Characterization of the Complete Genome of the Tupaia (Tree Shrew) Adenovirus

Abstract

The members of the family Adenoviridae are widely spread among vertebrate host species and normally cause acute but innocuous infections. Special attention is focused on adenoviruses because of their ability to transform host cells, their possible application in vector technology, and their phylogeny. The primary structure of the genome of Tupaia adenovirus (TAV), which infects Tupaia spp. (tree shrew) was determined. Tree shrews are taxonomically assumed to be at the base of the phylogenetic tree of mammals and are frequently used as laboratory animals in neurological and behavior research. The TAV genome is 33,501 bp in length with a G+C content of 49.96% and has 166-bp inverted terminal repeats. Analysis of the complete nucleotide sequence resulted in the identification of 109 open reading frames (ORFs) with a coding capacity of at least 40 amino acid residues. Thirty-eight of them are predicted to encode viral proteins based on the presence of transcription and translation signals and sequence and positional conservation. Thirty viral ORFs were found to show significant similarities to known adenoviral genes, arranged into discrete early and late genome regions as they are known from mastadenoviruses. Analysis of the nucleotide content of the TAV genome revealed a significant CG dinucleotide depletion at the genome ends that suggests methylation of these genomic regions during the viral life cycle. Phylogenetic analysis of the viral gene products, including penton and hexon proteins, viral protease, terminal protein, protein VIII, DNA polymerase, protein IVa2, and 100,000-molecular-weight protein, revealed that the evolutionary lineage of TAV forms a separate branch within the phylogenetic tree of the Mastadenovirus genus.

Adenoviruses were first isolated and characterized in 1953 as causative agents of acute respiratory diseases in humans (56). Adenoviruses are distributed worldwide and are commonly found in vertebrate host species, including humans. Until now about fifty different serotypes of human adenoviruses are known. The family Adenoviridae is composed of the two established genera Mastadenovirus, whose members infect mammals, and Aviadenovirus, whose members infect birds, and the two proposed genera Atadenovirus and Siadenovirus (5, 6, 18, 26, 31, 69, 71).

Tupaias, or tree shrews, belong to the family Tupaiidae and are originally distributed in Southeast Asia. They are used worldwide as laboratory animals in neurological and physiological research because their evolutionary lineage is supposed to have branched off at the base of the phylogenetic tree of mammals (35). To date, several herpesviruses (3), a rhabdovirus (36), a paramyxovirus (66), and an adenovirus (16) have been isolated from tree shrews (Tupaia spp.).

Tupaia adenovirus (TAV) was isolated from a degenerating kidney cell culture of an apparently healthy tree shrew (Tupaia belangeri) (16). The biological properties of this virus and the molecular features of its DNA and proteins had been studied in detail previously (7, 8, 16, 20, 21, 44, 61). An extensive host-range study with different animal and human cell lines has revealed that only Tupaia cells were susceptible to a productive TAV infection and viral replication (16). Tupaia baby kidney cells are the cells of choice for the efficient production of cell-free TAV (10⁸ PFU ml⁻¹ within 36 h postinfection) and plaque assay (16). Additionally, host-range studies have shown that TAV immortalizes skin fibroblasts of the New World monkey Callithrix jacchus (16). Purified TAV particles have a capsid diameter of 78 to 80 nm, and the molecular mass of viral DNA was found to be 21.5 × 10⁶ Da, corresponding to a genome length of 32.5 kbp as determined by contour length measurements (16, 44). The buoyant density of TAV DNA is 1.706 g ml⁻¹ as determined by isopycnic CsCl centrifugation (16). Physical maps of the double-stranded DNA genome of TAV were constructed for a variety of restriction endonucleases (44). The virion polypeptides of TAV, which were analyzed by sodium dodecyl sulfate-polyacrylamide slab gel electrophoresis, isoelectric focusing, and two-dimensional analysis, revealed the presence of 18 discernible polypeptides, 15 of which seemed to be different from those of known human and animal adenoviruses (20). The nucleotide sequences of the TAV genome termini have been determined. Inverted terminal repetitions with a length of 166 bp were found, which contain an A+T-rich highly conserved sequence that is present in all adenovirus genomes analyzed so far (8). Furthermore, the major late promoter (MLP) of TAV was identified, which harbors defined motifs conserved in MLPs of other adenoviruses (61). In addition, the nucleotide sequence of the left-hand region of the TAV genome has been determined (7, 21). Transcription signals, open reading frames (ORFs), and splice sites were assigned on the basis of relatedness to the E1A region of human adenoviruses 5, 7, and 12. A consensus sequence for encapsidation of adenoviral DNA that was based on the established packaging region of human adenovirus 16 was detected, and a viral gene encoding a polypeptide of 18 kDa corresponding to the E1A protein of the transforming region of human adenovirus serotypes was identified (7).

