VP4 of APV has been characterized as an inductor of apoptosis and the gene products of ORF-X of the other bird polyomaviruses show considerable homologies to this protein, a general mechanism of pathogenicity by expression of these proteins may exist (40). Further research is needed to characterize these proteins and respective knock-out mutants to verify this hypothesis. VP4 deletion mutants of APV have been considered as vaccine candidates to be used against the APV-induced disease in parrots (39). As efficient and safe vaccines are urgently needed against the diseases caused by APV and GHPV, the VP4 deletion mutants as live-attenuated vaccines should be tested in parallel to classical vaccines based on inactivated virus (69) or recombinant vaccines based on virus-like particles (37, 86).

In summary, on the basis of the differences observed in genome structure, gene protein coding assignments, functions of accessory proteins, and virus-host relationships, a revision of the actual classification of the polyomaviruses is proposed. Grouping of the mammalian and bird PVs into separate genera with the proposed designations Mastpolyomavirus and Avipolyomavirus, respectively, is suggested.