Peptide matching between Epstein–Barr virus and human proteins

Abstract

The human herpesvirus 4, also called Epstein–Barr virus (EBV), is a common virus in humans (Rickinson & Kieff, 2007 ; http://www.cdc.gov/ncidod/diseases/ebv.htm ). EBV infection, although usually asymptomatic (Klein et al. , 2010 ; Saha & Robertson, 2011 ), may cause mononucleosis, neoplasms, and autoimmune diseases (Cesarman, 2002 ; Farrell & Jarrett, 2011 ; Saha & Robertson, 2011 ; Gourzones et al. , 2012 ). However, the pathogenic contribution of EBV to tumors as well as the etiology of autoimmune diseases associated with EBV infection such as multiple sclerosis (Farrell & Jarrett, 2011 ; Niller et al. , 2011 ; Tselis, 2012 ) and systemic lupus erythematosus (SLE) remain unclear (Draborg et al. , 2012 ).

Other poorly understood issues are EBV immunoevasion, latency, and (re)activation (Jochum et al. , 2012 ; Kalla & Hammerschmidt, 2012 ; Severa et al. , 2013 ) as well as the EBV ubiquitous presence in the human population, with approximately 95% of adults worldwide infected ( http://www.cdc.gov/ncidod/diseases/ebv.htm ; Rickinson & Kieff, 2007 ). It seems that a number of EBV proteins can protect the virus from the host immune attack. Such EBV proteins, also called immunoevasins, function by targeting MHC class I and MHC class II antigen presentation pathways (Ressing et al. , 2008 ; Rowe & Zuo, 2010 ); for example, the EBV immunoevasin interleukin-10 homolog (IL-10H) protects infected B cells from immune recognition and elimination (Jochum et al. , 2012 ), thus contributing to EBV successful persistence and immune escape.

However, EBV immunoevasin IL-10H might contribute through other additional pathways to the EBV immunoevasion phenomenon given its sequence identity to the human IL-10. In fact, IL-10H is a striking example of the conservation of amino acid (aa) sequence between EBV and the human host (Moore et al. , 1990 ; Moore et al. , 2001 ; Yoon et al. , 2005 ), with viral IL-10H and human IL-10 sharing four continuous identical stretches ranging from 17 to 42 aa (e.g. MLRDLRDAFSRVKTFFQ, DNLLLKESLLEDFKGYLGCQALSEMIQFYLEEVMPQAENQDP, HVNSLGENLKTLRLRLRRCHRFLPCENKSKAVEQ, KNAFNKLQEKGIYKAMSEFDIFINYIEAYMT). The sequence alignment of EBV IL-10H (P0CAP9, IL10H_EBVG) and human IL-10 (P22301, IL10_HUMAN) shows a percent identity equal to 77.2 at the aa level, with 139 identical positions.

In this regard, we already observed that a striking level of sequence identity to the human host might act as a camouflage mechanism of infectious agents (Natale et al. , 2000 ; Tindle, 2002 ; Lucchese et al. , 2009 ; Capone et al. , 2013 ), because when high levels of sequence/structure identity are present between microbial and human molecules, the breaking of the self-tolerance mechanisms that prevent self-reactivity is highly improbable (Silverstein, 2001 ). Thus, the sharing of continuous aa sequences with host molecules may represent an elective microbial mechanism to escape immune recognition attack (Kanduc et al. , 2008 ; Trost et al. , 2010 ). Here, to further our understanding of how can EBV establish a persistent infection in the human host (Ressing et al. , 2008 ; Rowe & Zuo, 2010 ; Jochum et al. , 2012 ), we analyzed EBV proteins for aa sequence identity to human proteins to investigate whether other identity regions are present between EBV and the human host in addition to the above-cited sequence identities between EBV IL-10H and human IL-10.

To this aim, we examined the polyprotein derived from EBV, strain GD1, GenBank: AY961628.3, NCBI taxonomic identifier: 10376 (Zeng et al. , 2005 ), consisting of 68 proteins (total number of aa: 34 503) listed and described at http://www.ncbi.nlm.nih.gov/nuccore/AY961628 . Sequence identity analyses of EBV proteins to the human proteome were conducted using viral decapeptides as probes to scan the Homo sapiens proteome, searching for exact peptide matches. Each probe was shifted by one residue; that is, viral decapeptides sequentially overlapped by nine residues such as MVHVLERALL, VHVLERALLE, HVLERALLEQ, were used in scanning the human proteome for peptide matching using Protein International Resource (PIR) peptide match program ( pir.georgetown.edu/pirwww/search/peptide.shtml ) (Wu et al. , 2003 ). The human proteins containing viral matches were analyzed using UniProt database ( http://www.uniprot.org ) (UniProt Consortium, 2009 ). Fragments, duplicated sequences, and obsolete entries were filtered out manually.

