Isolation of a Bovine Plasma Fibronectin-Containin
http://www.100md.com
病菌学杂志 2005年第13期
Department of Pediatrics, University of Western Ontario, London, Ontario, Canada
Department of Pathobiology, University of Guelph, Guelph, Ontario, Canada
ABSTRACT
Bovine leukemia virus (BLV) is a deltaretrovirus that infects cattle worldwide. In agriculturally intensive regions, approximately 30% of dairy cows are BLV infected. Like the human T-cell leukemia virus (HTLV), there is a lengthy period of viral quiescence after initial infection with BLV. Unlike HTLV, BLV resides predominantly in B cells. Lymphoma is observed in less than 10% of BLV-infected adult cattle. Although viremia is undetectable in vivo, BLV-infected peripheral blood mononuclear cells readily become productive when cultured in vitro. Productivity is markedly diminished when cultures are supplemented with bovine plasma. This inhibitory activity of bovine plasma has been attributed to the "plasma blocking factor" (PBF). Here, we describe the purification of a PBF whose activity was resistant to heating to 65°C for 10 min and was attributable to a fibronectin-containing complex of approximately 320 kDa under nonreducing conditions. By use of two-dimensional polyacrylamide gel electrophoresis and matrix-assisted laser desorption ionization-time of flight (mass spectrometry), a protein with a size of 220 kDa and a pI of 5.4 was identified as a member of the fibronectin group of molecules. Both the purified protein and the commercially available bovine fibronectin inhibited BLV production in naturally infected peripheral blood mononuclear cells, although the fibronectin was less biologically active.
INTRODUCTION
Bovine leukemia virus (BLV), the etiological agent of lymphoma in adult cattle, is closely related to human T-cell leukemia virus (HTLV) types I and II. BLV and HTLV share features such as random integration into the host genome, a long period of viral latency, and rare development of disease (4, 5, 12). The HTLV/BLV genus is defined by the presence of a unique open reading frame called pX, located at the 3' end of the genome and coding for transregulatory proteins Tax and Rex (7, 8, 19, 22). BLV differs from similar retroviruses by preferentially infecting B cells.
Although there is no detectable viremia in fresh blood, BLV-infected cattle can be identified by the presence of a persistent anti-BLV antibody response (10, 18). tax/rex and, uncommonly, env transcripts have been found in BLV-infected individuals, indicating that latency is incomplete (16, 17). Presumably, there is a low level of viral structural protein or virion production sufficient to stimulate antibody production, but the production is below the sensitivity of current technologies or is very efficiently compartmentalized.
Previous studies have demonstrated that the in vivo inhibition of BLV expression can be overcome by the short-term culture of lymphocytes in vitro. The addition of B-cell mitogens, such as phytohemagglutinin (18, 20, 23) or immunoglobulin G (IgG) (28), has been demonstrated to further stimulate viral expression. Conversely, the addition of plasma to the short-term cultures inhibits viral expression to levels comparable to those of freshly isolated lymphocytes (14, 15, 24). Initially, it was thought that blocking activity was present in plasma from BLV-infected cattle only, but Taylor and Jacobs demonstrated that all bovine plasmas are able to block the expression of BLV, although plasma from BLV-infected individuals tended to have greater blocking activity (24). Therefore, the blocking activity cannot be attributed solely to anti-BLV antibody, since it is present in uninfected cattle. Similar blocking activity has been described in plasma from BLV-infected sheep (6) and plasma from humans infected with HTLV (28). Blocking activity can be demonstrated in serum as well but with lower activity than that in plasma. A "platelet-derived factor" made from platelet lysate effectively and irreversibly inhibited the blocking activity of plasma (25). Proteases, important in the coagulation cascade, could conceivably play a role in the diminished blocking activity observed in serum.
It is evident that a factor inhibitory to retroviral replication is constitutively expressed in cattle. This factor has been termed plasma BLV blocking factor (14) or plasma blocking factor (PBF) (24). Previous investigations demonstrated that PBF inhibited BLV transcription but not murine or feline leukemia viruses produced in monolayer cell cultures (14, 15). It was not cytotoxic (24, 27), nor was it an antibody or an interferon-like molecule (15). The inhibitory activity was resistant to freezing and thawing but was destroyed by heating at 56°C for 10 min, was at least partially precipitated by 20 to 33% saturated ammonium sulfate, and had a molecular size of 150 kDa (15).
Such blocking activity likely plays an important role in mediating viral latency and thus contributes to the low frequency and morbidity of lymphoma in BLV-infected adult cattle. Here, we describe experiments in which we first established reliable quantification of viral core protein production (p24) in cultured cells in the presence or absence of PBF, then separated PBF from other plasma proteins, and finally identified PBF as a unique fibronectin-containing complex of proteins capable of inhibiting expression of BLV p24.
MATERIALS AND METHODS
Animals. A BLV-seropositive (BLV+) Holstein cow with persistent lymphocytosis (PL) was housed according to the guidelines of the Canadian Council on Animal Care. PL is a sustained polyclonal increase in peripheral blood lymphocytes and is observed in approximately one-third of infected cattle. This animal was the source of plasma from which the PBF was derived as well as the source of mononuclear cells for all preliminary experiments to isolate PBF. Blood samples were obtained at a commercial dairy farm from six other BLV+ cows whose BLV serostatus was determined by repeated agar gel immunodiffusion tests.
Assessment of p24 production by cultured BLV+ mononuclear cells: BLV p24 bioassay. The enzyme-linked immunosorbent assay (ELISA) used to quantify production of the BLV p24 core protein was described by van den Heuvel et al. (26). Briefly, 5 x 106 blood mononuclear cells, obtained by density gradient centrifugation (specific gravity, 1.077) (Histopaque 1077; Sigma-Aldrich, Oakville, ON, Canada) of venous blood anticoagulated with acid citrate dextrose (0.091 M Na citrate dihydrate, 0.018 M citric acid monohydrate, 2.5% dextrose), were cultured for 24 h in enriched RPMI 1640 (supplemented with 10% fetal bovine serum [FBS], L-glutamine [0.3g/liter], and 0.02% amikacin) 24-well culture plates in the presence or absence of plasma proteins. After incubation, live cell counts were determined by trypan blue exclusion, and cells were lysed in 10% n-octyl-glucopyranoside (Sigma-Aldrich). The lysate was diluted to 1% with phosphate-buffered saline (PBS; 0.01 M, pH 7), clarified by centrifugation, and applied overnight to triplicate wells of a 96-well plate (Nunc Immunosorb; Life Technologies, Mississauga, ON, Canada) precoated with a monoclonal anti-BLV p24 antibody. Dilutions of purified p24 were added as reference samples. After the plates were washed, a mix of 200 μg each of 2'C1 and 4'F5 peroxidase-conjugated monoclonal anti-BLV p24 antibodies (gift of D. Portetelle) were applied for 1 h prior to color development with tetramethyl benzidine (Kirkegaard and Perry Laboratories, Gaithersburg, MD). Absorbance at 450 nm was measured, and the mean amount (ng) of BLV p24/105 live cells was calculated for each sample. Each experiment included a positive control (addition of defibrinated plasma to a final concentration of 25% [vol/vol]) and a negative control (10% FBS in RPMI 1640).
Specificity of PBF activity. Plasma extracted from citrated blood (pH 6.98) was centrifuged at 1,500 x g for 20 min at 20°C to pellet platelets, which had been shown to interfere with PBF activity (24). BLV p24 production in blood BLV+ mononuclear cells cultured in enriched RPMI 1640 alone or with plasma at concentrations ranging from 10 to 100% (vol/vol) was examined to determine the optimal amount of plasma necessary to block p24 production. As a 24-h culture resulted in clot formation, plasma was defibrinated by heating to 65°C for 15 min and then clarified by centrifugation at 1,500 x g for 20 min (pH 7.14). To confirm that blocking activity was not lost with the heating necessary to precipitate fibrin, the titration experiment described above was repeated using defibrinated plasma.
