Abstract
Epstein–Barr virus (EBV), also known as human herpesvirus 4, is
prevalent in all human populations. EBV mainly infects human B
lymphocytes and epithelial cells, and is therefore associated with
their various malignancies. To unravel the cellular mechanisms during
the infection, we constructed interspecies networks to investigate the
molecular cross-talk mechanisms between human B cells and EBV at the
first (0–24 hours) and second (8–72 hours) stages of EBV infection. We
first constructed a candidate genome-wide interspecies
genetic-and-epigenetic network (the candidate GIGEN) by big database
mining. We then pruned false positives in the candidate GIGEN to obtain
the real GIGENs at the first and second infection stages in the lytic
phase by their corresponding next-generation sequencing data through
dynamic interaction models, the system identification approach, and the
system order detection method. The real GIGENs are very complex and
comprise protein–protein interaction networks, gene/microRNA
(miRNA)/long non-coding RNA regulation networks, and host–virus
cross-talk networks. To understand the molecular cross-talk mechanisms
underlying EBV infection, we extracted the core GIGENs including
host–virus core networks and host–virus core pathways from the real
GIGENs using the principal network projection method. According to the
results, we found that the activities of epigenetics-associated human
proteins or genes were initially inhibited by viral proteins and
miRNAs, and human immune responses were then dysregulated by epigenetic
modification. We suggested that EBV exploits viral proteins and miRNAs,
such as EBNA1, BPLF1, BALF3, BVRF1 and miR-BART14, to develop its
defensive mechanism to defeat multiple immune attacks by the human
immune system, promotes virion production, and facilitates the
transportation of viral particles by activating the human genes NRP1
and CLIC5. Ultimately, we propose a therapeutic intervention comprising
thymoquinone, valpromide, and zebularine to act as inhibitors of
EBV-associated malignancies.
Introduction
The Epstein–Barr virus (EBV), also known as human herpesvirus 4
(HHV-4), was first identified in 1964 by Michael Epstein and Yvonne
Barr [[28]1]. Epstein and Barr investigated Burkitt’s lymphoma (BL),
and demonstrated that the malignant cells contained viral particles
with characteristic herpesvirus morphology; they subsequently reported
the first evidence of a tumor-associated virus in humans. EBV is a
ubiquitous virus that seriously infects more than 90% of the global
population. It mainly infects human B lymphocytes and epithelial cells,
and is therefore associated with a variety of their malignancies,
including BL, Hodgkin’s lymphoma (HL), gastric cancer (GC),
nasopharyngeal carcinoma (NPC), T/NK cell lymphoma, and AIDS- or
transplantation-associated lymphoma [[29]2]. Like other herpesviruses,
EBV exists in both latent and lytic phases with respect to viral gene
expression [[30]3, [31]4]. Upon infection, EBV establishes a lifelong
latency in the infected cells, predominantly in human B cells.
In the latent phase, the genomic DNA of EBV transforms into the episome
of the memory B cells, but only a limited subset of the viral latent
genes is expressed. Thus, the human immune system cannot target those
genes easily, which allows EBV to evade the human immune response; this
latent mode of infection is beneficial to EBV because it allows it to
persist. In contrast, during the lytic phase of infection, nearly all
the viral lytic genes of EBV are transcribed[[32]2]. It is essential
that the lytic cycle produces infectious viral particles, enabling the
spread of the virus from cell to cell and form host to host. In vitro
assays indicate that hypoxia, B cell receptor (BCR) stimulation, and
transforming growth factor-beta (TGF-β) can also induce a lytic
replication cycle under some circumstances[[33]5]. Lytic reactivation
causes a cascade of viral lytic genes expressed in a temporally
regulated manner in three stages: immediate–early (IE), early, and
late, which are accompanied by the replication of viral genomes and the
production of viral particles. Following EBV genome encapsidation, DNA
packaging, and virion release, new infectious virions can infect new
cells in the same host and new hosts[[34]6].
EBV, a double-stranded DNA virus, latent genomes can assemble into
chromatin structures with different histone and epigenetic modification
patterns that can regulate viral gene expression. These epigenetic
regulators include ubiquitin proteins, histone acetyltransferases,
deacetylases, and methyltransferases as well as DNA methyltransferases.
They also influence EBV, an oncogenic herpesvirus, pathogenesis by
evading human immune detection, resisting apoptosis, and driving human
cell carcinogenesis.
Because a study by Tina O'Grady et al. has reported the two-sided
next-generation sequencing (NGS)-based genome-wide time-course
expression data of human B cells and EBV during EBV infection [[35]7],
dynamic system modelling, applied to the molecular characterization of
transcription regulations, microRNA (miRNA) repressions, long
non-coding RNA (lncRNA) regulations and protein–protein interactions
(PPIs) (including their interactions with epigenetic enzymes), can be
solved to identify the genome-wide interspecies genetic-and-epigenetic
network (GIGEN) in this study. According to the gene expression
profiles of the viral immediate–early (IE) lytic genes BZLF1 and BRLF1,
the early lytic genes BMRF1, BBLF3, BBLF4, BGLF5, BNLF2A and BSLF1 and
the late viral genes BCRF1, BVRF2, BDLF1, BLLF1 and BCLF1 during the
infection[[36]7], we identified the GIGEN at first stage, where the IE
lytic genes and the early lytic genes are highly expressed, and the
GIGEN at second stage, where the early lytic genes and the late viral
genes are highly expressed. Furthermore, we determined more specific
interactions, regulations, and gene/protein functions between humans
and EBV by extracting the host–virus core networks (HVCNs) from the
GIGENs to provide more information on drug targets for multi-molecule
drug design. We then extracted the host–virus core pathways (HVCPs)
from the HVCNs to investigate the relationship between the defensive
and offensive human immune mechanisms and the antagonism strategies of
EBV from the perspective of the core signaling pathways in the GIGENs.
The HVCPs helped us understand in detail the significant events and
their corresponding molecular mechanisms, such as
genetic-and-epigenetic regulations and miRNA repressions, at the first
and second stages of infection. Finally, we discussed the potential
viral drug target proteins and miRNAs that are inferred from the HVCNs
and HVCPs, supported by the EBV-related literature review.
According to the results, the proposed potential multi-molecule drugs
can inhibit the switch from the latent phase to the lytic phase during
viral reactivation, and can suppress the expression of some critical
EBV lytic genes/proteins during EBV infection. This interrupts the
production of virions, interferes with the transportation of viral
particles, and destroys the viral defensive mechanisms.
Results
GIGENs of the first and second stages of the lytic phase of infection in
EBV-infected B cells
A flow chart of the strategy for constructing the GIGENs, HVCNs and
HVCPs in human B cells infected with EBV lytic infection from 0 to 72
hours post reactivation is shown in [37]Fig 1. According to the gene
expression profiles of the viral IE lytic genes, the early lytic genes
and the late viral genes in [38]Fig 2, we classified the lytic phase
into the first infection stage from 0 to 24 hours, where the IE lytic
genes and the early lytic genes are highly expressed, and the second
infection stage from 8 to 72 hours, where the early lytic genes and the
late viral genes are highly expressed. The GIGENs of the first and
second infection stages are shown in [39]Fig 3A and 3B, respectively.
The numbers of nodes and edges are recorded in [40]Table 1A and 1B,
respectively. Among these edges, three human TF complexes were
identified in the real GIGENs. The first was ARNT::AHR, which had 31
human TF-gene pairs at the first infection stage and 15 pairs at the
second infection stage; the second was HIF1A::ARNT, which had 16 human
TF-gene pairs at the first infection stage and 3 pairs at the second
infection stage; and the third was NFE2L1::MAFG, which had 38 human
TF-gene pairs at the first infection stage and 54 pairs at the second
infection stage. There were no remarkable differences in the number of
nodes between the first and second infection stages during the lytic
phase. Nevertheless, the edges of the real GIGENs between both
infection stages in [41]Table 1B revealed significant differences in
the human PPIs (first: 39,846/second: 28,325), interspecies PPIs
(first: 86/second: 45), and EBV-miRNAs to human-genes (first:
914/second: 620). Each edge at the first and second infection stages in
the real GIGENs is shown in [42]S1A Table. The results indicate that
there are more interactions in B cells, and between B cells and EBV,
which contribute to the enhancement of the transcriptional replication
of viral particles in B cells. EBV protects itself from silencing
through EBV miRNA, and inhibits some human biological processes, such
as the immune response, apoptosis, autophagy, and metabolism. We then
carried out DAVID analyses of target genes in GIGENs to evaluate the
specific functions between the first and second infection stages
([43]Table 2A and 2B, respectively) [[44]8].
Fig 1. Flow chart describing the constructing of the interspecies GIGEN
network and HVCP for the multiple drug targets and potential multi-molecule
drugs via the systems biology approach.
[45]Fig 1
[46]Open in a new tab
The blocks filled with gray indicate the input information exploited in
this process, obtained by big data mining to establish the candidate
GIGEN, NGS data to obtain the gene expression of human and EBV during
the lytic phase, genome-wide DNA methylation profiles to verify the
epigenetic regulation of DNA methylation of human genomes, and
literature information on multi-molecule drugs for multi-molecule drug
design based on the predicted drug targets. The blocks with gray frames
represent the systems biology approach exploited to provide the
identified information in our results; and the white blocks with solid
line frames contain the results obtained from these processes.
Fig 2. Changes of gene expression levels of typical lytic genes based on
classification from the literature review.
[47]Fig 2
[48]Open in a new tab
The NGS data of these typical lytic genes were sequenced by the Reads
Per Kilobase per Million mapped reads (RPKM) procedure at every
time-point, and are classified as immediate–early (IE), early, and late
stages during the lytic phase on the basis of the classification from
the literature review.
Fig 3.
[49]Fig 3
[50]Open in a new tab
Real interspecies GIGENs of the first (A) and second (B) infection
stages in the lytic phase. The nodes with red frames correspond to the
EBV proteins/TF/miRNAs; the nodes with blue frames indicate the human
receptors/proteins/TFs/miRNAs/lncRNAs; the edges in green denote the
PPIs of humans, EBV, and human-EBV; the edges in purple represent the
miRNA repressions of miRNAs on intraspecies and interspecies genes; the
edges in black represent the transcriptional regulations of TFs on
intraspecies and interspecies genes; the edges in orange signify the
lncRNA regulations of lncRNAs on human genes.
Table 1.
Numbers of nodes (A) and edges (B) in candidate GIGEN and the real
GIGENs at the first and second infection stages.
A. The number of nodes
Nodes Candidate GIGEN First infection stage Second infection stage
V_T 1 1 1
V_M 43 30 30
V_P 85 80 75
H_T 2,688 385 366
H_M 1,326 687 655
H_L 186 91 85
H_P 18,227 16,230 15,946
H_R 2,880 2,500 2,446
total 43,689 20,004 19,604
B. The number of edges
Edges Candidate GIGEN First infection stage Second infection stage
V_P ↹ V_P 301 58 44
V_T → V_M 5 0 1
V_M ┥V_G 67 1 2
H_P ↹ H_P 23,570,918 39,846 28,325
H_T → H_G 897,805 8,424 6,222
H_T → H_M 7,471 86 71
H_T → H_L 1,335 79 65
H_M ┥H_G 815,889 6,582 4,837
H_M ┥H_M 215 6 3
H_M ┥H_L 1,796 26 16
H_L→H_G 1,948 24 12
V_P ↹ H_P 5,135 86 45
V_T → H_G 1,252 15 14
V_M ┥H_G 39,558 914 620
V_M ┥H_M 39 2 4
V_M ┥H_L 175 6 7
H_T → V_G 680 6 4
H_T → V_M 675 6 4
H_M ┥V_G 1,708 1 4
H_M ┥V_M 10 1 1
total 25,346,982 56,169 40,301
[51]Open in a new tab
V_T: TFs of EBV, V_M: miRNAs of EBV, V_P: proteins of EBV, V_G: genes
of EBV, H_T: TFs of human B cells, H_M: miRNAs of human B cells, H_L:
lncRNAs of human B cells, H_P: proteins of human B cells, H_G: genes of
human B cells, H_R: receptors of human B cells, ↹: PPIs, →:
transcriptional regulations, ┥: miRNA repressions.
Table 2. Specific functional annotations of target genes in the real GIGENs
at the first infection stage (A) and at the second infection stage (B)
obtained by DAVID analysis.
A.Real GIGEN at the first infection stage
Functional annotation p-value
GO:0000082~G1/S transition of mitotic cell cycle 6.36E-12
GO:0019058~viral life cycle 3.3E-10
GO:0045815~positive regulation of gene expression, epigenetic 4.58E-10
GO:0006414~translational elongation 7.18E-10
GO:0006996~organelle organization 7.93E-10
GO:0043488~regulation of mRNA stability 4.5E-9
A.Real GIGEN at the second infection stage
Functional annotation p-value
GO:0006351~transcription, DNA-templated 7.22E-11
GO:0006614~SRP-dependent cotranslational protein targeting to membrane
1.23E-6
GO:0006334~nucleosome assembly 1.48E-5
GO:0042787~protein ubiquitination involved in ubiquitin-dependent
protein catabolic process 1.87E-5
GO:0097193~intrinsic apoptotic signaling pathway 3.85E-5
GO:0018279~protein N-linked glycosylation via asparagine 5.17E-5
[52]Open in a new tab
As indicated in [53]Table 2A, during the first stage of infection, EBV
begins lytic replication at the ori-Lyt site, the initial site for
viral lytic replication, and initiates the viral life cycle, which
involves decoding of genomic information, translation of viral mRNA by
human ribosomes, genome replication, and the assembly and release of
viral particles. The protein Zta encoded by BZLF1, an immediate–early
gene of EBV in the lytic phase, easily binds to the response elements
of hypermethylation; furthermore, there is an early viral protein
(SM/M) that enhances the posttranscriptional modification of EBV genes
[[54]9]. Moreover, the overexpression of an EBV protein kinase (BGLF4)
causes phosphorylation of the translational elongation factor, which
strengthens the output and stability of the nuclear mRNA
[[55]10–[56]12]. At this stage, human B cells enter the G1/S transition
of the cell cycle, in which DNA replication is initiated, and prepare
to undergo mitotic processing and, simultaneously, the formation of
some organelles, such as autophagosomes, ribosomes, and the
cytoskeleton.
During the second stage of infection, new virions are released from the
B cells to infect other uninfected B cells or epithelial cells.
[57]Table 2B indicates that after replication, nucleosomes are formed
to protect the integrity of the viral genome; they are then assembled,
packaged, and exported. Subsequently, the virions are dependent on
signal recognition particle (SRP) protein to help them target the
membrane. There are advanced glycation end product (AGE) receptors on
the cell surface; the complexes formed by products and receptors can
activate intracellular signaling pathways, such as the intrinsic
apoptotic signaling pathway or the ubiquitin-dependent protein
catabolic signaling pathway, to initiate reactions within the cells and
degrade target proteins.
HVCNs of the first and second stages of the lytic phase of infection in
EBV-infected B cells
Significant cellular processes of the HVCNs in the lytic replication cycle
Furthermore, in order to obtain the core cellular functions in humans
and EBV during lytic infection, we extracted HVCNs from the real GIGENs
at both stages of infection using the PNP method from the perspective
of the major network structure, as shown in [58]Fig 4A and 4B,
respectively; the numbers of nodes and edges are recorded in [59]Table
3A and 3B, respectively. Each edge of the HVCNs at the first and second
infection stages is shown in [60]S1B Table. Based on the identified
interactive parameters of the core PPINs in the HVCNs in Eq ([61]12) at
the first and second infection stages, the calculated statistical
significance (p value) of the edges in the core PPINs, is shown in
[62]S1C Table (See [63]Materials and Methods). The results reveal
98.33% and 99.68% statistically significant edges (p value≤ 0.05) at
the first and second infection stages, respectively. Moreover, in
contrast to the real GIGENs, we found no edges for human TF complexes
regulating the target genes in the HVCNs ([64]Fig 4). There was a
marked difference in the number of nodes among the EBV proteins (the
first infection stage: 34/the second infection stage: 12) during the
lytic phase in [65]Table 3A. This may account for the importance of the
early-expressed EBV genes, which are required for the cellular
functions of replication to help produce new viral particles, and the
simultaneous prevention of premature death in the human cells. However,
owing to the viral life cycle of EBV, late-expressed EBV genes may be
expressed less than early-expressed genes in preparation for entering
the latent phase, in which nearly all viral genes are silenced to evade
the human immune system. Furthermore, as shown in [66]Table 3B, there
were significant differences between the edges in both infection stages
in intraspecies PPIs (human, first: 510/second: 419; EBV, first:
42/second: 8), interspecies PPIs (first: 36/second: 8), human-TFs to
human-genes (first: 126/second: 67), human-TFs to human-miRNAs (first:
24/second: 7), and human-TFs to human-lncRNAs (first: 16/second: 4).
These data indicate that EBV mainly affects human PPIs through
protein–protein interaction with humans, and human-TFs further
transcriptionally regulate genes, miRNAs, and lncRNAs. To evaluate the
specific functions for the human genes during EBV infection, we also
analyzed target genes in HVCNs at both infection stages using DAVID,
and the results are presented in [67]Table 4A and 4B.
Fig 4.
[68]Fig 4
[69]Open in a new tab
HVCNs at the first (A) and second (B) infection stages in the lytic
phase. The nodes with red frames correspond to the EBV
proteins/TF/miRNAs; the nodes with blue frames indicate the human
receptors/proteins/TFs/miRNAs/lncRNAs; the edges in green denote the
PPIs of humans, EBV, and human-EBV; the edges in purple represent the
miRNA repressions of miRNAs on intraspecies and interspecies genes; the
edges in black represent the transcriptional regulations of TFs on
intraspecies and interspecies genes.
Table 3. Numbers of nodes (A) and edges (B) in HVCN at the first and second
stages of infection.
A. The number of nodes
Nodes First infection stage Second infection stage
V_T 1 0
V_M 10 11
V_P 34 12
H_T 20 14
H_M 96 118
H_L 14 14
H_P 166 148
H_R 82 95
total 423 412
B. The number of edges
Edges First infection stage Second infection stage
V_P ↹ V_P 42 8
V_T → V_M 0 0
V_M ┥V_G 1 0
H_P ↹ H_P 510 419
H_T → H_G 126 67
H_T → H_M 24 7
H_T → H_L 16 4
H_M ┥H_G 190 185
H_M ┥H_M 4 1
H_M ┥H_L 4 5
H_L→H_G 0 1
V_P ↹ H_P 36 8
V_T → H_G 3 0
V_M ┥H_G 22 23
V_M ┥H_M 1 4
V_M ┥H_L 1 6
H_T → V_G 0 1
H_T → V_M 1 2
H_M ┥V_G 1 1
H_M ┥V_M 1 1
total 983 743
[70]Open in a new tab
V_T: TFs of EBV, V_M: miRNAs of EBV, V_P: proteins of EBV, V_G: genes
of EBV, H_T: TFs of human, H_M: miRNAs of human, H_L: lncRNAs of human,
H_P: proteins of human, H_G: genes of human, H_R: receptors of human,
↹: PPIs, →: transcriptional regulations, ┥: miRNA repressions.
Table 4. Specific functional annotations of target genes in HVCNs at the
first infection stage (A) and at the second infection stage (B) obtained by
applying the DAVID analysis.
