There are several classes of endogenous murine leukemia viruses (MuLV) which can be defined by host range and interference patterns ( 1). Ecotropic MuLV, such as the Gross-AKR type, can only infect mouse cells ( 2). Xenotropic MuLV are able to infect a wide range of xenogenic cells but are almost totally incapable of infecting mouse cells ( 3). Amphotropic MuLV, recovered from wild mice, are able to infect both murine and non-murine cells ( 1, 4). The genetic information coding for xenotropic and amphotropic viruses is integrated into the chromosomal DNA of all inbred mouse strains ( 5) while the coding for ecotropic virus is present in the DNA of most but not all strains ( 6, 7). The physical structure of all three classes of MuLV is similar ( Figure 1). There are three structural genes: gag for "group specific antigen" codes for a 70,000 dalton polyprotein precursor which is cleaved to produce the internal proteins p30, p15, p12 and p10 ("p" for protein, the number referring to its size in 1,000's of daltons). Pol produces reverse transcriptase. Env codes for a glycosylated 90,000 dalton precursor which is split into gp70 and p15e, the "gp" indicating a glycoprotein and "e" standing for "envelope" to distinguish it from the core-associated p15 ( 8). Of the 70,000 daltons of gp70, 30,000 is carbohydrate. Gp70 is the external coat protein of the virion ( 9). It determines, in large part, the host range of the virus, presumably by mediating attachment to the target cell.
Three recent observations have further focused our attention on MuLV gp70. First, Jensen, Lerner and co-workers developed a method of fingerprint analysis of gp70 by chromatography of tryptic peptides of purified and iodinated gp70 ( 10). This technique has revealed a surprisingly extensive polymorphism of gp70 that may be as extensive as that of H-2. While the fingerprints of gp70 from different MuLV isolates fall into families which correspond to the three host range classes, the only identical patterns to date are those of the viruses coded by the Akv-1 and Akv-2 genes ( 11). The second finding concerns Hartley and Rowe's isolation of novel viruses from spontaneous lymphomas in several strains of mice which are infectious for both murine and xenogenic cells and form foci on mink lung cells ( 12). They are termed "MCF" for "mink cell focus-inducing." Isolates of this class of MuLV possess gp70 molecules which, by fingerprint analysis and neutralization testing, appear to be genetic recombinants between ecotropic and xenotropic viruses ( 11). The third observation is that gp70's which resemble MuLV proteins antigenically and by fingerprint analysis are found in the serum and on the cell surface of sperm and certain epithelial tissues in the absence of virus production ( 13, 14). Because the products of the many genes that must be present to code for various gp70's may function as differentiation antigens or be involved in cellular regulation and would be expected to mediate such functions at the cell surface, our group has undertaken the long-term study of the types of gp70 present on cell membranes and their genetic control.
The studies to be described in this chapter concern the expression on mouse lymphocytes of gp70 closely related to that of xenotropic MuLV. We have produced an antiserum which reacts with xenotropic MuLV by immunizing New Zealand White rabbits with SIRC (Serum Institute Rabbit Corneal) cells infected with xenotropic MuLV from NZB mice ( 15). By virus neutralization, which depends on antibody to gp70, the antiserum is quite specific for xenotropic virus ( Table 1). Not surprisingly, it also neutralizes the recombinant MCF virus. To determine exactly what antibody specificities are present in the serum we radioiodinated disrupted xenotropic MuLV, exposed it to the antiserum, and analyzed the precipitate on polyacrylamide gels ( Figure 2). With saturating amounts of antiserum, gp70, p30 and two smaller components are precipitated ( Figure 2a). When limiting amounts of reagent are used, only the gp70 is brought down, indicating that the highest titer is against gp70 ( Figure 2b).
To detect gp70 on the cell surface we have used the technique of flow microfluorometry as embodied by the fluorescence-activated cell sorter ( Figure 3) ( 16). The cells to be examined are entrained in the center of a rapidly flowing jet of liquid which is traversed by a laser beam. Molecules of fluorochrome present on the cells are excited by the laser beam and their fluorescence is detected by a photomultiplier tube perpendicular to both the liquid jet and the laser beam. The output is presented as a histogram of the number of cells (ordinate) in each of the 1,000 channels into which the fluorescence distribution (abscissa) is divided.