Purification of PBF. (i) Ammonium sulfate fractionation. Ammonium sulfate precipitation, a bulk separation technique which tends to precipitate large, negatively charged proteins first, was done according to an established protocol (9). Briefly, the plasma was chilled to 0°C, set in an ice bath, and stirred gently. Granular ammonium sulfate was added to bring the concentration to 20% saturated ammonium sulfate. The mixture was stirred for 10 min, allowed to equilibrate for a further 10 min, and finally centrifuged at 5,000 x g for 30 min at 4°C. The supernatant was aspirated and subjected to further ammonium sulfate precipitation, yielding eight fractions as summarized in Table 1. The protein pellets from each fraction were resuspended in water, concentrated in 15-ml tubes fitted with an ultrafilter with a 10-kDa-molecular-size cutoff (Centricon; Amicon Inc., Beverly, MA), washed three times in PBS, and stored briefly at 4°C until assayed for blocking activity. Protein concentrations in each of the eight fractions were determined by measuring the absorbancy at 280 nm. Twenty milligrams of each fraction was added to cultured BLV+ mononuclear cells and tested for the expression of BLV p24 by sandwich ELISA (sELISA). The fraction with the greatest inhibitory activity was used in a titration experiment to determine the efficiency of extraction.
(ii) Chromatography. The ammonium sulfate-precipitated plasma fractions which exhibited reductions in p24 relative to that of the medium control but equaling or exceeding the reductions induced by the plasma control were selected for further fractionation. The plasma fractions were pooled, filtered through a 0.22-μm filter, equilibrated with a 50 mM KH2PO4, 150 mM KCl, pH 7 buffer, and subjected to anion-exchange chromatography (Cibacron blue fast flow Sepharose; Amersham Biosciences, Baie d'Urfe, PQ, Canada). A stepwise gradient, ranging from 150 mM KCl to 1.5 M KCl, was used. Fractions were concentrated by ultrafiltration, washed three times with PBS, and tested for blocking activity in the BLV p24 sELISA. The bioactive fraction was adjusted to 10 mg/ml, and then 200-μl aliquots were separated by gel exclusion chromatography which had been precalibrated using molecular size standards of 25, 67, 158, 232, 440, and 669 kDa (Superose 6; Amersham Biosciences). Proteins were eluted with PBS (pH 7) at a flow rate of 0.3 ml/min. Fractions with blocking activity were pooled, concentrated, and reapplied to the same chromatographic column, and the fractions were titrated in the bioactivity assay.
(iii) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Twenty micrograms of each protein fraction mixed with 7 μl of SDS sample buffer (New England Biolabs, Mississauga, ON, Canada) containing 1 mM dithiothreitol was heated to 100°C for 5 min prior to being loaded on a 10% polyacrylamide gel. Electrophoresis was performed at 100 V for 40 min. Gels were stained with silver or Coomassie brilliant blue.
Identification of PBF proteins: two-dimensional gel electrophoresis and sequencing. Briefly, 20 μg of the purified plasma protein, bovine plasma-derived alpha 2 macroglobulin (2M) (gift of M. A. Hayes, University of Guelph, ON, Canada), or bovine plasma-derived fibronectin (bovine fibronectin [boFN]; Sigma-Aldrich) which had been desalted (Amersham Biosciences 2D cleanup kit) was resuspended in 125 μl rehydration buffer, applied to isoelectric focusing strips (Amersham Biosciences Immobiline Dry-strips; pH 4 to 7), and rehydrated for 12 h according to the manufacturer's directions (IGPhor isoelectric focusing system). Proteins were separated at 500 V for 30 min, then at 1,000 V for 30 min, and finally at 3,300 V for 10 h. Then strips were placed on a 5% polyacrylamide gel, and current was applied at 50 mA for 3.5 h at 4°C. The gels were silver stained. Stained spots were excised and sent for peptide mapping and amino acid sequencing by matrix-assisted laser desorption-ionization time of flight (mass spectrometry) analysis (Service Proteomique de l'Est du Québec, Ste. Foy, Quebec, Canada). Amino acid sequences were compared with those available in an automated database (GenBank).
Protein specificity. To identify the roles of individual proteins or peptides in mediating PBF activity, bovine serum albumin (BSA; Sigma-Aldrich), bovine plasma 2M, boFN (catalog no. F4759; Sigma-Aldrich), and human fibronectin (huFN) peptides, including plasma huFN (catalog no. F2006; Sigma-Aldrich); the integrin/cell binding fragment (RGD-containing peptide; catalog no. A9041; Sigma-Aldrich); the adhesion-promoting, heparin-binding peptide WQPPRARI (catalog no. F3667; Sigma-Aldrich); and the heparin-binding -chymotryptic fragment FN40 (Chemicon International, Temecula, CA) were applied to multiple cultures of bovine BLV+ mononuclear cells in a dose-response experiment and subsequently tested for p24 expression in a p24 ELISA. Similarly, 500 μg of each of these proteins or peptides was compared with purified PBF for the ability to inhibit viral transcription in mononuclear cells from six different animals from a commercial herd. The FN40 was used in a competitive assay in the presence of purified PBF to see if the inhibitory activity could be blocked in competition with the RGD site.
Statistical analysis. Data were summarized using descriptive statistics. Treatment group effects were assessed using one-way analysis of variance. Differences with a P of 0.05 were considered significant.
RESULTS
Specificity of PBF activity. Previous studies have shown that BLV is difficult to detect in freshly isolated blood cells but can easily be found after short-term cultures (24). The addition of plasma reduced viral production in a dose-dependent manner, as measured by sELISA (26). The results of three replicate experiments in which native plasma from a BLV+ animal (pH 6.98; 9 replicates) or defibrinated plasma from the same source (pH 7.14; 9 replicates) added to the culture medium of bovine peripheral blood mononuclear cells (PBMCs) ranged from 0 to 100% are shown in Fig. 1. A nonlinear decrease in BLV p24 expression was found when increasing concentrations of either defibrinated plasma or whole plasma were added to cultured blood mononuclear cells. Inhibition of BLV p24 expression was maximal at 25% plasma; further addition of plasma resulted in no greater inhibition. BLV p24 produced with no plasma present was 38 ± 6 ng/105 live cells (mean ± standard error [SE]), whereas with 25% plasma, expression was reduced to 4.2 ± 2 ng/105 live cells (mean ± SE). After 24 h of culture, the number of cells harvested represented 80% of cells applied; the viability averaged 75% ± 8.94% and did not differ significantly between treatments.
Purification of PBF. The largest amounts of plasma protein were precipitated with the addition of ammonium sulfate at concentrations of 21 to 30% and 31 to 40% of a saturated solution (Table 1). The greatest inhibition levels of BLV p24 expression were also found in these two fractions. The 41-to-50% fraction of plasma protein was stimulatory and induced a 125% increase in BLV p24 to an average of 45 ± 8 ng/105 live cells (mean ± SE). Figure 2 demonstrates both the bioactivity of the 21-to-30% fraction as it was titrated from 25 mg/ml to 0.39 mg/ml and the specificity as determined by the same titration of BSA. The application of BSA at any concentration, or less than 3 mg/ml of the 21-to-30% saturated ammonium sulfate fraction, had no effect on BLV p24 expression. Maximal reduction in BLV p24 expression was achieved with the addition of 12.5 mg/ml of the 21-to-30% fraction.
