Antibody Reactivity Patterns Research Articles

Drug-induced anti-histone autoantibodies display two patterns of reactivity with substructures of chromatin.

Increasing evidence suggests that autoantibodies in the rheumatic diseases are a consequence of immune selection by self-material, but the nature of the in vivo immunogen is unknown. Insight into this problem may be obtained by measuring autoantibody binding to various forms of a target antigen. Antihistone antibodies arising as a side effect of therapy with various drugs offer an opportunity to explore this premise because many forms of histone have been characterized and adapted to ELISA formats. Two patterns of antibody reactivity were observed. All 21 patients with symptomatic procainamide-induced lupus and 7 of 12 patients with quinidine-induced lupus had IgG antibodies reacting predominantly with the (H2A-H2B)-DNA complex and with chromatin. In contrast, antibodies in 19 of 24 patients taking procainamide without accompanying lupus-like symptoms did not show any pattern. The second pattern was observed in 18/19 chlorpromazine-treated patients and 14/17 patients with hydralazine-induced lupus in which IgM antibodies displayed more reactivity with DNA-free histones than with the corresponding histone-DNA complexes and almost no binding to H1-stripped chromatin. Absorption studies were entirely consistent with these results. Thus, the two patterns of reactivity with nucleosomal components reflect the molecular substructure of chromatin, suggesting that two processes underlie antihistone antibody induction by drugs. In one, IgG autoantibodies appear to be elicited by chromatin, whereas in the other, autoimmune tolerance to native chromatin appears largely intact, and IgM antibodies may be driven by DNA-free histone.

Variability of Penicillin-Binding Proteins from Penicillin-Sensitive Streptococcus pneumoniae

Penicillin-binding proteins (PBPs) from penicillin-susceptible strains of Streptococcus pneumoniae are believed to be fairly similar in contrast to PBPs occurring in resistant isolates. The antigenic variation of PBPs 1a and 2b in 65 penicillin-susceptible strains from different geographic areas and a wide variety of isolation sites was analyzed using a set of specific antisera and monoclonal antibodies. Three strains contained different antigenic variants of PBP 1a, and 50 strains contained one of three antigenic variants found in PBP 2b. Seven patterns of antibody reactivity could be defined; all but one were distinct from those found recently in resistant strains. In addition, electrophoretic mobilities of all six PBPs, compared after conventional SDS-PAGE and fluorography, revealed an unexpected variation of PBP-profiles even for strains of one sero-group. Few strains appeared identical to each other or to the laboratory reference strain R6.

Antigenic Variation of Penicillin-Binding Proteins from Penicillin-Resistant Clinical Strains of Streptococcus pneumoniae

Penicillin-resistant strains of Streptococcus pneumoniae that are isolated with increasing frequency worldwide contain low-affinity penicillin-binding proteins (PBPs). The relatedness of PBPs from 55 resistant strains isolated on three continents was investigated by testing the reactivity of antibodies specific for PBP 1a or 2b and by comparing the PBP patterns. Seventeen patterns of antibody reactivity could be distinguished, 12 of which were specific to one isolate. Most strains, including all German and South African strains, had a unique PBP profile. A few groups of Spanish and Finnish isolates were identified where the strains within each group shared the same PBP profile, the same antigenic variants of PBPs 1a and 2b, and the same serogroup, suggesting that they represent different clones of S. pneumoniae. The results demonstrated highly variable pathways of resistance development and confirmed that resistant strains have emerged independently in different locations.

Antigenic determinants responsible for the reactions of drug‐dependent antibodies with blood cells