The nucleotide sequences of the genomes of several human and animal adenoviruses have been completely determined (11, 12, 17, 18, 27, 33, 45, 46, 48, 51-55, 62, 70). These studies provided important insight into genome structure, gene content, gene arrangement, and evolution of adenoviruses. The genetic data obtained from these studies were fundamental and provided the basis for adenoviruses to be considered potential viral eukaryotic vectors for prophylactic and therapeutic application of foreign DNA associated with vaccination and gene delivery. During the last years, there was a strong increase in knowledge of the construction and usability of viral vectors, including adenoviruses (40, 47). Replication-deficient adenoviruses seem to be suitable gene delivery vectors for the genetic treatment of a variety of diseases. The ability of such vectors to mediate efficient expression of therapeutic genes in a broad spectrum of cell types, e.g., muscle, liver, and tumor cells, etc., constitutes an advantage over alternative gene transfer vectors (4, 40, 49, 50). The elucidation of the primary structure of the TAV genome by determination of the complete DNA nucleotide sequence of the viral genome, analysis of the TAV genome structure, gene content, coding capacity and strategy of the viral genome, and the phylogenetic classification of TAV are the goals of the present study. The results from this investigation contribute to the understanding of the evolution of adenoviruses.

MATERIALS AND METHODS

Viral DNA.

The propagation of TAV on Tupaia baby kidney cells and isolation of viral DNA was carried out as described previously (16).

Molecular cloning of TAV DNA.

The DNA of TAV was cleaved with restriction endonucleases BamHI, ClaI, ClaI/EcoRI, HindIII/EcoRI, BamHI/ClaI, and BamHI/EcoRI as described previously (44). The resulting DNA fragments were inserted into the corresponding sites of plasmid pAT153 (68) by T4 ligation. Competent Escherichia coli K-12 C600 (2) was used for transformation by previously described methods (14). Recombinant plasmids harboring specific DNA fragments of TAV were selected and identified by restriction endonuclease analysis and Southern blot hybridizations (63).

Enzymes and DNA isolation.

The restriction endonucleases were purchased from Roche Diagnostics GmbH (Mannheim, Germany). Incubations were carried out according to standard procedures for each enzyme. The recombinant plasmids harboring the DNA sequences of the TAV genome were extracted and purified by using the Qiagen (Hilden, Germany) plasmid midi prep procedure. The PCR products were purified by using MicroSpin S-300 HR columns (Pharmacia Biotech, Freiburg, Germany).

DNA sequencing.

The recombinant plasmids and the purified PCR products were sequenced by using the DyeDeoxy terminator Taq cycle sequencing technique (Ready Reaction DyeDeoxy terminator cycle sequencing kit; Applied Biosystems, Weiterstadt, Germany) and a 373A Extended DNA sequencer (Applied Biosystems) as described previously (65). Sequence Navigator software (version 1.01; Applied Biosystems) was used to assemble the nucleotide sequences obtained from individual sequencing reactions.

Computer-assisted analysis.

The strategies and criteria used to declare a TAV ORF to be coding for a viral protein were a minimum length of 120 bp, the presence of consensus promoter sequences, positional conservation, and homology to known genes. The identification of ORFs was performed with the program TRANSLATE of the PC/GENE program, release 6.85 (Intelligenetics Inc., Mountain View, Calif.). Homology searches of potential TAV proteins to known proteins were performed by using BLAST (Basic Local Alignment Search Tool) (1) and FSTPSCAN of PC/GENE to the SWISSPROT database, release 40. Protein alignments were carried out with the CLUSTAL program (28). Examination of the DNA sequence for transcription signals was performed by using the search features of the program EUKPROM of PC/GENE, release 6.85. Protein motif searches were performed with the program PROSITE of PC/GENE, release 6.85, and PROSITE, release 17.20, on the ExPASy molecular biology server (9, 29). Splice site predictions were performed with the Berkeley Drosophila Genome Project website. The phylogenetic trees were calculated by using the PC programs ClustalW (version 1.64b) (64), Kitsch (version 10.0), Protdist (version 10.0), and Treeview (version 1.5.2).

Nomenclature.

All ORFs that were identified in the TAV genome were numbered one after the other according to their appearance when the genome sequence was analyzed from bp 1 to 33,501. ORFs that were orientated to the right were marked with the letter R; ORFs that were orientated to the left were marked with the letter L. All ORFs that are predicted to code for viral proteins are listed in Table 1 along with relevant specifications and features.

TABLE 1.