A.HVCN at the first infection stage
Functional annotation p-value
GO:0042493~response to drug 1.79E-08
GO:0043388~positive regulation of DNA binding 1.37E-05
GO:1902895~positive regulation of pri-miRNA transcription from RNA
polymerase II promoter 6.55E-05
GO:2000378~negative regulation of reactive oxygen species metabolic
process 1.17E-04
GO:0051090~regulation of sequence-specific DNA binding transcription
factor activity 1.64E-04
GO:0016236~macroautophagy 2.14E-04
B.HVCN at the second infection stage
Functional annotation p-value
GO:0006919~activation of cysteine-type endopeptidase activity involved
in apoptotic process 2.37E-04
GO:0032212~positive regulation of telomere maintenance via telomerase
3.20E-04
GO:0051897~positive regulation of protein kinase B signaling 0.001802
GO:0006954~inflammatory response 0.006715
GO:0070374~positive regulation of ERK1 and ERK2 cascade 0.011796
GO:0015031~protein transport 0.02314
[71]Open in a new tab
[72]Table 4A shows two crucial cellular functions at the first
infection stage in response to drugs and macroautophagy. Drugs change
the activity of lytic genes and stimulate or induce reactivation of EBV
from the latent phase to the lytic phase. Faggioni et al. suggested
that autophagy is blocked in the late stage of degrading
microbiological infections during EBV replication [[73]13]. This block
enables EBV to hijack the autophagic vesicles for intracellular
transportation, thereby enhancing viral production[[74]14]. [75]Table
4B shows that protein transport and inflammatory response are two
specific cellular functions at the second stage of EBV infection. New
virions may be conveyed in the autophagic vesicles mentioned above and
transported to the plasma membrane. Upon membrane lysis, these virions
are released from the cell to infect other cells. During this process,
the human immediate defense system detects xenobiotics and elicits the
inflammatory response, which triggers the immune system against
infection by the new virions. For the purpose of adapting to EBV
infection, it would be useful to identify post-translation epigenetic
modifications in HVCNs that regulate certain intracellular signaling
pathways.
Discussion
HVCPs at the first and second stages of infection during the lytic
replication cycle
New virion production through host–virus cross-talk interactions at the first
stage of infection
Although extensive research has been carried out on the cross-talk
between virus and host, little is known about how cellular functions
are performed through host–virus PPIs and real genetic-and-epigenetic
network from the perspective of intracellular signaling transduction
pathways. Thus, by further using the significant changes in gene
expression between the first and second infection stages determined by
p-values and the PNP method, we extracted HVCPs during the lytic phase,
which is divided into the first and second infection stages as shown in
Figs [76]5 and [77]6, respectively. The EBV intraspecies connections,
including {LMP1, BKRF2}, {LMP1, BLLF2, EBNA3B}, {BCLF1, BFRF3}, and
{EBNA2, Zta}, and the interspecies connections, including {MAPK7,
BDLF1} and {PSMA3, BVRF1} in Figs [78]5 and [79]6 can be supported by
the literature [[80]15]. Human cells can affect the behavior of
proteins to accommodate rapidly varying circumstances, whereas EBV
genes, including EBV miRNA (also called miR-BARTs) have ways to
manipulate the operation of viral proteins and even human proteins for
the purposes of survival and the propagation of the progeny in this
microenvironment. EBV microRNAs in particular can evade the human
immune response through the anti-apoptosis strategies in the HVCPs
[[81]16]. These are commonly referred to as epigenetics or
post-translational modifications, and they include DNA methylation,
ubiquitination, acetylation, and deacetylation. Following lytic
reactivation of EBV from latent infection to lytic infection, initially
the viral IE lytic genes BZLF1 and BRLF1 are expressed. They then
collaboratively activate the promotors of the early lytic genes, which
encode the viral replication proteins. Next, viral genome replication
occurs in which the late viral genes are transcribed. The late EBV
genes encode certain structural proteins required for viral genome
encapsidation into infectious viral particles [[82]5].
Fig 5. HVCP in B cells infected with EBV at the first infection stage in the
lytic phase.
[83]Fig 5
[84]Open in a new tab
The solid lines indicate the protein–protein interactions; the dotted
lines denote the translocations, including protein translations and
miRNA transcriptions; the solid lines that end in arrows, bars, or
circles stand for positively transcriptional regulations, negatively
transcriptional regulations, and miRNA repressions, respectively; the
dash-dot lines represent the gene functions that are inhibited; the
bold lines mean the gene functions that are promoted; the short arrows
beside the gene functions signify susceptibility to repression or
enhancement.
Fig 6. HVCP in B cells infected with EBV at the second infection stage of the
lytic phase.
[85]Fig 6
[86]Open in a new tab
The solid lines indicate protein–protein interactions; the dotted lines
denote the translocations, including protein translations and miRNA
transcriptions; the solid lines that end in arrows, bars, or circles
stand for positively transcriptional regulations, negatively
transcriptional regulations, and miRNA repressions, respectively; the
dash-dot lines represent the gene functions that are inhibited; the
bold lines mean the gene functions that are promoted; the short arrows
beside the gene functions signify susceptibility to repression or
enhancement.
We further investigated the cellular mechanisms of the first infection
stage by dividing the HVCP in [87]Fig 5 into five parts, as shown in
[88]Fig 7A–7E. In [89]Fig 7A, human cells send a cell proliferation
signal via NFKB1 to induce the production of more immune cells, and an
immune-related signal to antagonize EBV lytic reactivation. In [90]Fig
7A, receptor CD46 receives the immunity and cell growth signals via
NFKB1 binding. It can be supported that the interaction between NFKB1
and CD46 in human B cells promotes immune responses [[91]17]. CD46 is a
co-stimulatory factor for the development of T-helper cells, and works
through IL-10 release, suppressing immune responses to prevent
autoimmunity. It has been proposed that the immune-evasive strategy of
EBV appears to rely strongly on IL-10, and EBV itself also codes an
IL-10 homologue, which is expressed during the lytic phase [[92]18]. In
[93]Fig 7A, the viral protein EBNA2 interacts with CD46 to exploit its
immune regulation property and directly induce an immunosuppressive
phenotype. This results in CD46 being significantly downregulated
(p-value = 5.73 × 10^−16) by NFKB1-mediated apoptosis, so the
downstream expression of the human transcription factor ETS1 is
downregulated (p-value = 3.76 × 10^−4). In [94]Fig 7A, the human genes
NAT1 and SNHG5 are positively regulated by ETS1, whereas the human
genes CD46 and RNF41 are negatively regulated by ETS1. The
significantly low expression of NAT1 (p-value = 2.93 × 10^−37) could be
due to the low activity of ETS1, the inhibition of viral IE protein,
Zta, the repression of viral miRNA, miR-BART1-3p, and DNA methylation
(p-value = 4.23 × 10^−5). The main function of NAT1, an
acetyltransferase protein that functions as a xenobiotic metabolizing
enzyme, is to impair xenobiotic substances by acetylation. The
significantly low expression of SNHG5 (p-value = 1.32 × 10^−13) could
also be due to the low activity of ETS1, the repression of human miRNA,
miR4465, and DNA methylation. SNHG5 is a long non-coding RNA that
suppresses the proliferation, migration, and invasion of infected
cells. It also prevents the translocation of cell growth factors from
the cytoplasm to the nucleus and interrupts viral translocation from
the nucleus to the cytoplasm. It can be supported that ETS1 was thought
to be involved in regulating chromosomal translocations in Human B cell
non-Hodgkin lymphoma [[95]19]. Therefore, it functions as a
translocation breakpoint in B cell lymphoma [[96]20, [97]21]. As
mentioned above, the gene of human receptor CD46 is downregulated by
the impairment of NFKB1, and DNA methylation (p-value = 3.89 × 10^−5),
so its effect on autophagy in immunity is reduced. Thus, the human body
naturally attempts to exploit the functions of genes (NAT1, SNHG5,
CD46) to defeat EBV during lytic reactivation, but EBV successfully
evades attack by the human immune system. Furthermore, the markedly
high expression of RNF41 (p-value = 8.21 × 10^−8) is due to the low
transcriptional inhibition of ETS1 and the low repression of miR98
(p-value = 4.84 × 10^−19). In [98]Fig 7A, RNF41 is degraded by
ubiquitination when it functions as a receptor in the autophagy
mechanism pathway blocked by EBV. It can be supported that RNF41 is a
RING finger-containing protein, and has been investigated for its
involvement in TLR-mediated responses, growth regulation, and
inflammatory responses by promoting the ubiquitination of target
proteins [[99]22, [100]23]. Thus, RNF41 could cause the degradation of
several proteins by ubiquitination during EBV infection.
Fig 7. Signaling pathways of the interspecies molecular mechanisms based on
the HVCP in [101]Fig 5 at the first infection stage during EBV infection.
[102]Fig 7
[103]Open in a new tab
(A) The core pathways promoting cell proliferation and the impairment
of immune information by the EBNA2-mediated pathway with receptor CD46;
(B) the pro-apoptotic human pathway blocked by EBV through
ubiquitination and acetylation; (C) the autophagy mechanism blocked by
EBV through the involvement of viral BALF4, BDLF4, EBNA3B, miR-BART14,
and miR-BART1-3p; (D) the complete progression of lytic production
through the impairment of pro-apoptosis and the promotion of viral
translocation and anti-apoptosis; (E) the promotion of the integrated
production of infectious virions by silencing autophagy and inhibiting
the expression of STAT3.
Furthermore, owing to interaction with CD46 in [104]Fig 7A, EBV evades
immune attack by NFKB1. The cell growth signal then induces an
immediate–early viral protein, Zta, for the transcriptional activation
of early lytic genes via the viral protein EBNA2, which appears to be
more efficient at upregulating genes that are involved in the
proliferation and survival of infected cells [[105]24]. The elevated
activity of EBNA2 (p-value = 2.01 × 10^−24) enables the proliferation
of infected cells. EBNA2 is repressed by the viral microRNA miR-BART5,
which collaborates with EBNA2 to control the transcriptional regulation
of lytic replicate genes during infected cell proliferation. However,
the immunity signal induces the EBV immune-evasive strategy. EBV
develops resistance to apoptosis by counteracting the pro-apoptotic
function of p53 with miR-BART5 [[106]25]. EBV-miR-BART5 has a role in
anti-apoptosis, and highly expressed EBV-BZLF1 (p-value = 0.023)
reduces inflammation, resulting in an innate immune response. In
[107]Fig 7A, EBV-miR-BART5 and EBV-BZLF1 are silenced by MIR346 and
MIR1233-1, respectively, but the low expression of MIR346 and MIR1233-1
demonstrates that B cells still require other proteins to inhibit the
progression of the lytic replication cycle. It can be supported that
the upregulated MIR346 promoted apoptosis [[108]26]. In [109]Fig 7A,
the human genes CLOCK and NAT1 are transcriptionally inhibited by Zta,
and repressed by miR-BART14 and miR-BART1-3p, respectively. CLOCK can
cause energy metabolism and induce the progression of apoptosis and
autophagy, but low CLOCK activity does not destroy viral proteins. The
following is a brief overview of [110]Fig 7A. In the first stage of EBV
infection, the human proteins CLOCK, NAT1, SNHG5, CD46, and RNF41 are
inhibited by epigenetic modifications, EBV proteins, miRNAs, or the low
activity of TFs, and the human body is unable to defeat EBV. At the
same time, EBV activates its defense mechanism to antagonize immune
attacks through anti-inflammation and anti-apoptosis responses, thereby
enhancing the proliferation of infected cells via EBV EBNA2.
[111]Fig 7B shows RELA, which is a subunit of NF-κB and is involved in
many biological processes such as inflammation, immunity,
differentiation, cell growth, tumorigenesis, and apoptosis. In [112]Fig
7B, the human receptor ERBB3 can transmit signals from RELA via PNMA5
to the human TFs ZEB1 and MYC, which induce pro-apoptosis and trigger
apoptosis by releasing mitochondrial products and inhibiting viral
anti-apoptosis. The result also shows that the human TFs MYC and ZEB1
are subjected to proteolysis via ubiquitination by the
ubiquitin-proteasome pathway related proteins MUL1 and UBE2E1,
respectively, which decreases the transcriptional regulation of the
target genes. Thus, at the first stage of infection, the low activity
of genes miR185 and MYC results in a lack of pro-apoptotic activity.
Human gene MYC is also controlled by DNA methylation (p-value = 7.95 ×
10^−5). Furthermore, the low activity of MYC at the first infection
stage affects the low expression of SLC25A6, which suffers from the
repression of human miR1244-1, the transcriptional silence of DNA
methylation (p-value = 1.14 × 10^−2), and the acetylation by
acetyltransferase protein KAT5. Therefore, SLC25A6 cannot trigger
apoptosis through the release of mitochondrial products, which
indirectly suppresses the progression of pro-apoptosis. Additionally,
the low expression of ZEB1 (p-value = 1.72 × 10^−13) is unable to
transcriptionally inhibit EBV-miR-BART1-3p. This gives rise to elevated
EBV-miR-BART1-3p expression, promoting anti-apoptosis against innate
immune responses and pro-apoptosis, and successfully repressing the
expression of the target gene NAT1. This protects EBV from acetylation
by NAT1 during lytic replication. ERBB3 has been reported to be
involved in mediating the regulations of acetylation and cell apoptosis
through its signaling pathway [[113]27, [114]28].
It has been reported that CHEK1 mediates cell cycle arrest in response
to DNA damage and suppresses the proliferation of infected cells
[[115]29]. SLC25A6 is ubiquitously expressed in all tissues, and is
involved in the regulation of cell viability and apoptosis triggering
[[116]30]. CARD9 plays an important regulatory role in cell apoptosis
and the innate immune response to a number of intracellular virus
[[117]31]. Thus, we suggested that the human proteins CHEK1, SLC25A6,
CARD9, and NAT1 form a signaling transduction pathway that mediates
cell apoptosis to promote the destruction of EBV. However, NAT1 is
repressed by EBV miR-BART1-3p, and the inactivation of NAT1 by
acetylation results in the activation of CHEK1 and SLC25A6, which are
subjected to acetylation by acetyltransferase proteins MGAT4B and KAT5,
respectively; MIB2 ubiquitinates CARD9. It is thought these epigenetic
modifications have the following three characteristics: CHEK1 cannot
suppress the expression of viral proteins in infected cells; SLC25A6
cannot translocate ADP into mitochondria and ATP into the cytoplasm,
decreasing the release of mitochondrial products and triggering
apoptosis; and CARD9 is unable to induce an immune response to defeat
EBV. The cellular conditions mentioned above indicate that EBV
miR-BART1-3p mediates the apoptotic dysfunction of human B cells for
the purpose of survival. The result can be supported that miR-BART1
directly targets cellular tumour suppressor to dysregulate cell
apoptosis. There are growing evidence supporting the pro-viral role of
caspase/apoptotic pathway in viral replication [[118]32–[119]34].
In [120]Fig 7C, human receptors RNF41, GRB2, and ATG5 interact with
each other and signal immune or apoptotic information using ligands
(JAK2, FOS, and CLOCK, respectively) to induce SLC25A6 via TFs, ESR1,
and MYC. SLC25A6 can then trigger apoptosis by releasing mitochondrial
products to counteract the invasion of EBV.
Human proteins ATG5 [[121]35], RAB7A [[122]36], RPS14 [[123]37], and
BECN1 [[124]38] have been reported to be involved in autophagic
response. Therefore, [125]Fig 7C showed that the human body exploits
ligands (JAK2, FOS, and CLOCK) to transmit autophagy mechanism signals
through receptors to induce autophagy-related proteins (ATG5, RAB7A,
and BECN1). However, EBV-miR-BART1-3p suppresses xenobiotic acetylation
by repressing the human gene NAT1 in [126]Fig 7C. The inactive
acetylation caused by NAT1 results in the activation of SLC25A6 and
autophagy-related RAB7A, which are subjected to acetylation by
acetyltransferase proteins KAT5 and CSGALNACT1, respectively ([127]Fig
7C). Furthermore, human proteins (RNF41, GRB2, and MYC) are degraded by
ubiquitination through the binding of ubiquitin-proteasome pathway
related proteins UBE2K, USP46, and MUL1, respectively. Therefore, we
suggested that the low expression of these human proteins leads to a
reduction in the transmission of apoptotic and autophagic signals.
[128]Fig 7C shows that EBV miR-BART14 represses the gene that encodes
the CLOCK ligand, which reduces its expression (p-value = 6.16 ×
10^−6); CLOCK decreases energy metabolism and prevents the signal to
autophagic-related ATG5. However, the low activity of ATG5 (p-value =
6.75 × 10^−32) is not only due to the low expression of its ligand,
CLOCK, but is also due to transcriptional inhibition by DNA methylation
and repression by high human miR24-1 activity (p-value = 9.67 × 10^−7).
This affects the formation and elongation of autophagosomes via ATG5,
which leads to a reduction in autophagy. MiR24-1 has been associated
with autophagic response [[129]39]. At the first stage of infection,
SLC25A6, which has low acetylation activity and DNA methylation
activity (p-value = 1.14 × 10^−2), is repressed by human miR1244-1 and
reduces transcriptional regulation via TFs, ESR1 and MYC. This leads to
an inability to induce apoptosis by releasing mitochondrial products,
and affects the autophagic pathway (RNF41, RPS14, SLC25A6, RBPMS, and
BECN1); therefore, it cannot fulfil the functions of BECN1 and mediate
the nucleation and maturation of autophagosomes.
In [130]Fig 7C, BECN1 is repressed by the elevated expression of
miR301B (p-value = 0.012), and is subjected to DNA methylation, which
reduces its effects on autophagy and immunity. The low expression of
autophagy-related RAB7A (p-value = 4.46 × 10^−8) is due to acetylation
and the repression of human miR4659A (p-value = 0.036). Downregulation
of RAB7A further weakens lysosomal degradation by reducing the number
of lysosomes and the fusion between autophagosomes and lysosomes. It
has been supported that miR4659A directly regulated genes involved in
autophagic response [[131]40]. Therefore, the reduction of RAB7A
indicates that the autophagy mechanism is blocked, as observed during
lytic EBV replication.
In [132]Fig 7C, upon EBV blocking of the progression of autophagy,
viral protein EBNA3B negatively interacts with human protein PSME3,
which promotes ubiquitination and proteasomal degradation, thereby
inhibiting apoptosis. We speculated that viral EBNA3B hijacks the
ubiquitination of RNF41 by negatively interacting with PSME3, so that
EBV induces other ubiquitin proteins to act at the first and second
infection stages in the lytic phase. EBNA3B positively interacts with
human protein RBPMS, which plays a role as a coactivator of
transcriptional activity; therefore, RBPMS helps EBV to employ
autophagy-related BECN1. Furthermore, viral receptor BALF4 is an
envelope glycoprotein that forms spikes at the surface of the virion
envelope, so BALF4 is essential for attachment to the autophagosome
surface. Therefore, we suggest that BALF4 is involved in the fusion of
EBV virions and autophagosome membranes leading to EBV transportation
to B cell membranes for lysis. Another viral protein, BDLF4, is
important for the EBV lytic replication cycle, but is currently
uncharacterized. We speculate that BDLF4 collaborates with BALF4 and
EBNA3B to block autophagy and hijack the autophagic vesicles, thereby
enhancing viral production and the intracellular transportation of
virions. Viral proteins BALF4, BDLF4, and EBNA3B, and viral miRNAs
miR-BART14 and miR-BART1-3p block autophagy in B cells and interfere
with viral antigen presentation to prevent their degradation (see
[133]Fig 7C). It has been reported that BALF4 [[134]41] and EBNA3B
[[135]42] are involved in autophagic response. Hence, blocking mediated
by EBV can occur at different stages of the autophagy mechanism
pathway, from the formation of autophagosomes to the degradation of
lysosomes. Therefore, an understanding of the relationship between EBV
and autophagy may aid the discovery of new approaches to manipulate EBV
infection and lytic replication through autophagy control.
