Deep RNA Sequencing Reveals Complex Transcriptional Landscape of a Bat Adenovirus

Abstract

Bat adenoviruses are a group of recently identified adenoviruses (AdVs) which are highly prevalent in bats yet share low similarity to known AdVs from other species. In this study, deep RNA sequencing was used to analyze the transcriptome at five time points following the infection of a bat AdV in a kidney cell line derived from a myotis bat species. Evidence of AdV replication was observed with the proportion of viral RNAs ranging from 0.01% at 6 h to 1.3% at 18 h. Further analysis of viral temporal gene expression revealed three replication stages, the early-stage genes encoding mainly host interaction proteins, the intermediate-stage genes for the DNA replication and assembly proteins, and the late-stage genes for most structural proteins. Several bat AdV genes were expressed at stages that differed from those of their counterpart genes previously reported for human AdV type 2. In addition, single-base resolution splice sites of several genes and promoter regions of all 30 viral genes were fully determined. Simultaneously, the temporal cellular gene expression profiles were identified. The most overrepresented functional categories of the differentially expressed genes were related to cellular immune response, transcription, translation, and DNA replication and repair. Taken together, the deep RNA sequencing provided a global, transcriptional profile of the novel bat AdV and the virus-host interactions which will be useful for the understanding and investigation of AdV replication, pathogenesis, and specific virus-bat interactions in future research.

INTRODUCTION

Adenoviruses (AdVs) are double-stranded DNA viruses with a nonenveloped icosahedral capsid, 60 to 100 nm in size, and a genome of 26 to 45 kb. They have a wide host range among vertebrates, including human and nonhuman primates (1–4). Human AdVs are important not only because of their high prevalence and their ability to cause clinical illnesses ranging from respiratory disease to gastroenteritis but also for their use as vectors for human gene therapy and vaccination (5–7). However, recent studies have shown that the presence of preexisting human AdV antibodies may hamper the clinical usage of the known human AdVs as effective gene transfer vectors (8, 9). Alternative AdVs distantly related to human AdVs may provide potential new candidates in this direction.

The human AdV replication cycle can be divided into three phases, early, intermediate, and late (10). The early phase begins immediately after virus infection of host cells and includes adsorption, penetration, transport of viral DNA through the nuclear pore complex into the nucleus, and expression of a set of early genes. The early genes are expressed from 1 h postinfection (p.i.), and their products mediate viral gene expression and DNA replication. Then two viral genes, IVa-2 and IX, are expressed from 6 h p.i. and represent an intermediate stage. After this, viral DNA replication starts and the late genes begin to express as well. The total infection cycle is completed 24 to 36 h after infection of host cells.

Bat AdVs (BtAdVs) are a group of AdVs discovered in bats (2, 11–13) that show low sequence similarity to known AdVs from other species, with only 23 to 71% amino acid identity to human AdV proteins (2). The low similarity of the BtAdVs highlights their potential as gene delivery vectors that are less likely to induce an immune response and less susceptible to neutralization in humans compared to human AdVs (14, 15). A better understanding of AdV infection may be helpful for understanding the interaction between the virus and the host and its future usage as a gene delivery vector. In our previous study, a novel BtAdV, BtAdV-TJM, was isolated from Myotis davidii (2). In this study, we analyzed the BtAdV-TJM transcriptome in a kidney cell line (BK) derived from M. davidii using a next-generation sequencing technique and provided the first genome-wide profile of BtAdV transcription in its host across 5 time points. Unique transcriptional features of BtAdV were revealed and confirmed by quantitative PCR and compared with those of human AdV-infected cells.

Although transcriptome data have been described for the megabat Pteropus alecto (16), this is the first bat transcriptome data set on virus-infected cells and the first from any microbat. This data set also represents the first bat transcriptome data set that describes both viral and host genes, providing information on host-virus interactions.

MATERIALS AND METHODS

Cell culture and viral infection.