By utilizing the computer associated with the system to determine the mean channel number of the fluorescence distribution after staining with the fluoresceinated antiserum, we obtain a highly reproducible measure of the amount of gp70 present on the cell. This is given in arbitrary, but constant fluorescence units. We have done this for non-SIRC cell lines infected with different viruses after staining with the antisera to NZB xenotropic MuLV-infected SIRC cells ( Table 2). There is 10 to 15 times more staining of cells infected with xenotropic virus than of cells infected with amphotropic or ecotropic virus, implying relative specificity for xenotropic MuLV-coded gp70. Despite the differences in fingerprint patterns of gp70 obtained from different xenotropic viruses, mink cells infected with xenotropic MuLV isolates from NZB, NZW, DBA/2, AKR, and BALB/c mice all stain brightly with the antiserum to the NZB virus, and antisera to the BALB xenotrope give results that are virtually identical to those obtained with the anti-NZB xenotropic MuLV serum.
The immunoglobulin was isolated from this antiserum, digested with pepsin to produce F(ab')2 fragments, fluoresceinated, and adsorbed with C57BL/6 spleen cells to remove heteroantibody. We have termed the gp70 antigen detected by this type of antiserum XenCSA, of Xenotropic envelope-related cell surface antigen.
The staining of NZB lymphocytes by this reagent is shown in Figure 4. The left panel shows a unimodal distribution of XenCSA on NZB thymocytes. In contrast, the right panel shows that NZB spleen cells display a bimodal distribution of XenCSA with a narrow peak of dull cells and a broad shoulder of more brightly staining cells. The dull peak is the T cells and the bright peak the B cells. This was shown by staining NZB spleen cells with anti-mouse immunoglobulin and separating the non-staining cell T cells from the surface immunoglobulin-bearing B cells by the cell sorter. The separated T cells were stained with anti-XenCSA and rerun on the sorter ( Figure 5). They gave the dotted profile with a peak which is superimposable on the dull peak of the unseparated spleen cells. The brighter cells are missing and hence must be the surface immunoglobulin-baring B cells which were removed. Thus the level of XenCSA expression by NZB lymphocytes depends upon their differentiation pathway. We have obtained the same result with NZW and DBA/2 spleen cells.
Although the antiserum to xenotropic MuLV is primarily directed against gp70, it was necessary to determine if this was the only cell surface molecule recognized by the antiserum. To this end, spleen cells from several strains were surface labeled with 125I, and detergent extracts of them were precipitated with the antisera and analyzed on polyacrylamide gels ( Figure 6). In each case, the only antigenic species recognized by the antisera had a molecular weight of 70,000 daltons. This reaction was abolished by absorption of the antiserum with xenotropic MuLV-infected mink cells but not uninfected mink cells. Similarly, only a 70,000 dalton protein was precipitated by the antiserum from iodinated cell surface extracts of the exogenously infected cell lines. The uptake of this molecule by a lentil lectin column indicates that it is a glycoprotein. The lack of precipitation by antibody to p30 indicates it is not the gag polyprotein.
Thus prepared, we set out on a tour through the subject of this book -- the inbred strains of mice. We found great variation in XenCSA expression ( 17). Examples of the three observed phenotypes are shown in Figure 7. DBA/2, NZB, and NZW are the most positive strains with thymocytes and both T and B cells of the spleen staining. A large group of strains resemble RIII with lower levels of staining of only a portion of spleen cells, presumably B cells, and no thymocyte staining. Another large group, represented by strain MA, had low levels of XenCSA expression on both thymocytes and spleen cells. The distributions of the averages of the computer-derived means of XenCSA staining for four or more mice of each of 99 inbred strains are shown in Figure 8. The triangles are the spleen cell means, the dots the thymocyte means, and the squares the sum of both. Both distributions show a break at 60 so we have arbitrarily scored means above 60 as high and below 60 as low for XenCSA expression. Strains expressing high XenCSA levels on both thymocyte and spleen cells are listed in Table 3.
In addition to NZB, NZW, and DBA/2, the C57BL/6 congenic for GIX and the C57BR-GIX+ mutant are XenCSA positive. Spleen high, thymus low strains are given in Table 4. These include AKR and C57Bl/6. Finally, the strains with low XenCSA levels on both populations are shown in Table 5. The prototype GIX positive strain, 129, and the non-mutant C57BR/cd are in this group. These data show that expression of XenCSA varies markedly between strains. XenCSA levels do not correlate with alleles of the Tla, Pca-1 or H-2 loci or with Gross cell surface antigen. XenCSA expression does show partial correlation with GIX. An illustration of the lack of H-2 effect and an indication of the reproducibility of the method is shown in Table 6. A/SnGrf and three H-2 congenics derived from this strain have very similar XenCSA levels. The same is true of C3H/DiSnGrf and the congenic C3H.SW. Two of the B10 congenics resemble the parent strain. B10.D2/(n) which was derived from the XenCSA negative B10.D2/(o) by further backcrossing to B10 is XenCSA positive, strongly suggesting that a mutation occurred during the derivation of this strain.
