Background
Vaccinia virus (VACV) is the best-studied member of the Orthopoxvirus genus of the poxvirus family. It has a wide host range and is able to infect cells of many different origins. VACV has played important roles in medicine and biomedical research. As VACV highly stimulates both the innate and adaptive arms of the immune system, it was used as the vaccine for eradication of smallpox and recently, the virus has been used as a live recombinant vaccine for the induction of protective immune response against many pathogens in experimental animals. VACV genome consists of a of 190 kbp dsDNA encoding over 200 proteins. The non-essential genes are used for the insertion of our gene of interest [
1]. The resultant recombinant virus (rVACV) usually expresses foreign genes without remarkable impact on viral infectivity. Recombinant proteins are correctly posttranslationally modified, properly localized or secreted from infected cells.
Flt3 ligand (FL) is a hematopoietic growth factor that plays an important role in the life cycle of several blood cells. It is produced by bone marrow stromal cells, T cells and endothelial cells and by a number of organs including spleen, ovary, testis, intestine and kidney. FL alone induces differentiation of macrophages in CD34+ cell culture and stimulates increase in dendritic cell numbers [
2‐
8]. When FL is administered to mice, hematopoietic stem cells and progenitors in the bone marrow and spleen are expanded and mobilized into the peripheral blood. FL increases beta-1-integrins or P-selectin expression and downregulates VCAM-1 on peripheral blood and folicular cells [
9‐
13]. Moreover, the ligand acts in synergy with other cytokines, including stem cell factor (SCF), granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukins 3, 6, 7, 11 and 15. Stimulation by FL leads to proliferation, differentiation, maintenance and long-term reconstitution of primitive hematopoietic cells (both lymphoid and myeloid progenitors) [
5,
14‐
16]. FL dramatically enhances the production of antibodies to soluble antigens
in vivo [
17]. Systemic inoculation enhances the production of IFN-γ, IL-12, GM-CSF and IL-5 which results in increase of cytotoxic T lymphocytes, natural killer cells and dendritic cells in blood [
18‐
20].
Human FL shares high homology with mouse FL in themino acid sequence, mainly in the extracellular part of the molecule, and is able to activate mouse Flt3 receptor [
21]. The human Flt3L gene encodes a 235-amino acid type I transmembrane protein consisting of four domains: 1) an N-terminal 26-residue signal peptide, 2) a 156-residue extracellular domain, 3) a 23-amino acids transmembrane domain, and 4) a 30-residue cytoplasmic domain [
4,
15,
22]. FL is expressed in membrane-bound and soluble forms. The cytokine is biologically active both in the transmembrane form and in the soluble form that is thought to be released into the circulation from the cell membrane by protease cleavage or is produced directly as the alternatively spliced soluble isoform [
15,
22‐
24]. The extracellular domain alone has been shown to be sufficient for bioactivity [
23]. FL exists in both monomeric and homodimeric forms. Soluble FL can be a noncovalently linked oligomer and contains six cysteine residues in each molecule that apparently form intramolecular disulfides. The integrity of the FL dimer seems to be essential for bioactivity; moreover, the fusion of two soluble FL molecules can increase the activity of the ligand [
25,
26]. FL belongs to the family of short chain helical cytokines where the three-dimensional structures of five members, i.e. interleukin-4 (IL-4), IL-2, IL-5, GM-CSF and MCSF, have been solved [
27]. The FL monomer has the most similar protein structure to IL-4 although the effects on blood cells are of different type [
25,
28].
The FL receptor, Flt3 (Fms-Like Tyrosine kinase 3), belongs to members of the class III receptor tyrosine kinase family of transmembrane glycoproteins and is structurally related to the c-kit (KIT), c-fms (FMS), and platelet-derived growth factor (PDGF) receptors. The receptor is expressed only in a limited number of tissues, including the human bone marrow, thymus, spleen, liver, and lymph nodes.
In this study, we examined the influence of FL production on the life cycle of recombinant virus. We constructed two recombinant vaccinia viruses of the Praha strain (clone P13) designed for the expression of the human gene encoding a soluble isoform of FL (sFL). The rVACVs were characterized and compared for their multiplication and virulence in vitro and in vivo, and for the ability to ensure the secretion of FL from infected cells.
