Altered cytochrome c display precedes apoptotic cell death in Drosophila.

Drosophila affords a genetically well-defined system to study apoptosis in vivo. It offers a powerful extension to in vitro models that have implicated a requirement for cytochrome c in caspase activation and apoptosis. We found that an overt alteration in cytochrome c anticipates programmed cell death (PCD) in Drosophila tissues, occurring at a time that considerably precedes other known indicators of apoptosis. The altered configuration is manifested by display of an otherwise hidden epitope and occurs without release of the protein into the cytosol. Conditional expression of the Drosophila death activators, reaper or grim, provoked apoptogenic cytochrome c display and, surprisingly, caspase activity was necessary and sufficient to induce this alteration. In cell-free studies, cytosolic caspase activation was triggered by mitochondria from apoptotic cells but identical preparations from healthy cells were inactive. Our observations provide compelling validation of an early role for altered cytochrome c in PCD and suggest propagation of apoptotic physiology through reciprocal, feed-forward amplification involving cytochrome c and caspases.

Regulated targeting of BAX to mitochondria.

The proapoptotic protein BAX contains a single predicted transmembrane domain at its COOH terminus. In unstimulated cells, BAX is located in the cytosol and in peripheral association with intracellular membranes including mitochondria, but inserts into mitochondrial membranes after a death signal. This failure to insert into mitochondrial membrane in the absence of a death signal correlates with repression of the transmembrane signal-anchor function of BAX by the NH2-terminal domain. Targeting can be instated by deleting the domain or by replacing the BAX transmembrane segment with that of BCL-2. In stimulated cells, the contribution of the NH2 terminus of BAX correlates with further exposure of this domain after membrane insertion of the protein. The peptidyl caspase inhibitor zVAD-fmk partly blocks the stimulated mitochondrial membrane insertion of BAX in vivo, which is consistent with the ability of apoptotic cell extracts to support mitochondrial targeting of BAX in vitro, dependent on activation of caspase(s). Taken together, our results suggest that regulated targeting of BAX to mitochondria in response to a death signal is mediated by discrete domains within the BAX polypeptide. The contribution of one or more caspases may reflect an initiation and/or amplification of this regulated targeting.

Mapping epitopes on a protein antigen by the proteolysis of antigen-antibody complexes.

A monoclonal antibody bound to a protein antigen decreases the rate of proteolytic cleavage of the antigen, having the greatest effect on those regions involved in antibody contact. Thus, an epitope can be identified by the ability of the antibody to protect one region of the antigen more than others from proteolysis. By means of this approach, two distinct epitopes, both conformationally well-ordered, were characterized on horse cytochrome c.

Characterization of the placental alkaline phosphatase-like (Nagao) isozyme on the surface of A431 human epidermoid carcinoma cells.

A431 human epidermoid carcinoma cells monophenotypically express the placental alkaline phosphatase (PLAP)-like enzyme shown by its catalytic and antigenic characteristics, properties which are shared by the Nagao isozyme. More specifically, it is L-leucine sensitive just as is the rare placental D-variant of PLAP and the testicular heat-stable enzyme. Collectively, these are all referred to as PLAP-like enzymes. The enzyme was localized to the surface of the plasma membrane since it was released in an active form by bromelain treatment of cells. The number of molecules per A431 cell was estimated by radioimmunoassay at 7.5 X 10(5), a value significantly higher than that observed for HeLa TCRC-1 cells (5 X 10(4) which express the S-variant of PLAP, also referred to as the Regan isozyme. The quantity of the enzyme was increased significantly (10-fold) by treating the cells with modulating agents including sodium butyrate, prednisolone, and hyperosmolar sodium chloride. The identification of a cell line such as A431 with enhanced expression in the amount of the PLAP-like enzyme and which can be further enhanced by modulating agents will facilitate studies of the differences and the similarities between this protein and other variants of PLAP. The A431 cell line now takes its place with other cell lines which are phenotypically restricted in their expression of alkaline phosphatase. Finally, the A431 cell line is also shown here to be a suitable model system for in vivo tumor studies such as immunolocalization.

Evidence for homology of normal and neoplastic human placental alkaline phosphatases as determined by monoclonal antibodies to the cancer-associated enzyme.

