Detection of autoantibodies in a quantitative immunoassay using recombinant ribonucleoprotein antigens.

A human cDNA expression library was screened with anti-ribonucleoprotein (RNP) antibodies from patients with connective tissue diseases. Three cDNA clones were isolated encoding 70 kD, A and B" ribonucleoprotein autoantigens which were expressed as beta-galactosidase fusion proteins. Antigens were purified and used to develop sensitive ELISAs suitable for the routine screening of large series of sera from patients with connective tissue diseases. More than 400 sera were tested both by ELISA and by immunoblotting. The ELISA was found to be at least as sensitive as immunoblotting and very specific. Anti-70 kD antibodies were found in 94% of patients with mixed connective tissue disease (MCTD), in 4% of patients with other connective tissue diseases but not in normal controls. Furthermore, the use of recombinant 70 kD antigen enabled us to discriminate between anti-70 kD antibodies present in anti-Sm and in anti-(U1) RNP sera. Recombinant A antigen contained at least two autoantibody-reactive sites; one unique for the A protein and another cross-reactive with anti-B" antibodies. Antibodies reactive with the unique site were found in 83% of MCTD patients, in 4% of patients with other connective tissue diseases and not in normal controls. Antibodies against the cross-reactive B" epitope present on A and B" recombinant antigens, were found in high titres in a small percentage of patients with systemic lupus erythematosus (SLE, 5%) and rheumatoid arthritis (RA, 2%).     

Antibodies against distinct nuclear matrix proteins are characteristic for mixed connective tissue disease.

Specific nuclear proteins, separated according to their molecular weight (mol. wt) by polyacrylamide gel electrophoresis (PAGE) and subsequently transferred to nitrocellulose sheets, are able to bind antibodies in sera from patients suffering from different types of connective tissue diseases. Antibodies against a characteristic set of nuclear protein antigens are found in sera from patients with mixed connective tissue disease (MCTD). Screening of 21 MCTD sera revealed a typical immunoblot pattern with major protein antigens of mol. wt 70,000 (20/21) (not identical with the Scl-70 antigen characteristic for scleroderma), mol. wt 31,000 (17/21), two proteins around mol. wt 23,000 (15/21) and two around mol. wt 19,000 (10/21). The 70,000, 23,000 and 19,000 antigens appeared to be rather insoluble nuclear proteins (i.e. components of the nuclear matrix). On behalf of their structural character they were present in nuclei from several types of cells but only in low amounts detectable in salt extracts of thymus acetone powder. The presence of antibodies directed against the mol. wt 70,000 antigen correlated strongly with the diagnosis of MCTD. This 70,000 antigen is not identical with the RNP antigen, a soluble ribonuclease sensitive ribonucleoprotein, since antibodies against nuclear RNP can be separated from anti-nuclear matrix antibodies by affinity chromatography using immobilized thymus salt extract. The distinct character of soluble nuclear RNP and structural nuclear matrix antigens is further supported by the fact that from 14 other anti-RNP sera obtained from patients with systemic lupus erythematosus (SLE), only three contained antibodies against the mol. wt 70,000 protein. Since the immunoblot pattern obtained with MCTD sera mostly was clearly distinguishable from the patterns obtained with sera from patients with related connective tissue diseases our results suggest that the immunoblotting technique might be useful as a diagnostic tool and support the concept of MCTD as a distinct entity.     

Recombinant 70-kD protein used for determination of autoantigenic epitopes recognized by anti-RNP sera.

A panel of high- and low-titre anti-RNP-positive patient sera was tested for reactivity with human snRNP proteins. The U1 snRNP-specific 70-kD peptide was found to be the most prominent RNP antigen recognized by high-titre anti-RNP sera, mainly found in patients with mixed connective tissue disease (MCTD). The reactivity with the 70-kD protein was further analysed with recombinant fusion proteins containing different segments of the protein. One major and two minor antigenic regions were found. Most patient sera, both with high- and low-titre anti-RNP antibodies, only recognized the major region. The most aminoterminal region, showing partial sequence homology with a mouse retroviral p30 gag protein, contained an epitope that was recognized by one serum only.     

T cell lines recognizing the 70-kD protein of U1 small nuclear ribonucleoprotein (U1snRNP).

