BINDING OF BACTERIAL LIPOPOLYSACCHARIDE TO MURINE LYMPHOCYTES* Diane M. Jacobs, Dianna B. Robertst, John H. Eldridge,* and Allen J. Rosenspireg Department of Microbiology State University of New York at Buffalo Buffalo, New York 14214 INTRODUCTION Lipopolysaccharide [LPS), a component of the outer membrane of gram- negative bacteria, possesses a wide spectrum of biological activities, including activation of cells of the immune system and modulation of the immune response.'-3 LPS stimulates cell division and immunoglobulin secretion in murine B-cells,'.' activates macrophage function and secretion of pharmacologically active materials:" and appears to have some effect on a small subset of T-~ells.'.'~-'~ For the past several years, my colleagues and I have sought to understand better the mechanism by which LPS activates cells, by evaluating the nature of the interaction of this bacterial constituent with murine lymphocytes. The selectivity of LPS in activating B- rather than T-lymphocytes would be consistent with a selectivity of interaction with a specific site on the plasma membrane of target cells; yet the hydrophobic nature of lipid A, the structural subunit responsible for most LPS biological activities, might favor a nonselective hydrophobic interaction between the lipid A moiety of LPS with membrane lipids [FIGURE 1). A few investigators's-'7 have demonstrated selective binding of LPS or lipid A to B-cells, whereas have not been able to detect differences in binding between T-cells and B-cells. We have used two approaches to study this question: immunofluorescence microscopy [IF)and radiobinding [RB), and found evidence for "specific" binding, which will be summarized below. In addition, we have examined the cell-bound LPS, and have some evidence that there is selectivity in binding of certain fractions of LPS to different murine cells. NATURE OF LPS BINDING TABLE 1 summarizes the major findings of this combined a p p r ~ a c h . ~ ~ - ~ ~ In immunofluorescence microscopy, cell-bound LPS is detected using a hapten sandwich technique in which the intermediate layer is antibody directed to the specific 0-antigens of the LPS we use: E. coli 055:B5. The use of double labeling techniques allows simultaneous detection of other membrane markers to deter- mine the surface phenotype of LPS-binding cells. Radiobinding assays quantify 'This research was supported by PHS Grant A1 16915 and 5T32 A1 07088. tPresent address: Division of Allied Health and Life Sciences, University of Texas at San $.Present address: Department of Microbiology, University of Alabama in Birmingham, §Present address: Sloan-Kettering Institute for Cancer Research, 145 Boston Post Road, Antonio, San Antonio, Texas 78285. University Station, Birmingham, Alabama 35294. Rye, New York 10580. 72 0077-8923/83/0409-0072 $1.75/0 0 1983. NYAS
Transcript
BINDING OF BACTERIAL LIPOPOLYSACCHARIDE TO MURINE LYMPHOCYTES*
Diane M. Jacobs, Dianna B. Robertst, John H. Eldridge,* and Allen J. Rosenspireg
Department of Microbiology State University of New York at Buffalo
Buffalo, New York 14214
INTRODUCTION
Lipopolysaccharide [LPS), a component of the outer membrane of gram- negative bacteria, possesses a wide spectrum of biological activities, including activation of cells of the immune system and modulation of the immune response.'-3 LPS stimulates cell division and immunoglobulin secretion in murine B-cells,'.' activates macrophage function and secretion of pharmacologically active materials:" and appears to have some effect on a small subset of T-~ells.'.'~-'~ For the past several years, my colleagues and I have sought to understand better the mechanism by which LPS activates cells, by evaluating the nature of the interaction of this bacterial constituent with murine lymphocytes. The selectivity of LPS in activating B- rather than T-lymphocytes would be consistent with a selectivity of interaction with a specific site on the plasma membrane of target cells; yet the hydrophobic nature of lipid A, the structural subunit responsible for most LPS biological activities, might favor a nonselective hydrophobic interaction between the lipid A moiety of LPS with membrane lipids [FIGURE 1). A few investigators's-'7 have demonstrated selective binding of LPS or lipid A to B-cells, whereas have not been able to detect differences in binding between T-cells and B-cells. We have used two approaches to study this question: immunofluorescence microscopy [IF) and radiobinding [RB), and found evidence for "specific" binding, which will be summarized below. In addition, we have examined the cell-bound LPS, and have some evidence that there is selectivity in binding of certain fractions of LPS to different murine cells.