We found out that the FL overexpression substantialy influenced the properties of rVACV. High production of FL resulted in decreased rVACV multiplication in macrophages and in mice. Biochemical and electron microscopic analysis of the recombinant virions revealed changes in the protein composition and incorporation of FL into the virion core. We have shown that the overexpression of the Flt3L gene in VACV results in attenuation of the virus in vivo.
Discussion
Several studies have shown that FL protein can inhibit tumor growth
in vivo and acts as a vaccine adjuvant. Instead of direct FL protein administration, the production of FL by a viral vector might result in prolonged FL presence in the body and in improved therapeutic activity [
30].
We constructed two recombinant VACVs derived from the Praha vaccine strain whose expression of FL was controlled by the natural early H5 promoter which is important for early antigen presentation, or by the synthetic E/L promoter with activity at both early and late times during VACV infection [
31,
32]. The original goal of our project was to use FL-expressing viruses for stimulation of antigen presenting cell activity and adaptive immunity in tumor bearing animals. Indeed, we have shown that immunization with VACV co-expressing sFL with the E7 protein of HPV16 as a tumor antigen inhibited the formation and growth of TC1 tumors in mice [
33]. In that study we observed that despite high expression of FL controlled by the synthetic E/L promoter
in vitro, the double recombinant P13-E/L-FL-SigE7LAMP induced only low levels of FL in the serum of inoculated mice. In an attempt to elucidate the nature of the inhibitory effect of FL overexpression controlled by the E/L promoter on the release of recombinant cytokine in mouse serum, we focused on a more detailed study of FL-producing recombinants.
The first step was to characterize the multiplication and FL production in vitro. We compared the replication of P13-H5-FL, P13-E/L-FL and control virus in CV1 cells by virus progeny titration. No distinct effect of recombinant protein expression on the replication was observed, apart from the moderately delayed release of P13-E/L-FL virus from infected cells. ELISA tests of media and cell lysates of infected cultures confirmed the generally accepted idea of the strength of natural H5 promoter of VACV which is active mainly in the early phase of infection and of the synthetic E/L promoter whose activity is increasing from the early to the late phase of infection.
Then we administered these viruses intraperitoneally to 6-weeks-old mice and measured the expression of FL
in vivo for several days after inoculation. The animals inoculated with P13-E/L-FL did not have highly elevated serum levels of FL, similarly to the previous experiments done with double recombinant P13-E/L-FL-SigE7LAMP. Moreover, we found out, by titration and by quantitative PCR of viral DNA in mouse ovaries, that the P13-E/L-FL virus was not able to multiply
in vivo. It was seemingly in contradiction with our previously published results where the expression of FL under the control of the E/L promoter did not affect the multiplication of the double recombinant P13-E/L-FL-SigE7LAMP in the mouse ovaries in comparison with the control virus and with the double recombinant P13-H5-FL-SigE7LAMP [
33]. When comparing the double recombinant P13-E/L-FL-SigE7LAMP with the single recombinant P13-E/L-FL
in vitro, we found the latter to produce higher levels of Flt3L in infected CV1 cells (not shown). Decrease in FL production by double recombinant
in vitro could be ascribed to the inactivation of the F7L locus as a result of the insertion of the SigE7LAMP gene, which is known to down-regulate the expression of the gene inserted in the TK locus [
34]. The overexpression of FL by the single recombinant was so high that it resulted not only in limited production of FL
in vivo but even in its decreased multiplication in mice. The block of P13-E/L-FL multiplication was confirmed by the examination of the independently derived recombinants and by the deletion of the E/L-FL expression cassette followed by reversion to the wt phenotype. We checked also the DNA pattern of the genome of all analyzed viruses using SalI, HindIII, PstI, XhoI and KpnI endonucleases and found no obvious differences in the restriction patterns (not shown). It was evident that the inhibition of virus multiplication
in vivo was caused by sFL overexpression under the control of the E/L promoter.
Attenuation of recombinant vaccinia vectors in consequence of the foreign gene insertion has been described for viruses expressing IL2 [
35], IL12 [
36] and IL15 [
37]. The multiplication was inhibited
in vivo and in lymphoid cell lines; however, the infected fibroblasts produced the control and cytokine-expressing viruses in equivalent titers.