Ten million M. davidii BK cells were cultured in a 3.3-cm well plate for 18 h at 37°C and maintained in RPMI 1640 medium supplemented with 15% fetal bovine serum (FBS). Cells were infected with bat adenovirus-TJM (BtAdV-TJM) at a multiplicity of infection (MOI) of 10 and incubated for 1 h at 37°C. The infected cells were then cultured at 37°C with fresh RPMI 1640 medium supplemented with 5% FBS. The RNA was extracted from the infected cells at 0, 6, 8, 12, and 18 h p.i. Experiments have been performed in triplicate.

RNA and cDNA preparation for RNA sequencing.

RNA was extracted using the RNeasy minikit according to the manufacturer's protocol (Qiagen). The mRNA was fragmented by addition of 5× fragmentation buffer after enrichment using the oligo(dT) magnetic beads. First-strand cDNA was synthesized using random hexamer primer and reverse transcriptase (Invitrogen). The second-strand cDNA was synthesized using RNase H (Invitrogen) and DNA polymerase I (New England BioLabs). The double-stranded cDNA was purified by a QIAquick PCR extraction kit (Qiagen), and end repair and adenine addition were performed before the ligation of sequencing adaptors. The required fragments were purified by agarose gel electrophoresis and enriched by PCR amplification. The library products were sequenced via Illumina HiSeq 2000.

Data collection and mapping of reads.

The above-mentioned libraries were sequenced as 49-mers using the standard Illumina GA pipeline (version 1.6). Sequencing adaptors, reads containing greater than 10% unknown bases, and low-quality reads were removed. Clean reads were mapped to both the BtAdV-TJM reference genome and the M. davidii reference genome (G. Zhang, S. Cowled, Z. Shi, Z. Huang, K. Bishop-Lilly, X. Fang, J. Wynne, Z. Xiong, M. Baker, W. Zhao, M. Tachedjian, Y. Zhu, P. Zhou, X. Jiang, J. Ng, L. Yang, L. Wu, J. Xiao, Y. Feng, Y. Chen, X. Sun, D. Fan, B. Broder, K. Frey, L.-F. Lin-Fa Wang, and J. Wang, personal communication) using SOAPaligner/soap2, with no more than three nucleotide mismatches allowed (17). To identify the different transcriptional patterns during BtAdV infection, we clustered gene transcription profiles using the reads per kb per million reads (RPKM) method, which was normally used for estimation of gene expression (18). Clean reads were used to calculate the RPKM value after subtraction of the very low background at time zero for each gene.

Promoter prediction and splice site verification.

To identify potential promoter regions, the 150 nucleotides (nt) upstream of the start codon of the 30 BtAdV genes was interrogated using the motif discovery tool MEME (19). To verify BtAdV splice sites predicted by the bioinformatics methods, PCR was performed using the cDNAs transcribed from the infected cells at 18 h p.i. with different sets of primers located within predicted exons of E1A, IVa-2, pol, E4 6/7, pTP, and 33K. PCR products shorter than those expected from unspliced mRNA were exacted from agarose gel and sequenced with the specific primers (Table 1).

Table 1.

