Journal of Clinical Laboratory Analysis 6:162-169 (1992)

Sensitive Enzyme lmmunoassay (Immune Complex Transfer Enzyme lmmunoassay) for (Antihuman T-cell Leukemia Virus

Type I) Immunoglobulin G in Serum Using a Synthetic Peptide, Ala-Cys-Env gp46(237-262), as Antigen Takeyuki Kohno,' lwane Sakoda,* and Eiji Ishikawa'

'Department of Biochemistry, Medical College of Miyazaki, and 2Miyazaki Blood Center, The Japanese Red Cross, Miyazaki, Japan

A sensitive enzyme immunoassay (immune complex transfer enzyme immunoassay) for (antihuman T-cell leukemia virus type I) IgG (anti-HTLV-l IgG) in serum using a synthetic peptide, Ala-Cys-env gp46(237-262), of HTLV- I is described. Anti-HTLV-l IgG in test serum, which had been incubated with excess of inac- tive p-D-galactosidase to eliminate interfer- ence by anti-p-D-galactosidase antibodies, was reacted simultaneously with 2,4-dinitro- phenyl-bovine serum albumin-Ala-Cys-env gp46(237-262) conjugate and Ala-Cys-env gp46(237-262)-p-D-galactosidase conjugate. The complex formed consisting of the three components was trapped onto polystyrene balls coated with affinity-purified (anti-2,4- dinitrophenyl group) IgG. After washing to eliminate nonspecific IgG in the test serum and excess of the p-D-galactosidase conjugate,

the complex was eluted from the polystyrene balls with ~N-2,4-dinitrophenyl-L-lysine and transferred to polystyrene balls coated with affinity-purified (anti-human IgG y-chain) IgG. p-0-galactosidase activity bound to the (anti- human IgG y-chain) IgG-coated polystyrene balls was assayed by fluorometry. This assay was more sensitive than other methods using HTLV-I as antigen, and most negative and positive sera were discriminated. However, some results appeared to be false positive or false negative, and the peptide, Ala-Cys- env gp46(237-262), was suggested to be useful, in combination with other peptides, for improving the reliability of serodiagnosis by separately demonstrating antibodies against as many different epitopes of HTLV-I as possible. 01992 ~itey-Liss, Inc.

Key words: antibody, adult T-cell leukemia, p-D-galactosidase, Western blotting, gelatin parti- cle agglutination, ELISA

INTRODUCTION

Human T-cell leukemia virus type I (HTLV-I), isolated from a patient with cutaneous T-cell lymphoma (l) , is etiologi- cally associated with adult T-cell leukemia (2) and adult T-cell cancers (3) . The virus is transmitted by blood transfusion (4) and breast feeding (5). For prevention of the virus transmis- sion, plasma or serum of blood donors and pregnant women has been tested for anti-HTLV-I antibodies by gelatin particle agglutination (6), enzyme-linked immunosorbent assay (7), fluoroimmunoassay (2), and Western blotting (8). HTLV-I used as antigen in these methods has been produced from particular cell lines such as TCL-Kan (9 ) and MT-2 (10) and purified by density gradient centrifugation (6) or by ultracentrifuga- tion (7). This is hazardous and inefficient. In addition, the specificity of these methods has not been tested by preincu- bation of serum with excess of HTLV-I as antigen.

To avoid the use of HTLV-I as antigen, a novel enzyme immunoassay (immune complex transfer enzyme immuno- assay) for anti-HTLV-I IgG has been developed using recom-

0 1992 Wiley-Liss, Inc.

binant gug p24( 14- 139)-env gp46(2 17-3 15) hybrid protein as antigen, which was safely and efficiently produced in Esch- erichia coli (1 1). This assay was 30-300-fold more sensitive than other methods, including Western blotting using HTLV-I, gelatin particle agglutination using HTLV-I, and enzyme-linked immunosorbent assay using the gag-env hybrid protein. Most of positive and negative samples were more clearly discrimi- nated than by the other methods. Anti-HTLV-I IgG was dem- onstrated in all samples, which were positive by Western blotting. Low levels of anti-HTLV-I IgG below those detected by Western blotting were suggested to be detectable. How- ever, the gag-env hybrid protein was soluble only in the pres- ence of detergents. This hampered the use of enzymes other than horseradish peroxidase as label, and the specificity could not be tested by preincubation of serum with excess of the

Received January 6, 1992; accepted January 8. 1992

Address reprint requests to Dr. E. Ishikawa, Department of Biochemistry, Medical College of Miyazaki, Kiyotake, Miyazaki 889- 16, Japan.