Listing of all TAV ORFs that are assumed to code for viral proteins^a

TAV ORF	Nucleotide positions	Length (aa)	Potential site(s) and signature(s) (amino acid position within protein sequence)	Most-homologous protein(s) (accession no.); possible function	% I/% S
003R (E1A)	451-828	126	pRB-binding motif (62-66), zinc finger motif (95-115)	HadV-B E1A protein (X03000); exon 1, transcription activator	21.8/27.9
007R (E1A)	896-1003	36		HadV-B E1A protein (X03000); exon 2, transcription activator	21.8/27.9
008R (E1BS)	1008-1418	137		E1BS_ADEB3 (AF030154); small T antigen	22.2/40.1
009R (E1BL)	1361-2533	391		E1BL_ADEB3 (NC001876); large T antigen, transformation-associated protein	31.4/38.0
011R (IX)	2562-2975	138	Leucine zipper pattern (84-105), putative DNA gyrase/topoisomerase subunit A (90-124)	HEX9_ADEP5 (NP108656); virion morphogenesis-associated protein	32.6/33.3
014L (IVa2)	2978-4333	452	NTP-binding site (173-180)	PIV2_ADE17 (NP597866); transcriptional activator	62.2/26.5
016L (pol)	4078-6978	967	Conserved region I (777-782), two zinc finger motifs (3-27, 823-857)	DPOL_ADECC (NP044189); DNA pol	65.4/24.2
018R	5240-5479	80		Y115_ADE07 (P03288)	51.2/14.9
023R	7034-7321	96
024L (pTP)	7332-8996	555	Sequence for protein-primed DNA replication (57-63), arginine-rich region (264-300), bipartite NLS (290-300), glutamic acid-rich region (304-349), NSGD motif (481-484)	TERM_ADEP5 (NP108659); terminal protein	64.2/24.2
028R (p52K)	9154-10209	352	Putative hydroxylmethylglutaryl-coenzyme A reductase motif (83-89)	L52_ADE17 (NC001460)	62.0/24.9
031R (pIIIa)	10237-11856	540		HEX3_ADE17 (NC002067); peripentonal hexon-associated protein	56.8/28.3
033R (III)	11923-13350	476	Fiber-interacting domain (234-245)	PEN3_ADEP3 (AAC99444); penton base protein	73.1/19.0
037R (pVII)	13359-13751	131	Protease cleavage site (20-24), alanine-rich region (50-102), arginine-rich region (72-130), bipartite NLS (96-112)	VCO7_ADECU (NC001734); major core protein	37.5/36.8
040R (V)	13821-14912	364	Three bipartite NLSs (69-85, 314-330, 319-335), arginine-rich region (315-335)	VCOM_ADE02 (P03267); minor core protein	43.1/25.8
045R (pX)	14943-15146	68	Arginine-rich region (2-34), two protease cleavage sites (25-29, 35-40), two bipartite NLSs (18-35, 20-37)	L2MU_ADECC (Q65953); core protein	62.0/19.7
047R (pVI)	15202-15942	247	Proline-rich region (175-214), two endoprotease cleavage sites (30-37, 233-240)	PIV6_ADEP5 (NP108666); hexon-associated protein	47.0/31.6
050R (hexon)	15997-18729	911	Glycine-rich region (118-142), glucosaminoglycane attachment site (381-384), yeast PIR protein repeat (404-421)	HEX_ADEP5 (NP108667); hexon protein	77.5/16.3
060R (protease)	18735-19346	204	Peptidase C5 (1-133)	ADEN_ADECU (P35990), protease	66.2/25.1
062L (T-DBP)	19390-20508	373	Two zinc binding motifs (129-200, 240-309)	DNB2_ADEP5 (NP108669); DBP	24.4/36.5
067L (DBP)	20540-21940	467	Two bipartite NLSs (62-65, 103-107), two zinc binding motifs (228-299, 338-407)	DNB2_ADEB3 (AAD09730); DBP	48.5/33.5
075R (100K)	21951-24005	685	Carboxyltransferase domain (580-607), glucosaminoglycane attachment site (484-487)	L100_ADEP5 (NP108670); hexon assembly-associated protein	58.3/28.5
078L	23003-23326	108		YL13_ADE41 (P23690)	25.7/13.3
080R (33K)	23827-24237	137	Proline-rich region	33K_ADEB10 (AAL86535); virion morphogenesis-associated protein	25.2/19.9
082R (pVIII)	24487-25167	227	Two protease cleavage sites (108-115, 153-160)	HEX8_ADEP5 (NP108672); hexon-associated protein	68.7/18.5
084R	25133-25993	287
086R	25965-26150	62
087R (E3 14.5)	26143-26529	129		E314_ADE05 (P04493)	28.0/38.6
088L (U exon)	26537-26725	63		U exon BAdV3 (54)	27.0/39.7
090R (fiber)	26703-28595	631	NLS (2-4), penton-base-interacting sequence (8-15), repeat/shaft region (201-422), last repeat motif (434-442), TLWT motif (458-461), head domain (459-631)	FIBP_ADEB2 (AF308811); fiber protein	29.5/42.1
095L (E434)	28824-29555	244	Bipartite NLS (193-209)	E434_ADE40 (Q64865)	33.9/32.7
096L	29539-29805	89
099L (E411k)	29808-30149	114		E411_ADE05 (P04489); nuclear binding protein	29.7/43.2
101L	30149-30463	105
102L	30473-30766	98
103L	30760-31062	101
104L (dut)	31065-31511	149		DUT_RAT (P70583)	51.5/17.2
				DUT_ADEG1 (Q89662); DUT	42.7/19.5
105R	31938-33242	435	Prokaryotic membrane lipoprotein lipid attachment site (133-143)