NATURE OF LPS BINDING
TABLE 1 summarizes the major findings of this combined a p p r ~ a c h . ~ ~ - ~ ~ In immunofluorescence microscopy, cell-bound LPS is detected using a hapten sandwich technique in which the intermediate layer is antibody directed to the specific 0-antigens of the LPS we use: E. coli 055:B5. The use of double labeling techniques allows simultaneous detection of other membrane markers to deter- mine the surface phenotype of LPS-binding cells. Radiobinding assays quantify
'This research was supported by PHS Grant A1 16915 and 5T32 A1 07088. tPresent address: Division of Allied Health and Life Sciences, University of Texas at San
$.Present address: Department of Microbiology, University of Alabama in Birmingham,
§Present address: Sloan-Kettering Institute for Cancer Research, 145 Boston Post Road,
FIGURE 1. Schematic representation of LPS. Reproduced from By permission of Raven Press.
the amount of iodinated LPS specifically bound to cell populations, and thus allow some analysis of the characteristics of the binding site.
Binding by either method is saturable with respect both to dose of LPS and time of exposure. In the IF assay, we have observed that the number of LPS' splenic lymphocytes reaches a plateau of 50-60% at 20-30 pg/ml LPS after 30 minutes of incubation at 0" . Radiolabeled LPS reaches equilibrium-binding in 50 minutes at 0" . Temperature dependence was demonstrated in both cases: binding occurred at lower concentrations of LPS at 37" C rather than at 0" C. but the plateau level of LPS' cells remained the same: with RB, more LPS was bound at the higher temperature. As these observations held true, whether or not cells were incubated in the presence of azide, the temperature dependence did not
TABLE 1
IMMUNOFLUORESCENCE MICROSCOPY AND RADIOBINDING.* CHARACTERISTICS OF LPS BINDING TO MURINE LYMPHOCVTES DETERMINED BY
Binding Detected bv:
Saturable: dose and time Yes Yes Temperature dependence Yes Yes Energy independent Yes Yes Inhibitable Yes Yes Correlation between LPS' cells and mitogenic response Yes ND
Selective for B-cells Yes ND Evidence for nonspecific binding Yes Yes Responder vs nonresponder cells Same Same
*LPS used in these experiments was extracted from lyophilized E. coli 055335 by the phenol-water procedure and purified by enzymatic treatment and ~hromatography.~~ It did not stimulate mitogenesis in cultures of C3H/HeJ spleen cells.
in lymphoid cells
74 Annals New York Academy of Sciences
reflect an energy-dependent binding event, but a temperature-dependent step that modulated the efficiency of binding.
Inhibition experiments were carried out to examine the specificity of binding and to determine the structural component responsible for binding. In RB experiments, unlabeled LPS of the same serotype clearly inhibited 25-4570 of the binding of labeled material; this inhibition allowed us to analyze the affinity and number of sites involved in specific binding, which are approximately 3 x lo7 M-' and 104/cell respectively. As the IF assay used antibody to the 0-specific antigen of E. coli 055:B5 to detect cell-bound LPS. a similar approach, examining inhibition by the same serotype could not be used. Nevertheless, we were able to detect inhibition of LPS binding when incubations were carried out with a mixture of 055:B5 and noncrossreactive LPS from S. typhimurium. Because the structural determinant common to these two LPS is lipid A, it is likely that this moiety plays a major role in binding to cells, as would be expected from its biological activity.
Double-labeling IF studies have shown that LPS binds selectively to B-cells, predominantly those of the mature subclass with the surface markers p+Iatd+. A small number of T-cells and null cells are also LPS'. The distribution of LPS' cells in various lymphoid organs parallels the mitogen response elicited by LPS in cell populations from the same sources. This distribution and the selectivity for B-cells suggest that the binding of LPS we detect is related to the activation of cells by this material, and the characteristics of binding are consistent with the presence of a plasma membrane receptor for this ligand.