The plasma protein fractions precipitated at 21 to 30% were used for further purification by anion exchange. Protein was eluted with 375 mM, 500 mM, and 1.5 M KCl. Maximum inhibition of BLV p24 expression was found in the fraction which eluted with 500 mM KCl (data not shown). Subsequent repeated titration of this fraction into PBMC cultures demonstrated that 1 mg or more of protein was necessary to significantly inhibit BLV p24 expression. The 500 mM KCl fraction from the anion-exchange chromatographic separation was then subjected to gel exclusion chromatography. At a flow rate of 0.3 ml/min, peak bioactivity was found in fractions that eluted between 20 and 26 min (data not shown). A summary of the purification steps and the efficiency of purification are shown in Table 2.
The fractions with peak bioactivity from the gel exclusion column were pooled, concentrated, and reapplied to a gel exclusion column calibrated for molecular size. The fraction with maximal inhibitory bioactivity had an estimated molecular size of 320 kDa. Figure 3A shows an SDS-PAGE analysis of each sequential fraction eluted from the column. The fraction with maximal bioactivity demonstrates that it consisted of three protein bands of less than 70 kDa and three protein bands with molecular sizes of 120, 180, and 225 kDa.
Identification of PBF. The bioactive fraction was further investigated using isoelectric focusing in the first dimension and SDS-PAGE analysis in the second dimension (Fig. 3B). A series of protein spots with pIs between 5.5 and 5.9 had a molecular size of 72,120 Da. These were identified as a heavy chain of IgG. A discrete protein spot at 69,440 Da and with a pI of 4.7 was identified as albumin. A series of protein spots at 188,240 Da with a pI of 4.6 but ranging up to 7.0 was sequenced and shown to have high homology to 2M. Sequencing of eight peptide fragments of a protein spot having a calculated molecular size of 220,400 Da and a pI of 4.9 revealed 100% homology with boFN. Purified bovine 2M and boFN were subjected to two-dimensional gel analysis. Two large-molecular-size spots and a 69-kDa spot corresponding to the albumin added as a stabilizer were found on the 2M gel (not shown). In Fig. 3C, commercial plasma boFN revealed a single protein band at approximately 250 kDa with a pI of 5.3 to 5.6.
Protein specificity. Commercially available purified boFN and bovine 2M were analyzed for inhibition of viral expression in the p24 sELISA. These proteins are purified by chromatography from plasma sources and have BSA added as a stabilizing agent. Figure 4 shows the dose response of bovine mononuclear cell production of p24 in response to culture with boFN. There was a dose response, but higher levels of commercial boFN than of purified PBF were required to achieve significant reduction. The bovine 2M had no effect on viral expression (data not shown).
To demonstrate that the purified PBF had a biological function in BLV-infected cells from other animals, mononuclear cells from six BLV+ donor cows with PL were cultured in medium alone or in medium with either 25% autologous plasma, 150 μg purified PBF, or 250 μg boFN (Fig. 5A). Twenty-five percent plasma, purified PBF, and boFN were each able to inhibit p24 production in cells from all six cows relative to the medium controls. Further experiments to investigate whether FN from human sources was able to mediate a blocking effect as suggested by Zandomeni et al. (8) were done using 500 μg of various human fibronectin peptides. As shown in Fig. 5B, while plasma, purified PBF, and boFN were able to reduce p24 expression to 25% of that of the reference medium in cultures of mononuclear cells isolated from six naturally infected BLV+ cows, neither huFN nor the integrin/cell-binding RGD peptide derived from huFN inhibited p24 production. FN40 competitively blocks the FN binding domain of 4 integrin (expressed on the cell membrane of most lymphocytes). FN40 that is added to a culture containing PBF competes for RGD integrin binding sites, resulting in loss of activity if the PBF also binds this site. We observed partial restoration of p24 expression, but the effect was not significant. Adhesion-promoting huFN peptides had no effect on p24 production (data not shown).
DISCUSSION
The aim of the current study was to isolate the factor present in bovine plasma which blocks BLV expression. We demonstrated that PBF was present in bovine plasma, that it was resistant to heating, and that its effect could be eliminated at a concentration of less than 22 mg/ml. Physicochemical fractionation by ammonium sulfate precipitation, followed by anion-exchange and gel exclusion chromatographies resulted in a highly concentrated product that retained biological activity. Under nonreducing conditions, blocking activity was associated with a 320-kDa protein complex which revealed six visible proteins under two-dimensional protein analysis. Amino acid sequencing showed these to be two forms of 2M and a protein of approximately 220 kDa with a pI of 4.9 and with 100% homology to FN. The remaining proteins were deduced to be albumin, immunoglobulin light chain, and IgG heavy chain. The molecular size determined under reduced conditions relative to that of the native protein suggested formation of an FN dimer complexed with 2M and immunoglobulins. The evidence presented here supports the possibility of an isoform of boFN with a pI of 4.9 as the putative plasma blocking factor.
We noted significant stimulation of BLV p24 production in plasma precipitated with 41 to 50% saturated ammonium sulfate, while most blocking activity was present in the 21-to-30% fraction reported previously (15). Commonly, 50% saturated ammonium sulfate is used to crudely fractionate immunoglobulin from plasma, suggesting that the peaks in blocking activity and immunoglobulin concentration were at least partially separated by this procedure. Others (28) have also reported that immunoglobulin stimulated BLV production, making immunoglobulin an unlikely candidate for PBF.
In contrast, commercially purified boFN did cause significant reduction in BLV p24 production, although PBF prepared for the present study was more potent. Two-dimensional analysis of boFN revealed only the high-molecular-weight protein, no smaller proteins, indicating that the pure boFN product was capable of blocking p24 production, but it is possible that another protein which complexes with FN also contributes to PBF activity. Differences in isolation methodologies (binding to denatured collagen, followed by extraction with 1 to 8 M urea) may account for the decreased bioactivity observed with the commercial product. Commercially purified bovine serum albumin and bovine 2M showed no inhibitory activity in our bioassay.
Fibronectin is a ubiquitous glycoprotein that has a multiplicity of roles, including embryogenesis, thrombosis, wound healing, and hemostasis. It has two forms: a soluble form in body fluids and an insoluble form in the extracellular matrix. Plasma FN is secreted as a dimer of 440 kDa, which is found in plasma at a concentration of 300 μg/ml (21). The 220- to 250-kDa monomers are bound together with two disulfide bonds near the carboxyl terminus. Fibronectin has a pI of 5.6 to 6.3, with peak activity at 6.1, but is reported to suffer rapid loss of activity at temperatures over 60 to 65°C (2). In contrast, the native fibronectin complex isolated here had a molecular size of 320 kDa, while the monomer was 220 kDa with a pI of 4.9. The fibronectin gene is composed of three types of repeating subunits (types I to III). Alternative splicing of type III repeats, inclusion or exclusion of two repeats known as EDI and EDII, and subdivision of a nonhomologous third variable region give rise to 20 permutations in huFN, of which 2 are plasma forms (11).
The amino and carboxy termini of FN are composed of repeated type I modules which bind fibrin and collagen. Type II modules are less frequent in the molecule and contribute to collagen binding. The most abundant module, of which there are 12 to 17 copies, is type III, which contains the RGD integrin binding site as well as sites for binding heparin and other integrins. Plasma FN differs from the insoluble extracellular matrix FN by having fewer type III modules and having a second RGD site. Curiously, despite greater than 80% homology at the amino acid level between huFN and boFN, huFN had no blocking activity in our bioassay. HuFN (2,386 amino acids [aa]; 260 kDa) differs from boFN (2,265 aa; 240 kDa) by having a 31-aa extension at the amino terminus and a 90-aa insert at aa 105 following the integrin-binding domain of FN, arginine, glycine, and aspartic acid (RGD). A human antibody to FN is reported to bind an epitope in the decapeptide RGDSPASSKP of bovine and marsupial FN but not huFN (13). Both bovine and human FN contain this decapeptide, but the bovine FN sequence is followed by a valine, while the human FN sequence is followed by an isoleucine, after which both are identical again until the human FN 90-aa insert. Since the antibody was able to distinguish boFN from huFN at the RGD integrin binding site, it is possible that huFN is unable to interact with the cell surface receptor and signal transduction apparatus present on bovine cells.