In an attempt to determine the antigenic combining sites of drug-dependent antibodies in patients with drug-related immune haemolysis, we have assessed the reactivity of 35 nomifensine-induced antibodies against human red blood cells (RBC) in the presence of 11 closely related compounds: nomifensine, its three main metabolites including their methoxylated analogues and diclofensine with its three main metabolites. Three types of antibody reactivity patterns could be differentiated: Firstly, antibodies most strongly reactive with nomifensine (n = 12); secondly, antibodies primarily directed against one of its main metabolites (n = 7), and thirdly, antibodies optimally reactive with its unknown metabolites (n = 16). The antibodies preferentially directed against nomifensine showed varying cross reactions with nomifensine-related compounds and in almost all cases also with diclofensine and/or its metabolites. Those antibodies which were optimally reactive with metabolites reacted only with nomifensine-derived compounds and only one of them was non-crossreactive. The third group of antibodies showed no (n = 12) or only weak (n = 4) cross reactions with nomifensine and/or its metabolites. Although none of the substances used bound firmly to RBC, certain blood groups have been identified in previous studies to be defined. RBC antigens were required for the reaction in approximately 40% of all antibodies, independent of their reactivity with the compounds. Thus, even when the specificity of some antibodies appeared to be predominantly controlled by certain structural features of the compounds, the actual antigenic combining site of each antibody was different and seemed to comprise parts of the drug-related determinants as well as different constituents on RBC membranes. These findings indicate firstly that RBCs function as 'carrier-like' macromolecules, since they are directly involved in the reaction; secondly, that the drugs and their metabolites act as 'pseudohaptens', in as much as they do not bind tightly to the cells; and, thirdly, that the determinants which govern the immune response appear to result from an accidental attachment rather than from a predetermined selection of antigenic membrane structures, since each antibody shows a unique reaction pattern.

Patterns of antibody reactivity to selected human immunodeficiency virus type 1 (HIV-1) gp160 epitopes in infected individuals grouped according to CD4+ cell levels

We examined sera from 160 HIV-infected individuals for antibodies reactive to HIV-1 gp160 epitopes defined by seven synthetic peptides. Seropositive individuals were placed into three groups based upon levels of circulating CD4+ cells. These groups consisted of individuals with (1) more than 400 CD4+ cells, (2) 200-400 CD4+ cells, and (3) fewer than 200 CD4+ cells/mm3. The percentage of sera containing antibodies reactive with two immunodominant gp160 epitopes (a.a. 304-321 and 600-611) was unchanged between groups, regardless of CD4 cell numbers. The percentage of sera containing antibodies reactive with weakly immunogenic gp160 epitopes, such as those defined by peptides 425-448 and 846-860, declined in the groups as CD4 values decreased. Our results suggest that the patterns of antibody reactivity to gp160 epitopes change as CD4 levels decline. A narrowing of the humoral immune response to epitopes on the envelope of HIV-1 appears to occur with disease progression.

Different roles of individual N-linked oligosaccharide chains in folding, assembly, and transport of the simian virus 5 hemagglutinin-neuraminidase.

The role of N-linked glycosylation in protein maturation and transport has been studied by using the simian virus 5 hemagglutinin-neuraminidase (HN) protein, a model class II integral membrane glycoprotein. The sites of N-linked glycosylation on HN were identified by eliminating each of the potential sites for N-linked glycosylation by oligonucleotide-directed mutagenesis on a cDNA clone. Expression of the mutant HN proteins in eucaryotic cells indicated that four sites are used in the HN glycoprotein for the addition of N-linked oligosaccharide chains. These functional glycosylation sites were systematically eliminated in various combinations from HN to form a panel of mutants in which the roles of individual carbohydrate chains and groups of carbohydrate chains could be analyzed. Alterations in the normal glycosylation pattern resulted in the impairment of HN protein folding and assembly which, in turn, affected the intracellular transport of HN. The severity of the consequences on HN maturation depended on both the number of deleted carbohydrate sites and their position in the HN molecule. Analysis of the reactivity pattern of HN conformation-specific monoclonal antibodies with the mutant HN proteins indicated that one specific carbohydrate chain plays a major role in promoting the correct folding of HN. Another carbohydrate chain, which is not essential for the initial folding of HN was found to play a role in preventing the aggregation of HN oligomers. The HN molecules which were misfolded, owing to their altered glycosylation pattern, were retained in the endoplasmic reticulum. Double-label immunofluorescence experiments indicate that misfolded HN and folded HN are segregated in the same cell. Misfolded HN forms disulfide-linked aggregates and is stably associated with the resident endoplasmic reticulum protein, GRP78-BiP, whereas wild-type HN forms a specific and transient complex with GRP78-BiP during its folding process.