Evidence can also be found for nonspecific binding by LPS. After the initial rapid rise to plateau levels of LPS' cells with increasing doses of LPS, the number of positive cells continues to increase gradually, and at the very high concentra- tion of 2 rng/ml. may be 90% of the total cells. We consider this increase to reflect nonspecific binding. It probably occurs at low doses as well, but is not distinguish- able in this assay system. It can, however, be detected using the RB assay, where binding is always measured in the presence and absence of unlabeled LPS. In this case, even at the very low doses of LPS used (usually 0.1 pg/ml], 55-7570 of the total LPS bound is not inhibitable by excess cold LPS, and is considered to be bound nonspecifically.
The above results are all consistent with the possibility that LPS selectively binds to lymphocytes by a biologically relevant specific binding site or receptor. With these results in mind, we compared the binding of LPS to lymphocytes of two strains of mice: C3H/St, which responds normally to LPS activation, and C3H/HeJ, which does not "recognize" any activation signals delivered by LPS and is therefore referred to as an LPS nonresponder. We could find no major differences between the strains with respect to the amount of LPS bound, the nature of the binding, or the number and nature of LPS-binding cells. It would therefore appear that the genetic lesion in this strain is not solely a result of altered ligand-binding properties of a cell membrane structure.
CHARACTERIZATION AND FATE OF CELL-BOUND LPS
'251-labeled LPS used in radiobinding experiments, described above, was prepared by chloramine T-iodination of LPS that was made reactive by addition of a phenyl group by reaction with p-hydroxy phenylacetic acid (pHPAA) in the presence of carbodiimide.'' We had previously found that LPS derivatized for iodination had biological activities similar to native material. Therefore, binding
Jacobs et 01.: Murine Lymphocytes 75
of radiolabeled LPS was assumed to measure the same material as the IF technique, and both were assumed to reflect the starting material. As this assay is more susceptible to misinterpretation when either the cell populations or the ligand are heterogenous or impure, we examined the bound radioactivity to confirm that it was, in fact, LPS and to ascertain whether that LPS that bound to cells was representative of native LPS. We analyzed '2SI-labeled LPS and the cell-bound radioactive material, using electrophoretic separation on disc gels of 15% acrylamide containing sodium dodecyl sulfate (SDS), according to the procedure of Laemm1Lz7
FIGURE 2. Fractionation on LPS of acrylamide gels. The gel on the left was loaded with 0.1 mg LPS and was stained for carbohydrate with Alcian blue. The gel on the right was loaded with 0.25 mg LPS and was stained with Sudan Black B. The dark smear on the bottom of the left-hand gel is an artifact of staining and obscures the band corresponding to that seen on the right-hand gel.
In aqueous solution, LPS exists as aggregates of many subunits of the structure diagrammed in FIGURE 1." In the presence of SDS. the large molecular aggregates are dissociated and will migrate through an SDS acrylamide Surprisingly, the subunits are heterogenous as seen in FIGURE 2. The pattern is the same when LPS is stained either with the carbohydrate stain, Alcian blue,31 or the lipid stain, Sudan Black B. The intensities are different, however, and the top appears to be carbohydrate-rich whereas the bottom bands appear to be lipid-rich. The pattern is similar to that reported by Palva and Makela,30 and Goldman and Leive." These investigators concluded, on the basis of the mobility of biosynthetically-labeled LPS, that heterogeneity is due to differences in molecular size, based on differences in the number of antigenic side-chain units per subunit. Thus, the number of 0-antigen side-chains in FIGURE 1 varies from none, to over forty.
76 Annals New York Academy of Sciences
Iodinated LPS was analyzed quantitatively by identical electrophoretic sepa- ration. The gel was sliced, and radioactivity in the fractions was determined in order to give the profile seen in FIGURE 3. The fastest moving peak was coincident with the dye front and had the same mobility as 12sII-pHPAA. We see that some amount of this small molecule used for attachment to LPS remains adsorbed but not covalently bound, and it was subsequently iodinated; or else a similar small molecule is a breakdown product of the iodinated derivative of LPS. We have assumed that all counts in this peak are not intact structural subunits of LPS. The remainder of the material in the LPS preparation resolves into several broad overlapping peaks. A peak always exists at Rr - 0.82 called peak 1. as well as at Rr = 0.29, and sometimes a small peak at R, - 0.15. For purposes of discussion, we have pooled the latter two components and called them peak 3. In different experiments, the radioactive material between these two major peaks may be resolved into a separate peak, or it may be a shoulder to peak 1, or a high plateau.