Enzyme lmmunoassay for Anti-HTLV-l IgG 163

was prepared from serum by fractionation with Na2S04 fol- lowed by passage through a column of diethylaminoethyl cel- lulose, and the amount of IgG was calculated from the absorbance at 280 nm (16).

gag-env hybrid protein. It was not easy to separate completely the gag-env hybrid pro9ein from proteins of Escherichia coli, antibodies against which are frequently present in human serum.

To overcome these difficulties, the recombinant protein has been replaced by synthetic peptides, including Cys-Arg-env gp46( 188-209) (1 2 ) , Cys-gag p 19( 100- 130) ( I3), and Cys- env gp46( 188-224) (14). The specificity can be tested by pre- incubation of serum with excess of the synthetic peptides, which are readily soluble in water. The sensitivities were 100-30,000-fold higher than those of other methods such as Western blotting using 4 core proteins of HTLV-I, gelatin par- ticle agglutination using HTLV-I, and enzyme-linked immu- nosorbent assay using the synthetic peptides and HTLV-I. Most negative and positive sera were discriminated. However, some results appeared to be false negative or false positive, and the use of as many synthetic peptides as possible was sug- gested to improve the reliability of serodiagnosis by separately demonstrating antibodies against as many different epitopes of HTLV-I as possible.

Recently, two synthetic peptides, env gp46( 19 1-2 15) and env gp46(242-257), of HTLV-I have been reported to react with 92% and loo%, respectively, of 52 seropositive Sam- ples by a conventional type of enzyme immunoassay, in which polyvinyl plates coated with the peptides were incubated with test serum (anti-HTLV-I IgG) and subsequently with alka- line phosphatase-conjugated goat antihuman IgG (15). This paper describes a sensitive enzyme immunoassay (immune complex transfer enzyme immunoassay) for anti-HTLV-I IgG using Ala-Cys-env gp46(237-262).

MATERIALS AND METHODS

Buffers

The regularly used buffers were 0.1 mol/liter sodium phos- phate buffer, pH 6.0, containing 5 mmol/liter EDTA (buffer A); 10 mmol/liter sodium phosphate buffer, pH 7.0, contain- ing 1 .O mmol/liter MgC12, 1 .O g/liter NaN3, and 1 .O g/liter bovine serum albumin (fraction V; Armour Pharmaceutical Co., Kankakee, IL) (buffer B); and 10 mmol/liter sodium phosphate buffer, pH 7.0, containing 1 .O g/liter bovine serum albumin (buffer C).

Ala-Cys-Env gp46(237-262) of HTLV-I

Ala-Cys-env @6(237-262) of HTLV-I (H2N-Ala-Cys-Trp237- His-Val-Leu-Tyr-Ser-Pro-Asn-Val-Ser-Val-Pro-Ser-Ser-Ser- Ser-Thr-Pro-Leu-Leu-Tyr-Pro-Ser-Leu-Ala-Le~262-COOH) was obtained from Peptide Institute, Inc. (Osaka, Japan).

An ti bodies

Rabbit (anti-2,4-dinitrophenyl-bovine serum albumin) serum was obtained from Shibayagi Co., Ltd. (Gumma, Japan). Rab- bit (antihuman IgG y-chain) IgG was obtained from Medical and Biological Laboratories Co., Ltd. (Nagoya, Japan). IgG

2,4-Dinitrophenyl-Bovine Serum Albumin

Thiol groups were introduced into bovine serum albumin molecules using N-succinimidyl-S-acetylmercaptoacetate ( 17) and were reacted with maleimide groups introduced into ~N-2,4-dinitrophenyl-L-lysine molecules using N-succinimidyl- 6-maleimidohexanoate ( 18). The amount of 2,4-dinitro- phenyl-bovine serum albumin was calculated from the absorbance at 280 and 360 nm ( 12). The average number of 2,4-dinitrophenyl groups introduced per albumin mole- cule was 6.8.

Affinity-Purification of Antibodies

2,4-Dinitrophenyl-bovine serum albumin ( 10 mg) and non- specific human IgG (10 mg) were coupled to CNBr-activated Sepharose 4B (1 .O g; Pharmacia LKB Biotechnology AB, Uppsala, Sweden) according to the instructions of Pharmacia LKB Biotechnology. (Anti-2,4-dinitrophenyl-bovine serum albumin) IgG and (antihuman IgG y-chain) IgG were affinity- purified by elution at pH 2.5 from columns of 2,4-dinitro- phenyl-bovine serum albumin-Sepharose 4B and nonspecific human IgG-Sepharose 4B, respectively (1 9).

