A COMPUTATIONAL STRATEGY TO INVESTIGATE

RELEVANT SIMILARITIES BETWEEN VIRUS

AND HUMAN PROTEINS

Local High Similarities between Herpes and Human Proteins

Anna Marabotti

Istituto di Tecnologie Biomediche, CNR, Segrate (MI), Italy

Istituto di Scienze dell’Alimentazione, CNR, Avellino, Italy

Corrado Cirielli, Daniela Agnese D’Arcangelo

Istituto Dermopatico dell’Immacolata, IDI-IRCCS, Rome, Italy

Claudia Giampietri

Department of Histology and Medical Embryology, “Sapienza” University of Rome, Italy

Francesco Facchiano

Dipartimento Ematologia, Oncologia e Medicina Molecolare, Istituto Superiore di Sanità, Rome, Italy

Antonio Facchiano

Istituto Dermopatico dell’Immacolata, IDI-IRCCS, Rome, Italy

Angelo M. Facchiano

Istituto di Scienze dell’Alimentazione, CNR, Avellino, Italy

Keywords: Proteome comparison, Molecular mimicry, Autoimmunity, Local similarity.

Abstract: Investigating primary sequence and structural features of viral proteins/genes has revealed molecular

mimicry and evolutionary relationship linking viruses to eukaryotes. The continuous improvement in

sequencing-techniques makes available almost daily the whole genome/proteome of several

microorganisms, making now possible systematic analyses of evolutionary correlations and accurate

phylogeny investigations. In the present study we set up a methodology to identify significant and relevant

similarities between viral and human proteomes. To this aim, the following steps were applied: i)

identification of local similarity corresponding to continuous identity over at least 8-residues long

fragments; ii) filtering results for statistical significance of the identified similarities, according to BLAST

parameters for short sequences; iii) additional filters applied to the BLAST outputs, to select specific

viruses. The present study indicates a novel accurate methodology to find relevant similarities among virus

and human proteomes, useful to further investigate pathogenic mechanisms underlying infectious and non-

infectious diseases.

1 INTRODUCTION

Several studies investigate genetic predisposition

toward human disorders (Baranzini, 2009; Rubstov,

2010). Mechanisms involving a genetic basis may

require events occurring at the germinal level

(giving hereditary disorders) or at somatic levels.

Environmental stimuli and life style, such as smoke

183

Marabotti A., Cirielli C., Agnese D’Arcangelo D., Giampietri C., Facchiano F., Facchiano A. and M. Facchiano A..

A COMPUTATIONAL STRATEGY TO INVESTIGATE RELEVANT SIMILARITIES BETWEEN VIRUS AND HUMAN PROTEINS - Local High Similarities

between Herpes and Human Proteins.

DOI: 10.5220/0003156801830188

In Proceedings of the International Conference on Bioinformatics Models, Methods and Algorithms (BIOINFORMATICS-2011), pages 183-188

ISBN: 978-989-8425-36-2

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

habit, professional chemical exposure, diet style or

exposure to pollution or to infectious agents, may

interfere at the genetic level on the system

homeostasis within the human body. In several cases

infective agents, besides the direct infection disease,

may trigger additional mechanisms responsible of

the onset of additional diseases. Virus infections are

often responsible of additional serious pathologies,

and in most cases the mechanisms underlying the

onset of these “secondary” pathologies are not fully

elucidated. For instance, Human Papilloma virus

(HPV) infection is known to strongly relate to the

occurrence of cervix cancer in humans (Leggatt,

2007); helicobacter pylori infection is strongly

related to gastric ulcer and gastric cancer (Cao,

2007); hepatitis B virus (HBV) infection has been

proposed to have a role in the development of

hepatocellular carcinoma (Neuveut, 2010); diabetes

has been related with rotavirus infection (Maklea,

2006) or coxsackie virus infection (Peng, 2006);