The response to VACV infection has been studied in several species [
38‐
40]. It has been illustrated recently in variola primate model that poxviruses productively infect large populations of circulating monocytes and macrophages in the lymph nodes, spleen and other tissues [
41]. For our study, we selected as an
in vitro target cell model the macrophage cell line J774.G8 which supports the growth of VACV. The multiplication of the P13-E/L-FL virus but not of the P13-H5-FL virus was restricted in confluent cultures of this cell line. Taken into consideration, this situation is likely to simulate the
in vivo state where the macrophages are terminally differentiated cells that rarely divide. We also determined the level of FL expression in J774.G8 cells by P13-H5-FL or P13-E/L-FL. Despite the sFL expression driven by promotors of different strengths, we found the same FL secretion level by either virus. It could mean that the inhibition of virus multiplication is not mediated by high level of extrinsic FL. This fact has been supported by the failure to find any Flt3 (CD135) molecules on the membrane of J774.G8 cells and by experimental addition of extrinsic Flt3L to J774.G8 cells that had no effect on the multiplication of the control P13-preS2S-βgal or P13-E7 viruses at any step of infection.The inhibition of multiplication of P13-E/L-FL in macrophages might be reversible since we showed the dependence of β-galactosidase production by P13-βgal-E/L-FL on the growing activity of cell cultures.
To exclude the possibility that apoptosis is responsible for inefficient multiplication of P13-E/L-FL in macrophages, the presence of the apoptosis marker Annexin-V and the cleavage of PARP were determined. For these experiments, we prepared the double recombinants expressing both the FL gene and GFP protein and used them for following up the early apoptosis marker (Annexin-V) in infected cells during virus replication. There were not significant differences among viruses in the infectivity (GFP positive cells amount) or in Annexin-V binding (not shown). We also tested the cleavage of PARP protein (late apoptosis marker) in macrophages infected by single recombinant viruses. The cleavage was obvious in cells infected by parental virus or P13-H5-FL and slightly in P13-E/L-FL infected macrophages. As a control, we added ara-C to macrophages, which caused PARP cleavage in all infected and non-infected macrophages (not shown). We concluded that the attenuation of P13-E/L-FL was not due to the enhanced apoptosis in macrophages.
After challenging the assumption that the antiviral state of macrophages is mediated by high levels of Flt3L produced by recombinants during infection, we considered the possibility of changes in the elementary protein composition of the virion itself. The strong expression of FL driven by the synthetic E/L promoter led to its incorporation into a core-associated protein fraction. This was due to the strong FL expression in BSC40 or HeLa cells used for virus stock preparation. The endoplasmic reticulum and Golgi apparatus of P13-E/L-FL infected cells contained a huge amount of FL protein. The foreign gene product has been reported previously to be trapped in the virion due to protein-protein interactions during the virion assembly process [
42‐
44]. Bereta et al. have indirectly shown that the rVV expressing CD40L gene contains biologically active CD40L protein in particles [
45] However, there is some selectivity in the encapsidation process. There is evidence of the incorporation of recombinant protein into one or more virion compartments. Vaccinia virus expressing the bacterial CAT gene incorporated the enzyme into the virus particle [
46]. The expression of the cytokine IL-12 led to the incorporation of about 0.01% of the total recombinant protein into an envelope fraction, HIV1 env (gp160-120) was tightly bound in protein-DNA complexes, and the enzyme beta-galactosidase was found exclusively in core-associated fraction [
47]. Transport and sorting of viral proteins directly from the endoplasmic reticulum into the growing immature virions using non-COPII vesicles [
48] could explain selective integration of the low-glycosylated FL into P13-E/L-FL virions. A similar selectivity for the integration of non-glycosylated form of a glycoprotein into the membrane of IMV particles was observed for A14 [
49]. However, FL was not incorporated into a membrane fraction, but it was tightly bound to core, although it was also exposed on the surface of IMV to offer the epitopes for anti-FL antibodies as verified by neutralization assays and proven by electron microscopy. Moreover, the localization of the recombinant proteins may be facilitated by specific protein-protein interactions. Variation in the localization of recombinant protein has been reported in double recombinants expressing the protein of interest and β-galactosidase [
47]. We also analysed the FL distribution also in double recombinant viruses P13-βgal-H5-FL and P13-βgal -E/L-FL and found no changes in comparison with the single recombinant viral particles.