Ala-Cys-Env gp46(237-262)-Bovine Serum Albumin

Maleimide groups were introduced into bovine serum albu- min molecules using N-succinimidyl-6-maleimidohexanoate and reacted with Ala-Cys-env gp46(237-262) (12). The aver- age number of Ala-Cys-env gp46(237-262) molecules con- jugated per albumin molecule was 8.7, which was calculated from the decrease in the number of maleimide groups (16). The amount of the conjugate was calculated from that of bovine serum albumin ( 12),

Protein-Coated Polystyrene Balls

Polystyrene balls (3.2 mm in diameter; Immuno Chemi- cal Inc., Okayama, Japan) were coated by physical adsorp- tion with affinity-purified (anti-2,4-dinitrophenyl-bovine serum albumin) IgG (0.1 g/liter), affinity-purified (anti-human IgG y-chain) IgG (0.1 g/liter) and Ala-Cys-env gp46(237-262)- bovine serum albumin (0.1 g/liter) (20). Affinity-purified (anti- 2,4-dinitrophenyl-bovine serum albumin) IgG-coated poly- styrene balls had been colored pink for discrimination from other polystyrene balls.

2,4-Dinitrophenyl-Bovine Serum Albumin-Ala-Cys- Env gp46(237-262) Conjugate

Maleimide groups were introduced into 2,4-dinitrophenyl- bovine serum albumin molecules using N-succinimidyl-

164 Kohno et al.

6-maleimidohexanoate and reacted with Ala-Cys-env gp46- (237-262) (12). The average number of Ala-Cys-env gp46- (237-262) molecules conjugated per albumin molecule was 3.9, which was calculated from the decrease in the number of maleimide groups (16). The amount of the conjugate was calculated from that of 2,4-dinitrophenyl-bovine serum albumin.

Ala-Cys-Enw gp46(237-262)-P-D-Galactosidase Conjugate

Maleimide-p- D-galactosidase Maleimide groups were introduced into molecules of

P-D-galactosidase (EC 3.2.1.23) from Escherichia coli using N,N’- 1,2-~henylenedimaleimide (16).

Ala-Cys-env gp46(237-262)-P-D-galactosidase conjugate

Maleimide-P-D-galactosidase (1.5 mg, 2.8 nmol) in 0.5 ml of buffer A was incubated with Ala-Cys-env gp46(237-262) (42 pg, 14 nmol) in 8 )11 of N,N-dimethylformamide at 4°C overnight. Subsequently, the reaction mixture was incubated with 20 ~1 of 0.1 mol/liter 2-mercaptoethylamine in buffer A at 30°C for 15 min and subjected to gel filtration on a col- umn (1.5 x 45 cm) of Ultrogel AcA 22 (IBF Biotechnics, Villeneuve-la-Garenne, France) using 10 mmoliliter sodium phosphate buffer, pH 7.0, containing 0.1 mol/liter NaCl, 1 .O mmol/liter MgC12, 1 .O g/liter NaN3, and 0.1 g/liter bovine serum albumin. The average number of Ala-Cys-env gp46 (237- 262) molecules conjugated per P-D-galactosidase molecule was 4.8, which was calculated from the decrease in the num- ber of maleimide groups (16). The amount of the conjugate was calculated from P-D-galactosidase activity (2 1).

Immune Complex Transfer Enzyme lmmunoassay

Test serum (20 pl) was incubated with 50 pg of inactive (3-D-galactosidase (P-Galactosidase Mutein; Boehringer Mann- heim GmbH, Mannheim, Federal Republic of Germany) in 30 )11 of buffer B containing 0.1 mol/liter NaCl at 20°C for 3 h to eliminate interference by anti-p-D-galactosidase antibod- ies (12). In some experiments, Ala-Cys-env gp46(237-262) (4.4 ng, 1.5 pmol) was added to test the specificity. After incubation, the reaction mixture was incubated simultaneously with 2,4-dinitrophenyl-bovine serum albumin-Ala-Cys-env gp46(237-262) conjugate (100 fmol) and Ala-Cys-env gp46- (237-262)-P-D-galactosidase conjugate (100 fmol) in 0.1 ml of buffer B containing 0.55 mol/liter NaCl at 20°C for 3 h. Two affinity-purified (anti-2,4-dinitrophenyl-bovine serum albumin) IgG-coated polystyrene balls, which had been colored pink, were added to the reaction mixture, and the incubation was continued at 20°C overnight.