[136]Fig 7D shows that BAX is a ligand that binds to human receptor
CHMP5, which is involved in the degradation of surface receptor
proteins and the formation of endocytic multivesicular bodies (MVBs).
BAX assigns endosomal cargo proteins for incorporation into MVBs, and
sometimes functions in membrane fission, such as the lysis of enveloped
viruses. Thus, when triggered by BAX, CHMP5 exports pro-apoptotic
signal molecules from B cells into the cytoplasm via endocytic MVBs and
endosomal trafficking. CHMP5 then signals directly to the transcription
factor AR, and indirectly to the TF GATA2 via JUN, which is involved in
the TLR pathway. CHMP5 therefore induces pro-apoptosis and the
interruption of translocation, and inhibits anti-apoptosis (as shown in
[137]Fig 7D). However, CHMP5 has low activity at the first infection
stage because it is degraded following ubiquitination by the ubiquitin
ligase protein HUWE1, which reduces the expression of CHMP5 so that the
low activity of TFs (GATA2 and AR) causes a decrease in the
transcriptional silencing of human genes TGFB1I1 and PTK7, which are
thereby highly expressed (p-value = 5.37 × 10^−3 and p-value = 2.96 ×
10^−13, respectively). It can be supported that ubiquitination of CHMP5
could be associated with induction of anti-apoptosis [[138]43]. This
promotes the anti-apoptosis of B cells infected with EBV, and protects
EBV from cell death.
In [139]Fig 7D, another pathway mediated by EBV induces anti-apoptosis
and reduces the interruption of viral translocation and pro-apoptosis.
APH1A, an endoprotease that catalyzes the intramembrane cleavage of
integral proteins and membrane protein ectodomain proteolysis
[[140]44], binds to the receptor KANK2, which controls cytoskeletal
formation by regulating actin polymerization and promoting cell
proliferation. However, the elevated expression of viral protein BNRF1
(p-value = 5.15 × 10^−4), a tegument protein that plays a role in the
suppression of human intrinsic defenses to enhance the activation and
transcription of early viral genes, negatively interacts with receptor
KANK2 to evade caspase-independent apoptosis. BNRF1 then interacts with
human DAXX, and may regulate apoptosis in the cytoplasm, thereby
disrupting the complex formed between DAXX and ATRX. Suppressing the
DAXX–ATRX-dependent deposition of histone H3.3 on the viral chromatin
allows viral transcription. The low expression of DAXX (p-value = 2.67
× 10^−3) is due to the disruption of viral BNRF1, the deacetylation by
a histone deacetylase protein (HDAC8), and the low activity of CHMP5.
In contrast, DAXX interacts negatively with CCDC136. The elevated
expression of CCDC136 (p-value = 3.64 × 10^−60) causes the inactivation
of AR, as mentioned above and depicted in [141]Fig 7D. Moreover, LMP1
can activate the NF-κB signaling pathway and induce anti-apoptosis
[[142]45] through a pathway involving TRAF3 and CCDC136 to reduce the
activity of AR. This induces the downstream target gene PTK7 to
initiate anti-apoptosis. The human body naturally enables the
pro-apoptotic influence of EBV proteins by CHMP5, which receives the
BAX pro-apoptotic signal via endocytosis, but the degradation of CHMP5
by ubiquitination and the viral protein BNRF1-mediated anti-apoptotic
pathway may cause the inactivation of human TFs (GATA2 and AR).
Consequently, pro-apoptosis and interruption of viral translocation are
impaired, promoting anti-apoptosis and the complete progression of
lytic replication at the first stage of EBV infection.
As shown in [143]Fig 7E, RAC3, a GTPase that belongs to the RAS
superfamily of small GTP-binding proteins, regulates a wide variety of
processes, including the control of cell growth, cytoskeletal
reorganization, the activation of protein kinases, differentiation,
movement, and lipid vesicle transport. PSAP is a mitochondrial
pro-apoptotic protein that forms a complex with BAX when apoptosis is
induced [[144]46]. RAC3 conveys signals, including those inducing cell
growth factor and the activation of protein kinases to PTK7; it plays a
role in anti-apoptosis, and is involved with the receptor for viral
protein, BHRF1. PSAP, a pro-apoptotic protein ligand, binds to a
receptor, TGFB1I1, and plays a role in anti-apoptosis against the
pro-apoptotic signal from PSAP. Therefore, the diminished expression of
PSAP (p-value = 2.73 × 10^−7) has a negative impact on highly expressed
TGFB1I1 (p-value = 5.37 × 10^−3). Thus, the pro-apoptotic signal
induces the anti-apoptotic function of TGFB1I1, which is also the
receptor for viral proteins BFLF2 and BNLF2A (as shown in [145]Fig 7E).
The cell growth signal from receptor PTK7 is first received by a viral
protein, BHRF1, which prevents the premature death of human cells
during virus production. The viral protein BFRF3 then participates in
the assembly of infectious particles by localizing on the outer surface
of the capsid shell, thereby forming a layer between the capsid and the
tegument. BFRF3 interacts with BCLF1 [[146]15], which self-assembles to
form an icosahedral capsid. The elevated expression of BCLF1 (p-value =
5.69 × 10^−4) promotes the protection of the viral genome by the
capsid. The anti-apoptotic signal induces the viral protein BNLF2A to
evade HLA class I-restricted T cell immunity, and prevent TAP-mediated
peptide transportation and subsequent loading. The function of viral
BNLF2A is to activate SCO2, a copper chaperone that transports copper
to the Cu site on cytochrome C oxidase subunit II (COX2), which assists
the inner mitochondrial membrane in aerobic ATP production. Although
SCO2 may trigger pro-apoptosis, it is degraded via ubiquitination by
the ubiquitin ligase protein, G2E3. Thus, the pro-apoptotic function of
SCO2 is inhibited so that EBV evades the immune response and apoptosis.
It can be supported that BNLF2A contributes to cell survival through an
immune evasion mechanism [[147]47].
In [148]Fig 7E, viral receptor BKRF2 is required for the fusion between
viral and plasma membranes leading to EBV entry into the human B cell.
Membrane fusion is mediated by the fusion machinery comprising gB (also
called BALF4), and the heterodimer gH (also called BXLF2) /gL (also
called BKRF2) may also be involved in the fusion between the virion
envelope and the outer nuclear membrane during virion morphogenesis.
Viral BKRF2 also interacts with viral BNLF2A, which can encode an
inhibitor of transporter associated with antigen processing (TAP) to
assist immune evasion. Additionally, viral BKRF2 interacts with a viral
membrane protein, LMP1, [[149]15] which acts as a CD40 functional
homolog to prevent the apoptosis of the infected B-lymphocytes and
drive their proliferation. LMP1 signaling leads to the upregulation of
anti-apoptotic proteins and provides cell growth signals in the
infected cells. LMP1 helps viral EBNA3B in immune evasion and hijacks
the autophagy mechanism via viral BLLF2, an uncharacterized viral
protein. It has been reported that LMP1 can regulate autophagy in
EBV-infected B cells [[150]48].
Furthermore, the host receptor TGFB1I1 interacts with viral BFLF2 and
plays a fundamental role in virion nuclear egress. NEC1 (also called
BFLF2) interacts with the newly formed capsid within the human nucleus
via the vertexes, and NEC2 (also called BFRF1) directs it to the inner
nuclear membrane. NEC1 then induces the budding of the capsid at the
inner nuclear membrane and its envelopment into the perinuclear space.
The NEC1/NEC2 complex has been reported to promote fusion between the
enveloped capsid and the outer nuclear membrane, and subsequently
release the viral capsid into the cytoplasm, where it binds to the
secondary budding sites in the human Golgi network [[151]49].
Therefore, the anti-apoptotic signal from human TGFB1I1 promotes the
egress of the nuclear EBV virion.
In [152]Fig 7E, BDRF1 receives the egress signal from its viral
receptor, BBRF3. BDRF1 is a component of the molecular motor that can
translocate viral genomic DNA into the empty capsid during DNA
packaging. BDRF1 forms a tripartite terminase complex together with
TRM1 (also called BALF3) and TRM2 in the human cytoplasm. Once the
complex reaches the human nucleus, it interacts with the vertex of the
capsid portal. This portal forms a ring in which genomic DNA is
translocated into the capsid. BDRF1 has RNase activity, which plays an
important role in the cleavage of concatemeric viral DNA into
unit-length genomes. It can be supported that BBRF3 is involved in
capsid budding via the inner nuclear membrane during egress, and
participates in penetration through the plasma membrane during lytic
infection [[153]50].
As shown in [154]Fig 7E, the human ligand FOXP transmits the
anti-apoptotic signals to MAPK7, and MAPK7, which is an
extracellular-signal-regulated kinase, promotes signaling transmission
in the downstream signaling processes. Viral BDLF1 is a structural
component of the icosahedral capsid. The capsid is composed of
pentamers and hexamers of major capsid protein (MCP), which are linked
together by heterotrimers called triplexes. These triplexes consist of
a single molecule of triplex protein 1 (TRX1, also called BORF1) and
two copies of triplex protein 2 (TRX2, also called BDLF1). Furthermore,
BORF1 is required for the efficient transportation of BDLF1 to the
nucleus, which is the site of capsid assembly. Thus, the signal
promotes highly expressed BDLF1 (p-value = 9.65 × 10^−4) to activate
the structural molecule, and BDLF1 positively interacts with BORF1 to
induce viral capsid assembly by transporting BDLF1 to the nucleus via
BORF1. BORF1 then interacts with highly expressed viral BALF3 (p-value
= 0.0486), a component of the molecular motor. BALF3 (also called TRM1)
functions with BDRF1 (also called TRM3) and TRM2, and they
collaboratively translocate EBV genomic DNA into the empty capsid
during DNA packaging. It has been reported that BALF3 has endonuclease
activity, and plays an essential role in the cleavage of concatemeric
viral DNA into unit-length genomes [[155]51].
As shown in [156]Fig 7E, viral proteins BALF3, BDRF1, and BFLF2
interact with another viral membrane protein, LMP2B. LMP2B
downregulates the functionality of LMP2A, which is able to block B cell
activation. It is possible that LMP2B works in cooperation with LMP1
via viral BLLF2. The elevated expression of LMP2B (p-value = 0.0142)
downregulates the LMP2A-mediated interruption of B cell signaling,
whereas LMP1 activates B cells through the NF-κB, AP-1, and JAK/STAT
pathways. Therefore, EBV can exploit LMP2B and LMP1, which
collaboratively interact with BLLF2 to work in concert with EBNA3B
[[157]15], so that it contributes to the immune-evasive transport of
complete virions via autophagic vesicles hijacked by EBV.
Furthermore, highly active BLLF2 (p-value = 1.89 × 10^−3) interacts
with a viral receptor, BLLF1, which initiates virion attachment to
human B cells. This attachment triggers the fusion of the virion and
the human membrane for the invasion of the human cell. However, in
[158]Fig 7E, BLLF1 is degraded via ubiquitination by the ubiquitin
ligase protein SIAH1, and BLLF1 prevents the progression of viral
replication from the aggression and interruption of uninfected human B
cell. Moreover, viral BLLF1 negatively interacts with human ubiquitin
ligase protein SIAH1, which decreases the degradation of ubiquitination
via SIAH1 and reduces the activity of SIAH1 (p-value = 8.42 × 10^−4).
SIAH1 negatively interacts with human histone deacetylase HDAC8, so
highly expressed HDAC8 (p-value = 6.39 × 10^−14) can deactivate viral
protein degradation. The mechanism has been exploited for antiviral
strategies [[159]52].
In [160]Fig 7E, highly expressed human protein RBPMS (p-value = 2.35 ×
10^−11) interacts with viral proteins EBNA3B, BALF4, and BDLF4, and
acts as a latent-lytic switch in EBV by negatively interacting with the
human TF, STAT3. STAT3 can transcriptionally activate cellular PCBP2,
which represses the expression of EBV lytic genes [[161]53].
Consequently, EBV not only utilizes viral proteins to control RBPMS and
further manipulate the activity of STAT3, but also uses viral
miR-BART1-3p to repress the expression of human NAT1, which carries out
acetylation, so that the reduction of acetylation via NAT1 induces the
inaction of STAT3. Owing to the negative regulation of human RBPMS,
acetylation by acetyltransferase LFNG, and DNA methylation (p-value =
8.62 × 10^−3), low activity STAT3 (p-value = 1.35 × 10^−23) is less
able to maintain latency, which contributes to the complete progression
of the EBV lytic replication cycle. STAT3 acetylation has also been
reported in EBV-infected B cells [[162]54]. The main perspective of
[163]Fig 7E is that EBV reduces autophagy in immunity and silences
human STAT3 through viral proteins and viral miRNA. This maintains the
integrated replication of infectious virions and also promotes
anti-apoptosis against immune responses, the cleavage of viral DNA into
unit-length genomes, viral DNA packaging, capsid and tegument assembly,
and the transportation of virions via autophagic vesicles.
Transportation of viral particles through host–virus cross-talk interactions
at the second stage of infection
The purpose of this study was to elucidate the pathogenesis mechanism
of EBV-infected human B cells at the second stage of the lytic phase of
infection ([164]Fig 6). This is split into three parts in [165]Fig
8A–8C to facilitate the investigation of the virion transportation
process and the lysis of viral particles for further detailed analysis
of cellular function.
Fig 8. Signaling pathways of the interspecies molecular mechanisms based on
the HVCP in [166]Fig 6 at the second infection stage of EBV infection.
[167]Fig 8
[168]Open in a new tab
(A) The core pathways of the enhancement of anti-apoptosis,
immunosuppression, and genetic diversity pathways by EBV for the
packaging, assembly, and transport of viral particles; (B) the
promotion of virion production, vesicle trafficking, release, and
anti-apoptosis pathways as a result of EBNA1-mediated PML disruption;
(C) the maintenance of virion transportation by decreasing the
repression of the lytic cycle and increasing anti-apoptosis activities.
[169]Fig 8A showed that CCL19 transmits the immunoregulatory signal to
TFEB via a human receptor, ESR2. TFEB activates the expression of CD40L
in T cells, thereby participating in T cell-dependent antibody
responses in the activated CD4(+) T cells. High activity TFEB (p-value
= 1.34 × 10^−9) can activate the expression of many lysosomal genes,
and enables the positive regulation of autophagy and pro-apoptosis
against EBV lytic infection. TFEB is able to directly induce FAM98B via
the human TF NFATC2 to suppress tRNA splicing, but human NFATC2 is
degraded via ubiquitination by a ubiquitin ligase protein, SIAH1, which
leads to the reduced expression of NFATC2 (p-value = 0.002). This
reduces the transcriptional inhibition of FAM98B. Thus, FAM98B enhances
the ability of tRNA splicing to assist some viral late lytic genes in
translation, and increases genetic diversity in packaging, assembly,
and transportation. However, TFEB indirectly interacts with the human
TF SPIB via HNRNPU, which is associated with pre-mRNA processing in the
nucleus. HNRNPU affects pre-mRNA metabolism and triggers the apoptotic
process, but the low activity of HNRNPU (p-value = 7.73 × 10^−63) is
due to ubiquitination by the ubiquitin ligase protein WWP1, which
reduces the apoptotic function and transcriptional regulation of SPIB.
The low activity of SPIB (p-value = 3.91 × 10^−24) reduces the
transcriptional inhibition of FAM98B. This causes FAM98B to counteract
the repression of human miR5586 and to enhance the ability of tRNA
splicing to increase viral and human genetic diversity in packaging,
assembly, and transport, which can assist virion production in EBV
infection. When SPIB expression is low, the transcriptional inhibition
of EBV protein BCRF1 and viral miR-BART1-3p is reduced. Highly
expressed viral BCRF1 (p-value = 1.97 × 10^−5) encodes the viral
homologue of IL-10 (vIL-10), which suppresses the immune response in
the EBV lytic phase [[170]55], whereas viral miR-BART1-3p has an
anti-apoptotic function against the innate immune response and
pro-apoptotic signals. Viral miR-BART1-3p reduces the repression of
human PRKACB, regulating various cellular processes such as cell
proliferation, the regulation of microtubule dynamics, envelope
disassembly and reassembly, and the regulation of intracellular
transport mechanisms and ion flux. Viral miR-BART1-3p increases the
expression of PRKACB (p-value = 4.08 × 10^−4), which normally carries
out envelope assembly and intracellular transport, and promotes viral
progeny production and subsequent lysis. The repressed SPIB has also
been reported to inhibit apoptosis in late lytic cycle of the
EBV-infect B cells [[171]56].
In [172]Fig 8A, GABRG1, which belongs to the ligand-gated ionic channel
family [[173]57], regulates the activity of ionic channels to transmit
the anti-apoptotic signal to its receptor, TNFRSF10D, which is also a
receptor for viral BNLF2B. TNFRSF10D, a member of the TNF-receptor
superfamily, contains a truncated cytoplasmic death domain, which is
incapable of inducing apoptosis but acts as an inhibitor to protect EBV
proteins from TRAIL-mediated apoptosis [[174]58]. TNFRSF10D interacts
with the viral protein BNLF2B, and is very similar to the EBV protein
BCRF1, which may evade the immune system and protect EBV. Viral BNLF2B
positively interacts with human PRKACB, and viral miR-BART1-3p reduces
the repression of human PRKACB, so EBV exploits PRKACB to promote
envelope assembly and the intracellular transport of virions. Thus,
viral BNLF2B transmits an anti-apoptotic signal that promotes viral
particle production and affects the transcriptional activity of the
human TF TP73 through human proteins HSD17B4, GLDN, and TUBGCP2. BAX is
upregulated by the transcriptional regulation of TP73, but it is also
subjected to the repression of human miR296, degradation via
ubiquitination by the ubiquitin-conjugating enzyme UBE2C, and DNA
methylation (p-value = 1.93 × 10^−3); these factors reduce the
pro-apoptotic function of human BAX. TNFRSF10D has been discovered to
contribute disease progression in EBV-infected cells [[175]59].
As shown in [176]Fig 8A, LMP1, functions as a CD40 homolog [[177]60],
negatively interacts with human IKBKB, phosphorylating the inhibitor in
the inhibitor-NF-κB complex, which causes the dissociation of the
inhibitor and the activation of NF-κB. However, the pro-apoptotic
function of human IKBKB is suppressed as a result of the negative
interaction with LMP1 and inactivation via acetylation by the
acetyltransferase GALNT7, so the expression of IKBKB is therefore low
(p-value = 3.54 × 10^−20). This directly affects the behavior of the
human pro-apoptotic protein BAX, which belongs to the BCL2 protein
family and acts as a pro-apoptotic regulator. During stress, BAX
experiences a conformational change that results in translocation to
the mitochondrion membrane, subsequently leading to the release of
cytochrome C, which triggers pro-apoptosis. However, owing to the
positive interaction with IKBKB (expressed at a low level), the
transcriptional regulation of TP73, repression by human miR296,
ubiquitination by the ubiquitin-conjugating enzyme UBE2C, and DNA
methylation (p-value = 1.93 × 10^−3), pro-apoptotic BAX is inhibited by
EBV proteins BNLF2B and LMP1. It can be supported that inhibition of
BAX leads to the repressed apoptosis in EBV-infected B cells [[178]61].