Primers used in this study

Gene	Primera	Sequence (5′–3′)	Application
E1A	E1A-1F	TACCGGCTGGTCTGATAATGTGG	RT-PCR
	E1A-1R	GTCATGGCCAAACCACTTCCTT	RT-PCR
	E1A-2F	CGGATGCCATTGGGAGAGT	qRT-PCR
	E1A-2R	CATGGCCAAACCACTTCCTT	qRT-PCR
IVa-2	IVa-2-1F	ACTCGTCCATGACAATGGCAAT	RT-PCR
	IVa-2-1R	TACAGGCTCATGGAGGAAAAAG	RT-PCR
	IVa-2-2F	GCCCCGAAAACACAGTCATC	qRT-PCR
	IVa-2-2R	ACCTCATTGGTCAGCTCAGCAT	qRT-PCR
33K	33K-1F	ATGGAGGACGAAACGGGCAGT	RT-PCR
	33K-1R	TTTTAGCAGAGAGCGGAGGGAG	RT-PCR
	33K-2F	CCCCCACCGCCTTTAAGATA	qRT-PCR
	33K-2R	CTGCGTCTCGGACTTGTGGTA	qRT-PCR
pTP	pTP-1F	AGTCTGAAACTCGGCGAGCCCG	RT-PCR
	pTP-1R	GAAACGCGGGCGTCATGGCTCT	RT-PCR
	pTP-2F	GGGATGGGCGAGTTGGA	qRT-PCR
	pTP-2F	CGTCCCCGCTGTTGTCAT	qRT-PCR
pol	pol-1F	GTACGGCGTGAGGGGGTTTTCA	RT-PCR
	Pol-1R	GAAACGCGGGCGTCATGGCTCT	RT-PCR
E4 6/7	E4 6/7-1F	CAATTTTATTGAACACACGTGTG	RT-PCR
	E4 6/7-1R	TGGACACCCCTCTCACGTGA	RT-PCR
100K	100K-1F	CGCCTTCCCTACCGACTTC	qRT-PCR
	100K-1R	GGCCAATCTGAGCAGGTAGGT	qRT-PCR
GAPDH	GAPDH-1F	TGACCCCTTCATTGACCTCAAC	qRT-PCR
	GAPDH-1R	TGACCGTGCCTTTGAACTTG	qRT-PCR

^aF, forward; R, reverse.

qRT-PCR.

To validate the results obtained by the RNA-seq analysis, quantitative real-time PCR (qRT-PCR) was performed on the same sets of RNA samples which had been used in the RNA-seq experiments. qRT-PCR primers of target genes were designed using Primer Express 3.0 (Applied Biosystems) (Table 1). In brief, 1 μg total RNA from each time point was reverse transcribed using the QuantiTect reverse transcription kit (Qiagen) and used for qPCR. Reactions were carried out using the SYBR GreenER qPCR Supermix Universal (Invitrogen) and an Applied Biosystems StepOne real-time qPCR instrument. For each reaction, 2 μl of 1:5 diluted cDNA and a final concentration of 200 nmol of each primer were used. The qPCR was performed using the following program: 90°C for 1 min, 40 cycles of 90°C for 15 s and 60°C for 1 min, and melting curve and plate read stages under default settings. Relative expression levels of the target genes were calculated using the comparative cycle threshold (C_T) method. The 0-h RNA was used as the reference sample. All data were normalized relative to the housekeeping gene GAPDH (glyceraldehyde-3-phosphate dehydrogenase).

RNA-seq data accession numbers.

The RNA-seq data obtained in this study have been deposited in the NCBI Gene Expression Omnibus (GEO) under the accession number GSE41275.

RESULTS

Viral mRNA expression in BtAdV-infected BK cells.

M. davidii BK cells were infected with BtAdV at an MOI of 10. Total RNA was isolated at 0, 6, 8, 12, and 18 h p.i., and cDNAs were prepared and sequenced. About 7 million reads were obtained at each time point, of which approximately 80% were able to be aligned to the viral and bat genomes. Following infection, there was a rapid increase in proportion of viral mRNA from 0.01% at 6 h, 0.04% at 8 h, and 0.27% at 12 h to 1.3% at 18 h and a gradual decrease in the cellular mRNA. At 18 h p.i., the viral reads mapped to 95% of the nucleotides of both strands of the BtAdV genome, demonstrating the rapid replication of BtAdV in BK cells. From the single-base resolution maps of the BtAdV transcriptome visualized with MochiView (20), the number of reads was found to be unevenly distributed among the viral genes, reflecting huge differences in gene expression levels (Fig. 1). All the previously predicted BtAdV protein-coding genes were covered by at least one read overlapping the open reading frame (ORF). The transcripts tended to occur in clusters on both strands, with slightly higher expression in the negative strand at 6 h and 8 h and more reads in the positive strand at 18 h.

