JOURNAL OF BIOLUMINESCENCE A N D CHEMILUMINESCENCE VOL 5 183-185
(1990)
Detection of Legionella pn eumophila Serogroup 1 Urinary Antigen Using an Enhanced Chemiluminescence ELISA Dhanraj Samuel, Timothy G. Harrison and Anthony
G.Taylor
Legionella Reference Unit, Division of Microbiological Reagents and Quality Control, London, UK
An enhanced chemiluminescence enzyme-linked immunosorbent assay has been developed for the detection of soluble antigen in the urine of patients with Legionnaires’ disease (LD). In the assay antigen(s) in the urine samples are captured by a rabbit anti-L. pneumophila antibody coated onto microtitre strips. A fluorescein-isothiocyanate (FITC) conjugate of the same antibody is then added which binds t o the captured antigen. Any immobilized FITC-labelled antibody is then detected with a horseradish peroxidase (HRP) conjugate of a monoclonal anti-FITC antibody. HRP activity is monitored after oxidation of luminol in the presence of H202and iodophenol. The resulting luminescence is recorded using a camera luminometer. Urine specimens were available for testing from 31 patients with evidence of ongoing L. pneumophila serogroup 1 infection. A positive result was obtained in the cases of 12/12 specimens from culture-proven LD patients, and 16/19 specimens from patients with serological evidence of LD. Thus the sensitivity is estimated t o be 28/31 (90%) The specificity was estimated using urine specimens from eight patients with non-L. pneumophila pneumonias of known aetiology. All eight specimens gave a negative result. Keywords: ELISA; Legionella; urinary antigen
INTRODUCTION
system. Using this detection system the assay is both highly sensitive and requires only small amounts of the limited quantities available of the suitable antisera.
The diagnosis of Legionnaires’ disease (LD) by the detection of urinary antigen was demonstrated, using an ELISA, by Tilton (1979) shortly after the causative organism, Legionella pneumophila, was first described by McDade and MATERIALSAND METHODS colleagues in 1977. However, this technique has not been widely used primarily because of the PBS: 0.1 mol/l phosphate buffer p H 7.3 containproblems associated with reproducibly preparing ing 2.7mmoM KCl and 0.14mol/l NaCl. PBSantisera suitable for capturing the urinary anti- azide: PBS containing 0.08% w/v NaN3. PBSgen. This paper presents preliminary data gained TW: PBS containing 0.05% v/v Tween-20. Rabbit anti-Legionella pneumophila serogroup using an ELISA which uses a novel enhanced chemiluminescence FITC-anti-FITC detection 1 (Lpl) IgG was a generous gift of Professor F. 0884-3996/90/030 18F03$05.OO 0 1990 by John Wiley & Sons, Ltd
Received J October I988 Revised 27 January 1989
184
D. SAMUEL, T.
G. HARRISON AND A. G. TAYLOR
Fehrenbach (Robert Koch-Institut, Berlin, luminol substrate and the metal mask completely enclosed by replacing the lid. The shutter was FRG). FITC-labelled anti-Lpl antibody was prepared removed exposing the ELISA strips to the essentially as described by Samuel et al. (1988) for photographic film. Various exposure timeswere protein-FITC conjugates. Briefly to 1 mg of used ranging from two seconds to three minutes. antibody (in 0.5 ml of 0.1 mol/l carbonate buffer, After exposure the shutter was re-inserted and pH 9.2, containing 0.1 mol/l NaCI), was added a the film developed for 35 seconds. The luminol substrate was prepared essentially total 7 . 8 ~ 1of FITC (5mg/ml in dry absolute ethanol) in three equal aliquots over a period of as described by Kricka and Thorpe (1986). 10 min with stirring at room temperature (RT). Sodium luminol (Sigma) was crystallized twice The reaction was allowed to proceed for a further from 5% NaOH as described by Stott and Kricka 30min before gel filtration on a PD 10 column (1987). The substrate solution was prepared, just (Pharmacia). The column was first blocked by before use, by adding 100pI of the enhancer applying 1 ml of 1% bovine serum albumin and solution (9mg of iodophenol in 1 ml of dimethylthen washed and equilibrated with Dulbecco A sulphoxide) to 10ml of luminol solution (2.5 mg phosphate buffered saline(pH 7.2). Fractions of sodium luminol in 10ml of 0.1 moll1 Tris (0.5 ml) containing the FITC-labelled immuno- buffer, pH 8.6 containing 3.1 pl of 30% v/v globulin, which eluted in the void volume of the H202). Urine samples were collected: from patients column, were collected, pooled (total approximately 2ml) and stored in glycerol (20% v/v) at with LD, where the diagnosis had been established by culture of Legionella 4°C until used. The production and characterization of the pneumophila serogroup 1 (12 cases) or serologically monoclonal anti-FITC antibody used in this study (19 cases) using standard techniques (Harrison has been described previously (Samuel et al., and Taylor, 1988); from patients with non-L. 1988). An HRP conjugate of this antibody was pneumophila pneumonias of known aetiology prepared according to the method of Wilson