This means that the human body loses one of the defensive mechanism
that can antagonize EBV lytic infection. [179]Fig 8A indicates that EBV
not only evades immune suppression by viral BCRF1 and inhibits
pro-apoptosis by viral BNLF2B and LMP1, but also exploits the
anti-apoptotic function to promote the propagation of EBV progeny via
viral miR-BART1-3p.
In [180]Fig 8B, highly expressed EBNA1 (p-value = 8.81 × 10^−3) is not
repressed by human miR127 in this system, and impairs DNA repair,
decreasing the activation of p53 and anti-apoptosis in response to DNA
damage. This leads to increased cell survival by the silencing of PML
proteins, and EBNA1 thereby protects virion production from apoptosis
and promotes lytic infection. The interaction between viral EBNA1 and
human CK2 kinase is important for EBNA1 to disrupt PML nuclear bodies
and degrade PML. EBNA1 increases the association of CK2 with PML,
thereby increasing the ability of CK2 to phosphorylate PML.
Phosphorylation is a modification that is well known to trigger the
polyubiquitylation and degradation of PML. It can be supported that PML
disruption and silencing by EBNA1 are defensive mechanisms by which EBV
may contribute to the advance of EBV-associated cancer [[181]62].
In [182]Fig 8B, disrupted PML can positively interact with NUP155, a
nucleoporin protein that plays an important role in the assembly and
function of the nuclear pore complex (NPC), which regulates the
movement of molecules across the nuclear envelope (NE) [[183]63].
NUP155 participates in the formation of the double membrane NE.
However, it positively interacts with human-disrupted PML and is
degraded by the ubiquitin protein USP3 via ubiquitination, leading to
the low expression of NUP155 (p-value = 9.79 × 10^−12). This affects
transporter activity, the structural constitution of nuclear pores, and
the binding and translocating ability of proteins during
nucleocytoplasmic transport. Therefore, the damaged DNA signal is not
successfully transported from the nucleus into the cytoplasm to induce
DNA repair and trigger pro-apoptosis to block EBV. NUP155 negatively
interacts with ABCA12, so the low expression of NUP155 results in the
elevated expression of ABCA12 (p-value = 5.57 × 10^−5). ABCA12 is a
member of the superfamily of ATP-binding cassette (ABC) transporters,
which transport various molecules across extra- and intracellular
membranes. ABCA12 interacts with HOOK1, a member of the hook family of
coiled-coil proteins, which bind to microtubules and organelles. HOOK1
links endocytic membrane trafficking to the microtubule cytoskeleton,
and its complex can promote vesicle trafficking. HOOK1 interacts with
TRIM3, a member of the cytoskeleton-associated recycling or transport
(CART) complex, through RASSF10 to collaboratively mediate vesicular
trafficking via TRIM3’s association with the CART complex and
cooperatively promote the transport of virions. TCTN1 is a member of
the family of secreted and transmembrane proteins that act as a barrier
preventing the diffusion of transmembrane proteins. However, TCTN1 is
inactivated via acetylation by the acetyltransferase NAA16, rendering
it incapable of controlling the diffusion of the transmembrane proteins
that allow EBV to hijack this function to assist virion transport.
TCTN1 expressed at a low level (p-value = 1.19 × 10^−14) causes the
reduction of signal transmission to the human TF YBX1 via TRIM46, and
therefore the low expression of YBX1 (p-value = 1.8 × 10^−19) leads to
a reduction of the transcriptional regulation of the human target gene,
ARRB2. The pro-apoptotic function of ARRB2 is inhibited by the
reduction of the transcriptional regulation of YBX1, ubiquitination by
the ubiquitin protein HERPUD1, and DNA methylation (p-value = 3.05 ×
10^−2). It can be supported that the decreased YBX1 contributes to
promote anti-apoptosis in EBV-infected cells [[184]64].
As shown in [185]Fig 8B, when ARRB2 is translocated to the plasma
membrane as a human receptor, it receives the signal from a ligand,
CCL23, which participates in immunoregulatory and inflammatory
processes. CCL23 transmits the signal in response to inflammation to
induce ARRB2 to carry out its pro-apoptotic function, but the activity
of ARRB2 is suppressed. It subsequently induces the operation of human
CARD8, which belongs to the caspase recruitment domain
(CARD)-containing family of proteins. CARD8 enables the activation and
expression of caspases or the NF-κB pathway, and may be a component of
the inflammasome, a protein complex that participates in the activation
of pro-inflammatory caspases. However, ARRB2 positively interacts with
CARD8, reducing the expression of CARD8 and inactivating it (p-value =
7.67 × 10^−19), and influences the inactivation of pro-inflammatory
responses so that it promotes viral immune evasion. Inactivated CARD8
causes the activation of human FAM98B. FAM98B can increase the ability
of tRNA splicing to assist the EBV lytic phase in the production of
viral particles. FAM98B drives the highly expressed human TF CTCFL
(p-value = 1.49 × 10^−54) to transcriptionally inhibit human miR130B.
Human miR130B promotes cell growth and self-renewal, and its normal
expression increases cell viability, reducing cell death and decreasing
the expression of apoptosis-related proteins [[186]65]. Nevertheless,
in spite of the fact that the inhibited miR130B may reduce cell
viability, and increase cell death and the expression of
apoptosis-related proteins, it also reduces the repression of TRIM3 by
miR130B. Thus, it also results in the elevated expression of TRIM3
(p-value = 7.92 × 10^−40), which helps EBV transport virions. MiR130B
has been associated with apoptosis in viral-infected cells [[187]66]. A
brief overview of [188]Fig 8B indicates that EBNA1-mediated PML
silencing and disruption are responsible for inducing the successful
progression of the EBV lytic cycle, which can promote virion
production, vesicle trafficking, intracellular transport, and
anti-apoptosis during lytic infection.
[189]Fig 8C features EIF2AK2 plays essential roles in the innate immune
response to viral infection, signal transduction regulation, apoptosis,
and cell proliferation; it exerts its influence on anti-viral activity
in a wide range of DNA and RNA viruses including EBV.
EIF2AK2, a serine/threonine protein kinase, is activated by
autophosphorylation [[190]67]. In [191]Fig 8C, EIF2AK2 binds to the
human receptor IL10RA, a receptor for interleukin 10 that is
structurally associated with interferon receptors. IL10RA has been
shown to mediate the immunosuppressive signal of interleukin 10,
thereby suppressing the biosynthesis of pro-inflammatory cytokines.
Hence, EIF2AK2 binds to IL10RA as a ligand. IL10RA is also a receptor
of viral BCRF1, and induces anti-inflammatory factors to protect EBV
proteins subjected to apoptotic attack by the immune system. BCRF1, a
late gene of the lytic phase, encodes viral interleukin-10 (vIL-10),
which is a human homolog of interleukin-10 (hIL-10). vIL-10 is
expressed in both the early and late phases of viral lytic production
when human B cells are infected with EBV. BCRF1 has certain cellular
functions that are similar to those of hIL-10 proteins, which generally
participate in immunosuppression. Highly expressed viral BCRF1 (p-value
= 1.97 × 10^−5) can suppress the cytokine synthesis of interleukins and
interferon, and weaken the human natural killer cell and cytotoxic
T-cell (CTL) responses so that EBV can subsequently establish latent
infection. It can be supported that BCRF1 may have a role in the
interaction between EBV and the human immune system [[192]68, [193]69].
As shown in [194]Fig 8C, BCRF1 interacts with viral BPLF1, a large
tegument protein (LTP) that plays numerous roles in the EBV viral
cycle. During EBV lysis, highly active BPLF1 (p-value = 3.5 × 10^−101)
remains associated with the capsid—whereas most of the tegument is
detached—and has a role in the transport of the capsid toward the human
membrane. As mentioned above, BPLF1 interacts with BALF3 and then BORF1
at the first infection stage. BALF3 interacts with BORF1 at the second
infection stage, which mainly helps EBV terminate some final steps of
viral replication during lytic infection, including DNA cleavage, DNA
packaging, assembly of the capsid and tegument, and intracellular
transport at the human membrane in preparation for virion lysis. Viral
BVRF1 acts as a checkpoint in the formation between the viral capsid
and the tegument. BVRF is a capsid vertex-specific component that is
involved in EBV DNA encapsidation, ensuring an accurate DNA genome
cleavage and stabilizing capsids. Moreover, viral BVRF1 can negatively
interact with human PSMA3 to avoid the degradation of EBV proteins, so
PSMA3 expresses with low activity (p-value = 1.28 × 10^−20). The
interaction has also been proposed in HIV-infected cells [[195]15].
In [196]Fig 8C, PSMA3 that is expressed at a low level positively
interacts with the human TF STAT3 (also at a low level of expression)
(p-value = 8.98 × 10^−13) via ST3GAL3, and STAT3 is subjected to
ubiquitination by ubiquitin protein USP33 so that human STAT3 has low
activity (p-value = 1.35 × 10^−23) with regard to reducing the
suppression of lytic infection and decreasing the transcriptional
inhibition of the human target gene NRP1. Human NRP1 is highly active
(p-value = 1 × 10^−80) and is able to transport and assist in EBV
virion translocation. STAT3 has also been associated with
transportation in viral-infected cells [[197]70].
LRRK2 acts as a ligand and is largely present in the cytoplasm, but is
also associated with the mitochondrial outer membrane, and positively
regulates autophagy through a calcium-dependent signaling pathway
[[198]71]. In [199]Fig 8C, it binds to a human receptor, NRP1, which
contains some specific protein domains that allows it to participate in
various types of signaling pathways that control cell migration, cell
survival, transport, and attraction. The blocking of autophagy by EBV
at the first infection stage causes the inactivation of LRRK2. The
inactivated LRRK2 negatively interacts with its receptor, NRP1, and
activates the transport function of NRP1. NRP1 transmits a signal to
the human TF RUNX2 through positive interaction with BDH1, and
subsequently VCAM1, both of which have high activity (p-value = 8.83 ×
10^−70 and p-value = 1.04 × 10^−39, respectively). However, human VCAM1
negatively interacts with RUNX2, which leads to the low expression of
RUNX2 (p-value = 5.62 × 10^−41), thereby decreasing the transcriptional
regulation of the human target gene AFG3L1P. AFG3L1P is a long
non-coding RNA, and is associated with mitochondrial fusion and the
import of proteins into the mitochondrial intermembrane space, which
may trigger apoptosis. Thus, the low expression of AFG3L1P (p-value =
3.34 × 10^−8) is due to the reduction of transcriptional regulation by
RUNX2 and the repression of EBV miR-BART14, so that human AFG3L1P is
unable to trigger apoptosis to defeat viral proteins owing to the
indirect influence of blocked autophagy by EBV and the direct effect of
the inhibition by EBV miR-BART14. RUNX2 has also been associated with
apoptosis in HBV-infected B cells [[200]37].
In [201]Fig 8C, the low-level expression of lncRNA AFG3L1P results in a
decrease of transcriptional inhibition of the human target gene CLIC5,
a member of the chloride intracellular channel (CLIC) family of
chloride ion channels. CLIC5, encoded by a human target gene, is
related to actin-based cytoskeletal structures; it is inserted into
membranes and forms poorly selective ion channels that may also
transport chloride ions. Hence, high activity CLIC5 (p-value = 7.32 ×
10^−5) influences the activity and formation of chloride channels to
enhance transport.
As shown in [202]Fig 8C, human proteins BAX and PCBP2 interact with
VCAM1 and ST3GAL3, respectively. Normally, the activation,
conformational change, and relocation of BAX from the cytosol to the
mitochondria cause the release of cytochrome C to the cytosol and
trigger apoptosis, which can function as an offensive mechanism against
EBV lytic infection [[203]72]. BAX interacts with PCBP2; PCBP2 can
regulate the susceptibility to lytic cycle activation signals and
interact with STAT3 via ST3GAL3, which can mediate the repression of
the EBV lytic cycle and maintain EBV latency [[204]73]. However, BAX is
degraded via ubiquitination by the ubiquitin-conjugating enzyme UBE2C,
which results in a reduction of the apoptotic effect on EBV proteins.
PCBP2 is also subjected to ubiquitination by UBE2C, repression of human
miR421 and EBV miR-BART10, and DNA methylation (p-value = 1.07 ×
10^−4), all of which reduce PCBP2 expression (p-value = 5.58 × 10^−70)
and decrease lytic cycle repression, contributing to EBV lytic
production and transportation. EBV miR-BART10 counteracts the
expression of low activity human miR421 (p-value = 2.59 × 10^−5), and
has an effect on the inhibition of apoptosis. It can be supported that
Bax has been associated with EBV lytic cycle gene, such as BALF1
[[205]74]. To summarize [206]Fig 8C, viral BCRF1 performs the
immunosuppressive mechanism that directly assists EBV in the final
processes of virion production, and indirectly influences the transport
of viral particles, the performance of anti-apoptotic function via EBV
miR-BART14 and miR-BART10, the persistence of the lytic production
cycle, and the progression of transportation by EBV miR-BART10. EBV
maintains the transportation of viral particles by decreasing the
repression of the lytic cycle and increasing anti-apoptosis activities.
Overview of the molecular mechanism of lytic infection from the first to the
second infection stages in human B cells infected with EBV
The suppressed expression of EBV latent antigens is important because
it allows EBV-related tumors to evade immune surveillance. EBV lytic
production is triggered by a variety of inducer treatments, and the
induced lytic genes lead to cytotoxic T lymphocyte (CTL) responses
[[207]75]. Once the resting memory B cells latently infected with EBV
are reactivated and enter the lytic phase, most of the EBV lytic genes
and a few of the latent genes are transcribed and expressed in the
lytic phase, and the human immune system simultaneously detects EBV
antigens. The human immune response is thereby triggered. The response
involves antigen-presenting cells, cytotoxic T cells, pro-inflammatory
cytokines, reactive oxygen species, and certain pro-apoptotic signals
that mediate immune mechanisms, including pro-apoptosis, autophagy, the
inflammatory response, the human intrinsic immune pathway, and the
extrinsic immune pathway in response to human B cells infected with EBV
in the lytic phase.
As shown in [208]Fig 9, EBV mediates certain defensive mechanisms
against the human immune response at the first infection stage. It has
been reported that EBNA2 and Notch intracellular domain (NICD) are
partially interchangeable [[209]76]. The nuclear-cytoplasmic transport
of EBNA2 [[210]77] could enable its association with Notch
extracellular domain (NECD) to mediate the interaction between EBNA2
and CD46. In [211]Fig 9, the viral protein EBNA2 evades immune
apoptosis by interacting with human receptor CD46 with its immune
inhibition property to induce an immunosuppressive phenotype, and then
reduces the operation of human autophagy in the immunity of CD46 and
the interruption of viral translocation of SNHG5. Viral IE protein Zta
interacts with EBNA2 [[212]15], transcriptionally activates EBV early
lytic genes, inhibits human acetylation of NAT1 with viral
miR-BART1-3p, and suppresses the human energy metabolism of CLOCK,
which may trigger apoptosis, with viral miR-BART14. Furthermore, BZLF1
itself has an anti-inflammatory response to the human immune system.
Viral miR-BART5 has an anti-apoptotic response that protects EBV from
human immune attacks. Moreover, the viral miRNA miR-BART1-3p plays a
role in anti-apoptosis, and may hijack the acetylation function of NAT1
by repressing the NAT1 gene, so that EBV exploits other
acetyltransferases to perform the acetylation at both infection stages.
In addition, viral protein EBNA3B may hijack the ubiquitination
function of RNF41 by negatively interacting with human protein PSME3,
so that EBV exploits other ubiquitin proteins to carry out
ubiquitination at both infection stages. EBV decreases the
transcriptional regulation of human TFs (ZEB1 and MYC) and the
expression of human receptors (RNF41 and GRB2) via ubiquitination to
indirectly inhibit the human pro-apoptotic function of MYC and reduce
the release of mitochondrial products of SLC25A6 that can trigger
pro-apoptosis. Thus, EBV can block the autophagy mechanism through the
indirect inhibition of autophagy-associated protein ATG5 with viral
miR-BART14, and through acetylation and ubiquitination, to directly and
indirectly restrict the other autophagy-related proteins (BECN1 and
RAB7A). The viral proteins EBNA3B, BALF4, and BDLF4 hijack the
autophagy-related autophagosomes, which facilitate intracellular
vesicle trafficking. Moreover, EBV reduces the pro-apoptosis and viral
release of CHMP5 and the translocational disruption of SNHG5, and
increases the anti-apoptotic function of TGFB1I1 and PTK7 through the
ubiquitination of CHMP5 and the suppression of human intrinsic defenses
by viral protein BNRF1, which also enhances the activation and
transcription of early viral genes. The anti-apoptotic signal induces
viral protein BNLF2A to evade HLA class I-restricted T cell immunity
and prevent TAP-mediated peptide transport and the subsequent loading.
In addition, the degradation of viral protein BLLF1 via ubiquitination
can protect viral replication from the aggression and interference of
uninfected human B cells.
Fig 9. Overview of molecular mechanisms in EBV lytic infection and the
significant network marker for potential multi-molecule drug design.
[213]Fig 9
[214]Open in a new tab
The red words indicate the potential drug target proteins for
multi-molecule drug design; the green words represent the molecular
mechanisms being hijacked by EBV; the blue and pink arrows denote the
cellular functions being inhibited or promoted, respectively; the
yellow arrows represent the progression from the first into the second
infection stage in the lytic phase. Upon EBV reactivation into the
lytic phase, the human immune system can detect EBV antigens, trigger
the human immune responses, and inhibit the progression of EBV lytic
replication. However, EBV mediates some defensive mechanisms against
human immune response, and thereby protects the complete virion
production and transportation from human immune interference at both
infection stages. Additionally, the transition stage of EBV is
dependent on the activation of late lytic genes and LMP1 function.
Under the protection of the defensive mechanism of EBV proteins and
miRNAs mentioned above, EBV can securely produce viral particles
without disturbance at the first infection stage. Viral protein EBNA2
can also promote infected cell proliferation. Viral protein BHRF1
prevents the premature death of the human cells during virus
production. Viral protein BFRF3 then participates in the assembly of
the infectious particles by decorating the outer surface of the capsid
shell, thereby forming a layer between the capsid and the tegument.
BFRF3 interacts with BCLF1, which self-assembles to form an icosahedral
capsid. Viral protein BDLF4 is important for the EBV lytic replication
cycle, and collaborates with BALF4 and EBNA3B so that they promote
viral production and the intracellular transportation of virions.
Membrane fusion is mediated by BALF4, and the heterodimer BXLF2/BKRF2
may also be involved in the fusion between the virion envelope and the
outer nuclear membrane during virion morphogenesis. LMP1 acts as a CD40
functional homolog to prevent the apoptosis of the infected B cells to
drive their proliferation. LMP1 signaling leads to the upregulation of
anti-apoptotic proteins and provides cell growth signals in the
infected cells. BFLF2 plays a fundamental role in EBV virion nuclear
egress. Viral receptor BBRF3, an essential lytic replication protein,
is an envelope glycoprotein and is crucial for virion assembly and
egress. BDRF1 can translocate viral genomic DNA into the empty capsid
during DNA packaging, and BDRF1 forms a tripartite terminase complex
together with BALF3 and TRM2 in the human cytoplasm. The viral
icosahedral capsid is composed of pentamers and hexamers of the major
capsid protein, which are linked together by triplexes. These triplexes
consist of BORF1 and BDLF1. Additionally, BORF1 is required for the
efficient transport of BDLF1 to the nucleus, which is the site of
capsid assembly. Viral BALF3 functions with BDRF1 and TRM2, and they
collaboratively translocate EBV genomic DNA into the empty capsid
during DNA packaging. Viral membrane proteins LMP2B and LMP1
collaboratively activate human B cells by interacting with BLLF2, and
both of them work in concert with EBNA3B so that EBNA3B can mediate the
immune evasive transport of complete virions via autophagic vesicles.