Fig 1.

BtAdV genome-wide transcriptome maps during the 18 h of infection. Numbers of sequence reads per nucleotide are displayed over the map of the entire BtAdV genome. The counts above the line map to the upper (rightward) DNA strand and counts below the line map to the lower (leftward) DNA strand. The genome map of the BtAdV is shown at the bottom for reference purposes.

Cluster analysis of BtAdV gene expression.

For determining the transcriptional profile of a single gene and the details of BtAdV transcription dynamics, the RPKM clustering analysis was performed in an unbiased manner. As shown in Fig. 2A, the resulting BtAdV gene clusters were divided into three main classes. The kinetics of gene expression within the three clusters are represented in line graphs shown in Fig. 2B. Cluster 1 included 15 genes that were expressed immediately after the infection, 9 of which were remarkable as their expression peaked at 6 h p.i. before declining. The expression of the other 6 genes in this cluster continued to increase from 0 h to 18 h p.i. Most of the cluster 1 genes were located at the right terminus of the upper strand and left terminus of the lower strand of the genome. The cluster 2 genes were expressed from 6 h p.i., their expression levels peaked at 8 h p.i., and their expression continued to increase through to 18 h p.i. This cluster included the transcripts of pTP, U exon, 22K, 33K, and IX genes located proximal to the terminus of the upper and lower strands of the genome. Cluster 3 genes were expressed from 8 h and 12 h p.i., with 6 genes expressed from 8 h p.i. and 4 expressed at 12 h p.i., and their expression levels peaked at 18 h p.i. The genes of this cluster corresponded mainly in the central region of the genome. Most of the BtAdV gene expression profiles were consistent with that of human AdV type 2, except a few genes which showed slightly different expression patterns. For example, the human AdV type 2 IVa-2 gene is expressed from the intermediate stage, while BtAdV IVa-2 was expressed at the early stage of infection (21). The structural protein genes 22K, 33K, and U exon were grouped into the late genes in the human AdV type 2 but grouped into the intermediate stage in the BtAdV. The pTP gene that acts as a protein primer is expressed at the early stage in the human AdV type 2 yet was expressed only at the intermediate stage in the BtAdV (Fig. 2B) (22). In total, the gene expression profiles of the BtAdV were similar to those of human AdV type 2. When clustering the BtAdV gene expression files by their function and compared with human AdV (Fig. 2C), the genes in cluster 1 corresponded to the human AdV type 2 homologs that were involved in the induction of the host cell entering the S phase or setting up an anti-host immune system. Genes in cluster 2 corresponded to the human AdV type 2 homologs that were involved in viral transcription and replication. Genes in cluster 3 corresponded to the human AdV type 2 homologs that were involved in the viral structural proteins and participate in virus assembly (23).

Fig 2.

Cluster analysis of the temporal expression of BtAdV mRNAs. (A) Heat map representation of the normalized read counts of BtAdV ORFs from 0 to 18 h. Colors from green to red indicate increased numbers of the normalized read counts of each ORF. (B) Normalized gene expression profiles are plotted and represent three main replication stages, genes belonging to (from top to bottom) the early stage, the intermediate stage, and the late stage. (C) Assigned functions of genes in the different expression classes.

Analysis of BtAdV splice sites.

RNA splicing is one characteristic feature of adenovirus. A series of alternative splice sites have been found in mastadenoviruses (24–26). Bioinformatics analysis has provided evidence for BtAdV mRNA splicing in E1A, E4 ORF6/7, 33K, IVa-2, pol, and pTP genes (2). In this study, only an E1A splicing site was detected by RNA-seq and RT-PCR. The splicing sites of the other 5 genes were detected by RT-PCR but not by RNA-seq due to the overlapping transcripts. Minor differences were detected at the splicing sites of E1A and 33K by RNA-seq and RT-PCR (Table 2). The accuracy of the E1A gene splice site prediction by RNA-seq is shown in Fig. 3.