The biological effects of FN vary with the cell type, expression of cell surface ligands for FN, and pleomorphism of the FN molecule. Little is known about the interaction of soluble FN and blood cells. However, adhesion of Jurkat T cells to the extracellular matrix FN increased NFB activation within 30 min, but this effect was progressively lost if incubation continued and NFB was deactivated by 12 h; the effect persisted for over 24 h (1). Interestingly, NFB proteins and Tax activate BLV transcription and may play a role in rescuing the virus from latency (3). Should a similar paradigm exist with bovine blood cells and soluble FN, then it is possible that FN acts by deactivating the nuclear transportation of NFB, thereby maintaining BLV in a latent state.
ACKNOWLEDGMENTS
This work was supported by the Natural Sciences and Engineering Research Council and the Ontario Ministry of Agriculture and Food. M. van den Heuvel was supported by a scholarship from the Dairy Farmers of Ontario.
The generosity of D. Portetelle and M. A. Hayes in sharing resources was very much appreciated.
REFERENCES
Bearz, A., G. Tell, A. Colombatti, S. Formisano, and C. Pucillo. 1998. Fibronectin binding promotes a PKC-dependent modulation of NF-kappa B in human T cells. Biochem. Biophys. Res. Commun. 243:732-737.
Bozzini, S., V. Falcone, P. G. Conaldi, L. Visai, L. Biancone, A. Dolei, A. Toniolo, and P. Speziale. 1998. Heparin-binding domain of human fibronectin binds HIV-1 gp120/160 and reduces virus infectivity. J. Med. Virol. 54:44-53.
Brooks, P. A., G. L. Cockerell, and J. K. Nyborg. 1998. Activation of BLV transcription by NF-kappa B and Tax. Virology 243:94-98.
Burny, A., C. Bruck, Y. Cleuter, D. Couez, J. Deschamps, D. Gregoire, J. Ghysdael, R. Kettmann, M. Mammerickx, and G. Marbaix. 1985. Bovine leukaemia virus and enzootic bovine leukosis. Onderstepoort. J. Vet. Res. 52:133-144.
Burny, A., Y. Cleuter, R. Kettmann, M. Mammerickx, G. Marbaix, D. Portetelle, A. Van den Broeke, L. Willems, and R. Thomas. 1987. Bovine leukaemia: facts and hypotheses derived from the study of an infectious cancer. Cancer Surv. 6:139-159.
Cornil, I., and D. Levy. 1989. In vivo inhibition of bovine leukemia virus (BLV) expression. Leukemia 3:159-161.
Derse, D. 1987. Bovine leukemia virus transcription is controlled by a virus-encoded trans-acting factor and by cis-acting response elements. J. Virol. 61:2462-2471.
Derse, D. 1988. trans-acting regulation of bovine leukemia virus mRNA processing. J. Virol. 62:1115-1119.
Doonan, S. 1996. Bulk purification by fractional precipitation, p. 135-144. In S. Doonan (ed.), Protein purification protocols. Humana Press, Totowa, N.J.
Ferrer, J. F. 1980. Bovine lymphosarcoma. Adv. Vet. Sci. Comp. Med. 24:1-68.
ffrench-Constant, C. 1995. Alternative splicing of fibronectin—many different proteins but few different functions. Exp. Cell Res. 221:261-271.
Gallo, R. C., and L. S. Nerurkar. 1989. Human retroviruses: their role in neoplasia and immunodeficiency. Ann. N. Y. Acad. Sci. 567:82-94.
Girard, D., Y. Raymond, P. Labbe, and J. L. Senecal. 1995. Characterization of a novel human antibody xenoreactive with fibronectin. Clin. Immunol. Immunopathol. 77:149-161.
Gupta, P., and J. F. Ferrer. 1982. Expression of bovine leukemia virus genome is blocked by a nonimmunoglobulin protein in plasma from infected cattle. Science 215:405-407.
Gupta, P., S. V. S. Kashmiri, and J. F. Ferrer. 1984. Transcriptional control of the bovine leukemia virus genome: role and characterization of a non-immunoglobulin plasma protein from bovine leukemia virus-infected cattle. J. Virol. 50:267-270.
Haas, L., T. Divers, and J. W. Casey. 1992. Bovine leukemia virus gene expression in vivo. J. Virol. 66:6223-6225.
Jensen, W. A., J. Rovnak, and G. L. Cockerell. 1991. In vivo transcription of the bovine leukemia virus tax/rex region in normal and neoplastic lymphocytes of cattle and sheep. J. Virol. 65:2484-2490.
Kettmann, R., G. Marbaix, Y. Cleuter, D. Portetelle, M. Mammerickx, and A. Burny. 1980. Genomic integration of bovine leukemia provirus and lack of viral RNA expression in the target cells of cattle with different responses to BLV infection. Leuk. Res. 4:509-519.
Lee, T. H., J. E. Coligan, J. G. Sodroski, W. A. Haseltine, S. Z. Salahuddin, F. Wong-Staal, R. C. Gallo, and M. Essex. 1984. Antigens encoded by the 3'-terminal region of human T-cell leukemia virus: evidence for a functional gene. Science 226:57-61.
Miller, J. M., L. D. Miller, C. Olson, and K. G. Gillette. 1969. Virus-like particles in phytohemagglutinin-stimulated lymphocyte cultures with reference to bovine lymphosarcoma. J. Natl. Cancer Inst. 43:1297-1305.
Retta, S. F., P. Ferraris, and G. Tarone. 1999. Purification of fibronectin from human plasma. Methods Mol. Biol. 96:119-124.
Slamon, D. J., K. Shimotohno, M. J. Cline, D. W. Golde, and I. S. Chen. 1984. Identification of the putative transforming protein of the human T-cell leukemia viruses HTLV-I and HTLV-II. Science 226:61-65.
Stock, N. D., and J. F. Ferrer. 1972. Replicating C-type virus in phytohemagglutinin-treated buffy-coat cultures of bovine origin. J. Natl. Cancer Inst. 48:985-996.
Taylor, J. A., and R. M. Jacobs. 1993. Effects of plasma and serum on the in vitro expression of bovine leukemia virus. Lab. Investig. 69:340-346.
Tsukiyama, K., M. Onuma, and H. Izawa. 1987. Effect of platelet-derived factor on expression of bovine leukemia virus genome. Arch. Virol. 96:89-96.
van den Heuvel, M., D. Portetelle, B. Jefferson, and R. M. Jacobs. 2003. Adaptation of a sandwich enzyme-linked immunosorbent assay to determine the concentration of bovine leukemia virus p24 and optimal conditions for p24 expression in short-term cultures of peripheral blood mononuclear cells. J. Virol. Methods 111:61-67.
Zandomeni, R. O., M. Carrera-Zandomeni, E. Esteban, W. Donawick, and J. F. Ferrer. 1992. Induction and inhibition of bovine leukaemia virus expression in naturally infected cells. J. Gen. Virol. 73:1915-1924.