or15 029 058 0.82 1.0 Av.Rf t t t t t 20.8 16.5 44.0 18.2 Av.%of
I I I I 1 CPM/Peok
R f FIGURE 3. Fractionation of '2sI-labeled LPS on polyacrylamide gels. Radioiodine was
introduced to LPS by chloramine T-iodination of LPS, which first had a reactive phenyl group attached by reaction with p-hydroxy phenylacetic acid (pHPAA) in the presence of carbodiimide. lZSI-labeled pHPAA was also prepared. Iodinated materials were treated with SDS and electrophoresed on 15% polyacrylamide gels by the procedure of Laamekn Two mm wide fractions were cut and counted, and the profile of counts are depicted here. R, values of major peaks of 30 gels were determined and averaged to give the figures above the curve. "Peak 3" refers to all material with an R, value less than 0.3. The area under each peak was determined, and the fraction of total counts in each peak was calculated. The average value of 11 gels is given.
Jacobs et a].: Murine Lymphocytes
TABLE 2
BY REACTIVE CELLS AND ANTIBODY PREFERENTIAL BINDING OF DIFFERENT LPS COMPONENTS
77
Adherent '*'I-LPS/Starting 9 - l abe led LPS in Peaks
*R, values of peaks 3. <0.3: 2.0.58; 1.0.82: pHPAA set equal to 1.0. Lymphocytes were the interface cells from a Ficoll-Hypaque separation of spleen cell suspension. Macrophages were the population of cells remaining adherent after a 24 hour incubation of spleen cells in glass tubes. Cells were incubated with 'ZSI-labeled LPS in RPMI-1640 for one hour at 0' C and washed in medium. All cells except for macrophages were transferred to clean glass tubes with fresh medium and centrifuged. The cell pellet or adherent macrophages were lysed in 0.1 ml distilled water. Material adherent toglass tubes after incubation and washing was analyzed in a similar fashion. Polystyrene tubes were coated with the IgG fraction of specific anti-O55:B5 as in the solid phase radioimmunoassay for LPS," incubated with iodinated LPS and washed. Control polystyrene tubes were incubated in the same buffer containing BSA before incubating with labeled material. Lysates and adherent material were treated with sample buffer containing SDS and run on 15% acrylamide gels containing SDS. The percentage of total counts of each gel distributed in each major peak was determined. The figures given are the ratios of adherent LPS/starting LPS for each peak. Those underlined are considered to be appreciably different from 1.00.
Counts in this region are called peak 2. The percentage of radioactivity in each peak was measured; the percentage of total counts, migrating in each gel, that falls into each peak is calculated. The average distribution of 'zsI-LPS in each fraction is also seen in FIGURE 3. About half the LPS is in peak 1 and the remainder is equally divided among peaks 2, 3, and low molecular weight material.
With this approach as a reference point, we carried out similar analyses of '251-LPS that bound to lymphocytes and other cell types. For each experiment, the distribution of counts in the peaks was determined for cell-bound LPS and the starting material. To determine if any components were preferentially bound, the ratio of adherent material to starting material for each peak was calculated. If a component was bound to the same degree as it was represented in the starting material, it would be expected to have a ratio of 1.0. A preferentially bound component would therefore have a ratio greater than 1.0 and other components would by default have ratios less than 1.0. A summary of these results is outlined in TABLE 2.
Splenic lymphocytes of either C3H/St or C3H/HeJ appear to preferentially bind components in peak 1. In contrast, adherent splenic macrophages preferen- tially bind peak 2. Erythrocytes, which are known to bind LPS but have no known biological response to it, preferentially bind both peak 1 and peak 2 and have a more pronounced negative preference for peak 3 than any cell type examined. To confirm that these apparent differences in preferential binding are not due solely to LPS binding to tubes, particularly in the case of macrophages. where cells are not transferred before being lysed. we also examined the LPS that bound to glass
7% Annals New York Academy of Sciences
tubes. This surface appears to have a preference for peak 1, even more pronounced than lymphocytes, and its profile is quite different from that of macrophages.