bacterial infections have been shown to be related to

heart diseases (Ott, 2006); myocarditis has been

proposed to have an autoimmune basis related to

auto-antibodies against cardiac myosin and based on

a mimicry process between cardiac myosin and the

beta-adrenergic receptor (Li, 2006); Herpes Virus

type 7 has been shown to be involved in initiation

and maintenance of Graves’ disease autoimmune

process (Leite, 2010); Herpes Virus type 1 infection

has been strongly related to autoimmune reaction

based on the auto-reactive T lymphocytes

recognizing shared epitopes (Zhao, 1998); herpes

antigens have been suggested to play a molecular

mimicry role for autoimmune-based diseases such as

psoriasis (Kirby, 2000; Mehraein, 2004) and

antibodies against Herpes viruses have been found in

immuno-deficient or auto-immune patients (Thomas,

2008). A recent study (Fumagalli, 2010) pointed out

that up to 8% of the human genome is of viral

source, representing the “fossil remnants of past

infections” via integration at several sites. Consistent

with these findings, a strong rational supports the

investigation of human-to-viral structural-functional

relationships based, at least in some cases, onto a

common-antigen sharing process. We have

previously suggested unexpected evolutionary

relations between human-lymphocytes CD4 receptor

and his counterpart HIV-capsid-constituent GP120

(Facchiano, 1995; Facchiano, 1996); allele

variability has been related to viral infection and

susceptibility, with a significant association of HIV

progression with patients HLA status (Limou, 2009;

Fellay, 2007). Based on significant proteins

similarities we have also suggested the occurrence of

molecular mimicry between pathogens and human

proteins possibly underlying different diseases

(Benvenga, 1995; Benvenga, 1999; Benvenga,

2003). According to such large body of evidence, it

is now accepted that eukaryotes have been, and still

are, under a strong selective pressure, driven, at least

in part, by viral/prokaryotes infections, with human

gene variants associated to pathogens–related

infections. The strong improvement of the

sequencing techniques is producing a continuous

novel identification of the entire genome/proteome

of several micro-organisms, including viral

pathogens. Such sequences, or at least those

available freely on the net, can be analyzed with a

number of different approaches and software to

investigate structural features as well as local or

general evolutionary correlations. In the current

study we present a novel accurate methodology

which allows to identify relevant high local

similarities between viral entire proteome and

human proteome.

2 METHODS

Protein sequences from viral and human source were

obtained by the RefSeq database to constitute our

viral and human proteome databases, which included

82918 and 39037 protein sequences, respectively.

The two databases were initially compared by using

self-developed proprietary PERL scripts. The

comparison was performed in two steps: in the first

step all 8 amino acids-long fragments were extracted

from the viral database; each virus fragment was

then searched in the human database. Any time a full

identity was found (8 identities out of 8), extensions

at the N- or C- terminus were investigated to verify

the occurrence of identity over a longer fragment.

The second step of the analysis was then carried

out with BLASTplus (available at the NCBI web

site: http://www.ncbi.nlm.nih.gov/) to verify the

statistical significance of the identities found. All

viral fragments selected within the first step were

compared to the human proteome database in order

to obtain scores and significance evaluation. The

“blastp-short” settings were applied, i.e. parameters

were optimized for query sequences shorter than 30

residues. Namely, the following settings were

applied: word_size=2, gapopen=9, gapextend=1,

matrix=PAM30, threshold=16, comp_based_stats=2.

Finally, a final filter based on the virus name was

applied to extract the similarities identified by

BLAST, concerning protein sequences from specific

viruses.

BIOINFORMATICS 2011 - International Conference on Bioinformatics Models, Methods and Algorithms

184

3 RESULTS AND DISCUSSION

During the first step of our search strategy, we

extracted all fragments of 8 amino acids from the

viral proteins. This minimum length was chosen as a

threshold to select fragments with a putative activity

as antigenic peptides, as discussed below. These

fragments were searched in the human proteome

database, and extended when possible to the

maximum length of identical segment. The search

identified a very high number of non redundant

fragments, i.e. 42993 having identical sequence in

viral and human proteins, with length of 8 or more

amino acids. In many cases, the sequences are

characterized by low sequence complexity, i.e.,

composed by only one or two different residues. The

statistical relevance of the similarity involving very

short sequences or low-complexity fragments is

often considered very low or of difficult evaluation.