Zandomeni, R. O., E. Esteban, M. Carrera-Zandomeni, and J. F. Ferrer. 1994. Host soluble factors with blocking and stimulating activity on the expression of the bovine leukemia virus. J. Infect. Dis. 170:787-794.(Marianne J. van den Heuve)
Department of Pathobiology, University of Guelph, Guelph, Ontario, Canada
ABSTRACT
Bovine leukemia virus (BLV) is a deltaretrovirus that infects cattle worldwide. In agriculturally intensive regions, approximately 30% of dairy cows are BLV infected. Like the human T-cell leukemia virus (HTLV), there is a lengthy period of viral quiescence after initial infection with BLV. Unlike HTLV, BLV resides predominantly in B cells. Lymphoma is observed in less than 10% of BLV-infected adult cattle. Although viremia is undetectable in vivo, BLV-infected peripheral blood mononuclear cells readily become productive when cultured in vitro. Productivity is markedly diminished when cultures are supplemented with bovine plasma. This inhibitory activity of bovine plasma has been attributed to the "plasma blocking factor" (PBF). Here, we describe the purification of a PBF whose activity was resistant to heating to 65°C for 10 min and was attributable to a fibronectin-containing complex of approximately 320 kDa under nonreducing conditions. By use of two-dimensional polyacrylamide gel electrophoresis and matrix-assisted laser desorption ionization-time of flight (mass spectrometry), a protein with a size of 220 kDa and a pI of 5.4 was identified as a member of the fibronectin group of molecules. Both the purified protein and the commercially available bovine fibronectin inhibited BLV production in naturally infected peripheral blood mononuclear cells, although the fibronectin was less biologically active.
INTRODUCTION
Bovine leukemia virus (BLV), the etiological agent of lymphoma in adult cattle, is closely related to human T-cell leukemia virus (HTLV) types I and II. BLV and HTLV share features such as random integration into the host genome, a long period of viral latency, and rare development of disease (4, 5, 12). The HTLV/BLV genus is defined by the presence of a unique open reading frame called pX, located at the 3' end of the genome and coding for transregulatory proteins Tax and Rex (7, 8, 19, 22). BLV differs from similar retroviruses by preferentially infecting B cells.
Although there is no detectable viremia in fresh blood, BLV-infected cattle can be identified by the presence of a persistent anti-BLV antibody response (10, 18). tax/rex and, uncommonly, env transcripts have been found in BLV-infected individuals, indicating that latency is incomplete (16, 17). Presumably, there is a low level of viral structural protein or virion production sufficient to stimulate antibody production, but the production is below the sensitivity of current technologies or is very efficiently compartmentalized.
Previous studies have demonstrated that the in vivo inhibition of BLV expression can be overcome by the short-term culture of lymphocytes in vitro. The addition of B-cell mitogens, such as phytohemagglutinin (18, 20, 23) or immunoglobulin G (IgG) (28), has been demonstrated to further stimulate viral expression. Conversely, the addition of plasma to the short-term cultures inhibits viral expression to levels comparable to those of freshly isolated lymphocytes (14, 15, 24). Initially, it was thought that blocking activity was present in plasma from BLV-infected cattle only, but Taylor and Jacobs demonstrated that all bovine plasmas are able to block the expression of BLV, although plasma from BLV-infected individuals tended to have greater blocking activity (24). Therefore, the blocking activity cannot be attributed solely to anti-BLV antibody, since it is present in uninfected cattle. Similar blocking activity has been described in plasma from BLV-infected sheep (6) and plasma from humans infected with HTLV (28). Blocking activity can be demonstrated in serum as well but with lower activity than that in plasma. A "platelet-derived factor" made from platelet lysate effectively and irreversibly inhibited the blocking activity of plasma (25). Proteases, important in the coagulation cascade, could conceivably play a role in the diminished blocking activity observed in serum.
It is evident that a factor inhibitory to retroviral replication is constitutively expressed in cattle. This factor has been termed plasma BLV blocking factor (14) or plasma blocking factor (PBF) (24). Previous investigations demonstrated that PBF inhibited BLV transcription but not murine or feline leukemia viruses produced in monolayer cell cultures (14, 15). It was not cytotoxic (24, 27), nor was it an antibody or an interferon-like molecule (15). The inhibitory activity was resistant to freezing and thawing but was destroyed by heating at 56°C for 10 min, was at least partially precipitated by 20 to 33% saturated ammonium sulfate, and had a molecular size of 150 kDa (15).
Such blocking activity likely plays an important role in mediating viral latency and thus contributes to the low frequency and morbidity of lymphoma in BLV-infected adult cattle. Here, we describe experiments in which we first established reliable quantification of viral core protein production (p24) in cultured cells in the presence or absence of PBF, then separated PBF from other plasma proteins, and finally identified PBF as a unique fibronectin-containing complex of proteins capable of inhibiting expression of BLV p24.
MATERIALS AND METHODS
Animals. A BLV-seropositive (BLV+) Holstein cow with persistent lymphocytosis (PL) was housed according to the guidelines of the Canadian Council on Animal Care. PL is a sustained polyclonal increase in peripheral blood lymphocytes and is observed in approximately one-third of infected cattle. This animal was the source of plasma from which the PBF was derived as well as the source of mononuclear cells for all preliminary experiments to isolate PBF. Blood samples were obtained at a commercial dairy farm from six other BLV+ cows whose BLV serostatus was determined by repeated agar gel immunodiffusion tests.
Assessment of p24 production by cultured BLV+ mononuclear cells: BLV p24 bioassay. The enzyme-linked immunosorbent assay (ELISA) used to quantify production of the BLV p24 core protein was described by van den Heuvel et al. (26). Briefly, 5 x 106 blood mononuclear cells, obtained by density gradient centrifugation (specific gravity, 1.077) (Histopaque 1077; Sigma-Aldrich, Oakville, ON, Canada) of venous blood anticoagulated with acid citrate dextrose (0.091 M Na citrate dihydrate, 0.018 M citric acid monohydrate, 2.5% dextrose), were cultured for 24 h in enriched RPMI 1640 (supplemented with 10% fetal bovine serum [FBS], L-glutamine [0.3g/liter], and 0.02% amikacin) 24-well culture plates in the presence or absence of plasma proteins. After incubation, live cell counts were determined by trypan blue exclusion, and cells were lysed in 10% n-octyl-glucopyranoside (Sigma-Aldrich). The lysate was diluted to 1% with phosphate-buffered saline (PBS; 0.01 M, pH 7), clarified by centrifugation, and applied overnight to triplicate wells of a 96-well plate (Nunc Immunosorb; Life Technologies, Mississauga, ON, Canada) precoated with a monoclonal anti-BLV p24 antibody. Dilutions of purified p24 were added as reference samples. After the plates were washed, a mix of 200 μg each of 2'C1 and 4'F5 peroxidase-conjugated monoclonal anti-BLV p24 antibodies (gift of D. Portetelle) were applied for 1 h prior to color development with tetramethyl benzidine (Kirkegaard and Perry Laboratories, Gaithersburg, MD). Absorbance at 450 nm was measured, and the mean amount (ng) of BLV p24/105 live cells was calculated for each sample. Each experiment included a positive control (addition of defibrinated plasma to a final concentration of 25% [vol/vol]) and a negative control (10% FBS in RPMI 1640).
Specificity of PBF activity. Plasma extracted from citrated blood (pH 6.98) was centrifuged at 1,500 x g for 20 min at 20°C to pellet platelets, which had been shown to interfere with PBF activity (24). BLV p24 production in blood BLV+ mononuclear cells cultured in enriched RPMI 1640 alone or with plasma at concentrations ranging from 10 to 100% (vol/vol) was examined to determine the optimal amount of plasma necessary to block p24 production. As a 24-h culture resulted in clot formation, plasma was defibrinated by heating to 65°C for 15 min and then clarified by centrifugation at 1,500 x g for 20 min (pH 7.14). To confirm that blocking activity was not lost with the heating necessary to precipitate fibrin, the titration experiment described above was repeated using defibrinated plasma.