The HeLa TCRC-1 human adenocarcinoma cell line expresses a form of alkaline phosphatase that is similar to the common S-variant of placental alkaline phosphatase (PLAP) on the basis of electrophoretic mobility, catalytic properties, and reactivity with polyclonal antibodies. More sensitive probes of changes in protein structure than polyclonal antibodies are monoclonal antibodies (MAbs) which recognize individual antigenic sites on molecules. Therefore, we produced MAbs to HeLa TCRC-1 cells and selected those which bound to the alkaline phosphatase expressed by the cancer cells. Seven MAbs were obtained and characterized by (a) fine specificity analysis using allelic variants of PLAP and other human alkaline phosphatase isozymes, (b) immunoglobulin isotype, and (c) relative binding affinities to PLAP from two sources, placental tissue and HeLa TCRC-1 cells. The seven MAbs bind the enzymes from both sources with equal affinity indicating a high degree of structural homology if not identity between the normal S-variant of PLAP and its cancer-associated counterpart. We note that most of the MAbs to cancer cell surface-bound PLAP express either Ig (immunoglobulin) G2a or IgG2b heavy-chain isotypes, a higher incidence of these classes of IgG than has been observed with the purified and soluble PLAP immunogen which yields MAbs predominantly of the IgG1 isotype. Finally, some of these antibodies, like the ones prepared from purified PLAP, recognize differences between allelic variants.

Release of intact, monomeric cytochrome c from apoptotic and necrotic cells.

Cytochrome c (Cyt c) has been shown to translocate from mitochondria to the cytoplasm of cells early in apoptosis. In this study sandwich ELISAs for Cyt c were used to determine if Cyt c is ultimately released from apoptotic and necrotic cells. Gel-filtration and cation-exchange chromatographies, in conjunction with immunoreactivity in ELISA, and Western blotting were employed to examine the integrity of the released Cyt c. The results show that Cyt c is released from both apoptotic and necrotic cells in an intact, monomeric form. The release of Cyt c from apoptotic splenocytes began within 2 h following apoptotic insult, while Cyt c was immediately released following induction of necrosis by heat shock. These findings may be relevant to understanding how Cyt c becomes a target for antibody production in some patients with systemic autoimmune diseases.

A cytosolic factor is required for mitochondrial cytochrome c efflux during apoptosis.

Treatment of HL-60 cells with staurosporine (STS) induced mitochondrial cytochrome c efflux into the cytosol, which was followed by caspase-3 activation and apoptosis. Consistent with these observations, in vitro experiments demonstrated that, except for cytochrome c, the cytosol of HL-60 cells contained sufficient amounts of all factors required for caspase-3 activation. In contrast, treatment of HCW-2 cells (an apoptotic-resistant HL-60 subclone) with STS failed to induce significant amounts of mitochondrial cytochrome c efflux, caspase-3 activation, and apoptosis. In vitro assays strongly suggested that a lack of cytochrome c in the cytosol was the primary limiting factor for caspase-3 activation in HCW-2 cells. To explore the mechanism which regulates mitochondrial cytochrome c efflux, we developed an in vitro assay which showed that cytosolic extracts from STS-treated, but not untreated, HL-60 cells contained an activity, which we designated 'CIF' (cytochrome c-efflux inducing factor), which rapidly induced cytochrome c efflux from HL-60 mitochondria. In contrast, there was no detectable CIF activity in STS-treated HCW-2 cells although the mitochondria from HCW-2 cells were responsive to the CIF activity from STS-treated HL-60 cells. These experiments have identified a novel activity, CIF, which is required for cytochrome c efflux and they indicate that the absence of CIF is the biochemical explanation for the impaired ability of HCW-2 cells to activate caspase-3 and undergo apoptosis.

Affinity consideration in the design of synthetic vaccines intended to elicit antibodies.

As a model for synthetic vaccines BALB/c mice were injected with a large cyanogen bromide cleaved fragment of horse cytochrome c, containing residues 1-80 of the 104 residue long polypeptide chain; then individual B cells specific for the peptide were challenged in vitro in splenic fragment cultures, with either the fragment or intact cytochrome c, both coupled to hemocyanin. The splenic environment in which the B cells were cultured contained hemocyanin-primed T cells, which provided equivalent T cell help for both the peptide and protein immunogens. In two experiments, intact cytochrome c-hemocyanin activated a total of only five peptide-primed B cells, compared to 66 that were activated by the peptide-hemocyanin conjugate. Furthermore, antibodies from the few cells that appeared to be activated by the protein did not bind the native protein with appreciable affinity in competitive ELISA. Of the mAbs elicited by the peptide, 51 were shown to have detectable affinity for native cytochrome c, but their affinity was dramatically less than that previously observed for antibodies elicited by protein-primed B cells and also less than that for the peptide. Thus, although the Ig receptors on many of the peptide-primed B cells did bind the protein to some extent, most such B cells were not activated. These results demonstrate that, in the development of synthetic vaccines, the affinity of a protein for peptide-primed antibodies (Ig receptors) is an important criterion to be considered. Qualitative examination of the binding of an anti-peptide antibody to a protein antigen, especially in denaturing conditions such as Western blots or in potentially denaturing conditions such as ELISA, is not an accurate indication of the efficacy of the peptide to prime B cells that can be activated upon challenge from the native protein.