In sera of patients with mixed connective tissue disease (MCTD) high titres of IgG autoantibodies to U1snRNP-specific proteins (70 kD, A, C) are found, suggesting an antigen-driven and T-cell-dependent process. In order to establish U1snRNP-specific T cell lines we cultured under various culture conditions mononuclear cells from MCTD patients and healthy donors with a highly purified UsnRNP preparation from HeLa cells. Nine T cell lines were established by limiting dilution cloning from two MCTD patients and five T cell lines from a healthy individual. All T cell lines expressed the TCR alpha beta/CD3 complex. Surprisingly, most of the T cells lines exhibited the CD8 phenotype. Irrespective of this phenotype, all T cell lines showed a proliferative response to an N-terminal part (aa 51-195) of recombinant U1-specific 70-kD protein. One CD8+ T cell clone exhibited cytotoxic activity against an autologous B cell line pulsed with snRNP or recombinant fragments (aa 51-95 and aa 51-88). Interestingly, two T cell lines proliferated in response to four recombinant polypeptides representing different parts of the U1snRNP 70-kD protein. Since regions of sequence homology are distributed over the 70-kD molecule, it is suggested that conserved motifs may be recognized by the T cell lines.     

Expression of autoantibodies to recombinant (U1) RNP-associated 70K antigen in systemic lupus erythematosus

To determine the specificity of antibodies to the (U1) ribonucleoprotein antigen in systemic lupus erythematosus (SLE), patient sera were tested for binding to a recombinant human 70K antigen. By solid-phase immunoassay, we detected anti-70K reactivity in sera from 31 of 96 patients with systemic lupus erythematosus (SLE), demonstrating that anti-70K antibodies may occur in patients with SLE as well as other clinical diagnoses. In sequential sera from 2 of these patients, we found that anti-70K binding varied dramatically over the course of disease. The changes in anti-70K antibody levels did not correlate with clinical events nor evolving antibody reactivity with the Sm-specific antigens.     

Serum antibody response to the 70,000-molecular-weight neisserial common antigen in humans infected by Neisseria gonorrhoeae.

We evaluated the presence of antibodies directed against a 70,000-molecular-weight (70K) common neisserial antigen in sera from patients with first or repeated gonococcal infections and in sera from healthy controls. Sera were taken as soon as possible after the onset of disease, and anti-70K antibodies were detected by Western blot (immunoblot). Results show that significantly fewer patients with gonococcal infection possessed anti-70K antibodies than controls (P less than 0.001). This suggests a possible role of anti-70K antibodies in natural immunity against Neisseria gonorrhoeae.     

Distribution and antigen specificity of anti-U1RNP antibodies in patients with systemic sclerosis.

Systemic sclerosis (SSc) is a generalized connective tissue disease which is characterized by the presence of several autoantibodies. To determine the prevalence and antigen specificity of anti-U1RNP antibodies (anti-U1RNP) in patients with SSc, serum samples from 223 patients with SSc, 117 patients with systemic lupus erythematosus (SLE), 18 patients with mixed connective tissue disease (MCTD) and 40 healthy control subjects were examined by indirect immunofluorescent analysis (IIF), double immunodiffusion, and immunoblotting using nuclear extract of HeLa cells. Eighteen of the 223 (8%) serum samples from patients with SSc were shown to be positive for anti-U1RNP. The frequency of anti-U1RNP positivity in limited cutaneous SSc (14%) was significantly higher than that in those with diffuse cutaneous SSc (3%). Anti-Sm antibodies were detected in patients with SLE positive for anti-U1RNP, but not in those with SSc positive for anti-U1RNP or those with MCTD. Immunoblotting demonstrated that anti-70-kD antibodies were detected more often in patients with SSc positive for anti-U1RNP and in those with MCTD than in those with SLE. Furthermore, anti-U1RNP was closely correlated with pulmonary fibrosis and joint involvement in patients with SSc. These results suggest that anti-70-kD antibodies are useful in the classification of patients with anti-U1RNP.     

Comparison of various preparations of nuclear antigens by hemagglutination inhibition (HAI)