Different Roles of Individual N-Linked Oligosaccharide Chains in Folding, Assembly, and Transport of the Simian Virus 5 Hemagglutinin-Neuraminidase

The role of N-linked glycosylation in protein maturation and transport has been studied by using the simian virus 5 hemagglutinin-neuraminidase (HN) protein, a model class II integral membrane glycoprotein. The sites of N-linked glycosylation on HN were identified by eliminating each of the potential sites for N-linked glycosylation by oligonucleotide-directed mutagenesis on a cDNA clone. Expression of the mutant HN proteins in eucaryotic cells indicated that four sites are used in the HN glycoprotein for the addition of N-linked oligosaccharide chains. These functional glycosylation sites were systematically eliminated in various combinations from HN to form a panel of mutants in which the roles of individual carbohydrate chains and groups of carbohydrate chains could be analyzed. Alterations in the normal glycosylation pattern resulted in the impairment of HN protein folding and assembly which, in turn, affected the intracellular transport of HN. The severity of the consequences on HN maturation depended on both the number of deleted carbohydrate sites and their position in the HN molecule. Analysis of the reactivity pattern of HN conformation-specific monoclonal antibodies with the mutant HN proteins indicated that one specific carbohydrate chain plays a major role in promoting the correct folding of HN. Another carbohydrate chain, which is not essential for the initial folding of HN was found to play a role in preventing the aggregation of HN oligomers. The HN molecules which were misfolded, owing to their altered glycosylation pattern, were retained in the endoplasmic reticulum. Double-label immunofluorescence experiments indicate that misfolded HN and folded HN are segregated in the same cell. Misfolded HN forms disulfide-linked aggregates and is stably associated with the resident endoplasmic reticulum protein, GRP78-BiP, whereas wild-type HN forms a specific and transient complex with GRP78-BiP during its folding process.

Detection of restricted predominant epitopes of Theiler's murine encephalomyelitis virus capsid proteins expressed in the λgt11 system: differential patterns of antibody reactivity among different mouse strains

Intracerebral injection of mice with Theiler's murine encephalomyelitis virus results in chronic demyelination in susceptible strains, and serves as a model system for the study of multiple sclerosis. The role of individual epitopes in the disease process remains to be elucidated. Random fragments of DNA from the viral capsid protein genome covering the coding regions from VP1, VP2, and VP3 have been expressed in the λgt11 vector system. Fusion proteins from the clones were expressed and probed with antibodies from both resistant and susceptible strains of mice. Each strain displays a distinctive pattern with certain fusion proteins recognized by all of the strains and others recognized uniquely by either the susceptible or the resistant strains.

Identification and partial characterization of Trypanosoma cruzi antigens recognized by T cells and immune sera from patients with Chagas' disease

Trypanosoma cruzi antigenic specificities involved in human T-cell and antibody responses were compared in chronic chagasic patients affected with cardiomyopathy (C) or with the indeterminate form (I), the asymptomatic form of chronic Chagas' disease. T-cell Western blotting (immunoblotting) was performed to identify the most active antigens in epimastigote extracts (EPI-Ag). The patterns of peripheral blood mononuclear cell (PBMC) and T-cell proliferative responses induced by fractions blotted to nitrocellulose were heterogeneous, but the computation of their frequency distribution disclosed some important antigen specificities. Molecules ranging from 100 to 150 kilodaltons (kDa) were frequently stimulatory to PBMC from I patients (5 of 8 cases) and were less so when confronted with C patient (1 of 7 cases) lymphocytes. In contrast, both groups of patients actively responded to fractions ranging from 48 to 57 and 28 to 32 kilodaltons (kDa). The Western immunoblotting patterns of antibody reactivity displayed by 17 C and 15 I patients were also similar, yielding outstanding staining in the molecular mass ranges of 70 to 80 and 43 to 57 kDa. The latter antigen complex was recognized by 100% of the 32 chronic Chagas' disease serum specimens tested and closely corresponded to the migratory position recognized by T cells of most patients tested. The identification of the active molecules contained in the 43- to 57-kDa region was sought, with a focus on GP57/51, an antigen with well-established serodiagnostic properties. Immunoblotting analysis of EPI-Ag with a monoclonal antibody to GP57/51 confirmed its presence within the predicted molecular weight region. Highly purified GP51 was then used to demonstrate directly its capacity to promote specific PBMC proliferative responses in vitro. Data computed from a survey with 12 patients have shown a linear correlation (r = 0.93) between PBMC responses to EPI-Ag and to purified GP51, suggesting that the immune response to this particular glycoprotein may be an important component of human immune responses against T. cruzi.