After incubation, the colored polystyrene balls were washed twice with 2 ml of buffer B containing 0.1 mol/liter NaCl

trophenyl-L-lysine (Tokyo Kasei Kogyo Co., Ltd., Tokyo, Japan) in the same buffer along with two affinity-purified (anti- human IgG y-chain) IgG-coated polystyrene balls at 20°C for 1 h. After incubation, the colored polystyrene balls were dis- carded, and the eluate and the (antihuman IgG y-chain) IgG- coated polystyrene balls were further incubated at 20°C for 2 h. After washing as described above, P-D-galactosidase activ- ity bound to the polystyrene balls was assayed at 30°C for 150 min by fluorometry using 4-methylumbelliferyl-~-D- galactoside as substrate (22). The fluorescence intensity was measured relative to 10 - * mol/liter 4-methyl umbelliferone in 0.1 moliliter glycine-NaOH buffer, pH 10.3.

Enzyme-Linked lmmunosorbent Assay (ELISA)

Rabbit (antihuman IgG y-chain) IgG was converted to the corresponding Fab‘ and conjugated to horseradish peroxidase using N-succinimidyl-6-maleimidohexanoate (23). The amount of the conjugate was calculated from peroxidase activity (16).

In the ELISA using Ala-Cys-env gp46(237-262), test serum (20 p1) was mixed with 0.13 ml of buffer C containing 0.46 mol/liter NaCl and 1 .O g/liter NaN3 and was incubated with an Ala-Cys-env gp46(237-262)-bovine serum albumin-coated polystyrene ball at 37°C for 3 h and at 4°C overnight. After incubation, the polystyrene ball was washed twice with 2 ml of 10 mmol/liter sodium phosphate buffer, pH 7 .O, containing 0.1 mol/liter NaCl and was incubated with 50 ng of (anti- human IgG y-chain) Fab’-peroxidase conjugate in 0.15 ml of buffer C containing 0.1 mollliter NaCl at 37°C for 3 h. The polystyrene ball was washed as described above, and bound peroxidase activity was assayed at 30°C for 10 min by fluorometry using 3-(4-hydroxyphenyl)propionic acid as hydrogen donor (24). The fluorescence intensity was measured relative to 1.0 mg/liter quinine in 50 mmol/ liter H2S04.

ELISA using HTLV-I as antigen was performed with a com- mercial kit (Eitest- ATL; Eisai Co. , Ltd. , Tokyo, Japan). Microtiter plates coated with HTLV-I produced by MT-2 cell line (10) were incubated with 20 pl of test sera and, after washing, with monoclonal (antihuman IgG Fc) antibody- alkaline phosphatase conjugate. Bound alkaline phosphatase activity was assayed by colorimetry using 4-nitrophenyl- phosphate as substrate.

Expression of the Detection Limit of Anti-HTLV-l IgG in Serum

The detection limit of anti-HTLV-I IgG in serum by enzyme immunoassays was expressed as the maximal dilution of serum containing anti-HTLV-I IgG with pooled normal serum, which gave a bound enzyme activity significantly in excess of that in the presence of normal serum (background). A significant difference from the background was confirmed by the t test

and incubated with 0.15 ml of 1.0 mmol/liter ~N-2,4-dini- (n = 5, P < 0.001).

Enzyme lmmunoassay for Anti-HTLV-l IgG 165

dinitrophenyl-bovine serum albumin and P-D-galactosidase and tested by a sensitive enzyme immunoassay (immune com- plex transfer enzyme immunoassay) as schematically shown in Figure 1 (the present assay).

p+-. + @ +

Anti-DNP- solid phase DNP-Ag

/ / / / /

17 0 0 DNP-lysine

Anti-lgG-solid phase

Ag-Enz

Fig. 1. DNP. 2.4-dinitrophenyl group; Ag, antigen; Ab, antibody; Enz, enzyme.

Immune complex transfer enzyme immunoassay for antibody IgC.

Gelatin Particle Agglutination

Detection of anti-HTLV-I antibodies by gelatin particle agglu- tination was performed using a commercial kit with HTLV-I produced by TCL-Kan cell line (9) as antigen (SERODIA- ATLA; Fujirebio Inc., Tokyo, Japan). Test serum was diluted eightfold or more with the diluent included in the kit. The diluted serum (25 pl) was mixed with the particle solution (25 p1) in U-shaped wells of microplates and allowed to stand at room temperature for 3 h.

Western Blotting

Western blotting using four core proteins (p 19, p24, p28, and p53) of HTLV-I produced by TCL-Kan cell line (9) was generously performed by Fujirebio Inc. (Tokyo, Japan).