A clinical laboratory carrying out tests for antinuclear antibodies requires an efficient, reliable preparation method to produce a high yield of nuclear antigens at low cost and a very sensitive, specific assay method for antigen activity. Various tissues were employed for preparation of small nuclear ribonucleoprotein (snRNP) and Sm antigens for these purposes. Fresh calf thymus cells and nuclei, commercially available calf and rabbit thymus acetone powders, fresh rat kidney and liver cells were used as sources of antigens prepared similarly by methods published previously. Preparations of antigens from whole calf thymus cell extracts were prepared with and without inhibitors to protease and RNase. snRNP and Sm antigens were assayed at each preparation step by hemagglutination inhibition (HAI). Using HAI it was possible to routinely assay snRNP and Sm at nanogram/ml quantities which was 10(6) fold more sensitive than Ouchterlony immunodiffusion. Results were expressed as relative specific activity as compared with calf thymus nuclear extract prepared by conventional methods. Protease and RNase inhibitors did not significantly increase yields. Thymus was the best source of snRNP and Sm. Fresh calf thymus extract produced a good, stable, reliable quantity of antigens, whereas calf and rabbit thymus acetone powders provided antigen at higher specific activity with less labor but slightly lower yields. Thus, considering the total cost of preparations, commercial sources may be superior to fresh sources in the clinical laboratory setting. These studies also revealed the utility of the sensitive HAI test not only in the clinical laboratory but also for further research endeavors.     

Characterization of B cell epitopes on the 16K antigen of Mycobacterium tuberculosis.

To characterize the antigenic parts of the 16K protein of Mycobacterium tuberculosis, overlapping peptides according to the amino acid sequence of the 16K protein were synthesized. In total, 14 peptides of 20 amino acids in length with an overlap of 10 amino acids and two additional decapeptides (amino acids 31-40 and 61-70) were tested with eight anti-16K MoAbs and human sera. The common recognition site of MoAbs F67-8 and F67-16 was LRPTFDTRLM (amino acids 31-40) and of MoAbs F159-1 and F159-11 DPDKDVDIMV (amino acids 61-70). However, for binding of the MoAbs to these peptides additional amino acids were required at either the N- or C-terminus, suggesting that some kind of conformation is required. The recognition sites of the MoAbs F23-41, F23-49, F24-2 and TB68 could not be identified using the peptides, indicating that the MoAbs only bound to conformational epitopes and not to peptides which may contain parts of these epitopes. The MoAbs bound to beta-galactosidase fusion proteins comprising parts of the 16K protein, indicating that some kind of native conformation is present on the recombinant proteins. Sera from 14 of 19 patients with tuberculosis and none from 19 controls reacted with the purified 16K protein. Sera from four of these 14 patients reacted with two overlapping peptides (amino acids 71-100). Apparently, antibodies in patients' sera against the 16K protein are predominantly directed against conformational epitopes.     

Antigenic domains on the U1 small nuclear ribonucleoprotein-associated 70K polypeptide: a comparison of regions selectively recognized by human and mouse autoantibodies and by monoclonal antibodies.

Antigenic regions on the U1 small nuclear ribonucleoprotein (snRNP)-associated 70K polypeptide recognized by human and mouse autoantibodies or by monoclonal antibodies were identified and compared. Using a set of 70K fusion proteins as antigen in enzyme-linked immunosorbent assay and immunoblotting revealed that serum autoantibodies of human and of MRL/Mp mouse origin recognized a common region of the 70K polypeptide. Monoclonal anti-70K antibodies derived from a patient with mixed connective tissue disease, from an autoimmune MRL/Mp mouse, and from a BALB/c mouse immunized with purified U1 snRNP were all shown to bind to a part of the 70K polypeptide rich in charged residues and different from the region recognized by most human and MRL/Mp mouse serum autoantibodies.     

Further characterization and subcellular localization of Sm and U1 ribonucleoprotein antigens

Sera from patients with systemic autoimmune diseases often contain antibodies against small nuclear ribonucleoprotein (snRNP) particles. Anti-Sm antibodies react with the entire set of U1, U2, U4, U5 and U6 (U1-U6) RNP particles whereas anti-(U1)RNP sera specifically recognize particles containing U1 RNA. Here we performed semi-quantitative immunoblotting using 16 human anti-Sm, 15 human anti-(U1)RNP sera and two mouse monoclonal antibodies to establish which snRNA-associated proteins carry antigenic determinants. Almost every (15/16) human anti-Sm sera recognized epitopes present on a 28-kDa (B/B') protein doublet and on a 16-kDa (D) polypeptide. Nine anti-(U1)RNP sera also recognized the B/B' doublet, but in all cases a much stronger reaction was observed with one or more of the specifically U1 RNA-associated 70 kDa, A or C antigens. With affinity-purified antibody fractions eluted from individual antigen bands on nitrocellulose blots it is shown that the anti-Sm-reactive polypeptides B/B' and D contain common epitopes. We also report the finding of one human anti-Sm serum with exclusive specificity for the B/B' doublet and a mouse monoclonal anti-Sm antibody recognizing only the D protein, indicating that these antigens also carry unique epitopes. In immunoprecipitation assays, purified anti-B/B' and -D antibodies react with (U1-U6) RNP while purified anti-70 kDa, anti-A and anti-C antibodies precipitate exclusively U1 RNP particles. Finally, we established the subcellular localization of Sm and U1 RNP antigens using a biochemical cell fractionation procedure. Part of the 70 kDa and B/B' antigens were found in a nuclease and high salt-resistant nuclear substructure, usually referred to as nuclear matrix, while the A and D antigens could be extracted completely from HeLa nuclei by ribonuclease treatment and subsequent high salt extraction.     