and (three Streptococcus pneumoniae, and one each of Haemophilus influenzae, Klebsiella pneumoniae, Nakane (1978). The ELISA for the detection of L. pneumophi- Branhamella catarrhulis, Mycoplasma pneumoniae la serogroup 1 urinary antigen was performed as and respiratory syncytial virus); and from three follows: flat-bottom polyethylene terephthalate apparently healthy laboratory staff. glycol ELISA eight-well strips (Costar) were coated with 100pliwell of anti-Lpl IgG antibody (10 pg/ml) in 0.01 mol/l PBS -azide for at least 48 h RESULTS at 4°C. Before use the wells were washed six times with PBS containing 0.05% (v/v) Tween-20 The light emmission produced by the HRP (PBS-TW). Urine samples (100 ~ 1were ) added to catalysed oxidation of luminol was recorded as a the wells and incubated at R T for one hour and photographic print. Initially the optimal exposure then washed six times in PBS-TW. FITC-labelled time was determined by examining a range of anti-Lpl IgG ( 9 0 ~ 1 )diluted 1/500 in PBS-TW prints produced with various exposure times. The containing 5% horse serum (HS) was then added print (and hence exposure time) used for the to each well and t h e strips were incubated for 1 h analysis of test results was chosen such that light at RT. The wells were again washed, as above, emission from wells containing urines known not and 8 5 ~ 1of anti-FITC-HRP diluted M O O in to contain antigen (i.e. urine from the three PBS-TW-HS, was added to each well. After a laboratory staff) were barely visible (see Fig. 1, further washing step, the immobilized anti-FITC- 10G). The optimum exposure was determined to HRP was detected by assaying for the enzyme as be 30 seconds and this was used for the remainder of the study. Urine samples from wells in which described below. The ELISA strips were transferred to a metal light emmission was clearly discernible was taken mask and placed inside the camera luminometer as containing L . pneumophila antigen. In the (Dynatech) over a sliding shutter. The lumino- majority of cases the presence of antigen(s) in meter was loaded with a Poloroid 612 (ASA urine was easily identified since the luminescence 20,000) instant photographic film. To each of the generated by the assay completely saturated the wells of the ELISA strip was added 100 pl of the Polaroid film and appeared as white dots on the
185
ELISA FOR L. PNEUMOPHlLA IN URINE
Figure 1. Photographic result of the chemilurninescent ELISA for Legionella pneumophda serogroup 1 urinary antigen Wells F10. G9 and G I 0 are urines from healthy
laboratory staff
(I e
negative)
prints (Fig. 1, lane C3-7). A few samples gave weaker (grey) but still discernible signals (Fig. 1, B4-6) and these were also considered to be positive. Antigen was detected in 12/12 urine samples taken from culture-proven LD patients, and in 16/19 patients with serological evidence of LD. Thus the sensitivity of the ELISA was estimated to be 28/31 (90%). The specificity was estimated using urine specimens from eight patients with penumonias of known aetiology, all of which gave negative results. DISCUSSION
In this paper, we describe an ELISA for the detection of a L. pneumophila serogroup 1 antigen in urine.The diagnosis of LD by detection of urinary antigen has several advantages over the more frequently used antibody estimation techniques. Urine is, with the exception of patients in oliguric renal failure, an easily and non-invasively collected specimen. In addition it has previously been reported that urinary antigen is detectable as early as 1-3 days after the onset of the illness (Kohler et al., 1984) thus allowing a rapid diagnosis to be established. This latter finding was confirmed by this study (data not shown) in the few instances where appropriately timed urine specimens were available for examination. The chemiluminescent and photographic method of monitoring HRP used here is particu-
larly suited to immunodiagnosis since a permanent and visual record is obtained. In addition, the chemiluminescent method reduces the time of the assay as compared to colorimetric immunoassay assays without loss in sensitivity. Although only qualitative results are obtained, this is not a disadvantage as the presence of antigen appears diagnostic for LD. A major problem with immunoassays for the detection of L . pneumophila urinary antigen is the preparation of rabbit antisera suitable for antigen capture. The difficulty of reproducing large quantities of suitable antiserum has hampered attempts to establish this type of ELISA as a routine laboratory assay (Harrison and Taylor, 1988). It is desirable therefore that antisera that appears to be suitable are conserved by efficient and economical use. The ELISA reported in this study is an indirect assay and as such is potentially very sensitive. In addition because it employs a novel detection system the limited supply of the anti-L. pneumophila antiserum can be used very economically for both antigen capture and detection.