Table 2.

BtAdV splice sites were verified by RT-PCR

Gene	Bioinformatics prediction (nt ranges)	Validation by RT-PCR (nt ranges)
E1A	391–883, 970–1130	391–883, 959–1130
IVA-2	3421–4736, 5036–5048	3421–4736, 5036–5048
pol	4521–7940, 12472–12480	4521–7940, 12472–12480
E4 6/7	28325–28567, 29303–29350	28325–28567, 29303–29350
pTP	7796–9616, 12472–12480	7796–9616, 12472–12480
33K	23676–23821, 23934–24282	23676–23820, 23990–24282

Fig 3.

Single-base resolution plot of the splice site of the E1A gene. The x axis represents the BtAdV E1A gene, and the y axis represents the number of reads mapping to the genome. Asterisks indicate the E1A splice donor and acceptor sites predicted by bioinformatics.

BtAdV promoter prediction.

The single-nucleotide resolution of the 5′ end of viral transcripts allowed us to analyze the promoters of gene transcription of different clusters. Here we used the motif discovery program MEME and analyzed the 150 bp upstream of the BtAdV gene start codons. As summarized in Table 3, the program detected an 8-nt consensus sequence which is an important promoter signature and required for efficient transcription of AdV type 2 (27).

Table 3.

Core promoter sequences of BtAdV genes

Gene	Temporal expressiona	Strandb	Start site (nt)c	Predicted promoter core sequenced	P valuee
E1A	E	−	34	AAAGAGAA	5.08E−05
E4ORF6/7	E	+	46	AAACAAAA	2.20E−05
E4ORFC	E	+	134	AAAGAAAA	7.81E−06
E4ORFD	E	+	9	AAAGACAA	1.88E−04
E4ORFA	E	+	42	ATAAACAA	1.26E−03
E4ORFB	E	+	18	AAAAAAAA	1.41E−05
E4 34K	E	+	46	AAACAAAA	2.20E−05
E1Bl	E	−	80	AAAGATAA	1.70E−04
E1Bs	E	−	62	ATACAAAA	1.78E−04
E3	E	−	20	AGAGAAAA	4.60E−04
DBP	E	−	9	AAAGAGGA	1.56E−04
pol	E	+	8	ACAGAAAA	2.44E−04
IVa2	E	−	68	AAAGAGAA	5.08E−05
pVIII	E	+	141	ATAAAAAA	1.70E−04
100K	I	−	92	AAGGAAGA	1.56E−04
pTP	I	+	10	AAAAAGAA	7.43E−05
U exon	I	+	92	AAGGATAA	4.60E−04
22K	I	−	102	AAGCAAAA	1.11E−04
33K	I	−	102	AAGCAAAA	1.11E−04
IX	I	+	83	GAAGAGGA	8.07E−04
52K	L	+	148	GAGGAAGA	8.07E−04
pIIIa	L	+	133	GAGGAGGA	1.48E−03
Penton	L	+	100	GTAGAGGA	2.07E−03
pV	L	+	78	AAAAAGAA	7.43E−05
pVI	L	−	28	AAGAAAAA	7.43E−05
Hexon	L	−	28	AAAAAAGA	7.43E−05
Protease	L	−	118	AAACAAAA	2.20E−05
Fiber	L	−	138	ACGCAAGA	1.78E−03
pVII	L	−	123	AAAGAAGA	5.08E−05
pX	L	−	17	AAGCAAAA	1.11E−04

^aE, early; I, intermediate; L, late.

^b+, positive strand; −, negative strand.

^cNucleotide position upstream of the start codon.

^dConsensus 8-nt sequence derived by the motif discovery program MEME.

^eThe P value of a site was computed from the match score with the position-specific scoring matrix for the motif.

Validation of BtAdV gene transcription by qPCR.