Purification of PBF. (i) Ammonium sulfate fractionation. Ammonium sulfate precipitation, a bulk separation technique which tends to precipitate large, negatively charged proteins first, was done according to an established protocol (9). Briefly, the plasma was chilled to 0°C, set in an ice bath, and stirred gently. Granular ammonium sulfate was added to bring the concentration to 20% saturated ammonium sulfate. The mixture was stirred for 10 min, allowed to equilibrate for a further 10 min, and finally centrifuged at 5,000 x g for 30 min at 4°C. The supernatant was aspirated and subjected to further ammonium sulfate precipitation, yielding eight fractions as summarized in Table 1. The protein pellets from each fraction were resuspended in water, concentrated in 15-ml tubes fitted with an ultrafilter with a 10-kDa-molecular-size cutoff (Centricon; Amicon Inc., Beverly, MA), washed three times in PBS, and stored briefly at 4°C until assayed for blocking activity. Protein concentrations in each of the eight fractions were determined by measuring the absorbancy at 280 nm. Twenty milligrams of each fraction was added to cultured BLV+ mononuclear cells and tested for the expression of BLV p24 by sandwich ELISA (sELISA). The fraction with the greatest inhibitory activity was used in a titration experiment to determine the efficiency of extraction.
(ii) Chromatography. The ammonium sulfate-precipitated plasma fractions which exhibited reductions in p24 relative to that of the medium control but equaling or exceeding the reductions induced by the plasma control were selected for further fractionation. The plasma fractions were pooled, filtered through a 0.22-μm filter, equilibrated with a 50 mM KH2PO4, 150 mM KCl, pH 7 buffer, and subjected to anion-exchange chromatography (Cibacron blue fast flow Sepharose; Amersham Biosciences, Baie d'Urfe, PQ, Canada). A stepwise gradient, ranging from 150 mM KCl to 1.5 M KCl, was used. Fractions were concentrated by ultrafiltration, washed three times with PBS, and tested for blocking activity in the BLV p24 sELISA. The bioactive fraction was adjusted to 10 mg/ml, and then 200-μl aliquots were separated by gel exclusion chromatography which had been precalibrated using molecular size standards of 25, 67, 158, 232, 440, and 669 kDa (Superose 6; Amersham Biosciences). Proteins were eluted with PBS (pH 7) at a flow rate of 0.3 ml/min. Fractions with blocking activity were pooled, concentrated, and reapplied to the same chromatographic column, and the fractions were titrated in the bioactivity assay.
(iii) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Twenty micrograms of each protein fraction mixed with 7 μl of SDS sample buffer (New England Biolabs, Mississauga, ON, Canada) containing 1 mM dithiothreitol was heated to 100°C for 5 min prior to being loaded on a 10% polyacrylamide gel. Electrophoresis was performed at 100 V for 40 min. Gels were stained with silver or Coomassie brilliant blue.
Identification of PBF proteins: two-dimensional gel electrophoresis and sequencing. Briefly, 20 μg of the purified plasma protein, bovine plasma-derived alpha 2 macroglobulin (2M) (gift of M. A. Hayes, University of Guelph, ON, Canada), or bovine plasma-derived fibronectin (bovine fibronectin [boFN]; Sigma-Aldrich) which had been desalted (Amersham Biosciences 2D cleanup kit) was resuspended in 125 μl rehydration buffer, applied to isoelectric focusing strips (Amersham Biosciences Immobiline Dry-strips; pH 4 to 7), and rehydrated for 12 h according to the manufacturer's directions (IGPhor isoelectric focusing system). Proteins were separated at 500 V for 30 min, then at 1,000 V for 30 min, and finally at 3,300 V for 10 h. Then strips were placed on a 5% polyacrylamide gel, and current was applied at 50 mA for 3.5 h at 4°C. The gels were silver stained. Stained spots were excised and sent for peptide mapping and amino acid sequencing by matrix-assisted laser desorption-ionization time of flight (mass spectrometry) analysis (Service Proteomique de l'Est du Québec, Ste. Foy, Quebec, Canada). Amino acid sequences were compared with those available in an automated database (GenBank).
Protein specificity. To identify the roles of individual proteins or peptides in mediating PBF activity, bovine serum albumin (BSA; Sigma-Aldrich), bovine plasma 2M, boFN (catalog no. F4759; Sigma-Aldrich), and human fibronectin (huFN) peptides, including plasma huFN (catalog no. F2006; Sigma-Aldrich); the integrin/cell binding fragment (RGD-containing peptide; catalog no. A9041; Sigma-Aldrich); the adhesion-promoting, heparin-binding peptide WQPPRARI (catalog no. F3667; Sigma-Aldrich); and the heparin-binding -chymotryptic fragment FN40 (Chemicon International, Temecula, CA) were applied to multiple cultures of bovine BLV+ mononuclear cells in a dose-response experiment and subsequently tested for p24 expression in a p24 ELISA. Similarly, 500 μg of each of these proteins or peptides was compared with purified PBF for the ability to inhibit viral transcription in mononuclear cells from six different animals from a commercial herd. The FN40 was used in a competitive assay in the presence of purified PBF to see if the inhibitory activity could be blocked in competition with the RGD site.
Statistical analysis. Data were summarized using descriptive statistics. Treatment group effects were assessed using one-way analysis of variance. Differences with a P of 0.05 were considered significant.
RESULTS
Specificity of PBF activity. Previous studies have shown that BLV is difficult to detect in freshly isolated blood cells but can easily be found after short-term cultures (24). The addition of plasma reduced viral production in a dose-dependent manner, as measured by sELISA (26). The results of three replicate experiments in which native plasma from a BLV+ animal (pH 6.98; 9 replicates) or defibrinated plasma from the same source (pH 7.14; 9 replicates) added to the culture medium of bovine peripheral blood mononuclear cells (PBMCs) ranged from 0 to 100% are shown in Fig. 1. A nonlinear decrease in BLV p24 expression was found when increasing concentrations of either defibrinated plasma or whole plasma were added to cultured blood mononuclear cells. Inhibition of BLV p24 expression was maximal at 25% plasma; further addition of plasma resulted in no greater inhibition. BLV p24 produced with no plasma present was 38 ± 6 ng/105 live cells (mean ± standard error [SE]), whereas with 25% plasma, expression was reduced to 4.2 ± 2 ng/105 live cells (mean ± SE). After 24 h of culture, the number of cells harvested represented 80% of cells applied; the viability averaged 75% ± 8.94% and did not differ significantly between treatments.
Purification of PBF. The largest amounts of plasma protein were precipitated with the addition of ammonium sulfate at concentrations of 21 to 30% and 31 to 40% of a saturated solution (Table 1). The greatest inhibition levels of BLV p24 expression were also found in these two fractions. The 41-to-50% fraction of plasma protein was stimulatory and induced a 125% increase in BLV p24 to an average of 45 ± 8 ng/105 live cells (mean ± SE). Figure 2 demonstrates both the bioactivity of the 21-to-30% fraction as it was titrated from 25 mg/ml to 0.39 mg/ml and the specificity as determined by the same titration of BSA. The application of BSA at any concentration, or less than 3 mg/ml of the 21-to-30% saturated ammonium sulfate fraction, had no effect on BLV p24 expression. Maximal reduction in BLV p24 expression was achieved with the addition of 12.5 mg/ml of the 21-to-30% fraction.