REFERENCES Harrison, T. G. and Taylor, A. G. (eds) (1988). A Laborarory Manual for Legionnaires’ Disease, John Wiley, Chichester. Kohler, R. B., Winn, W. C. and Wheat L. J. (1984). Onset and duration of urinary antigen excretion in Legionnaires’ disease. J . Clin. Microbiol., 20, 605-607. Kricka, L. J. and Thorpe, G. H. (1986). Photographic detection of chemiluminescent and bioluminescence reactions. Methods Enzymol., 133, 404-420. McDade, J . E., Shepard, C. C., Fraser, D . W., etal. (1977). Legionnaires’ disease: isolation of a bacterium and demonstration of its role in respiratory disease. N . Engl. J . Med., 297, 1197-1203. Samuel, D ., Patt, R. J. and Abuknesha, R. A. (1988). A sensitive method of detecting proteins on dot and Western blots using monoclonal antibody to FITC. J . Immunol. Methods, 107, 217-224. Stott, R . A. W . and Kricka, L. J . (1987). Purification of luminol for use in enhanced chemiluminescence immunoassay. In Bioluminescence and Chemiluminescence: New Perspectives, Scholmerich, J . , Andreesen, R., Kapp, A,, Ernst, M. and Woods, W. G. (Eds), John Wiley, Chichester, pp. 237-240. Tilton, R. C. (1979). Legionnaires disease antigen detected by enzyme-linked immunosorbent assay. Ann. Intern. Med. 90, 697-698. Wilson, M. B. and Nakane, P. K. (1978). Recent developments in the periodate method of conjugating horseradish peroxidase (HRPO) to antibodies. In Immunofluorescence and Related Staining Techniques, Knapp, W., Holubar, K. and Wick, G., (Eds), Elsevier, Amsterdam, pp. 215-224.
(1990)
Detection of Legionella pn eumophila Serogroup 1 Urinary Antigen Using an Enhanced Chemiluminescence ELISA Dhanraj Samuel, Timothy G. Harrison and Anthony
G.Taylor
Legionella Reference Unit, Division of Microbiological Reagents and Quality Control, London, UK
An enhanced chemiluminescence enzyme-linked immunosorbent assay has been developed for the detection of soluble antigen in the urine of patients with Legionnaires’ disease (LD). In the assay antigen(s) in the urine samples are captured by a rabbit anti-L. pneumophila antibody coated onto microtitre strips. A fluorescein-isothiocyanate (FITC) conjugate of the same antibody is then added which binds t o the captured antigen. Any immobilized FITC-labelled antibody is then detected with a horseradish peroxidase (HRP) conjugate of a monoclonal anti-FITC antibody. HRP activity is monitored after oxidation of luminol in the presence of H202and iodophenol. The resulting luminescence is recorded using a camera luminometer. Urine specimens were available for testing from 31 patients with evidence of ongoing L. pneumophila serogroup 1 infection. A positive result was obtained in the cases of 12/12 specimens from culture-proven LD patients, and 16/19 specimens from patients with serological evidence of LD. Thus the sensitivity is estimated t o be 28/31 (90%) The specificity was estimated using urine specimens from eight patients with non-L. pneumophila pneumonias of known aetiology. All eight specimens gave a negative result. Keywords: ELISA; Legionella; urinary antigen
INTRODUCTION
system. Using this detection system the assay is both highly sensitive and requires only small amounts of the limited quantities available of the suitable antisera.