The same RNA preparation as that for RNA-seq analyses was used for qPCR. The transcriptional activity was normalized to the activity of the housekeeping gene GAPDH, whose expression was not noticeably affected by virus infection. Three early genes, E1A, IVa-2, and 100K, two intermediate genes, pTP and 33K, and two late genes, hexon and penton, were tested by qPCR. The relative expression changes were expressed as log₂ of the C_T value. As shown in Fig. 4, for the seven genes tested, qPCR data were mostly consistent with those of RNA-seq. This suggests that RNA-seq can be used as a quantitative measurement of gene expression and that the RPKM values derived from each gene here are a good representation of relative gene expression levels.

Fig 4.

Comparison of mRNA expression levels detected by RNA-seq and qPCR. Graphs show the correlation of two methods for the seven viral genes, the early genes E1A, IVA-2, and 100K, the intermediate genes pTP and 33K, and the late genes hexon and penton. Solid squares indicate mRNA expression levels of viral genes determined by RNA-seq analysis, while the solid triangles indicate the relative expression levels of BtAdV genes as measured by qRT-PCR. All data generated by qRT-PCR were normalized against the housekeeping gene GAPDH. Data shown are the means of three replicates, and the error bars indicate standard deviations (SDs).

Cellular transcriptome profile following BtAdV infection.

The differentially expressed cellular genes identified by the RNA-seq analysis in BtAdV-infected cells were annotated based on the KEGG prediction and were classified into 10 defined categories (Fig. 5A). Among them, 1,969 bat genes were identified to be differentially expressed by more than 2-fold in BtAdV-infected cells compared to values in uninfected cells (Fig. 5B). The number of differentially expressed genes during the progression of the BtAdV infection is shown in Fig. 5C. There were 386 and 204 cellular genes up- and downregulated, respectively, at the early stage. Most of them were involved in cell cycle regulation and immune response. Some of them, such as cyclin G2, cyclin B1, cell division cycle 20 homolog, cell division cycle 37 homolog, and the transcript factor E2F, that were involved in the induction of cells entering into the S phage were significantly upregulated. Some immunity-related genes, such as chemokine ligand 12, integrin alpha 6, LIM domain kinase 1, and retinoic acid early transcript 1L, were downregulated, suggesting that the suppression of the host cell immune response has occurred at the early stage of infection (see Table S1 in the supplemental material). At the intermediate stage of BtAdV infection, numerous cellular genes involved in transcription, replication, and translation, including the genes encoding transcription factor A mitochondrial, muscleblind-like splicing regulator 2, centromere protein K, and transcription domain-associated protein, were upregulated. These genes were likely to play important roles for the replication of BtAdV. From 12 to 18 h p.i., the numbers of the upregulated and downregulated cellular genes were significantly increased and finally peaked at 18 h. The most conspicuous change was the downregulation of a large number of genes associated with intra- and extracellular structure. Numerous cellular genes involved in the actin filament assembly, microtubule organization, cell junction formation, and extracellular matrix were downregulated at the late stage of the infection. The deregulation of a large number of host genes likely plays important roles in the formation and release of viruses.

Fig 5.

Analysis of cellular gene changes during BtAdV infection. (A) KEGG function classification of differentially expressed genes in BtAdV-infected BK cells. (B) Heat map visualization of differentially expressed cellular genes. Genes sorted according to time points of expression change more than 2-fold. (C) Numbers of differentially expressed cellular genes at different time points after BtAdV infection.

DISCUSSION

The development of the next-generation sequencing techniques has allowed fast and high-throughput sequencing of genomes in a wide range of organisms from viruses to more-complex mammalian genomes (28–32). In this study, by using a deep RNA sequencing technique, we successfully identified the transcription profiles of BtAdV-TJM in the M. davidii BK cells at different time points. Temporal analyses based on the RPKM method indicated that the BtAdV gene transcription can be divided into early, intermediate, and late phases, a profile similar to that of human AdV type 2 (33, 34) despite the different host origin of BtAdV-TJM (2).