In the present study, we have shown that the integration of sFL into the IMV virions was associated with an altered composition of the virions. We observed that P13-E/L-FL virions contained H3L of a higher molecular weight than P13 or P13-H5-FL (lanes 1 and 4). The H3 protein is an immunodominant component of IMV, binds to heparansulphate and is found in IMV in two isoforms, (1-324 aa) and (48-324 aa) [
50]. Both the full length protein and N-terminally deleted form can be incorporated into the viral membrane [
51] although the specific functions of the H3 isoforms are yet not known. As the H3 protein is nonessential for virus multiplication in cell monolayers [
52], the integration of FL into the virion might modify the ratio of H3 isoforms. Similarly to H3, the second differently displayed virion component, the D8 protein, plays the role as the glycosaminoglycan (GAG) binding molecule [
53]. It has been shown, that the D8 protein integrated into the viral particle can be cleaved by trypsin without any decrease of virus infectivity [
54]. The D8 protein detected in band 5 could be a cleavage product of trypsin like protease. The unusual forms of both GAG binding structures resulting from high FL gene expression may influence the growth of virus in some cell types. Differences in these two membrane proteins in association with FL gene expression could imply that the presence of FL in the virus core also affects the composition of the IMV core. The pattern of core proteins showed multiple band changes. One dominant band contains the p28K protein encoded by L4R. This protein is a basic DNA binding protein and plays an essential role in virus replication. The protein fragment (33-251 aa) produced by Ala-Gly-specific cleavage is usually found in the virus core in the absence of its precursor [
55]. P13-E/L-FL virions contain less P28K than P13 or P13-H5-FL. As we did not analyze the adjacent bands, we cannot say whether or not decrease in the cleaved form of the L4 protein is associated with the presence of the P28K precursor in purified virions. There are reduced levels of early RNA and protein production in cells infected by L4R deficient vaccinia virus particles [
56], which could also explain decreased levels of mRNA throughout the course of macrophage infection by P13-E/L-FL virus (data not shown). P13-E/L-FL virions yielded additional bands containing host proteins. Tubulin β chain has been found previously in non-recombinant vaccinia virus IMV where it can form up to 0.7% of protein content [
57]. We found an increased amount of tubulin β chain in the core fraction together with increased FL in P13-E/L-FL. MALDI analysis of the core fraction yielded GAPDH. This enzyme, which is involved in many different cellular processes, has also been found incorporated in human immunodeficiency virus type 1 (HIV-1) virions [
58]. GAPDH binds actin filaments
in vitro [
59] and therefore could be incorporated in the complex with actin. Actin has been identified in virions of VACV [
57] several other DNA viruses and HIV.
Materials and methods
Plasmids
The pHUFLT3L plasmid containing the coding sequence of the soluble isoform of FL (ID number - U29874) was obtained from Immunex (now the part of Amgen Inc.). The plasmids pSC59-H5-FL and pSC59-E/L-FL have been described earlier [
33]. The expression plasmid pBSC-FL was prepared by ligation of an EcoRI fragment of pHUFLT3L plasmid carrying the FL coding sequence with the pBSC plasmid [
60] cleaved with the same enzyme. The pTK
+ plasmid was derived from pGS20 [
61] by excision of the EcoRI fragment. The pD357 plasmid [
62] containing an E.coli β-galactosidase gene under the control of the P7.5 promoter was used for replacement of the C23L and B29R genes by the β-galactosidase gene
Viruses
Vaccinia virus strain Praha, clone 13 [
63] was used as the parental virus. Single recombinants P13-H5-FL and P13-E/L-FL were prepared by the insertion of FL gene into thymidine kinase using plasmids pSC59-H5-FL and pSC59-E/L-FL, respectively, followed by selection in medium supplemented with bromodeoxyuridine. The double recombinants P13-βgal-H5-FL and P13-βgal-E/L-FL were prepared by the insertion of β-galactosidase into the FL-expressing single recombinants using pD357 plasmid and by the selection of the virus forming stable blue plaques after three purification steps. The revertant virus P13-ΔE/L-FL was prepared using the pTK
+ plasmid carrying the functional thymidine kinase gene; the selection occurred in 143B cells grown in HAT supplemented E-MEM medium. Virus P13-E7 carrying the E7 early protein of HPV16 has been described earlier [
64]. Virus P13-βgal-pS2S was derived from P13-pS2S [
65] using pD357 plasmid. Viruses were grown in BSC40 cells, purified by sucrose-gradient centrifugation [
66] and titrated in CV-1 cells. The number of virus particles was determined from the optical density measured at 260 nm using the formula 1 U of OD
260 nm = 1.2 × 10
10 viral particles/ml [
67]. The ratio was comparable for all the viruses used in the experiments.