^a% I/% S, percentage of amino acids identical (I) and similar (S) to the corresponding homologous proteins; HadV-B, human adenovirus B; ADEB3 (BadV3), bovine adenovirus type 3; ADEP5, porcine adenovirus type 5; ADE17, human adenovirus type 17; ADECC, canine adenovirus type 1; ADE07, human adenovirus type 7; ADEP3, porcine adenovirus type 3; ADECU, canine adenovirus type 1; ADE02, human adenovirus type 2; ADE41, human adenovirus type 41; ADEB10, bovine adenovirus type 10; ADEB2, bovine adenovirus type 2; ADE40, human adenovirus type 40; ADE05, human adenovirus type 5; ADEG1, fowl adenovirus type 1; NTP, nucleoside triphosphate. The homology values of the TAV E1A exons 003R and 007R are given for the complete corresponding protein of human adenovirus B (HAdV-B). The individual ORFs are numbered in the order in which they appear in the TAV genome, with the designation L for theoretical transcription to the left or R for theoretical transcription to the right. Missing numbers, represent TAV ORFs that were declared not to code for viral proteins and were therefore omitted from the table. Out of a total of 109 ORFs, 38 ORFs were selected to code for viral proteins according to the criteria described in Materials and Methods. The names written in parentheses in the first column correspond to the names of the corresponding homologous proteins of other adenoviruses. The sites and signatures found in the amino acid sequences of the potential TAV proteins were determined by the program PROSITE of the PC/GENE software package and the ExPASy Molecular Biology Server and by homology searches to known proteins of other adenoviruses. These sites and signatures are putative, and future experiments have to confirm their actual functionality.

Nucleotide sequence accession number.

The complete DNA sequence has been submitted to the GenBank database and given accession number AF258784.

RESULTS

Features of the TAV genome.

The nucleotide sequences of the partly overlapping endonuclease restriction fragments and PCR products displayed in Fig. 1A were determined by automated cycle sequencing and primer walking (65). The nucleotide sequences of all fragments were assembled to the complete DNA nucleotide sequence of the TAV genome according to the physical map shown in Fig. 1A, whose correctness was confirmed by amplification and sequencing of the genome regions of original TAV DNA around the endonuclease restriction sites. The TAV genome comprises 33,501 bp. Both DNA strands were sequenced independently until the identity of each nucleotide was undoubtedly determined. The average G+C content of the TAV genome was found to be 49.96%. As shown in Fig. 1B, the G+C distribution ranges between 40 and 60%, with slightly higher values in the center of the genome than at the genome termini.

FIG. 1.

(A) Physical map of the TAV genome consisting of BamHI, ClaI, EcoRI, and HindIII DNA fragments and schematic representation of the 12 cloned DNA fragments that were used for the determination of the complete nucleotide sequence of the TAV genome. The single fragments are named according to the restriction endonucleases that were used for their isolation. The numbers at both ends of each DNA fragment represent the nucleotide positions of the endonuclease restriction sites within the TAV genome. TAV-RE is a PCR product that was utilized for the determination of the nucleotide sequence of the right genome end. TAV-Tl and TAV-Tr were used by Brinckmann et al. for the determination of the genome termini (8). (B) Graphical presentation of the G+C content of the TAV genome

A characteristic feature of adenoviral genomes are inverted repetitive DNA sequences at the termini of the viral genome. The TAV genome possesses 166 bp inverted repetitive DNA sequences at the genome termini (8) that were verified in the context of this study.

Coding strategy of the TAV genome.

The analysis of the complete nucleotide sequence of the TAV genome resulted in the identification of 109 ORFs, with a possible coding capacity of 40 or more amino acid residues. The distribution of the ORFs on the two DNA strands is regular, with 55 ORFs oriented to the left and 54 ORFs oriented to the right. Thirty-eight ORFs, which are listed in Table 1, were selected to be actual coding sequences on the basis of a minimum length of 120 bp, the presence of transcription and translation signals, positional conservation, and homology to known genes. The remaining 71 ORFs, which were assumed not to encode viral proteins, overlap up to 100% with conserved ORFs. So far, they show no positional or sequence homologies on the DNA nucleotide or amino acid level to known genes and gene products of other adenoviruses.

Analyses of the amino acid sequences of the 38 potential TAV proteins revealed that 30 ORFs code for polypeptides with significant homologies to known proteins of other adenoviruses. These ORFs and their potential products were termed according to their homologous counterparts (Table 1).