However, cluster analysis showed that the transcription profiles of the seven BtAdV genes differed from those of human AdV type 2. The IVa-2 gene that was not present at 6 h p.i. in human AdV type 2 (35) was highly expressed in BtAdV at 6 h p.i. The pTP protein, which functions as a protein primer for initiation of DNA replication and was expressed early in human AdV type 2 infection, was expressed at 8 h p.i. in BtAdV and clustered as an intermediate gene. The 22K, 33K, and U exon genes that belong to the late genes in human AdV type 2 were clustered as the intermediate genes in BtAdV. The 100K and pVIII genes that belonged to the late genes in human AdV type 2 were clustered as the early genes in BtAdV. In total, it appears that a few BtAdV genes were transcribed earlier than they were in human AdV type 2, which accounts for both viral assembly and host innate immune antagonists, suggesting that BtAdV may replicate faster at least in bat cells. To confirm the kinetics of the BtAdV transcriptome more accurately, the inhibitors of DNA replication and protein synthesis should be used in the BtAdV infection experiments in the future.

Differences in gene expression profiles were observed between RNA-seq and qPCR. For example, the expression patterns of the E1A gene in the two methods did not perfectly match. The expression of two intermediate genes, pTP and 33K, and two late genes, hexon and penton, was detected at 6 and 8 h using qPCR but was not detected by RNA-seq analysis. A number of factors may influence the expression patterns of genes detected by the two methods. First, specifically to a single target, qPCR is more sensitive than RNA-seq. Second, different normalization procedures may affect mRNA levels calculated by the two technologies. Third, expression values obtained by RNA-seq may be affected by mutations present on BtAdV genes and reads with high mutation rates might not be detected during mapping analysis (18).

Bats are natural reservoirs of many emerging viruses, including severe acute respiratory syndrome (SARS)-like coronaviruses, henipaviruses, Ebola viruses, rabies viruses, and adenovirus (36, 37). These viruses can be fatal in other species but appear to cause no clinical signs of disease in bats under natural or excremental infection (36, 38, 39), implying a rapid innate immune system to control viral replication in bats. Several studies have demonstrated that bat type I and III interferons (IFN) have antiviral activity similar to that of other species (40, 41). Our study by RNA-seq revealed 79 immune genes differentially expressed more than 2-fold, among which six cellular innate immune genes involved in the interferon response were identified; alpha interferon (IFN-α)-inducible protein 6, IFN-induced protein with tetratricopeptide repeat 1, IFN-induced protein with tetratricopeptide repeat 2, IFN-induced transmembrane protein 3, IFN-induced protein 44-like, and IFN regulatory factor 9 were upregulated at 18 h p.i. This result is consistent with BtAdV infection triggering activation of the type I IFN pathway and subsequent activation of IFN-inducible genes. In addition, nuclear factor of kappa light polypeptide gene enhancer in B cells 1, which inhibits the NF-κB pathway, was downregulated at 18 h p.i. Consistent with the inhibition of the NF-κB pathway, the NF-κB-induced proinflammatory cytokines, interleukin 1 alpha and interleukin 8, were also suppressed following viral infection, as were three other proinflammatory cytokine-associated genes, interleukin 24, tumor necrosis factor alpha-induced protein 2, and tumor necrosis factor alpha-induced protein 3. Only one proinflammatory cytokine and a cytokine receptor, interleukin 15 and interleukin 4 receptor, respectively, were upregulated at 18 h p.i. It is worth it to mention that the level of mRNA may not at all reflect protein expression, since the two viral proteins E1BI and E4-33K, which selectively block the export of newly synthesized cellular mRNAs in human AdV2 (42), are also expressed early in BtAdV.

In summary, this study provides a comprehensive understanding of the BtAdV transcription profile and preliminary understanding of bat virus-host interactions, findings which have important implications for understanding AdV pathogenesis and their potential usage as gene delivery vectors.