The peptide-by-peptide comparison of EBV and Homo sapiens proteomes at the decapeptide level is presented in Table 1 . It can be seen that: (1) viral decapeptides repeatedly occur in 28 human proteins; (2) the peptide sharing also occurs at 11-, 12-, and 13-mer levels; (3) five of 68 EBV proteins (i.e. BDLF2, EBNA1, DEN, EBNA2, and Q3KSS2) are implicated in the viral vs. human peptide sharing. In particular, Table 1 highlights a heavy involvement of the Epstein–Barr nuclear antigen 1 (EBNA1) in the peptide sharing, with a clustering of identity regions in the glycine-rich region (GRR) allocated along the NH2 terminus of the 641-aa-long EBV EBNA1 sequence.

Biologically, the 28 human proteins (ARI1B, BD1L1, BMP2K, CHD5, CHTOP, DIAP3, FRM4A, FUS, FZD8, JUND, LAR1B, LARP1, MBD2, MLL4, NOVA2, NOXA1, ONEC3, PCSK6, RBM26, RBM27, RS2, SFR15, SHSA7, SKOR2, SMD1, TSN3, UNG, and ZN579) implicated in the sharing exert critical functions in crucial processes such as myelination, chromatin remodeling, RNA splicing, and proteolysis. For example:

1. The EBV BDLF2 _247–256 VYTLIPAVVI decapeptide is shared with human tetraspanin-3 (TSN3) that regulates the proliferation and migration of oligodendrocytes, a process essential for normal myelination and repair (Tiwari-Woodruff et al. , 2004 ).

2. The GAGAGGGAGG decapeptide is repeated eight times in EBNA1 and is also present in the neuro-oncological ventral antigen 2 (NOVA2). NOVA2 is a neuron-specific splicing factor that regulates neuronal migration (Yano et al. , 2010 ), is necessary for physiologic motor neuron firing (Ruggiu et al. , 2009 ), and has been identified as a target in autoimmune motor disease (Yang et al. , 1998 ).

3. The GAGGAGGAGAG 11-mer is present 12 times in EBNA1 and is shared with the human bi-orientation of chromosomes in cell division protein 1-like 1 (BD1L1) (Porter et al. , 2007 ).

4. The EBV EBNA1 _314–325 GGGAGAGGAGAG 12-mer is shared with the human AT-rich interactive domain–containing protein 1B (ARI1B). ARI1B is involved in gene transcriptional activation and repression by chromatin remodeling (Santen et al. , 2012 ). Of note, ARI1B is important in human brain development and function in general and in the development of corpus callosum in particular (Halgren et al. , 2012 ). Indeed, ARI1B alterations are associated with mental retardation, impairments in adaptative behavior, and speech disorders with expressive speech more severely affected than receptive function (Santen et al. , 2012 ).

5. The EBV EBNA1 _40–51 GRGRGRGRGRGG and the EBV EBNA2 _311–322 RGRGRGRGRGRG are common to small nuclear ribonucleoprotein Sm D1 (SMD1). SMD1 may act as a charged protein scaffold to promote snRNP assembly or strengthen snRNP–snRNP interactions through nonspecific electrostatic contacts with RNA. As a note of special importance, antinuclear antibodies with SMD1 specificity are developed in the autoimmune SLE disease (Poole et al. , 2006 ).

6. The EBV EBNA1 _202–214 AGAGGAGGAGGAG 13-mer is shared with proprotein convertase subtilisin/kexin type 6 (PCSK6). Of note, PCSK6 is associated with handedness in individuals with dyslexia (Scerri et al. , 2013 ).

7. The GGRGRGGSGG decapeptide is present three times in EBNA1 and is shared with RNA-binding protein FUS (FUS). Accumulation of FUS protein as cytoplasmic inclusions in neurons and glial cells in the central nervous system is the pathological hallmark of amyotrophic lateral sclerosis as well as certain subtypes of frontotemporal lobar degeneration (Takeuchi et al. , 2013 ).

8. The EBV EBNA2 _311–320 RGRGRGRGRG decapeptide is shared with the histone-lysine N-methyltransferase MLL4 or myeloid/lymphoid or mixed-lineage leukemia protein 4. MLL4 is a crucial player in cell viability and cell-cycle progression and is critical for tumor growth in vivo (Ansari et al. , 2012 ).

9. The EBV Q3KSS2 _84–95 KVVILGQDPYHG 12-mer is also present in human uracil–DNA glycosylase (UNG). UNG excises uracil residues from the DNA that can arise as a result of misincorporation of dUMP residues by DNA polymerase or due to deamination of cytosine. Defects in UNG are a cause of immunodeficiency with hyper-IgM type 5 (Kavli et al. , 2005 ).