The arrangement of the 38 ORFs encoding viral proteins is illustrated in Fig. 2. All ORFs with a significant homology to known genes of other adenoviruses are depicted, and those with no homologies to genes of other organisms are also represented. Conserved genes cover almost the complete nucleotide sequence of the TAV genome. The right genome terminus and the genomic region between the TAV ORFs pVIII and E3 14.5 seem to be locations for TAV-specific genes. Brinckmann et al. (7) have shown that the TAV E1A protein is encoded by two exons. Based on homology searches at the amino acid sequence level there is no evidence for further spliced genes in the TAV genome. Splice-donor and -acceptor site analysis at the nucleotide sequence level proposed the presence of further spliced TAV genes, as described in detail in Discussion; however, future studies have to confirm the actual presence of additional spliced genes in the TAV genome.

FIG. 2.

Potential coding strategy of the TAV genome. ORFs that code for proteins with significant homologies to known proteins of other adenoviruses are depicted with black arrows. ORFs with no homologies to known genes of other viruses and organisms but likely to code for a viral protein are pictured as white arrows. The orientation of each arrow corresponds to the direction of transcription. The conserved TAV ORFs are named according to known homologous adenoviral proteins. The two exons coding for the TAV E1A protein are linked by a thin line.

Conserved TAV proteins show the highest homologies to known proteins of various adenoviruses which are all members of the Mastadenovirus genus, with the exception of the TAV deoxyuridine 5′-triphosphate nucleotidohydrolase (DUT) homologue, which is most homologous to the corresponding proteins of avian adenoviruses (Table 1). No single Mastadenovirus species could be identified to be the most related to TAV. This is based on the fact that the highest homology values of the potential TAV proteins are evenly distributed among a large number of different adenoviruses. Figure 3A shows a graphic representation of 28 conserved putative TAV proteins sorted according to their highest homology values to human, porcine, bovine, canine, or avian adenoviruses.

FIG. 3.

(A) Graphic representation of 28 conserved putative TAV proteins arranged and classified according to their highest homology values to different adenoviruses. The TAV proteins are assigned to particular colors that represent particular adenovirus species. (B) Graphic representation of the homology value distribution of seven conserved TAV protein groups that are classified according to possible functions. Individual groups are marked by particular colors. The homology values that underlie this graphic are the highest detected homologies to known proteins regardless of the virus species the proteins belong to.

In addition, all of the conserved TAV ORFs are classified into seven groups according to their expected functional features and plotted according to their highest identity values to known adenoviral proteins (Fig. 3B). The penton (III) and hexon proteins were found to be the most-conserved polypeptides, with an identity of about 70 to 80% of the corresponding known adenoviral proteins. In contrast, the identity values at the amino acid level of TAV proteins encoded by genes in the E1, E3, and E4 genome regions do not exceed 40%.

The TAV genome harbors an ORF between nucleotides (nt) 31,065 and 31,511 coding for a potential protein that is most homologous to the DUT of avian adenoviruses. The highest identity value at the amino acid sequence level was 44.7% to the corresponding DUT of the fowl adenovirus type 8 (51).

Another interesting feature of the TAV genome is the apparent partial duplication of the viral E2A DNA-binding protein (DBP) gene locus. The DBP gene loci TAV T-DBP and TAV DBP are most homologous to the E2A DBPs of porcine adenovirus type 5 (48) and bovine adenovirus type 3 (54), respectively, and to each other. The alignment of the amino acid sequences of TAV T-DBP and TAV DBP revealed a significant homology of 56.1% between these two proteins. T-DBP possesses extensive deletions in the N-terminal region and is accordingly designated truncated DBP (T-DBP).

Gene arrangements of the TAV genome.

Thirty-eight TAV ORFs were predicted to code for viral proteins. The arrangement of these 38 ORFs in the TAV genome is given in Fig. 4 together with the corresponding gene arrangements of representative members of the four adenovirus genera Mastadenovirus, Aviadenovirus, Atadenovirus, and Siadenovirus. The viral genes that are conserved in all adenovirus genera are indicated in Fig. 4. Those gene loci that are only present in the genomes of the members of single genera are also labeled. The graphical representation shows that the gene content and arrangement of the TAV genome corresponds to that of members of the Mastadenovirus genus. The V and pIX ORFs, as well as the E1, E3, and E4 regions, are characteristic for mastadenoviruses and are positionally perfectly conserved in the TAV genome. The TAV ORFs 23R and T-DBP and some ORFs within the E3 and E4 regions, including 84R, 86R, 96L, 101L, 102L, and 103L, seem to be unique for TAV.

FIG. 4.