However, highly repetitive sequences are known to

play several key functions: for example, collagen

and sericin (the protein composing the silk) are

characterized by low-complexity sequences; further,

proline-rich peptides have low-complexity but

represent key sequences recognized by SH3 domain

(Yu, 1994); a 9 residues long poly-arginine has been

demonstrated to play a key role in cellular uptake

(Wender, 2000); proline-enrichment within peptide-

sequence may increase binding to mitochondrial

targets (Serasinghe, 2010); leucine-rich or lysine-

rich fragments have shown adhesive-, heparin-

binding and DNA-binding features. In order to

follow a statistics-driven approach, we decided to

proceed the analysis using only BLAST-defined

statistically significant identities; therefore, in the

second step of the present procedure we compared

any selected virus-fragment to the human proteome

by BLASTplus, using statistical significance settings

specifically chosen for short sequences. This second

step selected 1076 viral fragments which share

statistically significant similarity with human protein

sequences and constituted the VISHUM database

(VIral Similarities to HUMan fragments). This is

one of the products of the present study; it is a

growing database, planned to be updated

periodically for the presence of new available viral

sequences, as well as extended to sequences

extracted from different databases such as GenBank

and Uniprot. It presently contains the sequences and

the cross-references to the original viral and human

databases. Further investigation is now been carried

out on the VISHUM database; as an example, we

selected from the VISHUM database the herpesvirus

fragments and sorted the results by sequence

identity. The herpesvirus selected fragments

showing at least 8 continuous identities and at least

90% of sequence identity with human proteins are

listed in Table 1. In all cases, with just one

exception, the extension of the peptide is longer or

much longer than the minimum requested threshold.

While a simpler search strategy could be applied,

nevertheless we aimed at finding results significant

from both structural and functional point of view, to

select fragments with a full identity for a minimum

length of 8 amino acids (as minimum requirement

for a putative antigenic activity), further extended to

the maximum length of identity or similarity; the

BLAST evaluation of scores and significance was

then applied to each fragment to fix the best level of

extension.

To our knowledge this study reports for the first

time a systematic approach to investigate structural-

functional correlations of viral proteins with human

proteins. Choosing 8 consecutive residues, as the

minimum requested identity, represented the way to

wipe out all short identities difficult to evaluate from

the statistical and functional point of view, and

allowed to select those with high statistical

significance and with an immune-related biological

role. In fact, antigen presenting cells (APCs) carry 8-

residues long peptides to lymphocytes to elicit the

immune response. Additional post-filtering may

further improve the fragment selection procedure

and the further ongoing investigation of the

VISHUM database. For instance only human-

specific viral fragments may be selected to be

compared to human sequences or both human-

specific and human-non-specific fragments can be

selected to implement evolutionary-based analyses.

Moreover, our study can also be compared to data

collected from public databases related to human –

virus interactions. We present here data obtained by

a post-filtering procedure based on virus name (see

Table 1); other filters may be implemented within

this strategy, such as, for an example, a 3D-filter to

select fragments with a known 3D structure.

The procedure presented in the current study

allows to identify viral fragments sharing full

identity with human proteins, with a strong

statistical significance. The occurrence of highly

significant local identities may indicate specific

regions with common ancestry or regions where a

local evolutionary pressure or a molecular mimicry

process may have occurred. A very recent report

underlines the role of viral infections as possible

trigger of autoimmune disorders in genetically

predisposed individuals (Ji, 2010); hence developing

VISHUM database and effective procedures to

A COMPUTATIONAL STRATEGY TO INVESTIGATE RELEVANT SIMILARITIES BETWEEN VIRUS AND

HUMAN PROTEINS - Local High Similarities between Herpes and Human Proteins

185

Table 1: Results generated by the search strategy, filtered for human herpesviruses. The best hits are shown, in terms of

percentage of identity (at least 90%).