The diagnosis of Legionnaires’ disease (LD) by the detection of urinary antigen was demonstrated, using an ELISA, by Tilton (1979) shortly after the causative organism, Legionella pneumophila, was first described by McDade and MATERIALSAND METHODS colleagues in 1977. However, this technique has not been widely used primarily because of the PBS: 0.1 mol/l phosphate buffer p H 7.3 containproblems associated with reproducibly preparing ing 2.7mmoM KCl and 0.14mol/l NaCl. PBSantisera suitable for capturing the urinary anti- azide: PBS containing 0.08% w/v NaN3. PBSgen. This paper presents preliminary data gained TW: PBS containing 0.05% v/v Tween-20. Rabbit anti-Legionella pneumophila serogroup using an ELISA which uses a novel enhanced chemiluminescence FITC-anti-FITC detection 1 (Lpl) IgG was a generous gift of Professor F. 0884-3996/90/030 18F03$05.OO 0 1990 by John Wiley & Sons, Ltd
Received J October I988 Revised 27 January 1989
184
D. SAMUEL, T.
G. HARRISON AND A. G. TAYLOR
Fehrenbach (Robert Koch-Institut, Berlin, luminol substrate and the metal mask completely enclosed by replacing the lid. The shutter was FRG). FITC-labelled anti-Lpl antibody was prepared removed exposing the ELISA strips to the essentially as described by Samuel et al. (1988) for photographic film. Various exposure timeswere protein-FITC conjugates. Briefly to 1 mg of used ranging from two seconds to three minutes. antibody (in 0.5 ml of 0.1 mol/l carbonate buffer, After exposure the shutter was re-inserted and pH 9.2, containing 0.1 mol/l NaCI), was added a the film developed for 35 seconds. The luminol substrate was prepared essentially total 7 . 8 ~ 1of FITC (5mg/ml in dry absolute ethanol) in three equal aliquots over a period of as described by Kricka and Thorpe (1986). 10 min with stirring at room temperature (RT). Sodium luminol (Sigma) was crystallized twice The reaction was allowed to proceed for a further from 5% NaOH as described by Stott and Kricka 30min before gel filtration on a PD 10 column (1987). The substrate solution was prepared, just (Pharmacia). The column was first blocked by before use, by adding 100pI of the enhancer applying 1 ml of 1% bovine serum albumin and solution (9mg of iodophenol in 1 ml of dimethylthen washed and equilibrated with Dulbecco A sulphoxide) to 10ml of luminol solution (2.5 mg phosphate buffered saline(pH 7.2). Fractions of sodium luminol in 10ml of 0.1 moll1 Tris (0.5 ml) containing the FITC-labelled immuno- buffer, pH 8.6 containing 3.1 pl of 30% v/v globulin, which eluted in the void volume of the H202). Urine samples were collected: from patients column, were collected, pooled (total approximately 2ml) and stored in glycerol (20% v/v) at with LD, where the diagnosis had been established by culture of Legionella 4°C until used. The production and characterization of the pneumophila serogroup 1 (12 cases) or serologically monoclonal anti-FITC antibody used in this study (19 cases) using standard techniques (Harrison has been described previously (Samuel et al., and Taylor, 1988); from patients with non-L. 1988). An HRP conjugate of this antibody was pneumophila pneumonias of known aetiology prepared according to the method of Wilson and (three Streptococcus pneumoniae, and one each of Haemophilus influenzae, Klebsiella pneumoniae, Nakane (1978). The ELISA for the detection of L. pneumophi- Branhamella catarrhulis, Mycoplasma pneumoniae la serogroup 1 urinary antigen was performed as and respiratory syncytial virus); and from three follows: flat-bottom polyethylene terephthalate apparently healthy laboratory staff. glycol ELISA eight-well strips (Costar) were coated with 100pliwell of anti-Lpl IgG antibody (10 pg/ml) in 0.01 mol/l PBS -azide for at least 48 h RESULTS at 4°C. Before use the wells were washed six times with PBS containing 0.05% (v/v) Tween-20 The light emmission produced by the HRP (PBS-TW). Urine samples (100 ~ 1were ) added to catalysed oxidation of luminol was recorded as a the wells and incubated at R T for one hour and photographic print. Initially the optimal exposure then washed six times in PBS-TW. FITC-labelled time was determined by examining a range of anti-Lpl IgG ( 9 0 ~ 1 )diluted 1/500 in PBS-TW prints produced with various exposure times. The containing 5% horse serum (HS) was then added print (and hence exposure time) used for the to each well and t h e strips were incubated for 1 h analysis of test results was chosen such that light at RT. The wells were again washed, as above, emission from wells containing urines known not and 8 5 ~ 1of anti-FITC-HRP diluted M O O in to contain antigen (i.e. urine from the three PBS-TW-HS, was added to each well. After a laboratory staff) were barely visible (see Fig. 1, further washing step, the immobilized anti-FITC- 10G). The optimum exposure was determined to HRP was detected by assaying for the enzyme as be 30 seconds and this was used for the remainder of the study. Urine samples from wells in which described below. The ELISA strips were transferred to a metal light emmission was clearly discernible was taken mask and placed inside the camera luminometer as containing L . pneumophila antigen. In the (Dynatech) over a sliding shutter. The lumino- majority of cases the presence of antigen(s) in meter was loaded with a Poloroid 612 (ASA urine was easily identified since the luminescence 20,000) instant photographic film. To each of the generated by the assay completely saturated the wells of the ELISA strip was added 100 pl of the Polaroid film and appeared as white dots on the