Comparison of the gene content and genome organization of TAV with that of representatives of the Mastadenovirus, Aviadenovirus, Atadenovirus, and Siadenovirus genera of the family Adenoviridae. Potential protein-coding regions are depicted as arrows whose orientations correspond to the direction of transcription. Rectangles represent protein-coding exons. ORFs that are conserved in the genomes of all adenoviruses are illustrated in white; ORFs that are not conserved in all adenoviruses are drawn in black. Each genome is marked at 10-kbp intervals, and the inverted terminal repeats are drawn as black rectangles at the genome ends. The graphical presentation is in agreement with the data reported by Davison et al. (18). HadV-B, human adenovirus B; FadV-1, fowl adenovirus type 1; OadV-287, ovine adenovirus 287; FrAdV-1, frog adenovirus type 1.

CG dinucleotide distribution.

The complete genomes of TAV, fowl adenovirus A (11), human adenovirus A (62), canine adenovirus type 1 (46), and frog adenovirus (18) were analyzed by the program CPGPLOT (37) for the distribution of CG dinucleotides. The extensive CG depletion known to exist in the genomes of some adenoviruses (18) such as canine adenovirus type 1 and frog adenovirus and to a lesser extent in human adenoviruses could not be detected in the TAV genome. The largest part of the TAV genome contains more CG dinucleotides than expected from the overall G+C content, as also known for the genomes of avian adenoviruses. In contrast, it was found that the terminally located TAV E1 and E4 genome regions show a significant CG depletion.

Phylogenetic position of TAV.

Phylogenetic trees derived from the comparison of the penton protein, hexon protein, protease, terminal protein, protein VIII, DNA polymerase, protein IVa2, and 100,000-molecular-weight protein (100K protein) amino acid sequences of different adenoviruses are shown in Fig. 5. The phylogenetic trees show a subdivision into four main branches, representing the four adenovirus genera, Mastadenovirus, Aviadenovirus, Atadenovirus, and Siadenovirus. In all eight trees, TAV is unambiguously a member of the Mastadenovirus genus. In view of the eight phylogenetic trees, TAV seems to have similar phylogenetic distances to human, bovine, porcine, canine, and ovine adenoviruses.

FIG. 5.

Eight phylogenetic trees derived by comparison of the penton protein (A), hexon protein (B), protease (C), terminal protein (D), protein VIII (E), DNA polymerase (F), IVa2 protein (G), and 100K protein (H) amino acid sequences of different adenovirus species. The sequences of the individual proteins used to construct the phylogenetic trees were taken from the current GenBank and SWISSPROT databases. The lengths of the branches indicate the phylogenetic distances between the viruses. The scale bar represents 10 mutations per 100 sequence positions.

DISCUSSION

The primary structure of the TAV genome (33,501 bp) and its coding capacity were elucidated. The G+C content of the TAV genome was found to be 49.96%, which is in agreement with the average value known for the members of the Mastadenovirus genus.

An interesting phenomenon of eukaryotic genomes is the so-called CG depletion. It is assumed that the methylation of cytosine in CG dinucleotides followed by deamination to TG and fixation of the mutation during DNA replication is responsible for the diminished frequency of CG dinucleotides in eukaryotic genomes. This phenomenon of CG depletion was also identified in the genomes of several viruses, including the canine adenovirus types 1 and 2, ovine adenovirus 287, turkey adenovirus type 3, and frog adenovirus (18). The members of the proposed genus Siadenovirus, including frog adenovirus, show a very uniform CG depletion in their genomes. Most of the members of the genus Mastadenovirus, including human adenoviruses, show only a slight CG depletion, whereas the CG dinucleotide frequencies of the canine adenovirus types 1 and 2 correspond to that of siadenoviruses (18). Genomes of aviadenoviruses show no CG depletion at all. The CG dinucleotide distribution of the TAV genome is characterized by higher CG values than expected in the center and a significant CG depletion at the genome termini, where the E1 and the E4 regions are located, suggesting a strong methylation of these regions during the TAV replication cycle. However, the molecular events and underlying mechanisms of CG depletion in viral genomes, including that of adenoviruses, is still not understood in detail.

Thirty-eight out of 109 TAV ORFs were predicted to code for viral proteins. The putative translation products of 30 ORFs were found to have significant homology to known proteins of other adenoviruses, with the highest values found to belong to members of the Mastadenovirus genus. The genomes of mastadenoviruses comprise four early regions (E1 to E4) that are formed by genes that encode translation products with essential roles in the expression of other virus genes, replication of viral DNA, transformation of host cells, and influencing of the host immune response. Analysis of the TAV genome resulted in the identification of the TAV E1A, E1BS, and E1BL homologues in the 5′ terminal region of the viral genome. Characteristic sites of adenoviral E1A proteins are the retinoblastoma-susceptibility protein-binding motif (LXCXE) (19) and a zinc finger motif (CX₂CX₁₃CX₂C) (15). Both domains were also found in TAV E1A at amino acid positions (aa) 62 to 66 (LEEVE) and 95 to 115, respectively.