Peptide

Code

Length Virus name

BLAST results

Score E-value Number of matches Percentage of identity

10932 11 Human herpesvirus 5 26.2 3.8 11 100

18349 42 Human herpesvirus 4 type 2 87.8 2.00E-18 42 100

18349 42 Human herpesvirus 4 87.8 2.00E-18 42 100

22175 13 Human herpesvirus 4 type 2 30.4 0.19 12 100

22175 13 Human herpesvirus 4 30.4 0.19 12 100

22175 13 Human herpesvirus 8 28.5 0.8 11 100

28804 10 Human herpesvirus 8 25 6.4 10 100

34315 11 Human herpesvirus 8 26.6 2.5 11 100

34752 14 Human herpesvirus 6 35.8 0.006 14 100

34752 14 Human herpesvirus 7 35 0.01 14 100

35947 12 Human herpesvirus 8 28.5 0.66 11 100

4113 12 Human herpesvirus 8 28.5 0.72 12 100

41176 20 Human herpesvirus 8 44.7 1.00E-05 20 100

43174 14 Human herpesvirus 8 32.3 0.055 14 100

44922 9 Human herpesvirus 2 24.6 7.3 9 100

46915 11 Human herpesvirus 8 31.6 0.089 11 100

53339 31 Human herpesvirus 4 type 2 63.9 3.00E-11 31 100

53339 31 Human herpesvirus 4 63.9 3.00E-11 31 100

5707 9 Human herpesvirus 5 24.3 10 9 100

57304 13 Human herpesvirus 6 28.9 0.55 12 100

6005 8 Human herpesvirus 2 24.3 8.3 8 100

62235 24 Human herpesvirus 8 55.8 9.00E-09 24 100

55656 39 Human herpesvirus 4 type 2 45.8 9.00E-06 38 97

55656 39 Human herpesvirus 4 45.8 9.00E-06 38 97

58282 20 Human herpesvirus 8 43.5 4.00E-05 18 94

40584 15 Human herpesvirus 3 34.7 0.014 14 93

40584 15 Human herpesvirus 8 33.1 0.041 14 93

34752 14 Human herpesvirus 5 34.7 0.013 13 92

35527 15 Human herpesvirus 8 33.5 0.028 13 92

41386 14 Human herpesvirus 4 33.5 0.026 13 92

41386 14 Human herpesvirus 4 type 2 33.5 0.026 13 92

57304 13 Human herpesvirus 8 32.3 0.05 12 92

57304 13 Human herpesvirus 4 type 2 32.3 0.05 12 92

57304 13 Human herpesvirus 4 32.3 0.05 12 92

35158 23 Human herpesvirus 4 type 2 45.4 1.00E-05 21 91

35158 23 Human herpesvirus 4 45.4 1.00E-05 21 91

57304 13 Human herpesvirus 6 27.3 1.8 11 91

62235 24 Human herpesvirus 3 53.5 4.00E-08 22 91

16190 22 Human herpesvirus 8 48.5 1.00E-06 20 90

40505 21 Human herpesvirus 8 45.4 1.00E-05 19 90

41737 10 Human herpesvirus 4 type 2 26.6 2.1 9 90

41737 10 Human herpesvirus 4 26.6 2.1 9 90

investigate local protein regions may represent a step

forward the improved understating of mechanisms

underlying virus-related disorders.

4 CONCLUSIONS

The present study reports a novel procedure to

identify relevant similarities among virus and human

BIOINFORMATICS 2011 - International Conference on Bioinformatics Models, Methods and Algorithms

186

proteomes, useful to investigate structural/functional

features underlying infectious and non-infectious

diseases.

ACKNOWLEDGEMENTS

The present project was partially supported by

Programma Italia-USA “Oncoproteomica” (Ref.

527B/2A/5), Programma Italia-USA

“Farmacogenomica Oncologica” (Ref. 527/A/3A/5)

and by Biophysikalische Informations-Therapie

(Freiburg). We also thank Ruth Beisler for the

enthusiastic, stimulating and helpful support to the

entire project.

REFERENCES

Baranzini, S. E., The genetics of autoimmune diseases: a

networked perspective. Curr. Opin. Immunol. 2009.

21: 596-605.

Rubtsov, A. V., Rubtsova, K., Kappler, J. W., and P.

Marrack. Genetic and hormonal factors in female-

biased autoimmunity. Autoimmun. Rev. 2010. 9:494-8

Leggatt, G. R. and I. H Frazer. HPV vaccines: the

beginning of the end for cervical cancer. Curr. Opin.

Immunol. 2007. 19: 232-238.

Cao, X., Tsukamoto, T., Nozaki, K., Tanaka, H., Cao, L.,

Toyoda, T. et al.. 2007 Severity of gastritis determines

glandular stomach carcinogenesis in Helicobacter

pylori-infected Mongolian gerbils. Cancer Sci. 98:

478-483.

Neuveut, C., Wei, Y., and M. A. Buendia. Mechanisms of

HBV-related hepatocarcinogenesis. J. Hepatol. 2010.

52:594-604.

Makela, M., Vaarala, O., Hermann, R., Salminen, K.,

Vahlberg, T., Veijola, R. et al. 2006 Enteral virus

infections in early childhood and an enhanced type 1

diabetes-associated antibody response to dietary

insulin. J. Autoimmun. 27:54-61.

Peng, H. and W. Hagopian. 2006 Environmental factors in

the development of Type 1 diabetes. Rev. Endocr.

Metab. Disord. 7:149-162.

Ott, S. J., El Mokhtari, N. E., Musfeldt, M., Hellmig, S.,

Freitag, S., Rehman, A. et al. 2006 Detection of

diverse bacterial signatures in atherosclerotic lesions

of patients with coronary heart disease. Circulation

113:929-937.

Li, Y., Heuser, J. S., Cunningham, L. C., Kosanke, S. D.,

and M. W. Cunningham. 2006 Mimicry and Antibody-

Mediated Cell Signaling in Autoimmune Myocarditis.