Because the material we refer to as the carbohydrate-rich peak 3 has more repeating antigenic structures than the lipid-rich components, we also asked if antibody would more easily react with any of the components separated by this method. This point was evaluated by analyzing LPS that was bound to tubes that had been coated previously with the IgG fraction of specific anti-O55:B5." In comparison to tubes coated only with bovine serum albumin (BSA), specific antibody preferentially bound peak 3 components, which contain more antigenic side-chains.
Having looked at cell-bound LPS, we continued to analyze the fate of the LPS in/on these cells. The information we have on the rate of loss of cell-bound LPS is summarized in TABLE 3. In these experiments, cells were "loaded" by incubation with 1251-labeled LPS in the cold for one hour, washed, and transferred to clean tubes with fresh medium for incubation at 37O. Samples were taken at subsequent intervals, and the amount of radioactivity remaining on cells was determined. We
TABLE 3 LOSS OF CELL-BOUND LPS FROM MURINE CELLS*
Rate of Loss of 1261-labeled LPS tln in Hours Cell or Surface
*Cells were incubated with '2SI-labeled LPS in RPMI-1640 for one hour at 0" C. washed in medium, transferred to clean glass tubes with fresh medium and incubated for 72 hours at 37O C. Samples were taken at intervals, cells and supernatant fluid were separated by centrifugation, and radioactivity was determined.
saw a loss of lymphocyte-bound label to a level of about 50% of the initial label but little further loss after 24 hours. Macrophages retained the label for longer periods. A detailed analysis indicated that the initial rate of loss from lymphocytes and erythrocytes was fairly rapid. In contrast, macrophages lost label very slowly, possibly because it was internalized when the cells were transferred to 37'.
Differential binding of LPS components to different cell populations provides additional evidence for selectivity of LPS binding. Because not all cells bind each component to the same degree, it is unlikely that the cell surface structure(s) to which LPS binds in each population has identical characteristics. Furthermore, not all cells have a preferential affinity for the most hydrophobic or lipid-rich subunits in peak 1, so binding would not appear to be governed strictly by hydrophobic forces. The results also raise a quite different question about the structure of LPS aggregates in solution. The preferential binding of subunit components could be explained by segregation, such that some aggregates are more homogenous and are composed exclusively or preferentially of the compo- nents represented in the individual peaks. If all aggregates were composed of representative proportions of, for example, peaks 1,2, and 3, regardless of which
Jacobs et a].: Murine Lymphocytes 79
component interacted with the cell, all components would be part of the adhering aggregate and be detected on analysis. Alternatively, cells could be selectively removing components of certain structure from the large aggregates. Perhaps the most likely explanation for preferential adherence of individual components is that the individual subunits (or dimers or trimers) present in low concentrations in equilibrium with the micellar (aggregate] structures are the active interacting components. While they are being removed from solution by binding to the cell, they may be replaced by futher dissociation of the aggregates if the conditions for such dissociation are met. If this concept can be verified by experiments currently in progress in the laboratory, it will change the way we think about LPS interactions with cells and provide a fruitful analytical approach for examining the molecular nature of LPS interactions with cell surfaces.
CONCLUSION AND SUMMARY
Does LPS activate lymphocytes by binding to a specific cell-surface receptor or by nonspecific hydrophobic interaction with the plasma membrane? We examined this question by detecting cell-bound LPS using immunofluorescence microscopy and radiobinding techniques. LPS binding to splenic lymphocytes from C3H/St mice has characteristics of specific binding: saturability with respect to dose and time, selectivity for a subclass of B-cells, and a correlation between binding and mitogenesis. '251-labeled LPS bound to cells and analyzed quantita- tively by SDS-PAGE separated into 3 major components: peaks 1, 2, and 3 (1 equals the fastest moving]. Lymphocytes preferentially bound peak 1, murine RBC peaks 1 and 2, and macrophages peak 2. In contrast, specific antibody preferred peaks 2 and 3. Differential staining of gels suggested that peak 3 is carbohydrate-rich and peak 1 is lipid-rich. LPS was released from these cells at different rates. We conclude that selectivity of LPS binding may be reflected in preferential binding of LPS subunits of different size and/or composition, as well as differential retention of bound LPS.
ACKNOWLEDGMENTS
We thank Patricia Cotter and Swastika Majumdar for technical assistance in many of the experiments discussed here.