Differences in heavy chain amino acid sequences affecting the specificity of antibodies for variants of cytochrome c.

In a previous study [Goshorn et al. (1991) J. biol. Chem. 266, 2134-2142], several mAb specific for the same region on different cytochromes c were shown to have similar H or L chains. To determine the effect of differences in individual chains on antigenic variant specificity in the present study, chimeric mAb were prepared by recombining the H and L chains of mAb having the same or a different cytochrome c specificity. The H and L chains of two mAb to the region around residue 60 on horse cytochrome c (1F5.D1 and 2E5.G10) were functionally interchangeable even though the H chain differed by 11 amino acid residues in the complementarity-determining regions (CDR) and 15 amino acids overall in the variable regions. The L chains only differed by four amino acid residues in the CDR (five residues overall). Neither the H nor L chain of a mAb binding the same region of rat cytochrome c (6H2.B4) was functionally interchangeable with the chains of the two horse cytochrome c-specific mAb. The L chain of this mAb is very different from the other L chains which were derived from a different V kappa family, but the H chain is nearly as similar to the horse cytochrome c-specific H chains as they are to each other. Most of the differences occur in CDR3 and result from the use of a distinct DH segment. The results indicate that, in some cases, the specificity of a mAb for a particular variant of a protein Ag, at least in regard to the H chain, is determined by only a few amino acid differences. The differences in the sequences of the H chains of the three mAb in this study and in the structures of their specific Ag provide insight into a possible molecular basis for the specificity of these mAb.

Immunoglobulin gene joints compensate for reduced on-rates imposed by somatic mutations in a V(H) gene.

Affinity maturation in the B lymphocyte response to a protein epitope appears to be largely due to a decrease in the off-rate constant of the antibodies (Ab) resulting from somatic mutation without a significant increase in the on-rate constant. Here, we show by site-directed mutagenesis of a germline encoded single-chain Fv that somatic mutations frequently selected in the Ab response to mouse cytochrome c (CYT) at heavy (H) chain positions 31 and 58 actually cause a two and three-fold decrease, respectively, in the on-rate constant as well as a two and five-fold decrease, respectively, in the off-rate constant and together cause nearly an eight-fold decrease in the off-rate. However, additional selection for a tyrosine residue at position 96 in the V(kappa)-J(kappa) joint compensates for the decreased on-rate imposed by the somatic mutations. This allows for an increase in the affinity of Ab during the secondary response. Certain sequences at the V(H)-D-J(H) joint were also shown to maintain a normal on-rate constant in the context of the common H chain mutations and, in addition, to reduce the off-rate, thus increasing the affinity. The results support the idea that both faster on-rates and slower off-rates for B lymphocyte antigen-specific receptors are favored during the maturation of the Ab response to mouse CYT.

B lymphocyte recognition of the self antigen mouse cytochrome C in different mouse strains: targeting of the same dominant epitope by naturally-occurring cells expressing distinct VH genes.