185
ELISA FOR L. PNEUMOPHlLA IN URINE
Figure 1. Photographic result of the chemilurninescent ELISA for Legionella pneumophda serogroup 1 urinary antigen Wells F10. G9 and G I 0 are urines from healthy
laboratory staff
(I e
negative)
prints (Fig. 1, lane C3-7). A few samples gave weaker (grey) but still discernible signals (Fig. 1, B4-6) and these were also considered to be positive. Antigen was detected in 12/12 urine samples taken from culture-proven LD patients, and in 16/19 patients with serological evidence of LD. Thus the sensitivity of the ELISA was estimated to be 28/31 (90%). The specificity was estimated using urine specimens from eight patients with penumonias of known aetiology, all of which gave negative results. DISCUSSION
In this paper, we describe an ELISA for the detection of a L. pneumophila serogroup 1 antigen in urine.The diagnosis of LD by detection of urinary antigen has several advantages over the more frequently used antibody estimation techniques. Urine is, with the exception of patients in oliguric renal failure, an easily and non-invasively collected specimen. In addition it has previously been reported that urinary antigen is detectable as early as 1-3 days after the onset of the illness (Kohler et al., 1984) thus allowing a rapid diagnosis to be established. This latter finding was confirmed by this study (data not shown) in the few instances where appropriately timed urine specimens were available for examination. The chemiluminescent and photographic method of monitoring HRP used here is particu-
larly suited to immunodiagnosis since a permanent and visual record is obtained. In addition, the chemiluminescent method reduces the time of the assay as compared to colorimetric immunoassay assays without loss in sensitivity. Although only qualitative results are obtained, this is not a disadvantage as the presence of antigen appears diagnostic for LD. A major problem with immunoassays for the detection of L . pneumophila urinary antigen is the preparation of rabbit antisera suitable for antigen capture. The difficulty of reproducing large quantities of suitable antiserum has hampered attempts to establish this type of ELISA as a routine laboratory assay (Harrison and Taylor, 1988). It is desirable therefore that antisera that appears to be suitable are conserved by efficient and economical use. The ELISA reported in this study is an indirect assay and as such is potentially very sensitive. In addition because it employs a novel detection system the limited supply of the anti-L. pneumophila antiserum can be used very economically for both antigen capture and detection.
REFERENCES Harrison, T. G. and Taylor, A. G. (eds) (1988). A Laborarory Manual for Legionnaires’ Disease, John Wiley, Chichester. Kohler, R. B., Winn, W. C. and Wheat L. J. (1984). Onset and duration of urinary antigen excretion in Legionnaires’ disease. J . Clin. Microbiol., 20, 605-607. Kricka, L. J. and Thorpe, G. H. (1986). Photographic detection of chemiluminescent and bioluminescence reactions. Methods Enzymol., 133, 404-420. McDade, J . E., Shepard, C. C., Fraser, D . W., etal. (1977). Legionnaires’ disease: isolation of a bacterium and demonstration of its role in respiratory disease. N . Engl. J . Med., 297, 1197-1203. Samuel, D ., Patt, R. J. and Abuknesha, R. A. (1988). A sensitive method of detecting proteins on dot and Western blots using monoclonal antibody to FITC. J . Immunol. Methods, 107, 217-224. Stott, R . A. W . and Kricka, L. J . (1987). Purification of luminol for use in enhanced chemiluminescence immunoassay. In Bioluminescence and Chemiluminescence: New Perspectives, Scholmerich, J . , Andreesen, R., Kapp, A,, Ernst, M. and Woods, W. G. (Eds), John Wiley, Chichester, pp. 237-240. Tilton, R. C. (1979). Legionnaires disease antigen detected by enzyme-linked immunosorbent assay. Ann. Intern. Med. 90, 697-698. Wilson, M. B. and Nakane, P. K. (1978). Recent developments in the periodate method of conjugating horseradish peroxidase (HRPO) to antibodies. In Immunofluorescence and Related Staining Techniques, Knapp, W., Holubar, K. and Wick, G., (Eds), Elsevier, Amsterdam, pp. 215-224.