Expression of the human leukocyte adhesion molecule, LAM1. Identity with the TQ1 and Leu-8 differentiation antigens.

The LAM1 molecule is a member of the new family of cellular adhesion/homing molecules that contain a lectin-like domain at their amino-terminal end followed by an epidermal growth factor-like domain and short consensus repeat units like those found in C3/C4 binding proteins. Two mAb that react with the leukocyte adhesion molecule 1 (LAM1) were produced and used to examine the cell-surface expression of LAM1. The anti-LAM1 antibodies were reactive with the majority of blood lymphocytes, NK cells, neutrophils, and monocytes. LAM1 was also expressed by subpopulations of phenotypically immature and mature thymocytes. Blood lymphocytes rapidly modulated LAM1 from the cell surface during PMA exposure for 60 min. Coordinate with the loss of LAM1 from the cell surface, PMA-treated lymphocytes lost the ability to bind to lymph node high endothelial venules, indicating that expression of LAM1 may play a role in lymphocyte homing. Mitogen stimulation of blood T and B lymphocytes also resulted in decreased LAM1 expression, but at a slower rate. LAM1 was only weakly expressed by a minority of spleen lymphocytes. However, culturing spleen lymphocytes in media alone resulted in increased expression of LAM1 by a subpopulation of the cells (40 to 60%). Concomitant mitogen stimulation of spleen lymphocytes resulted initially in down-regulation of LAM1 expression followed by increased expression of LAM1 and then subsequent loss of LAM1 from the cell surface. The pattern of anti-LAM1 antibody reactivity was identical to that reported for the TQ1 and Leu-8 antibodies, and all of these antibodies reacted with cells transfected with the LAM1 cDNA. Thus, LAM1 is broadly expressed by leukocytes, and binding of LAM1 may participate in the process of leukocyte extravasation into lymphoid organs or sites of acute inflammation with subsequent loss of LAM1 from the cell surface.

Molecular machanics and antibody binding in the structural analysis of polycyclic aromatic hydrocarbon-diol-epoxide-DNA adducts

Analysis of polycyclic aromatic hydrocarbon (PAH)-DNA adducts using monoclonal antibodies raised against DNA that had been modified with (+-)-r-7-,t-8-dihydroxy-t-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene in an enzyme-linked immunosorbent assay, as well as analysis using human serum antibodies and antibodies raised in laboratory animals, have suggested the presence on these adducts of both common and unique immunological epitopes. The molecular mechanics studies reported here establish a model for the analysis of PAH-DNA adducts through the identification of energetically favored binding conformations and they further reveal structural alterations in DNA due to the presence of carcinogen adducts. The data explain the antibody reactivity patterns by defining different molecular presenting surfaces that are available for antibody binding. The preferred orientation of the aromatic portions of the adducts, which align either 3' or 5' in the minor groove, were found to be correlated with antibody reactivity patterns. Examination of the topographical characteristics of the adducts facilitated correlation of adduct-antibody recognition and adduct presenting surface. Significant differences were found between benzo[a]pyrene-diol-epoxide (BPDE)-DNA adducts, which align 5' in the minor groove, and benz[a]anthracene-diol-epoxide (BADE)-DNA and dibenz[a,c]anthracene-diol-epoxide-DNA adducts, which align 3' within the minor groove. Chrysene-diol-epoxide-DNA adducts were found to have only a weak preference for 5' alignment and therefore share topographical characteristics with both BPDE-DNA and BADE-DNA adducts.

Immunoblot Analysis of the Antibody Response to Murine Cytomegalovirus in Genetically Resistant and Susceptible Mice