Characterization of recombinant human autoantibody fragments directed toward the autoantigenic U1-70K protein

The U1-70K protein is specifically bound to stemloop I of the U1 small nuclear RNA contained in the U1 small nuclear ribonucleoprotein complex (U1 snRNP), which is involved in the splicing of pre-mRNA. All components of the U1 snRNP complex, including the U1-70K protein, are important autoantigens in patients with systemic lupus erythematosus (SLE) and mixed connective tissue disease (MCTD). Here we describe for the first time the selection and characterization of recombinant human anti-U1-70K single chain autoantibody fragments (anti-hU1-70K scFv) from autoimmune patient-derived phage display antibody libraries. All scFv specifically recognize parts of the hU1-70K protein and its apoptotic 40-kDa cleavage product. In Western blotting assays a number of scFv preferentially recognize the 40-kDa apoptotic cleavage fragment of the U1-70K protein, suggesting a possible involvement of this apoptotic cleavage product in the autoimmune response of patients. The germline gene usage of these recombinant autoantibodies was also determined. Using several U1-70K deletion and point mutants of both human (h) and Drosophila melanogaster (Dm) origin, it was established that the U1-70K epitope that is recognized by the anti-hU1-70K scFv is located within the RNA binding domain.     

Competitive enzyme immunoassay for antibodies to a 43,000-molecular-weight Francisella tularensis outer membrane protein for the diagnosis of tularemia.

Antibodies against a 43,000-molecular-weight Francisella tularensis outer membrane (OM) protein (43K protein) were measured in paired serum specimens from 23 patients with tularemia and matched against antibodies in sera from 25 patients with nontularemic infectious diseases and from 25 blood donors. Antibodies were measured by a competition enzyme-linked immunosorbent assay which tested the ability of human serum to compete with rabbit anti-43K protein antibodies for its binding to the F. tularensis OM coat. The sera from nontularemic patients and from blood donors, in dilutions of 1:16 and 1:64, respectively, showed no or very low levels of antibodies. All of the tularemia patients showed positive tests with the first, the second, or both of the serum specimens examined. For instance, with serum diluted 1:64, each of the serum specimens showed a sensitivity of 95.7% and a specificity of 96%. When used for antibody competition in Western blotting (immunoblotting), the rabbit anti-43K selectivity blocked the binding of human serum antibodies to the 43,000-molecular-weight protein. This protein was immunoaccessible in Formalin-killed F. tularensis. These data indicate an important role of the 43,000-molecular-weight OM protein in the immunobiology of tularemia and emphasize its potential usefulness as an antigen in serodiagnostic tests.     

Clinical significance of anti-RNP and anti-Sm autoantibodies as determined by immunoblotting and immunoprecipitation in sera from patients with connective tissue diseases.

Antibodies to Sm and RNP antigens have been detected by immunoblotting and immunoprecipitation of small nuclear ribonucleoproteins in 168 sera from patients with connective tissue diseases previously characterized by immunodiffusion. Anti-RNP and anti-Sm antibodies immunoprecipitated U1 and U1-U6 snRNA respectively. By immunoblotting anti-Sm reacted with B-B' and D polypeptides and we have distinguished two types of anti-RNP sera: 1) 'full spectrum' anti-RNP sera reacted with the 68 kD, A, C and B-B' polypeptides; 2) 'partially reactive' anti-RNP sera reacted with various combinations of these polypeptides but not the four of them. A strong specificity of anti-Sm antibodies for systemic lupus erythematosus (SLE) was found with all three methods but immunoblotting was more sensitive and detected anti-Sm in 76% of SLE sera. Sera containing a high titer of 'full spectrum' anti-RNP without anti-Sm activity were only detected in mixed connective tissue disease (MCTD) whereas anti-68 kD antibodies alone seemed to be less specific. This strong association between 'full spectrum' anti-RNP antibodies and MCTD supports the hypothesis that MCTD is a distinct clinical entity associated with a specific serologic marker.     