REFERENCES
1. 2. 3. 4. 5.
6. 7. 8. 9.
10.
MORRISON, D. C. & R. J. ULEVITCH. 1978. Am. I . Pathol. 93: 527-612. MORRISON, D. C. & 1. L. RYAN. 1979. Adv. Immunol. 28: 293-450. IACOBS. D. M. 1981. J. Immunopharm. 3: 119-132. GERY, I., 1. KRUCER & S. Z. SPIESEL. 1972. J. Immunol. 108: 1088-1091. ANDERSSON. 1.. F. MELCHERS. C. GALANOS & 0. LUDERITZ. 1973.1. Exp. Med. 1 3 7 943-
ANDERSON. J., G. MOLLER, & 0. SIBBERG. 1972. Cell. Immunol. I: 381-393. ANDERSSON, 1.. 0. S~OBERG & C. MOLLER. 1972. Eur. 1. Immunol. 2 349-353. ANDERSSON. I. . A. COUTINHO & F. MELCHERS. 1977.1. Exp. Med. 145: 1520-1530. ROSENSTREICH, D. L. & S. N. VOCEL. 1980. In Microbiology 1980. D. Schlesinger. Ed.:
WEINBERG, I . B.. H. A. CHAPMAN & 1. B. HIBBS. 1978.1. Immunol. 121: 72-80.
953.
11-15, American Society for Microbiology. Washington, D.C.
80
11. 12. 13. 14. 15. 16. 17.
18. 19.
20. 21. 22. 23. 24.
25. 26. 27. 28.
29. 30. 31. 32.
Annals New York Academy of Sciences
SKIDMORE. B. 1.. J. M. CHILLER & W. 0. WEIGLE. 1977. J. Immunol. 118: 274-281. KONDA, S., Y. NAKAO & R. T. SMITH. 1972.1. Exp. Med. 136: 1461-1477. GOODMAN. M. G., & W. 0. WEIGLE. 1979. J. Immunol. 122: 2548-2553. NORCROSS, M. A., & R. T. SMITH. 1977. J. Exp. Med. 145: 1299-1315. GREGORY, S. H.. D. H. ZIMMERMAN & M. KERN. 1980.1. Immunol. 125: 102-107. ZIMMERMAN. D. H.. S. GREGORY & M. KERN. 1977. J. Immunol. 119: 1018-1023. BONA, C., D. JUY, P. TRUFFA-BACHI & G. J. KAPLAN. 1976. 1. Microsc. Biol. Cell.
KABIR, S.. & D. L. ROSENSTREICH. 1977. Infect. Immun. 15: 156-164. M ~ L L E R , G.. J. ANDERSON, H. POHLIT & 0. SIOBERC. 1972. Clin. Exp. Immunol.
SYMONS, D. B. A. & C. A. CLARKSON. 1979. Immunology 38: 503-508. ELDRIDCE. J. H., F. J. SWARTZWELDER & D. M. JACOBS. Submitted for publication. JACOBS, D. M. & 1. H. ELDRIDGE. Submitted for publication. JACOBS, D. M. & A. J. ROSENSPIRE. 1980. Fed. Proc. Fed. Am. SOC. Exp. Biol. 39: 917. JACOBS, D. M. & A. 1. ROSENSPIRE. 1980. Fourth Int’l. Congress of Immunol., Paris,
MORRISON, D. C. & L. LEIVE. 1975. J. Biol. Chem. 250 2911-2919. GUTOWSKI. 1. A. & D. M. JACOBS. 1979. Immunol. Commun. 8(3): 347-364. LAEMMLI, U. 1970. Nature (London] 227: 680-685. GALANOS, C.. 0. LUDERITZ, E. T. RIETSCHEL & 0. WESTPHAL. 1977. In International
Review of Biochemistry: Biochemistry of Lipids 11. T. W. Goodwin. Ed.: Vol. 14 239-335. University Park Press. Baltimore.
25: 47-56.
13 89-99.
France.
GOLDMAN, R. C. & L. LEIVE. 1980. Eur. 1. Biochem. 187: 145-153. PALVA. T. & H. MAKELA. 1980. Eur. J. Biochem. 107 137-143. WARDI, A. H. & G. A. MICHOS. 1972. Anal. Biochem. 49: 607-609. JACOBS. D. M. In Recent Advances in Mucosal Immunity. W. Strober, Ed.. Raven Press.