The plasma protein fractions precipitated at 21 to 30% were used for further purification by anion exchange. Protein was eluted with 375 mM, 500 mM, and 1.5 M KCl. Maximum inhibition of BLV p24 expression was found in the fraction which eluted with 500 mM KCl (data not shown). Subsequent repeated titration of this fraction into PBMC cultures demonstrated that 1 mg or more of protein was necessary to significantly inhibit BLV p24 expression. The 500 mM KCl fraction from the anion-exchange chromatographic separation was then subjected to gel exclusion chromatography. At a flow rate of 0.3 ml/min, peak bioactivity was found in fractions that eluted between 20 and 26 min (data not shown). A summary of the purification steps and the efficiency of purification are shown in Table 2.
The fractions with peak bioactivity from the gel exclusion column were pooled, concentrated, and reapplied to a gel exclusion column calibrated for molecular size. The fraction with maximal inhibitory bioactivity had an estimated molecular size of 320 kDa. Figure 3A shows an SDS-PAGE analysis of each sequential fraction eluted from the column. The fraction with maximal bioactivity demonstrates that it consisted of three protein bands of less than 70 kDa and three protein bands with molecular sizes of 120, 180, and 225 kDa.
Identification of PBF. The bioactive fraction was further investigated using isoelectric focusing in the first dimension and SDS-PAGE analysis in the second dimension (Fig. 3B). A series of protein spots with pIs between 5.5 and 5.9 had a molecular size of 72,120 Da. These were identified as a heavy chain of IgG. A discrete protein spot at 69,440 Da and with a pI of 4.7 was identified as albumin. A series of protein spots at 188,240 Da with a pI of 4.6 but ranging up to 7.0 was sequenced and shown to have high homology to 2M. Sequencing of eight peptide fragments of a protein spot having a calculated molecular size of 220,400 Da and a pI of 4.9 revealed 100% homology with boFN. Purified bovine 2M and boFN were subjected to two-dimensional gel analysis. Two large-molecular-size spots and a 69-kDa spot corresponding to the albumin added as a stabilizer were found on the 2M gel (not shown). In Fig. 3C, commercial plasma boFN revealed a single protein band at approximately 250 kDa with a pI of 5.3 to 5.6.
Protein specificity. Commercially available purified boFN and bovine 2M were analyzed for inhibition of viral expression in the p24 sELISA. These proteins are purified by chromatography from plasma sources and have BSA added as a stabilizing agent. Figure 4 shows the dose response of bovine mononuclear cell production of p24 in response to culture with boFN. There was a dose response, but higher levels of commercial boFN than of purified PBF were required to achieve significant reduction. The bovine 2M had no effect on viral expression (data not shown).
To demonstrate that the purified PBF had a biological function in BLV-infected cells from other animals, mononuclear cells from six BLV+ donor cows with PL were cultured in medium alone or in medium with either 25% autologous plasma, 150 μg purified PBF, or 250 μg boFN (Fig. 5A). Twenty-five percent plasma, purified PBF, and boFN were each able to inhibit p24 production in cells from all six cows relative to the medium controls. Further experiments to investigate whether FN from human sources was able to mediate a blocking effect as suggested by Zandomeni et al. (8) were done using 500 μg of various human fibronectin peptides. As shown in Fig. 5B, while plasma, purified PBF, and boFN were able to reduce p24 expression to 25% of that of the reference medium in cultures of mononuclear cells isolated from six naturally infected BLV+ cows, neither huFN nor the integrin/cell-binding RGD peptide derived from huFN inhibited p24 production. FN40 competitively blocks the FN binding domain of 4 integrin (expressed on the cell membrane of most lymphocytes). FN40 that is added to a culture containing PBF competes for RGD integrin binding sites, resulting in loss of activity if the PBF also binds this site. We observed partial restoration of p24 expression, but the effect was not significant. Adhesion-promoting huFN peptides had no effect on p24 production (data not shown).
DISCUSSION
The aim of the current study was to isolate the factor present in bovine plasma which blocks BLV expression. We demonstrated that PBF was present in bovine plasma, that it was resistant to heating, and that its effect could be eliminated at a concentration of less than 22 mg/ml. Physicochemical fractionation by ammonium sulfate precipitation, followed by anion-exchange and gel exclusion chromatographies resulted in a highly concentrated product that retained biological activity. Under nonreducing conditions, blocking activity was associated with a 320-kDa protein complex which revealed six visible proteins under two-dimensional protein analysis. Amino acid sequencing showed these to be two forms of 2M and a protein of approximately 220 kDa with a pI of 4.9 and with 100% homology to FN. The remaining proteins were deduced to be albumin, immunoglobulin light chain, and IgG heavy chain. The molecular size determined under reduced conditions relative to that of the native protein suggested formation of an FN dimer complexed with 2M and immunoglobulins. The evidence presented here supports the possibility of an isoform of boFN with a pI of 4.9 as the putative plasma blocking factor.
We noted significant stimulation of BLV p24 production in plasma precipitated with 41 to 50% saturated ammonium sulfate, while most blocking activity was present in the 21-to-30% fraction reported previously (15). Commonly, 50% saturated ammonium sulfate is used to crudely fractionate immunoglobulin from plasma, suggesting that the peaks in blocking activity and immunoglobulin concentration were at least partially separated by this procedure. Others (28) have also reported that immunoglobulin stimulated BLV production, making immunoglobulin an unlikely candidate for PBF.
In contrast, commercially purified boFN did cause significant reduction in BLV p24 production, although PBF prepared for the present study was more potent. Two-dimensional analysis of boFN revealed only the high-molecular-weight protein, no smaller proteins, indicating that the pure boFN product was capable of blocking p24 production, but it is possible that another protein which complexes with FN also contributes to PBF activity. Differences in isolation methodologies (binding to denatured collagen, followed by extraction with 1 to 8 M urea) may account for the decreased bioactivity observed with the commercial product. Commercially purified bovine serum albumin and bovine 2M showed no inhibitory activity in our bioassay.
Fibronectin is a ubiquitous glycoprotein that has a multiplicity of roles, including embryogenesis, thrombosis, wound healing, and hemostasis. It has two forms: a soluble form in body fluids and an insoluble form in the extracellular matrix. Plasma FN is secreted as a dimer of 440 kDa, which is found in plasma at a concentration of 300 μg/ml (21). The 220- to 250-kDa monomers are bound together with two disulfide bonds near the carboxyl terminus. Fibronectin has a pI of 5.6 to 6.3, with peak activity at 6.1, but is reported to suffer rapid loss of activity at temperatures over 60 to 65°C (2). In contrast, the native fibronectin complex isolated here had a molecular size of 320 kDa, while the monomer was 220 kDa with a pI of 4.9. The fibronectin gene is composed of three types of repeating subunits (types I to III). Alternative splicing of type III repeats, inclusion or exclusion of two repeats known as EDI and EDII, and subdivision of a nonhomologous third variable region give rise to 20 permutations in huFN, of which 2 are plasma forms (11).
The amino and carboxy termini of FN are composed of repeated type I modules which bind fibrin and collagen. Type II modules are less frequent in the molecule and contribute to collagen binding. The most abundant module, of which there are 12 to 17 copies, is type III, which contains the RGD integrin binding site as well as sites for binding heparin and other integrins. Plasma FN differs from the insoluble extracellular matrix FN by having fewer type III modules and having a second RGD site. Curiously, despite greater than 80% homology at the amino acid level between huFN and boFN, huFN had no blocking activity in our bioassay. HuFN (2,386 amino acids [aa]; 260 kDa) differs from boFN (2,265 aa; 240 kDa) by having a 31-aa extension at the amino terminus and a 90-aa insert at aa 105 following the integrin-binding domain of FN, arginine, glycine, and aspartic acid (RGD). A human antibody to FN is reported to bind an epitope in the decapeptide RGDSPASSKP of bovine and marsupial FN but not huFN (13). Both bovine and human FN contain this decapeptide, but the bovine FN sequence is followed by a valine, while the human FN sequence is followed by an isoleucine, after which both are identical again until the human FN 90-aa insert. Since the antibody was able to distinguish boFN from huFN at the RGD integrin binding site, it is possible that huFN is unable to interact with the cell surface receptor and signal transduction apparatus present on bovine cells.