EB病毒潜伏膜蛋白2A抗原表位的融合表达及其意义

为了探讨含有EB病毒潜伏膜蛋白2A(LMP2A)抗原表位片断表达的融合蛋白在鼻咽癌血清学检测中的应用意义,通过重叠延伸PCR方法,合成了3对相互重叠的寡核苷酸引物,涵盖LMP2A的主要的4个抗原表位,将它们拼接在一起构建一个多肽融合基因,克隆到PGEX-4T-2载体中表达融合蛋白,以GST亲和层析柱法纯化融合蛋白,鉴定后以此为抗原检测鼻咽癌患者的血清。结果表明,获得了含EB病毒LMP2A主要的4个抗原表位的融合蛋白(EC2A),蛋白纯度达90%以上,ELISA结果显示鼻咽癌患者血清的检出率为77.9%,正常人群血清为阴性,与常规的VCA-IgA法进行比较,有9份(13.3%)血清VCA-IgA为阴性而EC2A-IgG检出阳性,为鼻咽癌的临床检测提供了新思路,也为后续的单克隆抗体制备奠定了基础。     

Antibodies to small ribonucleoprotein and to 73-kD heat shock protein: two distinct markers of mixed connective tissue disease.

We set out to discover whether antibodies to small ribonucleoprotein antigens (RNP) and to 73-kD heat shock protein (hsp 73) which have been proposed as markers of mixed connective tissue disease (MCTD) recognize different epitopes. MCTD serum was immunoadsorbed on hsp 73-coupled Sepharose and the affinity retained, and non-retained fractions were checked by immunoblotting for recognition of either purified bovine hsp 73 or calf thymus extract RNPs. The hsp 73 affinity-bound serum fraction recognized hsp 73 but not RNP antigens, the reverse being true for the non-retained fraction. We conclude that anti-hsp 73 and anti-RNPs are distinct markers of MCTD.     

Characterization of an HLA-DR4-restricted T cell clone recognizing a protein moiety of small nuclear ribonucleoproteins (UsnRNP).

In sera of patients with mixed connective tissue disease (MCTD, Sharp Syndrome) high titres of IgG autoantibodies to U1snRNP-specific proteins are found. The isolated occurrence of these autoantibodies is highly associated with the HLA-DR4 haplotype. snRNP-specific T cells are supposed to be involved in this autoantibody production. To address this question we cultured mononuclear cells from MCTD patients and healthy donors with a highly purified UsnRNP preparation from HeLa cells using bulk or limiting dilution cultures. Secondary responses to snRNP were detected only rarely with T cell lines from two patients and two controls, and turned out to be unstable during further expansion. One T cell line derived from a healthy individual retained its snRNP reactivity upon limiting dilution cloning and could be characterized in detail. The CD4+ T cell clone recognized native snRNP particles presented by monocytes in an HLA-DR4 (B1*0401)-restricted manner. Separation of the protein and RNA moieties of snRNP particles revealed that the T cell clone responded specifically to the protein fraction, but not to RNA and diverse control antigens. Sequencing of the T cell receptor alpha and beta chain cDNAs revealed that the clone used the V alpha 14.2 and V beta 14 elements. Upon antigen-specific and mitogenic stimulation the T cell clone showed a Th1-specific cytokine pattern, and did not provide helper activity for in vitro immunoglobulin production. This study demonstrate the presence of self-reactive snRNP-specific T cells in a healthy donor. The T cell clone may not represent a helper T cell for the formation of U1snRNP-specific autoantibodies.     

Comparison of human sera reactivities in immunoblots with recombinant human herpesvirus (HHV)-8 proteins associated with the latent (ORF73) and lytic (ORFs 65, K8.1A, and K8.1B) replicative cycles and in immunofluorescence assays with HHV-8-infected BCBL-1 cells