In summary, we find a significant peptide sharing between EBV and the human proteome that involves even peptides 13-mer long and is predominant at level of EBV EBNA1 protein. In fact, 42 of the total 47 decapeptides shared between EBV and human proteins are derived from EBNA1 (Table 1 ). Such a peptide commonality might explain the immunoevasive properties of EBNA1 protein, a viral antigen that has been reported to go undetected by the cell-mediated immune system (Münz, 2004 ). Indeed, it seems logical to hypothesize that human immunotolerance mechanism(s) should prevent attacks against a protein, EBNA1, endowed with such a high level of sequence identity to human proteins. However, in parallel, it has to be observed that EBNA1 antigen is also one of the most frequently recognized EBV antigens for CD4+ helper T cells (Long et al. , 2005 ), thus representing a prime target for T-cell-based immunotherapy (Tsang et al. , 2006 ). Interestingly, mapping of CD4+ epitopes within the primary sequences of EBNA1 reveals a specific epitope location in the EBNA1 COOH region (Long et al. , 2005 ; Tsang et al. , 2006 ). In molecular terms, data from Table 1 favor the view of EBNA1 protein sequence as hosting antigenicity and immunotolerance in two spatially distinct domains, with the antigenic portion allocated along the COOH terminus and the potentially immunotolerogenic GRR confined in the NH2 region. Accordingly, a thorough search through IEDB database (Peters et al. , 2005 ) highlights that the EBNA1 COOH terminus allocates 74 T-cell epitopes, while EBNA1 NH2 terminus GRR presents only one epitope (AGAGGGAGGAGAG, IEDB ID: 1432) (Petersen et al. , 1989 ) comprehending a 11-mer sequence reported in Table 1 (AGGGAGGAGAG).

This study suggests a role for EBNA1 GRR in immunoevasion and, in addition, may offer hints to explore the molecular basis underlying the delayed development of CD4+ T cell and humoral immune responses against EBV EBNA1 (Hislop et al., 2007 ; Long et al., 2013 ). Indeed, Levitskaya et al . ( 1995 ) demonstrated that the GRR allocated in the NH2 terminus of EBNA1 interferes with antigen processing and MHC class I-restricted presentation, possibly by inhibiting ubiquitin-/proteasome-dependent protein degradation (Levitskaya et al. , 1997 ). In parallel, Yin et al . ( 2003 ) showed that the GRR inhibits EBNA1 mRNA translation and proposed that minimizing translation of the EBNA1 transcript, cells expressing EBNA1 avoid cytotoxic T-cell recognition. However, the molecular mechanism by which the GRR repeat inhibits ubiquitin-/proteasome-dependent degradation remained unexplained (Levitskaya et al. , 1997 ). As a matter of fact, GRRs are critical domains in proteins such as TAR DNA-binding protein 43 (TDP–43), heterogeneous nuclear ribonucleoprotein A1 (ROA1), and the above-mentioned RNA-binding protein FUS (Rogelj et al. , 2011 ). TDP–43, ROA1, and FUS are proteins crucially involved in the regulation of different steps of gene expression, including transcription, splicing, mRNA transport, and translation (Strong et al. , 2007 ; Bekenstein & Soreq, 2012 ; Dormann and Haass, 2013 ). Then, it can be hypothesized that the EBNA1 GRR might interfere with GRRs present in host proteins, thus subverting crucial cellular functions, such as transcription and translation.

In the present context, it is also noteworthy that the extent and significance of the peptide overlap between EBNA1 and human proteins are even more relevant following sequence identity analyses using short peptide modules as scanning probes, being thousand the EBV matches in the human proteome (Kanduc et al. , 2008 ). As an example, the EBV EBNA1 heptapeptide GAGAGGG occurs 15 times along the NH2-terminal domain of EBV EBNA1 ( http://www.uniprot.org/uniprot/Q3KSS4 ). Of interest, the same heptapeptide GAGAGGG occurs in the human nuclear factor NF-kappa-B p105 (NFkB1) protein, where the Gly-rich heptapeptide GAGAGGG functions as a processing signal for the generation of the p50 subunit (Lin & Ghosh, 1996 ; Orian et al. , 1999 ). Hence, it might be postulated that the multiple EBNA1 GRRs could compete for the proteolytic reaction of NFkB1, a transcriptional factor crucial in the activation and function of mature B cell (Pohl et al. , 2002 ; Ruland & Mak, 2003 ).

More in general, the fact that short aa modules such as pentapeptides can represent minimal functional determinants in biological interactions and immune recognition (Kanduc, 2012a , b ; Kanduc, 2013 ) implies a wide array of physio(patho)logical viral–host relationships, being massive the level of peptide overlap between EBV and human proteins at the 5-mer level (Kanduc et al. , 2008 ). Hence, data reported in Table 1 not only might help understand EBV escape from immunosurveillance, but could also contribute to unfold the still obscure links between EBV infection and associated cancer diseases and autoimmune disorders (Farrell & Jarrett, 2011 ; Niller et al. , 2011 ; Draborg et al. , 2012 ; Tselis, 2012 ). Finally, the present phenetic analyses might open the way to innovative therapeutic approaches to fight/eradicate EBV infection and related pathologic sequelae (Kanduc et al. , 2007 ; Lucchese et al. , 2011 ; Capone et al. , 2013 ).

Authors' Contributions

All authors contributed to the computational analysis. D.K. proposed the original idea, interpreted the data, developed the research project, and wrote the manuscript. All authors discussed the results, and commented and revised the manuscript.

Acknowledgement

The authors declare that there are no conflict of interests.

References