The biological effects of FN vary with the cell type, expression of cell surface ligands for FN, and pleomorphism of the FN molecule. Little is known about the interaction of soluble FN and blood cells. However, adhesion of Jurkat T cells to the extracellular matrix FN increased NFB activation within 30 min, but this effect was progressively lost if incubation continued and NFB was deactivated by 12 h; the effect persisted for over 24 h (1). Interestingly, NFB proteins and Tax activate BLV transcription and may play a role in rescuing the virus from latency (3). Should a similar paradigm exist with bovine blood cells and soluble FN, then it is possible that FN acts by deactivating the nuclear transportation of NFB, thereby maintaining BLV in a latent state.
ACKNOWLEDGMENTS
This work was supported by the Natural Sciences and Engineering Research Council and the Ontario Ministry of Agriculture and Food. M. van den Heuvel was supported by a scholarship from the Dairy Farmers of Ontario.
The generosity of D. Portetelle and M. A. Hayes in sharing resources was very much appreciated.
REFERENCES
Bearz, A., G. Tell, A. Colombatti, S. Formisano, and C. Pucillo. 1998. Fibronectin binding promotes a PKC-dependent modulation of NF-kappa B in human T cells. Biochem. Biophys. Res. Commun. 243:732-737.
Bozzini, S., V. Falcone, P. G. Conaldi, L. Visai, L. Biancone, A. Dolei, A. Toniolo, and P. Speziale. 1998. Heparin-binding domain of human fibronectin binds HIV-1 gp120/160 and reduces virus infectivity. J. Med. Virol. 54:44-53.
Brooks, P. A., G. L. Cockerell, and J. K. Nyborg. 1998. Activation of BLV transcription by NF-kappa B and Tax. Virology 243:94-98.
Burny, A., C. Bruck, Y. Cleuter, D. Couez, J. Deschamps, D. Gregoire, J. Ghysdael, R. Kettmann, M. Mammerickx, and G. Marbaix. 1985. Bovine leukaemia virus and enzootic bovine leukosis. Onderstepoort. J. Vet. Res. 52:133-144.
Burny, A., Y. Cleuter, R. Kettmann, M. Mammerickx, G. Marbaix, D. Portetelle, A. Van den Broeke, L. Willems, and R. Thomas. 1987. Bovine leukaemia: facts and hypotheses derived from the study of an infectious cancer. Cancer Surv. 6:139-159.
Cornil, I., and D. Levy. 1989. In vivo inhibition of bovine leukemia virus (BLV) expression. Leukemia 3:159-161.
Derse, D. 1987. Bovine leukemia virus transcription is controlled by a virus-encoded trans-acting factor and by cis-acting response elements. J. Virol. 61:2462-2471.
Derse, D. 1988. trans-acting regulation of bovine leukemia virus mRNA processing. J. Virol. 62:1115-1119.
Doonan, S. 1996. Bulk purification by fractional precipitation, p. 135-144. In S. Doonan (ed.), Protein purification protocols. Humana Press, Totowa, N.J.
Ferrer, J. F. 1980. Bovine lymphosarcoma. Adv. Vet. Sci. Comp. Med. 24:1-68.
ffrench-Constant, C. 1995. Alternative splicing of fibronectin—many different proteins but few different functions. Exp. Cell Res. 221:261-271.
Gallo, R. C., and L. S. Nerurkar. 1989. Human retroviruses: their role in neoplasia and immunodeficiency. Ann. N. Y. Acad. Sci. 567:82-94.
Girard, D., Y. Raymond, P. Labbe, and J. L. Senecal. 1995. Characterization of a novel human antibody xenoreactive with fibronectin. Clin. Immunol. Immunopathol. 77:149-161.
Gupta, P., and J. F. Ferrer. 1982. Expression of bovine leukemia virus genome is blocked by a nonimmunoglobulin protein in plasma from infected cattle. Science 215:405-407.
Gupta, P., S. V. S. Kashmiri, and J. F. Ferrer. 1984. Transcriptional control of the bovine leukemia virus genome: role and characterization of a non-immunoglobulin plasma protein from bovine leukemia virus-infected cattle. J. Virol. 50:267-270.
Haas, L., T. Divers, and J. W. Casey. 1992. Bovine leukemia virus gene expression in vivo. J. Virol. 66:6223-6225.
Jensen, W. A., J. Rovnak, and G. L. Cockerell. 1991. In vivo transcription of the bovine leukemia virus tax/rex region in normal and neoplastic lymphocytes of cattle and sheep. J. Virol. 65:2484-2490.
Kettmann, R., G. Marbaix, Y. Cleuter, D. Portetelle, M. Mammerickx, and A. Burny. 1980. Genomic integration of bovine leukemia provirus and lack of viral RNA expression in the target cells of cattle with different responses to BLV infection. Leuk. Res. 4:509-519.
Lee, T. H., J. E. Coligan, J. G. Sodroski, W. A. Haseltine, S. Z. Salahuddin, F. Wong-Staal, R. C. Gallo, and M. Essex. 1984. Antigens encoded by the 3'-terminal region of human T-cell leukemia virus: evidence for a functional gene. Science 226:57-61.
Miller, J. M., L. D. Miller, C. Olson, and K. G. Gillette. 1969. Virus-like particles in phytohemagglutinin-stimulated lymphocyte cultures with reference to bovine lymphosarcoma. J. Natl. Cancer Inst. 43:1297-1305.
Retta, S. F., P. Ferraris, and G. Tarone. 1999. Purification of fibronectin from human plasma. Methods Mol. Biol. 96:119-124.
Slamon, D. J., K. Shimotohno, M. J. Cline, D. W. Golde, and I. S. Chen. 1984. Identification of the putative transforming protein of the human T-cell leukemia viruses HTLV-I and HTLV-II. Science 226:61-65.
Stock, N. D., and J. F. Ferrer. 1972. Replicating C-type virus in phytohemagglutinin-treated buffy-coat cultures of bovine origin. J. Natl. Cancer Inst. 48:985-996.
Taylor, J. A., and R. M. Jacobs. 1993. Effects of plasma and serum on the in vitro expression of bovine leukemia virus. Lab. Investig. 69:340-346.
Tsukiyama, K., M. Onuma, and H. Izawa. 1987. Effect of platelet-derived factor on expression of bovine leukemia virus genome. Arch. Virol. 96:89-96.
van den Heuvel, M., D. Portetelle, B. Jefferson, and R. M. Jacobs. 2003. Adaptation of a sandwich enzyme-linked immunosorbent assay to determine the concentration of bovine leukemia virus p24 and optimal conditions for p24 expression in short-term cultures of peripheral blood mononuclear cells. J. Virol. Methods 111:61-67.
Zandomeni, R. O., M. Carrera-Zandomeni, E. Esteban, W. Donawick, and J. F. Ferrer. 1992. Induction and inhibition of bovine leukaemia virus expression in naturally infected cells. J. Gen. Virol. 73:1915-1924.
Zandomeni, R. O., E. Esteban, M. Carrera-Zandomeni, and J. F. Ferrer. 1994. Host soluble factors with blocking and stimulating activity on the expression of the bovine leukemia virus. J. Infect. Dis. 170:787-794.(Marianne J. van den Heuve)