The murine of human cytomegalovirus infection was employed to analyse the kinetics of antibody production to murine cytomegalovirus (MCMV) structural and immediate early (IE) polypeptides following MCMV infection of genetically resistant and susceptible strains of mice. A total of 22 structural and six non-structural. IE proteins were identified. Analysis of immunoblots by densitometry identified four patterns of antibody reactivity to MCMV structural polypeptides during primary and secondary antibody responses over a period of 5 weeks post-infection (p.i.). Firstly, antibodies were strongly reactive with an 83K protein soon after infection, with levels which decreased with time: antibodies to a second group of viral proteins were also recognized soon after infection, but consistent levels of reactivity were maintained. Viral proteins that were recognized beyond day 14 p.i. or following a second MCMV infection formed the third group; the fourth consisted of viral proteins that were detected at variable times p.i. by antisera different mouse strains. The kinetics and intensity of the antibody response to individual viral proteins were influenced by virus dose, time p.i. and by a second MCMV infection. In addition, the genetic constitution of the host influenced the antibody response to MCMV proteins both quantitatively and qualitatively. In particular, sera from mice possessing C57BL but not BALB/c genes detected a 56K Mr viral during seroconversion. Antibody reactivity to this protein was shown to segregate among sera from CXB mice, with a strain distribution pattern which indicated a linkage with the b locus on chromosome 4. Finally, expression of the 56K protein was detected in B10. BR but not BALB/c embryo fibroblasts in vitro, with expression being a dominant trait in (BALB/c x B10.BR) F1 embryo fibroblasts. Thus, host genes may influence the expression of this structural viral protein.

Immunoglobulin-class-specific immune response to respiratory syncytial virus structural proteins in infants, children, and adults.

The protein specificities of IgG, IgM, and IgA antibodies induced during respiratory syncytial virus (RSV) infection in 74 patients (4 weeks to 81 years of age) were investigated using the technique of immunoblotting. Although the pattern of antibody reactivity varied among patients, most of the humoral immune response in all age groups was directed against the 48, 42, 35, and 27 K proteins. An infant's own antibody response was discernible in 55 of the 57 children below 1 year of age, despite the presence of maternally derived antibodies. Antibody against the 90 K surface glycoprotein was not detectable in those less than 1 year of age. Primary RSV infection induced antibodies only against a subset of RSV proteins. Although a broadening of the antibody response occurred with increasing age and in the course of reinfection, an immune response to all the viral structural proteins was observed rarely.

Patterns of serum IgM antibodies to GM1 and GD1a gangliosides in amyotrophic lateral sclerosis.

We studied the incidence and clinical correlates of serum antibodies to GM1 and GD1a gangliosides in patients with classical amyotrophic lateral sclerosis (ALS) and other "motor nerve" syndromes. Serum antibodies to GM1 and GD1a gangliosides were measured using enzyme-linked immunosorbent assays. Our results showed that polyclonal immunoglobulin M (IgM) antibodies to the GM1 or GD1a ganglioside or both were present at serum dilutions of 1:25 to 1:4,000 in 78% (57/73) of patients with ALS. Only 8% of normal controls had similar antibodies. The pattern of serum antibody reactivity correlated with the pattern of clinical involvement in our patients. Selective reactivity to GD1a ganglioside was common when upper motor neuron signs were prominent. IgM reactivity to GM1 ganglioside was common in ALS patients with prominent lower motor neuron signs. Most patients with motor neuropathies had serum reactivity to both GM1 and GD1a gangliosides. These results provide further evidence of ongoing autoimmune processes in ALS patients. There is a strong relationship between serum antiganglioside antibodies and patterns of clinical involvement in ALS.

Variability of Western blot patterns from sera of hemophiliacs determined with different human immunodeficiency virus antigens

Qualitative as well as quantitative changes in HIV-specific antibody titers may be of prognostic value in individuals at risk for developing AIDS. The Western blot (WB) method was used to monitor sequential serum samples from 12 patients with hemophilia in order to determine if any changes occurred in patterns of antibody reactivity. When sera were initially screened for HIV and antibody by ELISA, six patients were negative, five were positive, and one patient was variably negative to weakly positive. ELISA-positive sera produced typical WB band patterns, although these differed for a given sample when tested against a panel of five HIV antigens. Discrepancies were noted for p18, p55, and gp120 160 viral proteins. ELISA-negative samples were nonreactive by WB except for reactivity of some samples with p24 when one of the antigens was used. An equivocal WB pattern, i.e. distinct bands at p24 and an unidentified protein, was obtained from the patient with variable ELISA results. WB analysis of serum is a potentially useful tool for monitoring disease progression, however minor changes in methodology can lead to significant differences in the final result.

Subtyping of Legionella pneumophila serogroup 1 isolates by monoclonal antibody and plasmid techniques.