Previously we reported that, early in the antibody response of BALB/c mice to several cytochromes c (CYT) coupled to ovalbumin (OVA), B cells responding to the self antigen mouse CYT recognized a single site on mouse CYT and were in much higher frequency than B cells responding to foreign CYT. In the present study these B cells were shown by in vitro activation of primary splenocytes to be present in naive BALB/c mice, i.e. prior to exposure to exogenous CYT. The higher frequency of B cells responsive to self versus foreign CYT was also shown in this study to occur in the early antibody response to CYT-OVA in C57BL/6 mice. The same dominant site was recognized in BALB/c mice (IgHa), C57BL/6 mice (IgHb) and the congenic strains BC-17 (IgHa on the C57BL/6 background) and CB-20 (IgHb on the BALB/c background). However, anti-mouse CYT mAbs produced in IgHb mice were shown to derive from the VH gene 5.54.4 while mAbs in IgHa mice derive from the VH gene 19.1.2. The polypeptides encoded by these VH genes, which differ by only five amino acid residues, paired with polypeptides encoded by the same Vk genes (R9 and 2G5). In both VH 19.1.2- and VH 5.54.4-derived mAbs H3 and the Vk-Jk join were variable. The affinity for mouse CYT was reduced in the VH gene 5.54.4-derived mAb due to a faster off-rate constant. This difference in affinity may relate to the lower frequency of B cells responding to mouse CYT in C57BL/6 mice than in BALB/c mice. The results show that naturally-occurring CYT-specific autoreactive B cells occur normally in more than one mouse strain and that self antigen recognition by those cells appears to be atypical involving mostly the immunoglobulin V gene-encoded segments.

A single amino acid substitution in a cytochrome c T cell stimulatory peptide changes the MHC restriction element from one isotype (I-Ak) to another (I-Ek).

The binding sites of class II major histocompatibility complex (MHC) molecules can accommodate many seemingly diverse peptides. In the case of mouse class II molecules, it appears that in general, the I-A and I-E isotypes associate with different peptides. In this study we report an example where a single amino acid substitution in an I-Ak restricted peptide changes the restriction element to I-Ek. A T cell hybridoma, F6.A10, specific for the peptide 93-104 from mouse testicular cytochrome c (Mt cyt 93-104) was found to be restricted by I-Ak using class II molecule specific blocking monoclonal antibodies (mAb). The activation of this hybridoma by Mt cyt 93-104 was competitively inhibited by other peptides that bind to the I-Ak molecule but not by the peptide Mt cyt 93-104(A96) in which lysine at position 96 was substituted by alanine. This single amino acid substitution resulted in the ability of Mt cyt 93-104(A96) to activate the pigeon cytochrome c specific, I-Ek restricted, T cell hybridoma 2B4.11. The activation of 2B4.11 by Mt cyt 93-104(A96) was inhibited by peptides which bind to the I-Ek molecule but not by Mt cyt 93-104 and by mAb specific for I-Ek but not by mAb specific for I-Ak. These results suggest that the amino acid at position 96 may be an important anchor residue for both I-Ak and I-Ek binding but that peptides with different amino acid side chains are accommodated at that position by one or the other MHC class II isotype. Thus, in this particular case a single amino acid residue in the peptides determines the MHC class II isotype specificity.

Major and minor epitopes on the self antigen mouse cytochrome c mapped by site-directed mutagenesis.

Variants of rat (mouse) cytochrome c, prepared by site-directed mutagenesis or represented by closely-related cytochromes c from different species, were employed to map the functional boundaries of a number of mouse monoclonal antibodies (mAb) specific for the major antigenic region on the self antigen (Ag) around residue 62 and the minor antigenic region around residue 44. The recombinant mouse cytochromes c tested were, unlike the tissue-derived Ag, trimethylated at position 72, and included the wild-type which was acetylated at the amino terminus, a variant that was unacetylated at the amino terminus, and variants with the following single amino acid residue replacements: V11I (valine to isoleucine at position 11), Q12M, A15S, A44P, F46Y, D50A, T58I and G89E. Of these, only the A44P variant affected the binding of mAb to the region previously localized to the vicinity of residue 44, thus confirming that assignment. Loss of the acetyl group at the amino terminus affected the binding of most of the mAb to the region around residue 62. The other mutations had little, if any, affect on mAb binding. The epitopes of mAb binding the region around residue 62 were shown in this study to have similar functional boundaries. This site on the self Ag, which encompasses at least three discontinuous segments of the polypeptide chain, is comparable in size to epitopes on other protein Ag that have been mapped by X-ray crystallography and is similar to an epitope in the corresponding region of the foreign Ag, horse cytochrome c, that has been mapped by hydrogen-deuterium exchange. In addition to the mAb binding the regions around residues 44 and 62, a third group of mouse cytochrome c-specific mAb known to be broadly reactive with mammalian cytochromes c and that represents a minor portion of the mAb was tested for binding the site-directed mutants of mouse cytochrome c. None of these mAb were affected by the mutations, indicating the presence of at least one more antigenic region on the self Ag in an area not encompassed by these mutations that is structurally highly conserved.