RBPMS, interacting with viral proteins EBNA3B, BALF4, and BDLF4, acts
as the latent–lytic switch in EBV by negatively interacting with STAT3.
Inactivated STAT3 cannot transcriptionally activate cellular PCBP2, and
PCBP2 cannot repress the expression of EBV lytic genes. This increases
the transcriptional activation of EBV late lytic genes, which may
contribute to the complete progression of the EBV lytic production
cycle into the late stage. A viral membrane protein existing in both
infection stages may play a crucial role in transformation during the
lytic phase of viral production. EBV utilizes the significantly
different expression levels of LMP1 for two purposes during its life
cycle. First, in the latent infection of B cells, the steady-state
expression of LMP1 is important for maintaining the transformational
state. Second, in the EBV lytic production cycle, the induced LMP1
efficiently facilitates the release of the virions from the B cells.
Once EBV lytic replication has been induced, the expression of LMP1 is
upregulated. Here, we indicated that the gene products of LMP1 are
responsible for the efficient release of the virus from the B cells. It
is feasible that viral LMP1 enhances the envelopment, de-envelopment,
and re-envelopment pathways either at the nuclear membrane or in the
cytoplasm during the lytic phase [[215]78].
Viral membrane protein LMP1 promotes the transformation of human B
cells infected with EBV from the first into second infection stage, and
protects the infected B cells from apoptosis. EBV maintains the
defensive mechanism to protect the complete virion production and
transportation from human immune interference. Viral proteins (LMP1 and
BNLF2B) indirectly reduce the pro-apoptotic function of BAX. BNLF2B is
very similar to BCRF1, which may facilitate immunosuppression to
protect EBV. EBV increases the anti-apoptotic performance of viral
miR-BART1-3p and the immunosuppression of viral protein BCRF1 as an
immune evasive mechanism through the degradation of HNRNPU via
ubiquitination. Viral antigen EBNA1 can induce anti-apoptosis to
mediate the disruption and silencing of PML and block the pro-apoptotic
signal. The pro-apoptosis of ARRB2 is inhibited by the ubiquitination
and indirect suppression of EBNA1. Viral miR-BART14 can repress the
expression of AFG3L1P, which triggers pro-apoptosis. Viral miR-BART10
exploits the anti-apoptotic response to antagonize human pro-apoptosis,
and represses the expression of PCBP2 to protect virion production in
the lytic cycle from restriction. EBV also exploits the effects of
ubiquitination to degrade BAX, which has a role in pro-apoptosis, and
to reduce the expression of proteins STAT3 and PCBP2, which participate
in the transformation from the lytic phase to the latent phase.
EBV promotes transport to maintain the production of viral particles.
Viral BNLF2B and miR-BART1-3p exploit PRKACB to promote envelope
assembly and the intracellular transport of virions. EBV promotes tRNA
splicing of FAM98B through the degradation of TF NFATC2 via
ubiquitination to increase the expression of late lytic genes and the
genetic diversity in cell packaging, assembly, and cell transport. The
activated EBNA1 can assist virion production and indirectly promote
cell transportation of TRIM3. Viral miR-BART14 can assist CLIC5 in cell
transport by repressing the expression of AFG3L1P, and EBV can help
NRP1 activate cell transport through the degradation of STAT3 via
ubiquitination. Viral BPLF1 remains associated with the capsid—whereas
most of the tegument is detached—and has a role in the transport of the
capsid toward the plasma membrane. BALF3 interacts with BORF1 at the
second infection stage, which mainly helps EBV terminate some final
steps of viral production during lytic infection. Viral BVRF1 acts as a
checkpoint at the formation between the viral capsid and the tegument,
and is involved in EBV DNA encapsidation, maintaining an accurate DNA
genome cleavage to stabilize the capsids. These factors suggest that
EBV exploits viral proteins and miRNAs to develop its defensive
mechanism to defeat multiple immune attacks by the human immune system,
promotes virion production via viral proteins, and facilitates the
transportation of viral particles by activating the expression of
certain human genes. A better understanding of host–virus cross-talk
interactions could help in the design of new therapeutic drugs against
EBV-associated malignancies.
Network parameter-based pathway enrichment analysis and validation of HVCNs
In order to validate our results, we applied the well-proposed analysis
method to the two-sided time-course expression data to discover the top
lytic-cycle genes at the first and second infection stages. In 2012,
coexpression analysis has been applied to the expressions of EBV lytic
genes and human host genes across 201 RNA-seq experiments to identify
the potential human genes, which were upregulated by EBV lytic genes
during lytic reactivation [[216]79]. In this study, we applied
coexpression analysis to the two-sided time-course expression data
during the EBV infection to identify the top human genes (top 500
genes), positively related with first stage lytic cycle genes (upper
two panels in [217]Fig 2) or second stage lytic cycle genes (lower
panel in [218]Fig 2). By comparing top genes at first and second
infection stages with the identified 378 and 389 human core
genes/proteins /receptors/TFs in HVCNs (in Figs [219]5 and [220]6,
respectively) at the first and second infection stages during the lytic
phase, respectively, the result shows 3.6% and 2.2% consistency at the
two stages, respectively. It can be supported that only about 2% of the
coexpressed genes have related cellular functions [[221]80]. Recently,
there is still no genome-wide approach to discover the potential first
and second stage lytic cycle genes during the lytic phase.
In order to validate the network functions of the identified HVCNs
based on the network parameters, we applied network parameter-based
pathway enrichment analysis [[222]81] to the interaction parameters
[MATH:
αin(h) :MATH]
in (1) of the core human PPIs. The top five pathways and their
corresponding genes at the first and second infection stages are shown
in [223]Table 5 (full table shown in [224]S1D Table).
Table 5. Network parameter-based pathway enrichment analysis of human core
PPIs in HVCPs.
The result is sorted by the number of genes involved in the pathways.
The full table is shown in S1D Table.
First infection stage Second infection stage
Enriched Gene Sets of HVCP interacting with EBV proteins z-score Genes
Enriched Gene Sets of HVCP interacting with EBV proteins z-score Genes
NABA_MATRISOME_ASSOCIATED 1.887 BMP1, CLEC2D, CTSC, FGF2, FREM1, GH1,
IGF1, PLAU, S100A8, S100A9, SEMA4G REACTOME_IMMUNE_SYSTEM 5.172 CD40LG,
EIF2AK2, FBXO4, FOXO3, HCK, IKBKB, IL6ST, IRAK4, MALT1, MAP3K14, MAPK1,
MAPKAP1, NUP155, PCBP2, PIK3R3, PML, PRKACB, PSMA3, PTPN6, STAT3,
UBE2C, UBE3A, VCAM1
KEGG_P53_SIGNALING_PATHWAY 2.318 BAX, CDKN2A, CHEK1, IGF1, RCHY1,
SIAH1, TNFRSF10B, TP73 REACTOME_SIGNALING_BY_GPCR 5.028 CCL19, CCL3L1,
CXCL13, EGFR, GNB5, GNG5, GPBAR1, GPR65, MAPK1, MCHR1, NPFFR2, OR10A6,
OR3A2, OR9A2, PDE1B, PIK3R3, PRKACB, PTGER1
BIOCARTA_HIVNEF_PATHWAY 3.010 DAXX, MAPK8, NFKB1, RELA, TRAF1, XIAP
REACTOME_ADAPTIVE_IMMUNE_SYSTEM 5.497 CD40LG, FBXO4, FOXO3, IKBKB,
MALT1, MAP3K14, MAPKAP1, PIK3R3, PRKACB, PSMA3, PTPN6, UBE2C, UBE3A,
VCAM1
BIOCARTA_CBL_PATHWAY 1.789 GRB2, PRKCB, SH3GLB1, SH3GLB2
REACTOME_DEVELOPMENTAL_BIOLOGY 5.497 ALCAM, EGFR, FOXO3, GFRA2,
KIAA1598, MAPK1, NCOA1, NRP1, RDX, SMAD2, TCF4
PID_ER_NONGENOMIC_PATHWAY 1.768 ESR1, ESR2, GRB2, IGF1
KEGG_NEUROTROPHIN_SIGNALING_PATHWAY 3.643 ARHGDIB, BAX, FOXO3, IKBKB,
IRAK4, MAPK1, PIK3R3, TP73
[225]Open in a new tab
At the first infection stage, it has been reported that the tightly
controlled ECM homeostasis is essential for life-threatening
pathological conditions in B cells, resulted from the sustained
dysregulation [[226]82]. p53 has been known to be essential for the
regulation of miRNAs and activation of histone deacetylase (HDAC)
inhibitor (HDACi)-induced early EBV lytic infection in B cell lines
[[227]83, [228]84]. Like HIV Nef protein, EBV proteins BHRF1 and BZLF1
have been reported to protect EBV-infected cells from cell death by
inhibiting the expression of TNFR1 or FAS at the first infection stage
during the lytic phase [[229]85, [230]86]. Since it has been previously
shown that Cbl is a negative regulator of early EBV lytic induction and
promotes the degradation of LMP2A and LMP2A associated proteins
[[231]87], we suggested that the induced CBL pathway ensures the
progression of the first lytic infection to the second infection stage
in EBV-infected B cells. The activation of early lytic cycle is
associated with the dynamic interaction between human autoimmunity and
EBV proteins through estrogen signaling pathway [[232]88].
At the second infection stage, it has been proposed that late lytic
gene products, such as BCRF1 [[233]89], and G-protein-coupled receptor
(GPCR) pathway [[234]90] are involved in immune evasion and adaptive
immune evasion mechanisms, which lead to high-level resistance of
late-lytically infected B cells to nature killer cell killing
[[235]91]. In EBV-infected B cells, late lytic proteins, such as BGLF2,
are found to activate AP1-p38 MAPK signaling pathway [[236]92,
[237]93], which is associated with cell growth, apoptosis, and
differentiation. Neurotrophin and NGF signaling pathways are thought to
contribute to non-Hodgkin [[238]94] and Hodgkin lymphomas [[239]95],
respectively, which are highly associated with the patients with EBV
infection. Late lytic genes, such as BPLF1, were found to help drive B
cell immortalization and increase the incidence of lymphomagenesis
[[240]96]. Therefore, the results in HVCNs at the first and second
infection stages were supported.
Drug target proteins and multi-molecule drug design
In this section we will explore potential drug target proteins and
multi-molecule drug design, aimed at human B cells infected with the
EBV lytic phase. EBV is a γ- herpesvirus that has dual roles in its
life cycle. One is its latent infestation of human cells with only some
latent genes being expressed, and the other is reactivated lytic
infection, which contributes to new virion production and
transportation. Because most viral proteins are expressed during the
lytic phase, EBV has developed defensive mechanisms to antagonize the
human immune system. Therefore, our therapeutic strategies are aimed at
applying multi-molecule drugs to the cells infected with EBV for the
purpose of blocking reactivation to the lytic phase, interrupting the
viral production of virions, interfering with the transportation of
viral particles, and destroying viral defensive mechanisms.
The reactivation of EBV to the lytic phase provides an opportunity to
promote EBV-dependent viral cell killing; the method, called induced
lytic therapy, requires drugs and other agents that can induce EBV
reactivation without causing unacceptable cytotoxic substances to form
in normal cells. Current induced lytic strategies use a protein kinase
(PK) encoded by the EBV early lytic gene BGLF4, which can transform the
nucleoside analog ganciclovir (GCV) into cytotoxic drugs that can
promote apoptosis and kill viral cells. Phosphorylated GCV can also be
transferred to the adjacent cells via gap junctions, so the activation
of ganciclovir phosphorylation could result in the “bystander killing”
of a number of viral cells and even normal cells. Ganciclovir, the
classic anti-viral drug, can suppress the activity of viral cell DNA
polymerase and eradicate the virus-infected cells. However, ganciclovir
in the phosphorylation form can lead to much greater cell death due to
bystander killing, and can also inhibit the activity of human cell DNA
polymerase. Moreover, ganciclovir is only effective against EBV lytic
infected cells. Because EBV-positive tumor cells are mainly in the EBV
latent infection phase, ganciclovir is not useful for treating
EBV-positive tumors by itself [[241]97, [242]98].
We considered viral proteins EBNA2 and Zta to have the primary role in
the initiation of the EBV lytic phase. We forecast EBNA2 and Zta as the
potential drug targets in the progression of the EBV lytic phase. EBNA2
can efficiently upregulate the genes involved in infected cell
proliferation and survival, and it can evade human immune attacks by
interacting with CD46, as shown in [243]Fig 8A. Moreover, the switch
between the latent phase and the lytic phase of EBV infection can be
induced by the expression of the immediate–early gene product Zta,
which is a transcription factor and is capable of inducing the entire
program of EBV lytic gene expression [[244]99, [245]100]. We also
suggested that EBV membrane proteins LMP1 and LMP2B, and EBV nuclear
antigen EBNA1 are potential drug targets, because they not only
participate in reactivation from the latent phase to the lytic phase,
but also in the defensive mechanisms of virion production at both
infection stages, as shown in Figs [246]5 and [247]6, respectively.
LMP2B works in cooperation with LMP1 via viral BLLF2 to enhance the
activity of human B cells and facilitate the production and
transportation of viral particles. In addition, among all EBV-encoded
proteins, LMP1 performs the function of anti-apoptosis at both
infection stages during the lytic phase, and plays a central role in
the propagation of EBV-associated lymphoma. The conventional treatment
for EBV-associated malignancies cannot prevent tumor metastasis,
recurrence, and disease progression, so we considered that therapeutic
strategies targeting EBV-encoded proteins may increase the cure rate
and provide a clinical benefit [[248]101]. Viral protein EBNA1, another
prime target for therapeutic intervention, acts as the major switch
that regulates EBV gene activity and activates EBV dormancy in humans.
EBNA1 is essential for the virus to reproduce via anti-apoptosis.
Knocking out EBNA1 could therefore destroy EBV and manipulate the
growth of EBV-associated cancer [[249]102].
To develop anti-EBV drugs and construct the drug databases for drugs to
target EBV proteins, we began a complex screening process to discover a
small molecule that could chemically bind to viral proteins and inhibit
their abilities to perform. Thus, we carried out drug mining of the
literature to design multi-molecule drugs that were appropriate for the
potential drug targets. Ismail et al. carried out an in vitro
investigation of the activity of the potent herbal extract drug
thymoquinone in EBV [[250]103]. Thymoquinone (TQ) was tested for
cytotoxicity in human Burkitt’s lymphoma cells and certain other
EBV-related lymphomas. Thymoquinone was found to efficiently inhibit
the RNA expression of viral genes EBNA2, LMP1, and EBNA1. In
particular, the optimal EBNA2 expression level indicated that EBNA2
might make the main contribution to thymoquinone potency against
EBV-infected cells [[251]103]. The researcher’s results suggest that
thymoquinone has the potential to efficiently inhibit the growth of
EBV-infected B cells. Valpromide (VPM), which is not an HDAC inhibitor,
is another drug that can block EBV reactivation. VPM can prevent the
gene expression of viral BZLF1, which mediates lytic reactivation. VPM
cannot activate the expression of some cellular immediate–early genes
including FOS and EGR1, which are upstream of the EBV lytic cycle, but
it can reduce their activities. Thus, VPM can selectively suppress both
viral and cellular gene expression. VPM represents a new class of
antiviral agents that prevent the initiation of the EBV lytic phase.
VPM will be useful in exploring the mechanism of EBV lytic reactivation
and may have therapeutic potential [[252]104]. Zebularine (Zeb) is a
DNA methyltransferase inhibitor (DNMTi) that induces the expression of
E-cadherin, which is encoded by a cellular gene that is frequently
silenced by hypermethylation in cancers. Zebularine can decrease the
upregulation of viral genes LMP2A, LMP2B, and EBNA2 to prevent the
switch from the latent phase to the lytic phase upon cross-linking with
lytic inducers such as B-cell receptors. Zebularine could also be used
to treat EBV-associated tumors, because it does not induce the switch
from the latent phase to the lytic phase that may cause secondary
EBV-related malignancies [[253]105]. Consequently, we integrated these
drugs as the potential multi-molecule drugs, as shown in [254]Fig 10A,
for the predicted drug targets. We considered that the multi-molecule
drugs can inhibit the activities of viral proteins and thus play a
potent role in counteracting the reactivation from latency to the lytic
phase during the EBV lytic infection of human B cells. They are
therefore suitable candidates for further development as inhibitors of
EBV-associated malignancies.
Fig 10.
[255]Fig 10
[256]Open in a new tab
(A) Multi-molecule drug design for the predicted drug targets; (B)
Protein structures of EBV interleukin-10 and human interleukin-10.
Thymoquinone (TQ) inhibits the RNA expression of viral EBNA2, LMP1, and
EBNA1; valpromide (VPM) can prevent the gene expression of viral BZLF1;
zebularine (Zeb) can decrease the upregulation of viral LMP2A, LMP2B,
and EBNA2. These drugs can block reactivation, interrupt the viral
production of virions, interfere with the transportation of viral
particles, and destroy viral defensive mechanisms.
Moreover, the hijack of epigenetic regulation by EBV is essential for
the establishment of life-long latency and the facilitation of
malignant B cell transformation. Understanding how specific epigenetic
regulations promote the progression of lymphomas, autoimmunity, and
EBV-associated cancers is fundamental to the design of novel
therapeutic interventions for the cure of certain often fatal symptoms
[[257]106]. Thus, from the HVCPs in Figs [258]7 and [259]8, we can
predict other drug targets comprising EBV proteins and the four viral
miRNAs. Certain viral proteins (BALF4, EBNA3B, and BDLF4) may block the
autophagy mechanism and hijack the autophagic vesicles for virion
transportation; BNRF1 can inhibit human intrinsic defenses to enhance
the activation and transcription of early viral lytic genes; BNLF2A can
reduce the effect of antigen-presenting cells, and thereby evade T cell
immune responses; BLLF2 mediates the interaction between EBV membrane
proteins LMP1 and LMP2B, and may participate in the hijack of the
autophagy mechanism; BNLF2B is similar to viral BCRF1, and BNLF2B and
BCRF1 can perform the function of immune evasion to protect EBV from
human immune surveillance. Therefore, novel multi-molecule drugs based
on targets including EBV proteins and, particularly, EBV miRNAs, will
interrupt the development of the EBV lytic phase and facilitate the
rapid detection of viral lytic proteins.
Furthermore, EBV can exploit certain viral-encoded proteins that are
homologues of human proteins to interact with human receptors with the
objective of evading and even inhibiting the human immune response. We
carried out analyses to determine the structural similarities between
viral BCRF1 and human cellular IL10 using the Protein Data Bank in
Europe (PDBe) database and its tool, PDBeFold. The structure alignment
results show 94% amino acid sequence identity and 71% secondary
structure identity between vIL-10 and hIL-10. As shown in [260]Fig 10B,
viral BCRF1 is a homologue of human IL10. It is a functionally critical
cytokine regulator of immune tolerance, and interacts with human
receptor IL10RA in the second stage of infection shown in [261]Fig 6.
This potentially crucial form of EBV immune modulation and immune
suppression is due to EBV-encoded proteins called virokines, which are
similar to human cytokines [[262]107]. It may become very important to
discover more interactions between EBV homologues and human proteins.