We have begun to investigate the genetic control of XenCSA expression. One portion of these studies utilized recombinant inbred strains, derived by brother-sister matings of the F2 cross of two inbred strains. We have examined 24 such strains derived from the strongly XenCSA positive DBA/2 (D2) and the low XenCSA C57BL/6 (B6) strains (Morse, H.C. III, Taylor, B. et al., in preparation). Xen CSA expression by the (B6D2)F1 was intermediate between the parental levels. The XenCSA levels of the recombinant inbred lines ranged from BXD-2, which resembles B6, to BXD-14, which matches the D2 profile ( Figure 9). Many of the lines, despite being homozygous at 99% of their genetic loci, were intermediate between the parental profiles. Averages of the log mean fluorescence for 6 to 11 samples of thymus and spleen for each line are shown in Figure 10. The stippled areas include the mean and 95% confidence limits of the parental strain XenCSA levels. Spleen values, in the lower panel, show partial correlation with thymus levels. For the purpose of genetic analyses, lines with both spleen and thymus values within the range of one of the parents were classified as being of that type. The remainder were assigned to the parental type whose range contained the majority of the individual mean fluorescence values. By this latter criterion, BXD-1 and 12 are considered D2-like and BXD-23, 8, 21, 15 and 6 are scored as B6-like. Comparison of these somewhat arbitrary XenCSA assignments with the alleles at the 60 loci for which these lines have been characterized shows a close association with the Fv-1 type ( Table 7). The D2 allele, Fv-1n, is associated with high XenCSA levels and the B6 type, Fv-1b, with low XenCSA expression. The correlation with Gpd-1, which is 1 map unit from Fv-1, is similar. Three of the four discordant lines, 5, 8 and 21, have intermediate XenCSA levels which makes their classification uncertain. BXD-25 appears clearly D2-like despite being Fv-1b and is being studied further to see if it is a true recombinant. Taking the data at face value gives an apparent recombination frequency of 5% between Fv-1 and the gene controlling XenCSA. To confirm this linkage we have examined 54 (B6D2)F1 x D2 backcross mice. Again, a spectrum of XenCSA expression is observed. Twenty-six of the 54 were in the D2 range, 5 in the B6 range, and 23 were intermediate between the parental strains. The result of Gpd-1 typing all and Fv-1 typing 28 of these mice was the same as with the recombinant inbred lines: an apparent interval of 7 map units between Fv-1 and the gene influencing XenCSA.
There were three apparent recombinants between Fv-1 and the XenCSA controlling gene in this group. One, which typed as Fv-1n but had low XenCSA expression, was progeny tested by mating with D2. It did not breed true since all six offspring scored high for XenCSA. With respect to the DBA/2 system, we conclude that the effect of this locus is regulatory rather than structural since XenCSA can still be expressed in the presence of the low allele. The data at hand do not allow us to distinguish between a pleiotropic effect of Fv-1 or the effects of a closely linked gene which regulates XenCSA independently of Fv-1. We feel that the closeness of the linkage to Fv-1 makes quasi-linkage, as described for Fv-1-incompatible crosses in the GIX system ( 18), unlikely.
Finally, we would like to mention several observations made during the genetic analysis of infectious xenotropic MuLV and XenCSA in NZB mice which illustrate the problems and complexities of XenCSA typing (Chused, T.M., et al., in preparation). As our background strain we have used NFS, an inbred Swiss, Fv-1n, mouse from which we have not recovered infectious ecotropic or xenotropic virus. Seventy-five (NZB x NFS)F1 x NFS backcross mice were tested for virus production and XenCSA expression. In agreement with the studies of Datta and Schwartz ( 19), we also find that NZB contains two independently segregating xenotropic viruses. The first produces large numbers of foci when spleen cells are cocultivated with mink lung cells carrying the mouse sarcoma genome but not the leukemia virus genome ( 20), which we term the S+L- virus. Sixty-two percent of the backcross mice each produced about 40% of the number of foci given by the parental NZB spleen cells in a clear-cut, all-or-none fashion indicating semi-dominant inheritance. In subsequent backcross generations 51.5% of the progeny have been S+L- positive. The second virus does not give foci in the mink S+L- cells but grows in uninfected mink cells and is detected by cytoplasmic immunofluorescence with antibody to p30. It can only be tested for in the absence of the S+L- virus which also grows in mink cells. Fourteen of the 29 S+L- negative backcross mice produced the second virus which we call "FA." The most interesting result is that 84% of the backcross mice were XenCSA positive.