Kinetic and genetic bases for the heteroclitic recognition of mouse cytochrome c by mouse anti-pigeon cytochrome c monoclonal antibodies.

The B lymphocyte response to pigeon cytochrome c (CYT) in BALB/c mice was previously shown to initiate as a heteroclitic response specific for the self antigen mouse CYT. As the immune response progressed, the mAb that were produced became less heteroclitic and often bound pigeon CYT with higher affinity than they bound mouse CYT [Minnerath, J. et al., 1995. Proc. Natl. Acad. Sci. USA 92, 12379-12383]. To study the basis for heterocliticity and its loss in this system, the H and L chain amino acid sequences of anti-pigeon CYT mAb obtained from the primary and secondary Ab responses were first compared. The most frequent somatic mutations and Ig gene joints were then introduced into an engineered single-chain Fv (scFv) that expressed the germline-encoded V(H) and V(L) amino acid sequences. The effects of those changes on the on- rate, off-rate, and affinity constants in binding both mouse and pigeon CYT were determined by surface plasmon resonance. In this system, the heterocliticity of primary mAb was largely due to a decreased on-rate constant for binding pigeon CYT relative to mouse CYT. Various combinations of the three frequently occurring H chain somatic mutations (H31, H56, and H58) increased the on-rate constant to maximal levels. Only one of the mutations (H58) decreased the off-rate constant when in combination with the other mutations and the effect of H58 was greater for scFv binding mouse CYT than pigeon CYT. Consequently, the mutated scFv and many secondary mAb remained heteroclitic, although their affinities for pigeon CYT increased. Secondary mAb that were no longer heteroclitic expressed non-canonical amino acid sequences in the V(H)-D-J(H) joint in the context of the canonical V genes or expressed different V genes. In addition to providing insight into the molecular basis for heterocliticity, our findings confirm that both faster on-rate and slower off-rate constants are favored during affinity maturation of the Ab response.

Antibody-detected folding: kinetics of surface epitope formation are distinct from other folding phases.

The rate of macromolecular surface formation in yeast iso-2 cytochrome c and its site-specific mutant, N52I iso-2, has been studied using a monoclonal antibody that recognizes a tertiary epitope including K58 and H39. The results indicate that epitope refolding occurs after fast folding but prior to slow folding, in contrast to horse cytochrome c where surface formation occurs early. The antibody-detected (ad) kinetic phase accompanying epitope formation has k(ad) = 0.2 s(-1) and is approximately 40-fold slower than the fastest detectable event in the folding of yeast iso-2 cytochrome c (k2f approximately 8 s(-1)), but occurs prior to the absorbance- and fluorescence-detected slow folding steps (k1a approximately 0.06 s(-1); k1b approximately 0.09 s(-1)). N5I iso-2 cytochrome c exhibits similar kinetic behavior with respect to epitope formation. A detailed dissection of the mechanistic differences between the folding pathways of horse and yeast cytochromes c identifies possible reasons for the slow surface formation in the latter. Our results suggest that non-native ligation involving H33 or H39 during refolding may slow down the formation of the tertiary epitope in iso-2 cytochrome c. This study illustrates that surface formation can be coupled to early events in protein folding. Thus, the rate of macromolecular surface formation is fine tuned by the residues that make up the surface and the interactions they entertain during refolding.

Monoclonal antibodies block the trypsin cleavage site on human placental alkaline phosphatase.

Three of eleven monoclonal antibodies (mAbs) to human placental alkaline phosphatase (PLAP) were shown to block cleavage by trypsin at the only proteolytically sensitive site on the native molecule. These results illustrate the potential importance of using mAbs to restrict proteolysis of proteins, in general, and serve as a novel means to identify the relative locations of antigenic determinants.

Monoclonal antibodies block the bromelain-mediated release of human placental alkaline phosphatase from cultured cancer cells.

Certain monoclonal antibodies (mAbs) to human placental alkaline phosphatase (PLAP) block bromelain cleavage of a 2-kDa segment from each of the two polypeptide chains of PLAP. These mAbs also prevent the release of PLAP from cultured cancer cell surfaces by bromelain. Such proteolysis-blocking mAbs serve as tools to specifically modify the molecular topography of cell surfaces by protease treatment.

Monoclonal antibodies to different epitopes on a cell-surface enzyme, human placental alkaline phosphatase, effect different patterns of labeling with protein A-colloidal gold.