Cell lines
CV-1 and BSC-40 African green monkey kidney cell lines were grown in Modified E-MEM medium (EPL, SEVAPHARMA, Prague) containing bovine serum growth-active proteins but no complete serum [
68]. Human embryonic kidney 293T cells [
69]kindly provided by J.A. Kleinschmidt, DKFZ, Heidelberg, Germany, human HeLa cell line and J774.G8 mouse macrophage cell line were grown in DMEM (PAA Laboratories, Linz, Austria) supplemented with 10% fetal bovine serum (FBS; PAA Laboratories). The cell lines MOLM9 and 32D were grown in RPMI-1640 medium (Sigma, Saint Louis, MO) supplemented with 10% FBS. The 143B cell line was maintained in E-MEM medium (SEVAPHARMA, Prague) supplemented with 10% FBS. Each medium contained 2 mM L-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin.
Mice
Six-week-old C57Bl/6 (H-2b) female mice were obtained from Charles River, Germany. Animals were maintained under standard conditions at the National Institute of Public Health (Prague). The experiments were performed in compliance with Acts Nos. 246/92 and 77/2004 on animal protection against cruelty and Decree No. 311/97 of the Ministry of Health of the Czech Republic, on the care and use of experimental animals. Mice were injected intraperitoneally (i.p.) with 0.5 ml PBS containing sonicated suspension of sucrose-purified particles of rVACV. Mice were anesthetized with halothane (Narcotane, Léciva, Praha) and carotid blood was collected at indicated time intervals.
ELISA
FL was quantified with an Flt3 ligand ELISA detection kit (R&D Systems GmBH, Wiesbaden-Nordenstadt, Germany) using the capture mouse monoclonal antibody MAB608 (100 ng/well), biotinylated detection goat polyclonal antibody BAF308 (7.5 ng/well), streptavidin-HPR (1:250) or avidin-HPR (1:1000) complex, both obtained from Pharmingen (BD Biosciences, Erembodegem, Belgium), and TMB substrate solution for visualization of the reaction. Samples were measured by an ELISA reader at 450 nm. Standard Flt3 ligand protein (PeproTech EC Ltd, London, UK) was diluted to 500-7,5 pg/ml. Detection of VACV-specific antibodies has been described earlier [
70].
SDS-PAGE and western blot
Infected cells or purified viral particles were extracted with denaturing, reducing sample buffer [
71]. Samples were separated by SDS-PAGE in 10% or 12% gels. Proteins were blotted onto a nitrocellulose membrane (Hybond-C Extra, Amersham) and after blocking with 10% skimmed dry milk in PBS, the membrane was incubated with primary antibody BAF308 (anti-FL, R&D Systems GmBH, Wiesbaden-Nordenstadt, Germany) diluted 1:500 or convalescent mouse serum (anti-VACV) diluted 1:50 - 1:100. After washing, the membrane was incubated with rabbit anti-mouse IgG horseradish-peroxidase-conjugated secondary antibody (Sigma-Aldrich, Steinheim, Germany). Proteins were visualized with the ECL Plus system (Amersham).
Mass spectrometry and protein identification
Electrophoretic gels were stained with Coomassie blue. Selected spots on the preparative gels were excised and destained using 50% acetonitrile in 25 mM ammonium bicarbonate, dehydrated with 200 μl of acetonitrile for 5 min at 30°C and then vacuum-dried (SpeedVac, Thermo Scientific, Waltham, Ma). Gel pieces were rehydrated and proteins were digested for 8 hours at 37°C with 30 ng/μl trypsin (Trypsin Gold Mass Spectrometry Grade, Promega, Madison, WI) in 25 mM ammonium bicarbonate. After digestion, peptides were extracted from gel pieces using step by step extraction with an acetonitrile gradient (15%-60% acetonitrile with 1% trifluoroacetic acid) using sonicator (Elma, Singen, Germany) cooled with ice cubes. Extracted peptides were concentrated in SpeedVac. MALDI mass spectrometry (MALDI/MS) peptide mass fingerprint analysis was used to characterize the digests. The MALDI/MS was performed in a Refelex IV MALDI-TOF mass spectrometer (Bruker). Data were processed by proteomic software Mascot.