Another strategy is to induce the transformation from the latent form
of infection to the lytic phase, thereby causing EBV-associated cell
death. However, the period of latency in the EBV life cycle is much
longer than the period of lytic reactivation and production, and the
expression of latent EBV proteins is limited; these two aspects of
latency may affect the accuracy and effectiveness of multi-molecule
drugs. Drugs that affect the EBV latent phase can prevent the mutations
in EBV-infected cells that cause EBV-associated malignancies, such as
EBV-positive tumor cells, which are primarily initiated in the latent
form of EBV infection.
Conclusion
The World Health Organization (WHO) has defined EBV as a Class I
carcinogen, and it is estimated to result in a small but significant
portion of all human cancers. EBV may persist in the human body for
decades and cause infected cells to become cancerous. It is estimated
that EBV leads to nearly 400,000 cases of cancer each year, including
Burkitt’s lymphoma, Hodgkin’s lymphoma, gastric carcinoma, and
nasopharyngeal carcinoma. To persist in human B cells, EBV has evolved
various strategies to manipulate the human immune response, including
restricting immune cell functions, blocking apoptotic pathways, and
interfering with antigen presentation functions. In this study, we
investigated the pathogenic mechanisms by which dysregulations and
dysfunctions of human B cells and immune modulation by EBV can
contribute to the development of EBV lytic replication and production.
The reactivation of EBV-infected human B cells into the lytic phase
initiates viral replication and infectious virion production. At the
same time, a number of EBV lytic genes are successively activated and
expressed, so the human immune system can detect and respond to
infection-associated viral proteins. Viral nuclear antigen EBNA2 can
impair human immune information by interacting with receptor CD46,
which then promotes the proliferation of infected cells and increases
the transcription of the viral immediate–early lytic gene BZLF1. Zta
can activate early lytic genes and induce viral anti-apoptosis. EBV can
exploit the epigenetic changes due to ubiquitination and acetylation to
block the pro-apoptotic pathway via viral miR-BART1-3p, which can
inhibit the expression of NAT1 and mediate human proteins and genes
relevant to pro-apoptosis that are subjected to DNA methylation,
acetylation, and ubiquitination. The human body’s natural response is
to exploit the autophagy mechanism to eliminate EBV proteins. However,
EBV blocks autophagy and hijacks the autophagic vesicles to facilitate
viral transport. The human body uses CHMP5 and lncRNA SNHG5 to induce
pro-apoptosis and interrupt viral translocation, respectively. The
epigenetic effects of ubiquitination, deacetylation, and DNA
methylation influence these pathways, and viral BNRF1 destroys the DAXX
complex in the pro-apoptotic pathway. Thus, these host/virus mechanisms
contribute to the complete progression from lytic production through to
the impairment of pro-apoptosis and the promotion of viral
translocation and anti-apoptosis. In the first stage of infection, EBV
protects the infected human B cells from elimination by autophagy and
keeps the infected B cells in the lytic phase until the integrated
production and transportation of new infectious virions is complete.
Thus, EBV inhibits and hijacks the function of autophagy via BECN1, and
silences the expression of STAT3, which can repress the expression and
activity of lytic genes upon STAT3 in the activated status. These
defensive mechanisms of EBV result in the persistence of lytic
infection and the promotion of viral production. Viral membrane
proteins LMP1, BNLF2B, and miR-BART1-3p can perform the function of
anti-apoptosis, and viral BCRF1 enables immunosuppression by releasing
viral IL10 to escape immune surveillance. Furthermore, EBV exploits
transcriptionally activated FAM98B, which can increase tRNA splicing,
thereby creating more genetic diversity in the proteins involved in the
packaging, assembly, and transportation of new viral particles. The
disruption of PML by viral EBNA1 contributes to the inhibition of
pro-apoptosis. EBNA1 has the ability to promote virion production, and
indirectly influences vesicle trafficking and virion release by TRIM3
and viral membrane protein LMP1. It is necessary to maintain virion
transportation by decreasing the repression of the lytic cycle and
increasing anti-apoptosis. The epigenetic modifications of
ubiquitination and DNA methylation, and the repression of PCBP2 by
viral miR-BART10 decrease the ability of PCBP2 to inhibit the lytic
phase, and viral miR-BART14 represses the expression of lncRNA AFG3L1P,
which can trigger pro-apoptosis. EBV indirectly exploits the expressed
NRP1 and CLIC5 to promote the transportation of new virions.
Real GIGENs, HVCNs, and HVCPs obtained through dynamic system models,
big database mining, and NGS data provide a new perspective. Compared
with literature reviews, the approach reported here provides more
information on interspecies and intraspecies protein–protein
interactions and signaling transduction pathways from receptors to TFs.
It also provides information on the transcription regulation of target
genes in the context of epigenetic miRNAs and lncRNAs regulation. Here,
we propose a multi-molecule drug design based on our exploration of the
EBV drug target proteins from HVCNs. These multi-molecule drugs
([263]Fig 10A) can inhibit reactivation from the latent form to the
lytic form in the EBV life cycle. They can also reduce the abilities of
some critical EBV lytic genes/proteins, which can reactivate the lytic
phase, the viral production of virions, the transportation of viral
particles, and EBV defensive mechanisms during lytic infection. In the
future, we hope to discover and collect more information about lncRNAs
and open reading frames (ORF), and predict links between humans and EBV
to improve on the models and results reported in this study. Thus, it
may be possible to discover other novel cross-talk mechanisms between
humans and EBV, as well as further EBV pathogenic mechanisms occurring
during lytic reactivation, lytic infection, and even the latent phase
and its associated carcinogenesis.
Materials and methods
Overview of the construction for interspecies GIGENs in human B cells
infected with EBV during the lytic production phase
A flow chart of the progression for constructing the GIGENs, the HVCNs,
and the HVCPs in human B cells infected with EBV at the first and
second infection stages in the lytic phase is shown in [264]Fig 1. The
GIGENs were composed of human/EBV gene/miRNA/lncRNA regulatory networks
(GRNs), human/EBV protein–protein interaction networks (PPINs), the
interspecies PPINs, and the interspecies GRNs. The GIGENs, the HVCNs,
and the HVCPs were constructed in the following steps: (1) big data
mining and data preprocessing to establish the candidate GIGEN; (2)
identification of the real GIGENs by pruning false positives from the
candidate GIGEN using the system identification approach and the system
order detection scheme with the genome-wide NGS data for human B cells
and EBV at the first and second stages during the EBV lytic reactivated
infection; and (3) extraction of the HVCNs by applying the principal
network projection (PNP) method to the real GIGENs at the first and
second infection stages. These procedures were used to identify the
crucial and specific interspecies mechanisms at both infection stages
during the EBV lytic phase.
Big data mining and data preprocessing of NGS data for humans and EBV, and
DNA methylation profiles for humans
NGS datasets were obtained from the Gene Expression Omnibus (GEO) at
the National Center for Biotechnology Information (NCBI). A study by
Tina O'Grady et al. demonstrated that EBV reactivation includes the
ordered induction of approximately 90 viral genes that are involved in
the production of infectious virions[[265]7]. They found extensive
bidirectional transcription stretching across nearly the entire genome,
and estimated that probably hundreds more EBV genes are expressed
during EBV reactivation than was previously thought. They also
suggested that the viral genome during EBV reactivation might be much
more complex than had been suspected, and changed our view of the
virion production process during the EBV lytic phase. We obtained the
NGS data from this study containing both the human (hg19 assembly) and
the Akata EBV genomes with the time course through the GEO series with
accession number [266]GSE52490[[267]7]. The raw data of the NGS dataset
comprises two parts. One part involves the gene expression profiles of
human B cells with EBV lytic infection at 0, 1/12, 0.5, 1, 2, 4, 8, 24,
and 48 hours post reactivation. The other part involves the gene
expression profiles of EBV in the lytic phase at 0, 1/12, 0.5, 1, 2, 4,
8, 24, and 48 hours post reactivation. It contains 44,446 human probes
and 134 EBV probes. According to the gene expression profiles of the
viral IE lytic genes BZLF1 and BRLF1 ([268]Fig 2, top panel), the early
lytic genes BMRF1, BBLF3, BBLF4, BGLF5, BNLF2A and BSLF1 ([269]Fig 2,
middle panel) and the late viral genes BCRF1, BVRF2, BDLF1, BLLF1 and
BCLF1 ([270]Fig 2, bottom panel) during the infection, we classified
the lytic phase into the first infection stage from 0 to 24 hours,
where the IE lytic genes and the early lytic genes are highly
expressed, and the second infection stage from 8 to 72 hours, where the
early lytic genes and the late viral genes are highly expressed. We
subsequently applied analysis of variance (ANOVA) to the NGS data for
human and EBV mRNA expression to evaluate the p-value for the
differential expression data for the first and second infection stages.
The candidate GIGEN was constructed through big data mining from
numerous databases that contain many experimental data and
bioinformatic (computational) predictions. The human candidate PPIN was
obtained from BioGRID[[271]108], DIP[[272]109], BIND[[273]110],
IntAct[[274]111], and VirusMINT[[275]112]. The human candidate GRN
comprised transcription factors (TFs)/ TF complex-regulating genes,
lncRNA-regulating genes, and miRNA-repressing genes, which were
available at HTRIdb[[276]113], ITFP[[277]114], TargetScan
([278]http://www.targetscan.org/), and CircuitsDB 2[[279]115]. The
interspecies candidate PPIN and EBV candidate PPIN required
interspecies and intraspecies interactions, which were obtained from
VirusMentha[[280]116], CDFD,
([281]http://www.cdfd.org.in/labpages/computational_biology.html),
Virhostome[[282]117], IMEx[[283]118], and PSICQUIC[[284]119]. The
interspecies candidate GRN and EBV candidate GRN involved interspecies
and intraspecies TF-regulating genes and miRNA-repressing genes that
were collected from VIRmiRNA[[285]120], ViRBase[[286]121],
miRecords[[287]122], starBase v2.0[[288]123], and miRTarBase[[289]124].
To support the inference of human target genes that are subjected to
epigenetic regulation of DNA methylation based on the results of system
identification, we exploited the genome-wide DNA methylation profiles
of B cells and immortalized B cells ([290]GSE41957)[[291]125] that were
uninfected and infected with EBV, respectively (with a sample size of
6). We applied the ANOVA statistics to these DNA methylation data.
As a result, in the intraspecies candidate PPIN, we obtained 301 EBV
PPI pairs and 23,570,918 human PPI pairs; in the interspecies candidate
PPIN, we obtained 5,135 human-EBV PPI pairs. In the intraspecies
candidate GRN, we obtained 5 EBV TF-gene pairs, 67 EBV miRNA-gene
pairs, 906,611 human TF-gene pairs, 817,900 human miRNA-gene pairs, and
1,948 human lncRNA-gene pairs; in the interspecies candidate GRN, we
obtained 1,252 EBV TF-human gene pairs, 39,772 EBV miRNA-human gene
pairs, 1,355 human TF-EBV gene pairs, and 1,718 human miRNA-EBV gene
pairs. Among the intraspecies human candidate GRNs there are three
human TF complexes. The first is ARNT::AHR, which has 6,368 human
TF-gene pairs; the second is HIF1A::ARNT, which has 1,011 human TF-gene
pairs; and the third is NFE2L1::MAFG, which has 5,787 human TF-gene
pairs. In conclusion, we built a candidate GIGEN comprising the various
candidate pairs mentioned above, and we then detected the real GIGENs
by pruning the false positives from the corresponding candidate GIGEN
via the system identification approach and the system order detection
scheme using the genome-wide NGS data for human B cells and EBV at both
stages of infection during the EBV lytic phase.
Dynamic models of the GIGEN for human B cells and EBV during the lytic
infection process
The candidate GIGEN comprised the experimental and computational
predictions, which would have resulted in a number of false-positive
interactions and regulations. It was, therefore, necessary to prune the
false positives of the candidate GIGEN to construct real GIGENs using
the genome-wide NGS data at the first and second infection stages for
human B cells and EBV through the system identification approach and
the system order detection scheme. We then extracted the core GIGENs
using the PNP scheme to characterize the principal biological
mechanisms of the GIGENs.
The PPI of human-protein i in the candidate PPIN can be described by
the following stochastic dynamic equation:
[MATH: pi(h)(t+1)=pi(h)(t)+∑n=1Niαin(h)pn(h)(t)pi(h)(t)+∑j=1Jiγij(h)pj(v)(t)pi(h)(t)−σi(h)pi(h)(t)+λi(h)gi(h)(t)+βi(h)+εi(h)(t),fori=1,2,…,I,−σi(h)≤0,<
mspace width="0.25em">andλi(h)≥0 :MATH]
(1)
where
[MATH: pi(h)(t) :MATH]
,
[MATH: pn(h)(t) :MATH]
,
[MATH: gi(h)(t) :MATH]
, and
[MATH: pj(v)(t) :MATH]
indicate the expression levels of human-protein i, human-protein n,
human-gene i, and EBV-protein j at time t, respectively;
[MATH:
αin(h) :MATH]
and
[MATH:
γij(h) :MATH]
represent the interactive abilities between human-protein n and
human-protein i, and between EBV-protein j and human-protein i,
respectively;
[MATH: ‑σi(h) :MATH]
,
[MATH: λi(h) :MATH]
, and
[MATH: βi(h) :MATH]
denote the degradation rate, the translation effect, and the basal
level of human-protein i, respectively. The basal level
[MATH: βi(h) :MATH]
denotes interactions with unknown factors, for example, acetylation and
ubiquitination. N[i] and J[i] represent the numbers of human proteins
and EBV proteins interacting with human-protein i in the candidate
GIGEN, respectively; and
[MATH: εi(h)(t)
:MATH]
is the stochastic noise of human-protein i owing to model uncertainty
or other uncertain factors at time t. Note that the biological
interaction mechanism of human proteins in (1) involves the
intraspecies human PPIs represented by
[MATH: ∑n=1Niαin
(h)pn(h)(t)pi(h)(t)
:MATH]
, and the interspecies PPIs represented by
[MATH: ∑j=1Jiγij
(h)pj(v)(t)pi(h)(t)
:MATH]
.
The PPI of EBV-protein j in the candidate PPIN can be described by the
following stochastic dynamic equation:
[MATH: pj(v)(t+1)=pj(v)(t)+∑m=1Mjαjm(v)pm(v)(t)pj(v)(t)+∑i=1Ijγji(v)pi(h)(t)pj(v)(t)−σj(v)pj(v)(t)+λj(v)gj(v)(t)+βj(v)+εj(v)(t),forj=1,2,…,J,−σj(v)≤0,<
mspace width="0.25em">andλj(v)≥0 :MATH]
(2)
where
[MATH: pj(v)(t) :MATH]
,
[MATH: pm(v)(t) :MATH]
,
[MATH: gj(v)(t) :MATH]
, and
[MATH: pi(h)(t) :MATH]
represent the expression levels of EBV-protein j, EBV-protein m,
EBV-gene j, and human-protein i at time t, respectively;
[MATH:
αjm(v) :MATH]
and
[MATH:
γji(v) :MATH]
show the interactive abilities between EBV-protein m and EBV-protein j,
and between human-protein i and EBV-protein j, respectively;
[MATH: ‑σj(v) :MATH]
,
[MATH: λj(v) :MATH]
, and
[MATH: βj(v) :MATH]
correspond to the degradation rate, the translation effect, and the
basal level of EBV-protein j, respectively; M[j] and I[j] represent the
numbers of EBV proteins and human proteins interacting with EBV-protein
j in the candidate GIGEN, respectively; and
[MATH: εj(v)(t)
:MATH]
is the stochastic noise of EBV-protein j owing to model uncertainty or
other uncertain factors at time t. Note that the biological interaction
mechanism of EBV proteins in (2) involves the intraspecies EBV PPIs
represented by
[MATH: ∑m=1Miαjm
(v)pm(v)(t)pj(v)(t)
:MATH]
, and the interspecies PPIs represented by
[MATH: ∑i=1Iiγji
(v)pi(h)(t)pj(v)(t)
:MATH]
.
The GRN of human-gene k in the candidate GRN can be described by the
following stochastic dynamic equation:
[MATH: gk(h)(t+1)=gk(h)(t)+∑i=1Ikaki(h)pi(h)(t)+∑i'=1Ik'∑i''=1Ik''ζk(Ik''(i'−1)+i'')(h)pi<
/mi>'(h)(t)pi''(h)(t)
−∑r=1Rkbkr(h)wr(h)(t)gk(h)(t)+∑l=1Lkckl(h)ol(h)(t)+∑j=1Jkdkj(h)pj(v)(t)−∑q=1Qkekq(h)wq(v)(t)gk(h)(t)−μk(h)gk(h)(t)+δk(h)+ωk(h)(t),fork=1,2,…,K,−bkr(h)≤0,<
mspace
width="0.25em">−ekq(h)≤0,<
mspace width="0.25em">and−μk(h)≤0 :MATH]
(3)
where
[MATH: gk(h)(t)
:MATH]
,
[MATH: pi(h)(t)
:MATH]
,
[MATH:
pi'(h)(t)pi
''(h)(t)
:MATH]
,
[MATH: wr(h)(t)
:MATH]
,
[MATH: ol(h)(t)
:MATH]
,
[MATH: pj(v)(t)
:MATH]
, and
[MATH: wq(v)(t)
:MATH]
indicate the expression levels of human-gene k, human-TF i, human-TF
complex i'::i'', human-miRNA r, human-lncRNA l, EBV-TF j, and EBV-miRNA
q at time t, respectively; human-TF complex i'::i'' is composed of
human-TF i' and human-TF i'';
[MATH:
aki(h) :MATH]
,
[MATH: ζk(Ik''(i'−1)+i'')(h) :MATH]
,
[MATH:
‑bkr(h) :MATH]
,
[MATH:
ckl(h) :MATH]
,
[MATH:
dkj(h) :MATH]
, and
[MATH:
‑ekq(h) :MATH]
represent the regulatory abilities of human-TF i regulation, human-TF
complex i'::i'' regulation, human-miRNA r repression, human-lncRNA l
regulation, EBV-TF j regulation, and EBV-miRNA q repression on
human-gene k, respectively; and
[MATH: ‑μk(h) :MATH]
and
[MATH: δk(h) :MATH]
denote the degradation rate and the basal level of human-gene k,
respectively. Remarkably, regarding the regulation ability
[MATH: ζk(Ik''(i'−1)+i'')(h) :MATH]
of the human TF complex on the human gene k, the index
[MATH:
Ik‘’(i'−1)+i''
:MATH]
assures the appropriate coordinate of the regulation ability
[MATH: ζk(Ik''(i'−1)+i'')(h) :MATH]
of the human TF complex
[MATH:
pi’(h)(t)pi’’(h)(t)
:MATH]
in the human GRN of the system matrix of the human gene k, i.e., the
regulation abilities of human TF complexes on the human gene k can be
arranged as a one-row matrix as follows:
[MATH:
ζk1(h),ζk2(h),…,ζk(Ik'')(h),ζk(Ik''+1)(h),ζk(Ik''+2)(h),…,ζk(2Ik'')(h),ζk(2Ik''+1)(h),ζk(2Ik''+2)(h),…,ζk(3Ik'')(h),…,ζk(Ik''(i'−1)+i'')(h),…,ζk(Ik''Ik')(h) :MATH]
The basal level
[MATH: δk(h) :MATH]
denotes regulations from other unknown regulators. I[k],
[MATH: Ik’
:MATH]
,
[MATH:
Ik’’ :MATH]
, R[k], L[k], J[k], and Q[k] represent the numbers of human TFs, human
TF complex subunit i', human TF complex subunit i'', human miRNAs,
human lncRNAs, EBV TFs, and EBV miRNAs regulating human-gene k in the
candidate GIGEN, respectively; and
[MATH: ωk(h)(t)
:MATH]
is the stochastic noise of human-gene k owing to model uncertainty or
other uncertain factors at time t—for example, methylation and histone
modification. Note that the biological regulatory mechanism of human
genes in (3) involves human-TF transcription regulations represented by
[MATH: ∑i=1Iaki(h)pi(h)(t) :MATH]
, human-TF complex transcription regulations represented by
[MATH: ∑i’=<
/mo>1I’∑i’’<
/mo>=1I’’<
/mrow> :MATH]
[MATH: ζk(Ik''(i'−1)+i'')(h) :MATH]
[MATH:
pi’(h)(t)pi’’(h)(t) :MATH]
, human-miRNA repressions represented by
[MATH: ‑∑r=1Rbkr(h)wr(h)(t)gk(h)(t)
:MATH]
, human-lncRNA regulations represented by
[MATH: ∑l=1Lckl(h)ol(h)(t) :MATH]
, EBV-TF transcription regulations represented by
[MATH: ∑j=1Jdkl(h)pj(v)(t) :MATH]
, and EBV-miRNA repressions represented by
[MATH: ‑∑q=1Qekq(h)wq(v)(t)gk(h)(t) :MATH]
. It seems reasonable to suppose that DNA methylation may have a robust
connection with gene expression changes and be associated with the
transcriptional activity of human genes. DNA methylation influences the
dynamics and stability of RNA polymerase II elongation, so that
intragenic DNA methylation coordinates differential gene expression via
alternative promoters or splicing. Thus, we supposed that the
differential changes of basal level
[MATH: δk(h) :MATH]
of human-gene k between the first and second infection stage in Eq
([292]3) were mainly due to DNA methylation during the EBV lytic
infection.