New York. In press.
DISCUSSION OF THE PAPER
M. BLAKE (Rockfeller University, New York, N.Y.): I would like to know the nature of the LPS preparation you are using. Have you ruled out the possibility that cell digestion of LPS is occurring? DeVol and Gilchrist suggested that LPS was being digested by the cell and then released as lower molecular weight forms.
D. M. JACOBS (State University of New York, Buffalo, N.Y.): I will answer your second question first. All of the binding studies that I have shown have been done with cells incubated at 0°C in the presence of sodium azide for no more than one hour. Whether or not digestion can take place under these conditions I will leave up to you to decide. The LPS that we use is prepared by a phenol-water extraction of lyophilized E. coli 055:B5 that we make ourselves and that we further purify with RNAse, Pronase treatment and column chromatography. Our LPS prepara- tion does not stimulate mitogenic responses in C3H/HeJ spleen cell cultures.
J. R. MCGHEE (The University of Alabama in Birmingham): You mentioned experiments regarding LPS binding to B-cells in baby mice, and you looked at relatively immature B-cells versus more mature B-cells with surface IgM and IgD. Have you also looked at LPS binding in CBA/N, X-linked immunodeficient mice?
JACOBS: No. The maturity was based on the surface phenotype; when lympho-
Jacobs et al.: Murine Lymphocytes 81
cytes from newborns were studied, we found no LPS-binding B-cells. We have not looked at any other mouse strains. Of course, other obvious experiments would involve LPS binding to cells from germfree mice.
M. D. COOPER (The University of Alabama in Birmingham, Ala): Could you discuss the subclasses of B-cells that bind LPS? For example, the triggering studies would suggest that B-cells have those receptors as soon as they are formed. As soon as they are transformed from pre-B-cells or as soon as you see surface immunoglobulin-positive cells, you find cells that can be triggered to plasma cell differentiation with LPS. Is this lost later on, or, is it just a subset that does not relate to differentiation? Finally, has anyone looked at Peyer's patch B-cells, for example. for their binding of LPS?
JACOBS: We have looked at Peyer's patch cells. This organ has the smallest number of LPS' cells: approximately 4 % of total lymphocytes. We have only 11% pi cells in the Peyer's patches. These cell preparations were not made by enzyme dissociation, but by teasing, so we may have lost cells. LPS will activate B-cells as soon as p' cells appear; but this is at the same time at which 6 and Ia begin to appear. If you look at both markers, when the LPS' cells appear-we have looked at six days-we have a small fraction of 6' and Ia-positive cells as well as a larger fraction of pi cells. Comparing neonates of 6 days, which is the time at which the LPS' B-cells start to appear, as determined by double labeling experiments, we see LPS' pi cells in the same numbers and at about the same time as LPS' 6+ and LPS' Ia' cells. The other problem of interpreting the triggering data with the time of appearance of a marker-and I think Melchers and others have also found this-is that if one puts B-cells in culture, two days are required for maturation to LPS-responsiveness. If one puts cells into culture with LPS, one finds responses three days later. and yet the cells may have taken two days to mature in culture.
QUESTION: Have you looked to see whether there is any difference in the fluidity of the membranes after addition of LPS?
JACOBS: No, but others have found a difference in fluidity, depending on whether you use rough LPS, which is predominantly lipid-A, or smooth LPS, which also has carbohydrate. There is a problem with those experiments, however, because most people do not use pure populations of cells and one does not really know what one is looking at if you use a mixture.
QUESTION: How do you interpret your data with the macrophages binding of this LPS? Everyone uses E. coli 055 lipopolysaccharide by tradition. The structure however, has been determined recently, and it contains colitose and branched sugars. This structure is a bacterium that is frequently cultured from food. It would be better to use LPS from bacteria that are not normally in contact with the mouse. Your data indicates that peak 2 bound better to the macrophages. Could this binding be due to recognition from the macrophages on the 0-polysaccharide side-chain, rather than binding by the lipid-A moiety?
JACOBS: Yes. We have no additional data on the nature of the binding other than that which I presented by the polyacrylamide gel electrophoresis.