Beta-galactosidase assay
Beta-galactosidase activity was determined according to Miller [
72]. The samples of infected cells were frozen and thawed and then centrifuged to remove cellular debris. Beta-galactosidase activity of cell extracts was measured by a colorimetric assay using o-nitrophenyl β-D-galactopyranoside (ONPG). The absorbance of samples was determined at 450 nm.
Virus neutralization
Sonicated, sucrose purified virus particles were incubated in a minimal volume of PBS, with 0.5 μg of rabbit polyclonal antibody against FL (MBL, Woburn, MA) or with rabbit and mouse anti-VACV serum or negative serum. After a 1-hour-incubation at 37°C, the viruses were used for the infection of confluent cell layers or for the preparation of electron-microscopy samples.
Electron microscopy
Metal grids were freshly coated with a Formvar (polyvinylformal, Serva) membrane. Five to ten μl of viral suspension (sonicated or antibody-treated) were absorbed to the grid for 10 min and then washed twice with water and twice with 1% phosphowolframic acid, pH 9.0, each time for 1 min. The samples were observed by transmission electron microscope JEM1011 (JEOL, Tokyo, Japan) at indicated magnifications.
Fluorescent microscopy
The cell monolayer grown on a round glass plate was infected with virus at a MOI of 2 at 37°C for 30 minutes. The inoculum was then removed and cells were cultivated in DMEM. At time intervals not longer than 12 hours post infection (h.p.i.), the cells were washed with cold PBS, fixed for 10 minutes in 4% paraformaldehyde (PA) and permeabilized for 10 minutes in 2% PA with 1% Triton-X100. The remaining PA was neutralized by incubation of samples with 0.1 M glycine in PBS for 10 minutes. After blocking in 10% skimmed dry milk in PBS for 30 minutes, samples were incubated with anti-FL MAB608 (diluted 1:500) and with anti-GM130 or with anti-calreticulin (both diluted 1:100; Santa Cruz Biotechnology) in 5% dry milk in PBS for 30 minutes, washed five times (PBS, 0.2% Tween 20) and incubated with secondary antibody against mouse IgG labeled with Alexa fluor 488 and/or with antibody against rabbit IgG labeled with Alexa fluor 546 (Invitrogen, USA), diluted 1:500 in 5% dry milk in PBS for 30 minutes and counterstained with 1 μg/ml propidium iodide or DAPI (Sigma-Aldrich Gmbh, Munich, Germany). The washed plates were observed by Nikon E600 fluorescence microscope for green, red and blue signals at a magnification of 1000×.
Fractionation of viral proteins
Virion samples containing equal amounts of proteins (determined by the Bradford protein assay, BioRad) were incubated in 50 mM Tris-buffer with 10 mM MgCl
2, pH 8.5, at 37°C for 30 minutes and subsequently supplemented with detergents for solubilization of two membrane and two core fractions (soluble lipid envelope fraction "M1", protein-matrix-like "M2", soluble core "C1" and DNA-core fraction "C2") [
47]. As the supplements for the membrane and core fractions, 1% NP-40 or 1% NP-40 plus 50 mM DTT and 0.5%DOC plus 0.1% SDS or Laemmli buffer were used, respectively. Each fraction was separated from insoluble proteins by centrifugation at 13000 g for 10 min.
Quantitative PCR (Q-PCR)
Mice were anaesthetized with halothane (Narcotan, Léciva, Praha) and sacrificed. The ovaries were dissected, washed in PBS and homogenized. DNA was extracted using DNeasy Tissue Kit (Qiagen). Real-time quantitative PCR was performed as described previously [
33].
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
KZ participated in the design of the study, performed most of the experiments and drafted the manuscript. PH prepared the succrose purified virus. JK carried out the quantitative PCR. LK performed ELISA for detection of anti-vaccinia antibodies and provided valuable background for manipulation with vaccinia virus, cell cultures and mice. MS carried out MALDI. SN conceived of the study, participated in its design and coordination and helped in elaboration of the manuscript. All authors read and approved the final manuscript.