We are now deriving several lines from the first backcross group by successive backcrosses of males with females of the virus negative NFS strain. Table 8 shows the mean XenCSA levels in the virus positive and virus negative members of each generation of backcross progeny. For reference, the NFS background strain has a mean XenCSA level of 40 with the 95% confidence limits being 4 to 77. There are several points to be made: First, although it is not indicated by the data in the table, the variation in XenCSA levels among both virus positive and virus negative segregants of these crosses is much greater than that observed in the inbred lines we have examined. Second, production of virus is associated with higher XenCSA levels. Third, there are mating-to-mating variations in mean XenCSA levels. The S+L- line being selected for average XenCSA level has shown cyclic variation over four generations. This is suggestive of maternal antibody or some other suppressive factor. Fourth, the S+L- line being selected for low XenCSA has so far not bred true, again suggesting an influence of factor(s) other than the viral genome on XenCSA expression. Fifth, in the line which carries both viruses and was derived from one of the first five backcross mice which expressed XenCSA levels as high as NZB, all 63 progeny through N4 are XenCSA positive, implying a very large number of XenCSA producing genes or an unusual mechanism. The same phenomenon was observed in N4 of the S+L- line where the mean XenCSA level of the virus-negative segregants was 100, even though the virus-negative mice of the preceding generation did not have significantly elevated XenCSA expression. We have not been able to demonstrate infectious spread of XenCSA when tested by several different methods and do not have an explanation for this phenomenon. Sixth, there is a line at N3 which expresses XenCSA in the absence of both viruses. We have not yet tested enough progeny to determine the segregation ratio. Thus NZB mice contain at least three phenotypically distinct semi-dominant genes that affect XenCSA expression by lymphocytes.
We anticipate that future studies will determine the genetic control of XenCSA expression and its relationship to cell differentiation. We are also investigating its role in the autoimmune disease of NZB mice.
1. Hartley, J.W., and Rowe, W.P. (1976). J. Virol. 19: 19.
See also
PubMed.
2. Gross, L. (1951). Proc. Soc. Exp. Biol. Med. 78: 342.
See also
PubMed.
3. Levy, J.A., and Pincus, T. (1970). Science 170: 326.
See also
PubMed.
4. Rasheed, S., Gardner, M.B., and Chan, E. (1976). J. Virol. 19: 13.
See also
PubMed.
5. Callahan, R., Benveniste, R.E., Lieber, M.M., and Todaro, G.J. (1974). J. Virol. 14: 1394.
See also
PubMed.
6. Chattopadhyay, S.K., Rowe, W.P., Teich, N.M., and Lowy, D.R. (1975). Proc. Natl. Acad. Sci. USA 72: 906.
See also
MGI.
7. Chattopadhyay, S.K., Lowy, D.R., Teich, N.M., and Rowe, W.P. (1975). Cold Spring Harbor Symp. Quant. Biol. 39: 1085.
See also
PubMed.
8. Bolognesi, D.P., Montelaro, R.C., Frank, H., and Schafer, W. (1978). Science 199: 183.
See also
PubMed.
9. Witte, O.N., Weissman, I.L, and Kaplan, H.S. (1973). Proc. Natl. Acad. Sci. USA 70: 36.
See also
PubMed.
10. Elder, J.H., Jensen, F.C., Bryant, M.L., and Lerner, R.A. (1977). Nature 267: 23.
See also
PubMed.
11. Elder, J.H., Gautsch, J.W., Jensen, F.C., Lerner, R.A., Hartley, J.W., and Rowe, W.P. (1977). Proc. Natl. Acad. Sci. USA 74: 4676.
See also
PubMed.
12. Hartley, J.W., Wolford, N.K., Old, L.J., and Rowe, W.P. (1977). Proc. Natl. Acad. Sci. USA 74: 789.
See also
MGI.
13. DelVillano, B., Nave, B., Croker, B.P., Lerner, R.A., and Dixon, F.J. (1975). J. Exp. Med. 141: 172.
14. McClintock, P.R., Ihle, J.N., and Jospeh, D.R. (1977). J. Exp. Med. 146: 422.
See also
PubMed.
15. Morse, H.C., Chused, T.M., Boeh-Truitt, M., Mathieson, B.J., Sharrow, S.S., and Hartley, J. W. (1979). J. Immunol. 122: 443.
See also
MGI.
16. Loken, M.R., and Herzenberg, L.A. (1975). Ann. N.Y. Acad. Sci. 254: 163.
See also
PubMed.
17. Morse, H.C., Chused, T.M., Hartley, J.W., Mathieson, B.J., Sharrow, S.O., and Taylor, B.A. (1979). J. Exp. Med. 149: 1183.
See also
MGI.
18. Stockert, E., Boyse, E.A., Sato, H., and Itakura, K. (1976). Proc. Natl. Acad. Sci. 73: 2077.
See also
MGI.
19. Datta, S.K., and Schwartz, R.S. (1976). Nature 263: 412.
See also
PubMed.
20. Peebles, P.T. (1975). Virol. 67: 288.
See also
PubMed.