Two monoclonal antibodies (mAbs) to different epitopes on human placental alkaline phosphatase (PLAP), both of the immunoglobulin G2a heavy-chain class and having similar affinities for PLAP, were compared for their ability to label the enzyme on the HeLa cell surface. In one type of experiment employing [125I]-labeled mAbs, the results demonstrated quantitative differences in binding of the mAbs to the cells. At saturating levels, the number of molecules of mAb E5 bound to the cells was almost eight times the number of mAb B10 molecules bound. In another type of experiment, mAbs were indirectly visualized on the cell surface using protein A tagged with colloidal gold particles in transmission electron microscopy. Only one of the antibodies (E5) displayed a clustered distribution of PLAP that previously had been observed with rabbit polyclonal antibodies and goat anti-rabbit IgG-labeled gold (J Histochem Cytochem 33:1227, 1985). The other antibody (B10) showed less frequent and more scattered labeling; three to four times more gold particles were visualized in each cluster on cells bound by mAb E5 compared to cells bound by B10. These results are consistent with the idea that not all epitopes on a membrane-bound antigen may be equally accessible for antibody binding. Even identical epitopes on different PLAP molecules are not equally hindered by other membrane components, since at least some of the PLAP molecules are labeled by the more sterically hindered mAb B10. Quantification of the number of gold particles employing the more abundantly bound mAb E5 provides an average estimate of seven to eight molecules of PLAP in each cluster. Because of inefficiencies in labeling, however, this value is probably lower than the real number.

Clustered distribution of human placental alkaline phosphatase on the surface of both placental and cancer cells. Electron microscopic observations using gold-labeled antibodies.

The cell-surface distribution of human placental alkaline phosphatase (PLAP) on cultured cancer cells, A431 and HeLa TCRC-1, and on normal syncytial cells of placental tissue was examined in immunoelectron transmission microscopy using the gold-labeling technique. Chemically fixed cells were reacted with affinity-purified rabbit polyclonal antibodies to PLAP, and the antibodies were visualized using gold particles tagged with goat antirabbit IgG. On all cells PLAP was observed in clusters distributed throughout the membrane surface, including microvilli, but it was not expressed in desmosomes or along other dense regions on the membrane. Previous histochemical and immunochemical techniques failed to demonstrate clusters. The results show that (1) the gold-labeling technique allows a more precise localization of PLAP on the cell surface than previously employed methods, and (2) the distribution of the enzyme is the same on cultured cancer cells and on normal placental syncytial cells. The clustered distribution of PLAP is thus a general phenomenon and is probably influenced by the physiological function of the enzyme, which has yet to be defined.

Characterization of monoclonal antibodies to an immunodominant protein of the etiologic agent of human granulocytic ehrlichiosis.

Immunodominant proteins in the range of 42-45 kD are important for the serodiagnosis of human granulocytic ehrlichiosis (HGE). Antigens from human isolates of the etiologic agent of HGE cultivated in HL-60 cells were used to immunize BALB/c mice and generate a panel of hybridomas secreting monoclonal antibodies. Using an enzyme immunoassay, an immunofluorescent assay (IFA), and Western blotting, we showed that culture supernatants and ascites of these hybridomas were reactive with human isolates of the etiologic agent of HGE, Ehrlichia equi and E. phagocytophila. Following screening and subcloning, we selected three stable hybridomas, R1B10, R5E4, and R5A9, which were determined to be of the isotypes IgG3, IgG1, and IgG2a, respectively. These results suggest that the epitopes of the 42-45-kD protein recognized by these three monoclonal antibodies are conserved among E. equi, E. phagocytophila, and the etiologic agent of HGE. Western blot analysis showed reactivity with the 44-kD protein of human isolates of the HGE agent. None of the monoclonal antibodies were reactive with HL-60 cells that were not infected with the HGE agent. No cross-reactivity with related intracellular pathogens could be detected when undiluted supernatants from hybridoma cultures were allowed to react by IFA with antigens from E. chaffeensis, E. risticii, E. platys, Rickettsia rickettsii, R. prowazekii, or Coxiella burnetii. The additivity index of two antibodies, R5E4 and R1B10 was near zero, suggesting that these two antibodies may compete for the same epitope of the 44-kD protein, while monoclonal antibody R5A9 appears to interact with a different epitope. The antibodies secreted by these hybridomas may be useful as immunologic agents in serodiagnostic, immunohistochemical, and other studies of the etiologic agent of HGE.