The precise composition of the TAV E3 region, whose members encode proteins that are involved in counteracting the host immune response, is questionable. TAV E3 14.5 is most homologous to the early E3 14.5-kDa protein of human adenovirus type 5, but it is not known whether the two adjacent and unique TAV ORFs, 84R and 86R, are also part of the E3 region. Furthermore, it is proposed that the TAV ORFs E434, 96L, E411k, 101L, 102L, 103L, DUT, and 105R form the E4 region.

The DNA polymerase (pol), the terminal protein (pTP) precursor, and the DBP genes are components of the E2 region, and the pertinent proteins are necessary for viral DNA replication. The TAV pol ORF encodes a putative protein of 967 aa. The conserved region I was found between aa 777 and 782, and two possible zinc finger motifs (CX₂CX₉CX₉HH and CX₃CX₂CX₂₃CX₂C) were identified at aa 3 to 27 and 823 to 857, respectively (48, 53). The pol genes of many adenoviruses are composed of two exons; however, thorough homology searches and analysis of predicted splice sites have led to the conclusion that the TAV pol gene is probably not spliced.

The TAV pTP ORF encodes the TAV homologue of the adenoviral pTP precursor between nt positions 8996 and 7332. A predicted splice-acceptor site is located at nt positions 9149 to 9109. Future experiments have to confirm whether or not the TAV pTP gene is composed of different exons. All known pTPs have the sequence motif YSRLRYT that plays an important role in protein-primed DNA replication initiation (32). In TAV pTP, the corresponding well-conserved sequence YSRLKYH was found at aa 57 to 63. Furthermore, the nuclear localization signal (NLS) RLPIRRRQRRR between aa 290 and 300 (48, 53) and the NSGD motif (aa 481 to 484), which is involved in the initiation of DNA replication (60), are also well conserved in TAV pTP.

The presence of the TAV ORFs T-DBP (truncated DBP) and DBP, which are both homologous to the E2A DBP of other adenoviruses, is a unique feature of TAV, so far. In regard of the tandem-like arrangement of the two TAV ORFs, it is conceivable that the T-DBP ORF is the result of a duplication event of the DBP ORF followed by deletion of large parts of the 5′ coding region of the T-DBP gene. The TAV DBP contains two conserved NLSs in the N terminus at aa 62 to 65 (RRKL) and 103 to 107 (KKRRK). However, both signals were absent from the N terminus of TAV T-DBP. Furthermore, there are two conserved zinc-binding motifs that are characteristic of adenoviral DBPs (53). Both motifs were also found in TAV DBP at aa 228 to 299 (CXHX₅₂CX₁₅C) and 338 to 407 (CXCX₅₁CX₁₄C) and in TAV T-DBP at aa 129 to 200 (CXHX₅₂CX₁₅C) and 240 to 309 (CXCX₅₁CX₁₄C), respectively. A noteworthy finding in this context is that the initiation codons of TAV T-DBP and TAV DBP are both located within predicted splice-acceptor-site motifs, raising the possibility that an as-yet-unknown exon could encode the real N terminus of at least one or both proteins and could provide the two missing NLSs for TAV T-DBP.

The two genes that encode the IX and the IVa2 proteins in human adenoviruses are designated intermediate genes (59). Protein IX is the so-called minor capsid component required for packaging of the viral DNA (23) and involved in activating the MLP (42). The IVa2 protein was also found to enhance the activity of the MLP (67). Homologues of the IX and IVa2 genes were also found in the TAV genome. In addition, the potential NTP-binding site GPTGCGKS (24), which is characteristic of adenoviral IVa2 proteins, was also identified in TAV IVa2 at aa 173 to 180.

The late regions of adenoviral genomes comprise mainly structural protein genes whose transcription categorically starts from the MLP followed by processing of the transcript into several late mRNAs (59). The nucleotide sequence and composition of the TAV MLP was analyzed by Song et al. (61) and verified during this study by nucleotide sequencing and homology searches to known MLPs of other adenoviruses. The TAV MLP is predicted to cover nt positions 4846 to 5040. An inverted CAAT box and the canonical TATA box were identified at nt 4853 to 4857 and nt 4891 to 4897, respectively. A conserved upstream promoter element (58), which is involved in binding of the gene-specific transcription factor USF, was found between nt positions 4873 and 4878. The positions of two downstream activating elements called DE1 and DE2 (39) were predicted between nt 5006 and 5016 and nt 5021 and 5036, respectively. The initiator element described by Lu et al. (41) could not be identified by sequence comparison analysis within the TAV MLP.

The genes coding for p52K, pIIIa, III, pVII, V, pX, VI, hexon, protease, 100K protein, 33K, pVIII, and fiber are main constituents of the late genome regions of mastadenoviruses and are also present in the TAV genome. The TAV III gene, which encodes the penton base protein, contains none of the known motifs that are involved in the attachment of the virus to cellular receptors (34, 74). In contrast, the highly conserved fiber-interacting domain of penton base proteins (10) was found in TAV III between aa 234 and 245 (SRLNNLLGIRKR).