J. Immunol. 177:8234-8240.

Leite, J. L., Bufalo, N. E., Santos, R. B., Romaldini, J. H.,

and L. S. Ward. 2010 Herpesvirus type 7 infection

may play an important role in individuals with a

genetic profile of susceptibility to Graves' disease.

Eur. J. Endocrinol. 162:315-321.

Zhao, Z. S., Granucci, F., Yeh, L., Schaffer, P. A., and H.

Cantor. 1998. Molecular Mimicry by Herpes Simplex

Virus-Type 1: Autoimmune Disease After Viral

Infection. Science 279:1344-1347.

Kirby, B., Al-Jiffri, O., Cooper, R. J., Corbitt, G., Klapper,

P. E., and C. E. Griffiths. 2000. Investigation of

cytomegalovirus and human herpes viruses 6 and 7 as

possible causative antigens in psoriasis. Acta Derm.

Venereol. 80:404-406.

Mehraein, Y., Lennerz, C., Ehlhardt, S., Zang, K. D. and

H. Madry. 2004. Replicative multivirus infection with

cytomegalovirus, herpes simplex virus 1, and

parvovirus B19, and latent Epstein-Barr virus infection

in the synovial tissue of a psoriatic arthritis patient. J.

Clin. Virol. 31:25-31.

Thomas, D., Karachaliou, F., Kallergi, K.,

Vlachopapadopoulou, E., Antonaki, G., Chatzimarkou,

F. et al. 2008. Herpes virus antibodies seroprevalence

in children with autoimmune thyroid disease.

Endocrine 33:171-175.

Fumagalli, M., Pozzoli, U., Cagliani, R., Comi, G. P.,

Bresolin, N., Clerici, M., and M. Sironi. 2010.

Genome-Wide Identification of Susceptibility Alleles

for Viral Infections through a Population Genetics

Approach. PLoS Genet. 6:e1000849.

Facchiano A. 1995. Investigating Hypothetical Products of

Non-coding Frames (HyPNoFs), J. Mol. Evol. 40:570-

577.

Facchiano A. 1996. Coding in noncoding frames. Trends

in Genetics 12:168-169.

Limou, S., Le Clerc, S., Coulonges, C., Carpentier, W.,

Dina C., Delaneau, O., et al. 2009. Genomewide

association study of an AIDS-nonprogression cohort

emphasizes the role played by HLA genes (ANRS

Genomewide Association Study 02). J. Infect. Dis.

199:419-426.

Fellay, J., Shianna, K. V., Ge D., Colombo, S.,

Ledergerber, B., Weale, M. et al. 2007. A whole-

genome association study of major determinants for

host control of HIV-1. Science 317:944-947.

Benvenga, S., and A. Facchiano, 1995. Homology of atrial

natriuretic protein with the proteins associated with

amyloidosis. J. Int. Medicine 237:525-526.

Benvenga, S., Alesci, S., Trimarchi, F., and A. Facchiano.

Homologies of the thyroid sodium-iodide symporter

with bacterial and viral proteins. Endocrinol. Invest.

1999. 22: 535-540.

Benvenga, S., Trimarchi, F., and A. Facchiano. 2003.

Cogan’s syndrome as an autoimmune disease. The

Lancet 361:530-53l.

Yu, H., Chen, J. K., Feng, S., Dalgarno, D. C., Brauer, A.

W., and S. L. Schreiber. 1994. Structural basis for the

binding of proline-rich peptides to SH3 domains. Cell

76:933-945.

Wender, P. A., Mitchell, D. J., Pattabiraman, K., Pelkey,

E. T., Steinman, L., and J. B. Rothbard. The design,

synthesis, and evaluation of molecules that enable or

enhance cellular uptake: peptoid molecular

A COMPUTATIONAL STRATEGY TO INVESTIGATE RELEVANT SIMILARITIES BETWEEN VIRUS AND

HUMAN PROTEINS - Local High Similarities between Herpes and Human Proteins

187

transporters. Proc. Natl. Acad. Sci. U S A. 2000. 97:

13003-13008.

Serasinghe, M. N., Seneviratne, A. M., Smrcka, A. V., and

Y. Yoon. 2010. Identification and characterization of

unique proline-rich peptides binding to the

mitochondrial fission protein hFis1. J. Biol. Chem.

285:620-630.

Ji, Q., Perchellet A., and Goverman J. M. 2010 Viral

infection triggers central nervous system

autoimmunity via activation of CD8+ T cells

expressing dual TCRs. Nature Immunology 11: 628-

635

BIOINFORMATICS 2011 - International Conference on Bioinformatics Models, Methods and Algorithms

188