The GRN of EBV-gene s in the candidate GRN can be described by the
following stochastic dynamic equation:
[MATH: gs(v)(t+1)=gs(v)(t)+∑i=1Isasi(v)pi(h)(t)+∑i'=1Is'∑i''=1Is''ζs(Is''(i'−1)+i'')(v)pi<
/mi>'(h)(t)pi''(h)(t)
−∑r=1Rsbsr(v)wr(h)(t)gs(v)(t)+∑l=1Lscsl(v)ol(h)(t)+∑j=1Jsdsj(v)pj(v)(t)−∑q=1Qsesq(v)wq(v)(t)gs(v)(t)−μs(v)gs(v)(t)+δs(v)+ωs(v)(t),
fors=1,2,…,S,−bsr(v)≤0,<
mspace
width="0.25em">−esq(v)≤0,<
mspace width="0.25em">and−μs(v)≤0 :MATH]
(4)
where
[MATH: gs(v)(t)
:MATH]
signifies the expression level of EBV-gene s at time t;
[MATH:
asi(v) :MATH]
,
[MATH: ζs(Is''(i'−1)+i'')(v) :MATH]
,
[MATH:
‑bsr(v) :MATH]
,
[MATH:
csl(v) :MATH]
,
[MATH:
dsj(v) :MATH]
, and
[MATH:
‑esq(v) :MATH]
show the regulatory abilities of human-TF i regulation, human-TF
complex i'::i'' regulation, human-miRNA r repression, human-lncRNA l
regulation, EBV-TF j regulation, and EBV-miRNA q repression on EBV-gene
s, respectively;
[MATH: ‑μs(v) :MATH]
and
[MATH: δs(v) :MATH]
correspond to the degradation rate and the basal level of EBV-gene s,
respectively; and
[MATH: ωs(v)(t)
:MATH]
is the stochastic noise of EBV-gene s owing to model uncertainty or
other uncertain factors at time t. Note that the biological regulatory
mechanism of EBV genes in (4) involves human-TF transcription
regulations represented by
[MATH: ‑∑i=1Isasi
(v)pi(h)(t)
:MATH]
, human-TF complex transcription regulations represented by
[MATH: ∑i’=1Is’∑i’’−1Is’’ :MATH]
[MATH: ζs(Is''(i'−1)+i'')(v) :MATH]
[MATH:
pi’(h)(t)pi’’(h)(t)
:MATH]
, human-miRNA repressions represented by
[MATH: ‑∑r=1Rsbsr
(v)wr(h)(t)gs(v)(t)
:MATH]
, human-lncRNA regulations represented by
[MATH: ∑l=1Lscsl
(v)ol(h)(t) :MATH]
, EBV-TF transcription regulations represented by
[MATH: ∑j=1Jsdsj
(v)pj(v)(t)
:MATH]
, and EBV-miRNA repressions represented by
[MATH: ‑∑q=1Qsesq
(v)wq(v)(t)gs(v)(t)
:MATH]
.
The GRN of human-lncRNA z in the candidate GRN can be described by the
following stochastic dynamic equation:
[MATH: gz(L)(t+1)=gz(L)(t)+∑i=1Izazi(L)pi(h)(t)+∑i'=1Iz'∑i''=1Iz''ζz(Iz''(i'−1)+i'')(L)pi<
/mi>'(h)(t)pi''(h)(t)
−∑r=1Rzbzr(L)wr(h)(t)gz(L)(t)+∑l=1Lzczl(L)ol(h)(t)+∑j=1Jzdzj(L)pj(v)(t)−∑q=1Qzezq(L)wq(v)(t)gz(L)(t)−μz(L)gz(L)(t)+δz(L)+ωz(L)(t),
forz=1,2,…,Z,−bzr(L)≤0,<
mspace
width="0.25em">−ezq(L)≤0,<
mspace width="0.25em">and−μz(L)≤0 :MATH]
(5)
where
[MATH: gz(L)(t)
:MATH]
stands for the expression level of human-lncRNA z at time t;
[MATH:
azi(L) :MATH]
,
[MATH: ζz(Iz''(i'−1)+i'')(L) :MATH]
,
[MATH:
‑bzr(L) :MATH]
,
[MATH:
czl(L) :MATH]
,
[MATH:
dzj(L) :MATH]
, and
[MATH:
‑ezq(L) :MATH]
indicate the regulatory abilities of human-TF i regulation, human-TF
complex i'::i'' regulation, human-miRNA r repression, human-lncRNA l
regulation, EBV-TF j regulation, and EBV-miRNA q repression on
human-lncRNA z, respectively;
[MATH: ‑μz(L) :MATH]
and
[MATH: δz(L) :MATH]
denote the degradation rate and the basal level of human-lncRNA z,
respectively; and
[MATH: ωz(L)(t)
:MATH]
is the stochastic noise of human-lncRNA z owing to model uncertainty or
other uncertain factors at time t. Note that the biological regulatory
mechanism of human lncRNAs in (5) involves human-TF transcription
regulations represented by
[MATH: ∑i=1Izazi
(L)pi(h)(t) :MATH]
, human-TF complex transcription regulations represented by
[MATH: ∑i’=<
/mo>1Iz’∑i’’<
/mo>=1Iz'' :MATH]
[MATH: ζz(Iz''(i'−1)+i'')(L) :MATH]
[MATH:
pi’(h)(t)pi’’(h)(t) :MATH]
, human-miRNA repressions represented by
[MATH: −∑r=1Rzbzr
(L)wr(h)(t)gz(L)(t) :MATH]
, human-lncRNA regulations represented by
[MATH: ∑l=1<
/mn>Lzczl(L)ol(h)(t) :MATH]
, EBV-TF transcription regulations represented by
[MATH: ∑j=1Jzdzj
(L)pj(v)(t) :MATH]
, and EBV-miRNA repressions represented by
[MATH: −∑q=1Qzezq
(L)wq(v)(t)gz(L)(t)
:MATH]
.
The GRN of human-miRNA f in the candidate GRN can be described by the
following stochastic dynamic equation:
[MATH: xf(h)(t+1)=xf(h)(t)+∑i=1Ifa¯fi(h)pi(h)(t)+∑i'=1If'∑i''=1If''ζ¯f(If''(i'−1)+i'')(h)pi<
/mi>'(h)(t)pi''(h)(t)
−∑r=1Rfb¯fr(h)wr(h)(t)xf(h)(t)+∑j=1Jfd¯fj(h)pj(v)(t)−∑q=1Qfe¯fq(h)wq(v)(t)xf(h)(t)−ρf(h)xf(h)(t)+ηf(h)+ψf(h)(t),
forf=1,2,…,F,−b¯fr(h)≤0,<
mspace width="0.25em">−e¯fq(h)≤0,<
mspace width="0.25em">and−ρf(h)≤0 :MATH]
(6)
where
[MATH: xf(h)(t) :MATH]
represents the expression level of human-miRNA f at time t;
[MATH: a¯fi(h) :MATH]
,
[MATH: ζ¯f(If''(i'−1)+i'')(h) :MATH]
,
[MATH: −b¯fr(h) :MATH]
,
[MATH: d¯fj(h) :MATH]
, and
[MATH: −e¯fq(h) :MATH]
mean the regulatory abilities of human-TF i regulation, human-TF
complex i'::i'' regulation, human-miRNA r repression, EBV-TF j
regulation, and EBV-miRNA q repression on human-miRNA f, respectively;
[MATH: ‑ρf(h) :MATH]
and
[MATH: ηf(h) :MATH]
signify the degradation rate and the basal level of human-miRNA f,
respectively; and
[MATH: ψf(h)(t) :MATH]
is the stochastic noise of human-miRNA f owing to model uncertainty or
other uncertain factors at time t. Note that the biological regulatory
mechanism of human miRNAs in (6) involves human-TF transcription
regulations represented by
[MATH: ∑i=1<
/mn>Ifa¯fi(h)pi(h)(t) :MATH]
, human-TF complex transcription regulations represented by
[MATH: ∑i'=<
/mo>1If'∑i''<
/mo>=1If''<
mover accent="true">ζ¯f(If''(i'−1)+i'')(h)pi<
/mi>'(h)(t)pi''(h)(t) :MATH]
, human-miRNA repressions represented by
[MATH: −∑r=1<
/mn>Rfb¯fr(h)wr(h)(t)xf(h)(t) :MATH]
, EBV-TF transcription regulations represented by
[MATH: ∑j=1<
/mn>Jfd¯fj(h)pj(v)(t) :MATH]
, and EBV-miRNA repressions represented by
[MATH: −∑q=1<
/mn>Qfe¯fq(h)wq(v)(t)xf(h)(t) :MATH]
.
The GRN of EBV-miRNA u in the candidate GRN can be described by the
following stochastic dynamic equation:
[MATH: xu(v)(t+1)=xu(v)(t)+∑i=1Iua¯ui(v)pi(h)(t)+∑i'=1Iu'∑i''=1Iu''ζ¯u(Iu''(i'−1)+i'')(v)pi<
/mi>'(h)(t)pi''(h)(t)
−∑r=1Rub¯ur(v)wr(h)(t)xu(v)(t)+∑j=1Jud¯uj(v)pj(v)(t)−∑q=1Que¯uq(v)wq(v)(t)xu(v)(t)−ρu(v)xu(v)(t)+ηu(v)+ψu(v)(t),
foru=1,2,…,U,−b¯ur(v)≤0,<
mspace width="0.25em">−e¯uq(v)≤0,<
mspace width="0.25em">and−ρu(v)≤0 :MATH]
(7)
where
[MATH: xu(v)(t)
:MATH]
shows the expression level of EBV-miRNA u at time t;
[MATH: a¯ui(v) :MATH]
,
[MATH: ζ¯u(Iu''(i'−1)+i'')(v) :MATH]
,
[MATH: −b¯ur(v) :MATH]
,
[MATH: d¯uj(v) :MATH]
, and
[MATH: −e¯uq(v) :MATH]
correspond to the regulatory abilities of human-TF i regulation,
human-TF complex i'::i'' regulation, human-miRNA r repression, EBV-TF j
regulation, and EBV-miRNA q repression on EBV-miRNA u, respectively;
[MATH: ‑ρu(v) :MATH]
and
[MATH: ηu(v) :MATH]
stand for the degradation rate and the basal level of EBV-miRNA u,
respectively; and
[MATH: ψu(v)(t)
:MATH]
is the stochastic noise of EBV-miRNA u owing to model uncertainty or
other uncertain factors at time t. Note that the biological regulatory
mechanism of EBV miRNAs in (7) involves human-TF transcription
regulations represented by
[MATH: ∑i=1<
/mn>Iua¯ui(v)pi(h)(t) :MATH]
, human-TF complex transcription regulations represented by
[MATH: ∑i'=<
/mo>1Iu'∑i''<
/mo>=1Iu''<
mover accent="true">ζ¯u(Iu''(i'−1)+i'')(v)pi<
/mi>'(h)(t)pi''(h)(t) :MATH]
, human-miRNA repressions represented by
[MATH: −∑r=1<
/mn>Rub¯ur(v)wr(h)(t)xu(v)(t) :MATH]
, EBV-TF transcription regulations represented by
[MATH: ∑j=1<
/mn>Jud¯uj(v)pj(v)(t) :MATH]
, and EBV-miRNA repressions represented by
[MATH: −∑q=1<
/mn>Que¯uq(v)wq(v)(t)xu(v)(t) :MATH]
.
System identification approach for the dynamic models of GIGEN
After establishing the stochastic dynamic model Eqs ([293]1)–([294]7)
for the characterization of the molecular mechanism in the GIGEN, we
identified the interactive parameters of PPIN in (1) and (2), and the
regulatory parameters of GRN in (3)–(7) using the system identification
approach to solve the parameter estimation problems for the purpose of
pruning the false positives under infection conditions. Thus, we
rewrote human PPIN Eq ([295]1) as the following linear regression:
[MATH: pi(h)(t+1)=[p1(h)(t)pi(h)(t)⋯pNi(h)(t)pi(h)(t)p1
(v)(t)pi(h)(t)
⋯pJi(v)(t)pi(h)(t)gi
(h)(t)pi
(h)(t)1][
αi1(h)⋮αi
Ni(h)γi1(h)⋮γi
Ji(h)λi(h)1−σi(h)βi(h)]+εi(h)(t),fori=1,2,…,I,−σi(h)≤0,<
mspace width="0.25em">andλi(h)≥0 :MATH]
(8)
which can be simplified to:
[MATH: pi(h)(t+1)=ϕiHP(t)θiHP+εi(h)(t),fori=1,2,…,I,−σi(h)≤0,<
mspace width="0.25em">andλi(h)≥0
:MATH]
(9)
where
[MATH:
ϕiHP(t)
:MATH]
indicates the regression vector obtained from the corresponding
expression data, and
[MATH:
θiHP :MATH]
denotes the unknown interaction parameter vector of human-protein i in
the human PPIN to be estimated. Eq ([296]9) could be augmented for Y[i]
data points of human-protein i as follows:
[MATH:
[p
i(h)(t2)
pi(h)(t3)⋮pi(h)(tYi
+1)]=[ϕiHP(t1)
ϕiHP(t2)⋮ϕiHP(tYi
)]θiHP+[<
mi>εi(h)(t1)
εi(h)(t2)⋮εi(h)(tYi
)],fori=1,2,…,I,−σi(h)≤0,<
mspace width="0.25em">andλi(h)≥0
:MATH]
(10)
where Y[i] is the number of data points of protein expression. Thus, we
defined the notations
[MATH: Pi(h) :MATH]
,
[MATH: ΦiHP<
/msubsup> :MATH]
, and
[MATH: ΞiHP<
/msubsup> :MATH]
to represent Eq ([297]10) as follows:
[MATH: Pi(h)=ΦiHP<
/msubsup>θiHP
+ΞiHP<
/msubsup>,fori=1,2,…,I,−σi(h)≤0,<
mspace width="0.25em">andλi(h)≥0
:MATH]
(11)
where
[MATH: Pi(h)=[pi(h)(t2)
pi(h)(t3)⋮pi(h)(tYi
+1)],ΦiHP<
/msubsup>=[ϕiHP(t1)
ϕiHP(t2)⋮ϕiHP(tYi
)],ΞiHP<
/msubsup>=[εi(h)(t1)
εi(h)(t2)⋮εi(h)(tYi
)] :MATH]
.
Next, we formulated the parameter estimation of
[MATH:
θiHP :MATH]
as the following constrained least square equation:
[MATH: minθi
HP1
2‖ΦiHP<
/msubsup>θiHP
−Pi(h)‖
22subjecttoAθiHP≤b :MATH]
(12)
where
[MATH:
A=[0⋯00⋯0<
/mtd>︷Ni+Ji−100010]
,b=[01] :MATH]
.
By solving the parameter estimation in (12) with the help of the lsqlin
function in the MATLAB optimization toolbox, we acquired the
interaction parameters in human PPIN Eq ([298]1), and concurrently
ensured that the human-protein translation rate
[MATH: λi(h) :MATH]
was a non-negative value and the human-protein degradation rate
[MATH: ‑σi(h) :MATH]
was a non-positive value; that is to say
[MATH: λi(h)≥0
:MATH]
and
[MATH: ‑σi(h)≤0
:MATH]
. Similarly, system identification approach for the other Eqs
([299]2)–([300]7) in the GIGEN is shown in [301]S1 Text.
To obtain accurate results in the system identification approach, it is
necessary to interpolate some extra data points (5 times the number of
parameters in the corresponding parameter vectors:
[MATH:
θiHP :MATH]
in human PPIN,
[MATH:
θjVP :MATH]
in EBV PPIN,
[MATH:
θkHG :MATH]
in human-gene GRN,
[MATH:
θsVG :MATH]
in EBV-gene GRN,
[MATH:
θzHL :MATH]
in human-lncRNA GRN,
[MATH:
τfHM :MATH]
in human-miRNA GRN, and
[MATH:
τuVM :MATH]
in the EBV-miRNA GRN to be estimated) via the cubic spline method
mentioned above, which solves the parameter estimation problem by
preventing overfitting owing to insufficient data points. Therefore,
the solutions to constrained least square parameter estimation Eq
([302]12), (S5), (S10), (S15), (S20), (S25), and (S30) can be obtained
with NGS expression data using the lsqlin function in the MATLAB
optimization toolbox for the optimal estimation of the parameters in
these estimation equations. There remains an unsettled question: the
genome-wide mRNA microarray expression measurement cannot describe the
protein behavior in human B cells and EBV, but the corresponding mRNA
abundance can explain over 73% of the variance in protein abundance
[[303]126]; that is to say, protein behavior can be described by the
corresponding gene expression. As a result, the NGS gene expression
data can replace the protein expression data, thereby contributing to
the solution of constrained least square parameter estimation Eq
([304]12), (S5), (S10), (S15), (S20), (S25), and (S30).
System order detection scheme for the dynamic models of GIGEN
The candidate GIGEN constructed by database mining with computational
and experimental predictions contained many false positives for the
interactive and regulatory parameters. Therefore, we applied the system
order detection scheme to the human PPIN model in (11), the EBV PPIN
model in (S4), the human-gene GRN model in (S9), the EBV-gene GRN model
in (S14), the human-lncRNA GRN model in (S19), the human-miRNA GRN
model in (S24), and the EBV-miRNA GRN model in (S29) to prune the false
positives in the candidate GIGEN using the NGS human B cell and EBV
data at the first and second infection stages. According to the Akaike
information criterion (AIC), the insignificant parameters in the models
of GIGEN were deleted so that we ultimately acquired the real GIGENs at
the first and second infection stages during the EBV lytic phase. In
the human PPIN model (11), the AIC of human-protein i can be defined as
follows [[305]121]:
[MATH: AICiHP(Ni,Ji)=log(1Yi(Pi(h)−ΦiHP<
/msubsup>θ^iHP)T(Pi(h)−ΦiHP<
/msubsup>θ^iHP))+2(<
mrow>Ni+Ji)Yi :MATH]
(13)
()where
[MATH: θ^iHP :MATH]
indicates the estimated parameters of human-protein i obtained from the
solution of parameter estimation Eq ([306]12), and the estimated
residual error is
[MATH: κ^HP,i2=1
Yi(Pi(h)−ΦiHP<
/msubsup>θ^iHP)T(Pi(h)−ΦiHP<
/msubsup>θ^iHP) :MATH]
.