A group of environmental and clinical Legionella pneumophila serogroup 1 isolates was subtyped by monoclonal antibody dot immunoblotting and plasmid analysis. Monoclonal antibody analysis defined seven subtypes within three major groups. Plasmid analysis (including restriction endonuclease digestion) revealed 10 subtypes. By combining plasmid and monoclonal techniques, all 16 strains were shown to be distinct. Plasmid profiles and monoclonal antibody reactivities of selected strains were stable despite serial passage (greater than 100 times). No plasmid-associated antigen was defined by this panel of monoclonal antibodies. The observed dissociation of plasmid profiles and monoclonal antibody reactivity patterns suggests that accurate epidemiologic typing of L. pneumophila serogroup 1 strains will require use of both techniques.

Cellular and humoral immune responses to synthetic peptides deduced from the amino‐acid sequences of Epstein‐Barr virus‐encoded proteins in EBV‐transformed cells

Ten synthetic peptides containing 18-22 residues deduced from the amino-acid sequences of the EBV-encoded latent-infection-associated membrane protein (LMP) and the 2 principal nuclear antigens, EBNA-1 and EBNA-2, were tested for their ability to induce lymphokine release from sensitized T-cells of EBV-seropositive donors, as measured by the leukocyte migration inhibition assay (LMI). Only one of the 10 free peptides induced EBV-specific LMI. After Sepharose-coupling, 4 additional peptides were regularly active. In parallel, the sera of the same and other donors were screened for synthetic peptide-binding antibodies, as measured by an ELISA assay. Antibodies to 9 of the 10 peptides were detected in 25-80% of EBV-antibody-positive, but not in EBV-antibody-negative sera. A comparison of the two responses indicates that the humoral immune system tends to react with more epitopes on a given protein than the cellular immune system. Furthermore, the antibody reactivity pattern to different epitopes is more variable from individual to individual than the T-cell response. Also, the epitopes detected by antibodies and sensitized T-cells are often not identical.

The humoral immune response of sheep to antigens from larvae of the sheep blowfly ( Lucilia cuprina)

The sheep blowfly, Lucilia cuprina, is a myiasis-causing insect whose larvae evoke an immune response in sheep. By means of an immuno-dot blot and Western immuno-blot assays it has been demonstrated that sheep experimentally infected with larvae produce antibodies against a wide array of components from all three larval instars, with each instar displaying a differing set of antigens. The electrophoretic profiles of the proteins in various larval extracts and the patterns of antibody reactivity were very different. Of the extracts tested (1st, 2nd and 3rd instar larval excretions/secretions and visceral homogenates, extracts of 3rd instar salivary glands, mid guts, haemolymph and cuticle) the most intense antibody reaction was detected against the salivary gland extract: preparations of larval excretions/secretions and from the larval mid gut also reacted strongly. In contrast a cuticle extract reacted minimally with infected sheep sera.

Epizootiology of Hantavirus infections in Baltimore: isolation of a virus from Norway rats, and characteristics of infected rat populations.

Rats trapped from 14 locations in Baltimore, Maryland, were shown to have antibody to Hantavirus. Antibody prevalence rates were higher in residential locations than in parks. Infected rats were obtained continuously over a six-year period from 1980-1985 at five locations, indicating the enzootic nature of this infection in urban rats in the United States. Prevalence of antibody and geometric mean titers increased with rat mass and sexual maturity, suggesting that infections in rats are acquired through age-related mechanisms. Three isolates of Hantavirus were obtained from rats. One of these isolates was shown by serologic tests and monoclonal antibody reactivity patterns to be similar, if not identical, to rat virus previously obtained from Philadelphia. The presence of rats in many urban settings and the high prevalence of Hantavirus infection in this species suggest that human exposure to this virus is occurring in the United States, although human disease from this exposure has not been recognized.

Pontiac fever outbreak associated with a cooling tower.

In late April 1984, an outbreak of Pontiac fever was investigated in an office building in lower Manhattan (New York City). The outbreak was characterized by a high attack rate (78 per cent overall); the predominant symptoms were myalgias, chills, fatigue, fever, and headache. There was a clustering of cases in an office that was air cooled by a dedicated cooling tower separate from the remainder of the building. A high concentration of live L. Pneumophila cells in the cooling tower was quantified. Airborne spread via settle plates placed along the air intake system and within the office was demonstrated. Legionella pneumophila serogroup 1 antigen was found in the urine of two cases, and identical monoclonal antibody reactivity patterns of isolates from all sources was observed. Difficulty was experienced in eliminating the organism from the tower.