The development of reliable, sensitive, and specific serological methods for the detection of human herpesvirus-8 (HHV-8) antibodies is critical for a thorough understanding of HHV-8 prevalence and pathogenesis. To evaluate the potential usefulness of HHV-8 proteins in measuring the responses against both latent and lytic antigens, we selected 1 latent [open reading frame (ORF) 73] antigen and 3 HHV-8 lytic antigens (ORFs 65, K8.1A, and K8.1B) previously identified as immunogenic [Virology (1998) 243, 208-217]. Full-length genomic ORF 73 and full-length ORFs 65, K8.1A, and K8.1B from the cDNA clones were cloned, expressed in bacterial and baculovirus-insect cell expression systems, and purified as GST fusion proteins. These recombinant proteins were used in Western blot reactions to test sera from 104 human immunodeficiency virus (HIV)+/Kaposi's sarcoma (KS)+ homosexual men, 77 HIV+/KS- homosexual men, and 84 age-matched HIV-/KS- men. These sera were also tested in immunofluorescence assays (IFAs) with uninduced and 12-O-tetradecanoylphorbol-13-acetate-induced B cell lymphoma-1 cells to detect antibodies against latency-associated nuclear antigens (LANA) and antibodies against lytic antigens (cytoplasmic fluorescence). These sera exhibited differential reactivities reflecting different titers of antibodies against HHV-8 proteins, and variable reactivities were seen more commonly with the sera from HIV-/KS- adult men. In the Western blot assay, 89% (93 of 104) of HIV+/KS + sera, 60% (46 of 77) of HIV+/KS- sera, and 7% (6 of 84) HIV+/KS- sera were reactive with both latent and lytic recombinant antigens. Western blot reactions with ORF 73 protein were more sensitive than LANA-IFA results. The lytic IFA and lytic Western blot (ORFs 65 and K8.1A) assays were more sensitive than the ORF 73 Western blots and LANA-IFA. With an exception of 2 sera from the HIV-/KS- group, all sera positive for lytic IFA antibodies and ORF 65 and K8.1A antibodies were also positive for latent antibodies. With few exceptions, sera positive for ORF 65 antibodies were also positive for K8.1A antibodies, and sera recognized the K8.1A protein more often than the K8.1B protein. There is a high degree of concordance between IFA and Western blot reactions, suggesting that this panel of HHV-8 recombinant proteins could detect a majority of the HHV-8-seropositive individuals. These results suggest that IFA followed by confirmation with the Western blot reactions with a panel of latent and lytic immunogenic antigens would provide a reliable, sensitive, and specific method for the detection of HHV-8 antibodies.     

Comparative study of the antigenic and immunogenic properties of native and recombinant Marburg virus VP35 proteins]

Antigenic structure of Marburg virus (MBG) VP35 was compared with that of the recombinant VP35 (f35) expressed in a prokaryotic system. For this purpose, a gene encoding the full-length VP35 was cloned in vector pQE31(QIAGEN) and expressed at about 70 mg/liter culture fluid in Escherichia coli JM103 as a recombinant fusion protein f35. BALB/c mice were immunized with soluble f35 or purified inactivated virions of MBG. Antibodies cross-reacting with VP35 and f35 antigens were detected by ELISA and Western Blot analysis in immune sera. Serum from a convalescent after Marburg disease and polyclonal antibodies from animals immunized with MBG recognized f35 and the MBG VP35. VP35 and its recombinant analog induced the production of specific antiviral antibodies in mice, which cross-reacted with the studied antigens. Competitive EIA showed that VP35 and f35 cross-inhibit the antigenic reactivity with polyclonal antibodies of immune sera. Antigenic structure of f35 protein corresponded to antigenic structure of MBG VP35 protein.     

Capsular and somatic antigens of Klebsiella bacilli.

Capsular and somatic antigens were determined in 100 Kl. oxytoca strains isolated from patients with the respiratory tract inflammations. K antigens were assigned by the capsula swelling test using 77 specific anti-capsular sera. Most frequent were: K14, K2, K55, K8 and K16 antigens. Positive reaction was noted with 64 strains in 2 or more sera. Somatic antigens of Klebsiella oxytoca bacilli were tested by the test tube agglutination reaction. Of 63 strains tested with anti-01 Kl. pneumoniae and Kl. oxytoca sera, all reacted positively in anti-0 Kl. oxytoca serum and 77% strains in anti-01 Kl. pneumoniae serum. Of 29 strains agglutinating in anti-03 sera, 65.5% agglutinated with anti-0 Kl. oxytoca serum and 76% with anti-03 Kl. pneumoniae. The results have revealed that Kl. oxytoca strains investigated have more complicated capsular antigens and different frequency of their occurrence. The most commonly encountered somatic antigen is antigen 01, next in turn is antigen 03, these antigens proved nonidentical with their 01 and 03 counterparts in Kl. pneumoniae strains.