TAV pVII, a putative core protein, holds a highly conserved cleavage site for the viral protease between aa 20 and 24 (MYGGA) that was also found in pVII proteins of other adenoviruses (for examples, see references 48 and 53). Two additional potential cleavage sites are proposed for TAV pX, a further putative core protein, at aa 25 to 29 (LSGGR) and 35 to 40 (LKGGF) (72). Porcine adenovirus type 5 pVIII protein harbors two further viral protease cleavage sites (48). Both sites were identified within TAV pVIII at aa 108 to 115 (LAGGGHTS) and 153 to 160 (LAGGSRSS). Moreover, two amino acid sequence motifs in TAV pVI at aa 30 to 37 (MNGGAFSW) and 233 to 240 (IVGVGVRP) were identified that also corresponded well to consensus protease cleavage site sequences (57). Human adenovirus type 2 pVI protein includes a peptide (GVQSLKRRRCF) in the C terminus that is required by the adenoviral protease as a cofactor for its activity (43, 73). This peptide is well conserved between aa 237 and 247 (GVRPIKRRRCY) of TAV pVI.

The adenoviral fiber protein contains an NLS (30) and a penton-base-interacting sequence (10) in its N terminus. The amino acid sequence KRR (aa 2 to 4) in the putative TAV fiber protein could be an NLS, but its function is not yet verified. The penton-base-interacting sequence INPVYPYG at aa 8 to 15 is well conserved. In addition, a repeat shaft region between aa 201 and 422 (25) and a head domain between aa 459 and 631 were identified. The last complete repeat motif (KLGXGLXFD/N) (13) before the head domain was well conserved at aa 434 to 442 (KLGTGLSFG) in the TAV fiber protein. The TLWT motif, which indicates the N terminus of the fiber head (13), was also present in the TAV fiber protein at aa 458 to 461.

The TAV ORF 080R was designated 33K because of its homology to the 33K protein of bovine adenovirus type 10. Analysis of the TAV 33K ORF revealed a predicted splice-donor site in the 3′ terminus, raising the possibility that this genome region codes for at least two proteins with identical N termini.

The U exon was first identified during analysis of the human adenovirus type 40 genome (17), and it is present in almost all adenoviruses examined so far. The U exon is a putative coding region starting with an initiation codon and ending with a splice-donor site, and it is located upstream of the fiber gene with a transcription direction opposite that of the latter one. The U exon is proposed to encode the N terminus of an unknown viral protein; however, up to now, downstream exons are not identified in any adenovirus genome. The TAV ORF 088L product showed a significant amino acid homology of 27% identity and 39.7% similarity with the U exon protein of the bovine adenovirus type 3 (54).

The TAV ORF 104L in the E4 genome region that is homologous to DUT deserves special attention. Corresponding genes are also present in the left-hand genome regions of aviadenoviruses and in the E4 regions of some mastadenoviruses. Surprisingly, TAV DUT is most homologous and most related to DUTs of avian adenoviruses. The origin of the TAV DUT ORF is unclear. It is conceivable that an ancestral adenovirus has captured a DUT gene from its host. However, it is not known whether this virus was a common ancestor of avian and mastadenoviruses or whether this special gene capture has only taken place in the evolutionary lineage of TAV.

In some adenovirus genomes, low-molecular-mass virus-associated RNAs (VA RNAs) are encoded, transcribed by RNA polymerase III, and needed for down-regulation of interferon-mediated antiviral response (22) and translation of late viral mRNA (38). On the basis of known VA RNA nucleotide sequences of avian and human adenoviruses it was not possible to identify analogous VA RNAs in the TAV genome by homology analysis with the programs BLAST (Basic Local Alignment Search Tool) (1) and NALIGN of PC/GENE, release 6.85.

Gene content and arrangement comparisons revealed that the TAV genome holds 17 ORFs, mainly involved in DNA replication and virion formation, that are conserved in the genomes of all adenoviruses examined so far and that are therefore assumed to already be present in the genome of a common ancestor (18). The gene arrangement of TAV is almost identical to that of mastadenoviruses. ORFs unique to TAV are TAV T-DBP, the two ORFs 084R and 086R, which are located in the E3 region, and ORFs 096L, 101L, 102L, 103L, and 105R in the E4 genome region, underlining the phenomenon of evolution of lineage- and species-specific genes at the genome termini and in the E3 region of mastadenoviruses. Phylogenetic analysis of eight conserved adenoviral proteins confirmed the classification of TAV as a member of the Mastadenovirus genus. However, it was not possible to declare one adenovirus species to be the most related to TAV, which seems to have similar phylogenetic distances to human, bovine, canine, ovine, and porcine adenoviruses. One can assume that TAV claims a special evolutionary branch within the Mastadenovirus genus. This is in agreement with the taxonomic position of its natural host, whose phylogenetic lineage is also supposed to form a separate branch of the evolutionary tree of mammals, suggesting a coevolutionary development of TAV and Tupaias.