Because the parameter estimation in (12) has less overfit for large
enough data points, we applied cubic spline interpolation to the
time-course data to increase the number of data points Y[i]. In (13),
the first quantity measures the model fit, and the second quantity
penalizes for overfit. Philosophically, AIC in (13) is an estimate of
the expected relative distance between the fitted model and the unknown
true mechanism that actually generated the observed data [[307]127].
Therefore, the true number (or order) of regulations and interactions
is obtained by minimizing AIC in (13). Those insignificant regulations
and interactions are considered as false positives to be deleted from
the candidate network. According to system identification theory, the
AIC is a tradeoff between estimated residual error and
parameter-associated error, and will be minimal at the real system
order (i.e., the number of parameters). The minimum
[MATH:
AICiH
P :MATH]
in (13) can be solved for number
[MATH:
Ni*+Ji* :MATH]
of the real PPIs of protein i in the human PPIN. The insignificant
interactions of
[MATH: Ni*
:MATH]
and
[MATH: Ji*
:MATH]
should be deleted as false positives from PPIs of protein i. We then
obtained the real human PPIN by applying a similar procedure one
protein at a time. Similarly, system order detection scheme for the
other equations (S4), (S9), (S14), (S19), (S24), and (S29) in the GIGEN
is shown in [308]S2 Text. Furthermore, in order to evaluate statistical
significance of real human PPIN, student’s t-test was used to calculate
the statistical significance (p value) of the parameters in (1) in the
real human PPI network under the null hypothesis H[0]:
[MATH:
αin(h)=0
:MATH]
and
[MATH:
γij(h)=0
:MATH]
in (1) [[309]128].
Therefore, we were able to identify the real GIGENs at the first and
second infection stages (Figs [310]3 and [311]4, respectively) during
EBV lytic infection; once we had applied the system identification
approach and the system order detection scheme, we obtained the GIGENs
by pruning the false positives of the candidate GIGEN using NGS human B
cell and EBV data. Information concerning the number of nodes and edges
of the candidate GIGEN from databases, and the number of nodes and
edges of the real GIGENs at the first and second infection stages are
presented in [312]Table 1A and 1B, respectively. However, the real
GIGENs shown in Figs [313]3 and [314]4 were too complicated to allow us
to investigate the lytic replication, production, and cytolytic
mechanisms in humans and EBV during lytic infection. Therefore, we
extracted the HVCNs, which contain the principal network structures of
the real networks, from the real GIGENs at both infection stages in the
EBV lytic phase using the PNP method.
Extracting the core network from the real GIGEN using the PNP method
It is essential to establish an integrated system network matrix H of a
real GIGEN before applying the PNP method to extract the core GIGEN
from the real GIGEN. Furthermore, the system network matrix H includes
the whole estimated system parameters in the real GIGEN as follows:
[MATH:
H=[0
Hvp,vp0
Hvp,hp00<
/mn>0Hhp,vp0Hhp,hp<
mtd>00Hvm,vm
Hvm,vp<
msub>Hvm,hm
Hv<
mi>m,hp
Hvm,hc0
Hvg,vmH
vg,vpHvg,h<
/mi>mHvg,hpHvg,hcH<
/mi>vg,hlHh<
mi>m,vm
Hhm,vpH
hm,hmHhm,h<
/mi>pHhm,hc0Hhg,vmHhg,vpH<
/mi>hg,hmHhg,hpHhg,hcHhg
,hlHhl,v<
/mi>mHhl,vpHhl,hmH<
/mi>hl,hpHhl,hcHhl,hl]∈ℝ(2J+<
/mo>2I+Q+R+L)×(J+I+Q+R+L+I'I'
') :MATH]
[MATH: whereHvp,
vp=[
α^11(v)⋯α^1J(v)⋮α^jm(v)⋮α^J1(v)⋯α^JJ(v)],Hvp,hp=[γ^11(v)⋯γ^1I(v)⋮γ^ji(v)⋮γ^J1(v)⋯γ^JI(v)],Hhp,vp=[γ^11(h)⋯γ^1J(h)⋮γ^ij(h)⋮γ^I1(h)⋯γ^IJ(h)],Hhp,hp=[α^11(h)⋯α^1I(h)⋮α^in(h)⋮α^I1(h)⋯α^II(h)],Hvm,vm=[−e¯^1
1(v)⋯−e¯^1
Q(v)⋮−e¯^u
q(v)⋮−e¯^Q
1(v)⋯−e¯^Q
Q(v)],Hvm,vp=[d¯^1
1(v)⋯d¯^1
J(v)⋮d¯^u
j(v)⋮d¯^Q
1(v)⋯d¯^Q
J(v)],Hvm,hm=[−b¯^1
1(v)⋯−b¯^1
R(v)⋮−b¯^u
r(v)⋮−b¯^Q
1(v)⋯−b¯^Q
R(v)],Hvm,hp=[a¯^1
1(v)⋯a¯^1
I(v)⋮a¯^u
i(v)⋮a¯^Q
1(v)⋯a¯^Q
I(v)],Hvm,hc=[ζ¯^1
1(v)⋯ζ¯^1
I'I''(v)⋮ζ¯^u
(I''(i'−1)+i'')(v)⋮ζ¯^Q
1(v)⋯ζ¯^Q
I'I''(v)],Hvg,vm=[−e^11(v)⋯−e^1Q(v)⋮−e^sq(v)⋮−e^J1(v)⋯−e^JQ(v)],Hvg,vp=[d^11(v)⋯d^1J(v)⋮d^sj(v)⋮d^J1(v)⋯d^JJ(v)],Hvg,hm=[−b^11(v)⋯−b^1R(v)⋮−b^sr(v)⋮−b^J1(v)⋯−b^JR(v)],Hvg,hp=[a^11(v)⋯a^1I(v)⋮a^si(v)⋮a^J1(v)⋯a^JI(v)],Hvg,hc=[ζ^11(v)⋯ζ^1I'I''(v)⋮ζ^s(I''(i'−1)+i'')(v)⋮ζ^J1(v)⋯ζ^JI'I''(v)],Hvg,hl=[c^11(v)⋯c^1L(v)⋮c^sl(v)⋮c^J1(v)⋯c^JL(v)],Hhm,vm=[−e¯^1
1(h)⋯−e¯^1
Q(h)⋮−e¯^f
q(h)⋮−e¯^R
1(h)⋯−e¯^R
Q(h)],Hhm,vp=[d¯^1
1(h)⋯d¯^1
J(h)⋮d¯^f
j(h)⋮d¯^R
1(h)⋯d¯^R
J(h)],Hhm,hm=[−b¯^1
1(h)⋯−b¯^1
R(h)⋮−b¯^f
r(h)⋮−b¯^R
1(h)⋯−b¯^R
R(h)], :MATH]
[MATH: Hhm
,hp=[<
mrow>a¯^1
1(h)⋯a¯^1
I(h)⋮a¯^f
i(h)⋮a¯^R
1(h)⋯a¯^R
I(h)],Hhm,hc=[ζ¯^1
1(h)⋯ζ¯^1
I'I''(h)⋮ζ¯^f
(I''(i'−1)+i'')(h)⋮ζ¯^R
1(h)⋯ζ¯^R
I'I''(h)],Hhg,vm=[−e^11(h)⋯−e^1Q(h)⋮−e^kq(h)⋮−e^I1(h)⋯−e^IQ(h)],Hhg,vp=[d^11(h)⋯d^1J(h)⋮d^kj(h)⋮d^I1(h)⋯d^IJ(h)],Hhg,hm=[−b^11(h)⋯−b^1R(h)⋮−b^kr(h)⋮−b^I1(h)⋯−b^IR(h)],Hhg,hp=[a^11(h)⋯a^1I(h)⋮a^ki(h)⋮a^I1(h)⋯a^II(h)],Hhg,hc=[ζ^11(h)⋯ζ^1I'I''(h)⋮ζ^k(I''(i'−1)+i'')(h)⋮ζ^I1(h)⋯ζ^II'I''(h)],Hhg,hl=[c^11(h)⋯c^1L(h)⋮c^kl(h)⋮c^I1(h)⋯c^IL(h)],Hhl,vm=[−e^11(L)⋯−e^1Q(L)⋮−e^zq(L)⋮−e^L1(L)⋯−e^LQ(L)],Hhl,vp=[d^11(L)⋯d^1J(L)⋮d^zj(L)⋮d^L1(L)⋯d^LJ(L)],Hhl,hm=[−b^11(L)⋯−b^1R(L)⋮−b^zr(L)⋮−b^L1(L)⋯−b^LR(L)],Hhl,hp=[a^11(L)⋯a^1I(L)⋮a^zi(L)⋮a^L1(L)⋯a^LI(L)],Hhl,hc=[ζ^11(L)⋯ζ^1I'I''(L)⋮ζ^z(I''(i'−1)+i'')(L)⋮ζ^L1(L)⋯ζ^LI'I''(L)],Hhl,hl=[c^11(L)⋯c^1L(L)⋮c^zl(L)⋮c^L1(L)⋯c^LL(L)] :MATH]
where
[MATH: α^in(h) :MATH]
and
[MATH: γ^ij(h) :MATH]
are the corresponding components in
[MATH: θ^iHP :MATH]
obtained by solving parameter estimation Eq ([315]12) and system order
detection Eq ([316]13);
[MATH: α^jm(v) :MATH]
and
[MATH: γ^ji(v) :MATH]
are the corresponding components in
[MATH: θ^jVP :MATH]
obtained by solving parameter estimation equation (S5) and system order
detection equation (S31);
[MATH: a^ki(h),ζ^k(I''(i'−1)+i'')(h),−b^kr(h),c^kl(h),d^kj(h),and−e^kq(h) :MATH]
are the corresponding components in
[MATH: θ^kHG :MATH]
obtained by solving parameter estimation equation (S10) and system
order detection equation (S32);
[MATH: a^si(v),ζ^s(I''(i'−1)+i'')(v),−b^sr(v),c^sl(v),d^sj(v),and−e^sq(v) :MATH]
are the corresponding components in
[MATH: θ^sVG :MATH]
obtained by solving parameter estimation equation (S15) and system
order detection equation (S33);
[MATH: a^zi(L),ζ^z(I''(i'−1)+i'')(L),−b^zr(L),c^zl(L),d^zj(L),and−e^zq(L) :MATH]
are the corresponding components in
[MATH: θ^zHL :MATH]
obtained by solving parameter estimation equation (S20) and system
order detection equation (S34);
[MATH: a¯^f
i(h),ζ¯^f
(I''(i'−1)+i'')(h),−b¯^f
r(h),d¯^f
j(h),and−e¯^f
q(h) :MATH]
are the corresponding components in
[MATH: τ^fHM :MATH]
obtained by solving parameter estimation equation (S25) and system
order detection equation (S35); and
[MATH: a¯^u
i(v),ζ¯^u
(I''(i'−1)+i'')(v),−b¯^u
r(v),d¯^u
j(v),and−e¯^u
q(v) :MATH]
are the corresponding components in
[MATH: τ^uVM :MATH]
obtained by solving parameter estimation equation (S30) and system
order detection equation (S36).
[MATH: α^in(h) :MATH]
and
[MATH: α^jm(v) :MATH]
indicate the intraspecies interactive abilities of human and EBV PPINs
during EBV infection, respectively;
[MATH: γ^ij(h) :MATH]
and
[MATH: γ^ji(v) :MATH]
denote the intraspecies interactive abilities of human-protein i and
EBV-protein j in human and EBV PPINs;
[MATH: a^ki(h),a^si(v),a^zi(L),a¯^f
i(h),anda¯^u
i(v) :MATH]
represent the abilities of human-TF i to regulate the transcription of
human-gene k, EBV-gene s, human-lncRNA z, human-miRNA f, and EBV-miRNA
u, respectively, in human-gene GRN, EBV-gene GRN, human-lncRNA GRN,
human-miRNA GRN, and EBV-miRNA GRN during EBV infection, respectively;
[MATH: ζ^k(I''(i'−1)+i'')(h),ζ^s(I''(i'−1)+i'')(v),ζ^z(I''(i'−1)+i'')(L),ζ¯^f
(I''(i'−1)+i'')(h),andζ¯^u
(I''(i'−1)+i'')(v) :MATH]
represent the abilities of human-TF complex i'::i'' to regulate the
transcription of human-gene k, EBV-gene s, human-lncRNA z, human-miRNA
f, and EBV-miRNA u, respectively, in human-gene GRN, EBV-gene GRN,
human-lncRNA GRN, human-miRNA GRN, and EBV-miRNA GRN during EBV
infection, respectively;
[MATH: d^kj(h),d^sj(v),d^zj(L),d¯^f
j(h),andd¯^u
j(v) :MATH]
represent the abilities of EBV-TF s to regulate the transcription of
human-gene k, EBV-gene s, human-lncRNA z, human-miRNA f, and EBV-miRNA
u, respectively, in human-gene GRN, EBV-gene GRN, human-lncRNA GRN,
human-miRNA GRN, and EBV-miRNA GRN during EBV infection, respectively;
[MATH: c^kl(h),c^sl(v),andc^zl(L) :MATH]
represent the abilities of human-lncRNA z to regulate the transcription
of human-gene k, EBV-gene s, and human-lncRNA z, respectively, in
human-gene GRN, EBV-gene GRN, and human-lncRNA GRN during EBV
infection, respectively;
[MATH: ‑b^kr(h),−b^sr(v),−b^zr(L),−b¯^f
r(h),and−b¯^u
r(v) :MATH]
correspond to the abilities of human-miRNA r to inhibit human-gene k,
EBV-gene s, human-lncRNA z, human-miRNA f, and EBV-miRNA u,
respectively, in human-gene GRN, EBV-gene GRN, human-lncRNA GRN,
human-miRNA GRN, and EBV-miRNA GRN during EBV infection, respectively;
and
[MATH: ‑e^kq(h),−e^sq(v),−e^zq(L),−e¯^f
q(h),and−e¯^u
q(v) :MATH]
represent the abilities of EBV-miRNA q to inhibit human-gene k,
EBV-gene s, human-lncRNA z, human-miRNA f, and EBV-miRNA u,
respectively, in human-gene GRN, EBV-gene GRN, human-lncRNA GRN,
human-miRNA GRN, and EBV-miRNA GRN during EBV infection, respectively.
The estimated weights (i.e., parameters) of the network links in
intraspecies PPINs, intraspecies GRNs, interspecies PPINs, and
interspecies GRNs therefore constitute the system network matrix H of
the real GIGEN. In the network matrix H, the corresponding parameter is
zero if a link does not appear in the candidate GIGEN or has been
pruned via the AIC. We then extracted the core network of the real
GIGEN by applying PNP to network matrix H [[317]129]. PNP is achieved
on the basis of the singular value decomposition of H in the following
equation:
[MATH:
H=UDVT :MATH]
(14)
where
[MATH:
U∈ℝ(<
mn>2J+2I+Q+R+L)×(J+I+Q+R+L+I'I'') :MATH]
,
[MATH:
V∈ℝ(<
mi>J+I+Q+R+L+I'I'')×(J+I+Q+R+L+I'I''
) :MATH]
, and D = diag (d[1],…,d[s],…,d[J+I+Q+R+L+I'I'']), in which the
J+I+Q+R+L+I'I'' singular values of H are in descending order, i.e.,
d[1]≥…≥ d[s] ≥…≥ d[J+I+Q+R+L+I'I'']. Notably, diag (d[1], d[2])
indicates the diagonal matrix of d[1] and d[2]. Moreover, we defined
the eigenexpression fraction (E[s]) for the normalization of singular
values as follows:
[MATH:
Es=ds2∑s=1J+I+Q<
mo>+R+L+I'I''d<
mi>s2 :MATH]
(15)
It is necessary to maintain the system energy of the whole network
structure. Therefore, we chose the top K singular vectors of network
matrix H with the minimum K resulting in
[MATH:
∑s=1KEs≥85
% :MATH]
to represent at least 85% of the energy of the core network structure
of the GIGEN, which constituted the top K principal components. Next,
we defined the projection (T) of network matrix H for the top K
singular vectors of U and V as follows:
[MATH: TR(r,s)=hr,<
/mo>:×v:<
mo>,sandTL(l,s)=h:,lT×u<
mrow>:,s,fors=1,⋯,Kr=1<
/mn>,⋯,(2J+2I+Q+R+L),andl=1,⋯,(J+I+Q+R+L+I'I'') :MATH]
(16)
where h[r,:], v[:,s], h[:,l], and u[:,s] denote the rth row of H, the
sth column of V, the lth column of H, and the sth column of U,
respectively. Finally, we defined and applied the 2-norm projection
value of each node, including gene, miRNA, lncRNA, protein, and protein
complex in the real GIGEN for the top K right singular vectors and the
top K left singular vectors as follows:
[MATH: DR(r)=[∑s=1<
/mn>KTR<
msup>(r,s)2<
mo>]12andDL(l)=[∑s=1<
/mn>KTL<
msup>(l,s)2<
mo>]12<
mtr>forr=1,…,(2J+2I+Q+R+L)andl=1,…,(J+I+Q+R+L+I'I'') :MATH]
(17)
It is implied that if the D[R](r) or D[L](l) projection values approach
zero, the contributions of the corresponding rth or lth nodes,
respectively, are insignificant and almost independent of the core
network structure composed of the top K singular vectors. Consequently,
we built core networks comprising the core proteins, genes, and miRNAs
by selecting the proteins, genes and miRNAs with the top projection
values in (17) from the TF receptors and their associated genes and
miRNAs. The human and EBV HVCNs extracted at the first and second
infection stage from the real GIGENs by the above PNP method are
presented in Figs [318]5 and [319]6, respectively, and the information
concerning the number of nodes and edges of the HVCNs at the first and
second infection stage are exhibited in [320]Table 3A and 3B,
respectively.
Supporting information
S1 Text. System identification approach for the other Eqs
([321]2)–([322]7) in GIGEN.
(DOCX)
[323]Click here for additional data file.^ (139.8KB, docx)
S2 Text. System order detection scheme for the other Eqations (S4),
(S9), (S14), (S19), (S24), and (S29) in GIGEN.
(DOCX)
[324]Click here for additional data file.^ (57.8KB, docx)
S1 Table. Each edge in the real GIGENs and HVCNs and virus-interacting
host pathways.
S1A Table and S1B Table show each edge of the real GIGENs and HVCNs.
[325]S1C Table reveals the p-value of each edge in the PPINs of the
HVCNs. S1D Table is the identified virus-interacting host pathways at
the first and second infection stages.
(XLSX)
[326]Click here for additional data file.^ (2.6MB, xlsx)
Data Availability
All relevant data are within the paper and its Supporting Information
files.
Funding Statement
The work was supported by the Ministry of Science and Technology of
Taiwan under grant No. MOST 105-2811-E-007-041.
References