Serum Samples

Serum samples were obtained from healthy subjects aged 16-79 years in Miyazaki, Japan. Approximately 7% of healthy subjects were HTLV-I carriers.

RESULTS

Ala-Cys-env gp46(237-262) of HTLV-I, chemically syn- thesized, was conjugated through cysteine residue to 2,4-

Sensitivity

Four sera from HTLV-I carriers were serially diluted with normal serum and were subjected to the present assay using Ala-Cys-env gp46(237-262), ELISA using Ala-Cys-env gp46(237-262), ELISA using HTLV-I, gelatin particle agglu- tination using HTLV-I, and Western blotting using four core proteins (p19, p24, p28, and p53) of HTLV-I (Fig. 2, Table 1). The present assay was 30- 10,000-fold more sensitive than the other methods.

Assay Variation

The variation in the present assay was examined using serum samples, with which three different levels of fluorescence inten- sity for bound P-D-galactosidase activity (8.1, 85, and 820 for within assay and 9.6,92, and 980 for between assay) were obtained. The coefficients of variation for within assay and between assay were 6.1-8.4% (n = 15) and 8.8-1 1% (n = lo), respectively.

A

a, 5 c w -0

3 0 m

0 C

8 E 3 U -

I I! , , , , , , , $0' 1 0.:

1 o4 1 o3 1 o2 10'

Dilution of Serum Containing Anti-HTLV-l IgG with Normal Serum (-fold)

Fig. 2. Sensitivity of various methods. Four sera from HTLV-I carriers were seriaily diluted with normal serum and subjected to the immune com- plex transfer enzyme immunoassay using Ala-Cys-enw gp46(237-262) (the present assay) (open symbols with solid lines), ELISA using Ala-Cys-env gp46(237-262) (closed symbols with solid lines) and ELISA using HTLV-I (solid symbols with dashed lines).

166 Kohno et al.

TABLE 1. Sensitivity of Various Methodsa

Detection of anti-HTLV-1 antibodies or IgG [Dilution of serum containing anti-HTLV-I antibodies with normal serum (-fold)] Serum

Method No. 3 x l o 4 I x 104 3 x 103 1 x 10' 3 x 10 1 x 10 3

Present assay using 1 + + + + + + + Ala-Cys-envgp46(237-262) 2 + + + + + + +

+ + + + + + + + + + +

* + + * + +

* + + + * + +

* + + + * +

* + + + - + +

- 3 4 - -

~ - - - Gelatin particle 1 agglutination using 2 - - - -

- - - HTLV-I 3 4

Western blotting using 1 four core proteins of 2

~ - - -

- - - - - - - -

~ - - HTLV-I 3 4 - - - - -

"Four sera from HTLV-I carriers were diluted with normal serum and subjected to the three methods indicated.

Usefulness of the Present Assay with Ala-Cys-Em

Usefulness of the present assay with Ala-Cys-env gp46(237- 262) was examined using 3 18 sera, which had been tested by other methods, including gelatin particle agglutination, ELISA with HTLV-I, and Western blotting. Sera for which the fluo- rescence intensity for bound P-D-galactosidase activity in the present assay was over 7.0 were tentatively taken as positive, and other sera were taken as negative. The results are shown in Figure 3 and Tables 2 and 3.

The 3 18 sera consisted of three groups tested by gelatin particle agglutination, which has been most widely used for screening of HTLV-I carriers in Japan: 125 negative sera (group 1, sera 1-125); 91 strongly positive sera (group 2, sera 126-216), in which the maximal dilution with the diluent included in the kit to cause gelatin particle agglutination was 128 to 16,384-fold; and 102 weakly positive sera (group 3, sera 217-318), in which the maximal dilution was eight-

In group 1 ( 125 sera, sera 1 - 125), most of sera were con- sistently negative by gelatin particle agglutination, ELISA using HTLV-I, and the present assay. However, nine sera (sera 1 16- 124) appeared to be false positive by ELISA using HTLV-I or the present assay.

Ingroup 2 (91 sera, sera 126-216), 86 sera(sera 126-211) were consistently positive by gelatin particle agglutination, ELISA using HTLV-I, and the present assay. However, five sera (sera 2 12-2 16) appeared to be false negative by the pres- ent assay or false positive by gelatin particle agglutination.

Group 3 (102 sera, sera 217-318) was divided into two groups by Western blotting: 3 1 positive sera (group 3- 1 , sera 2 17-247) and 7 1 negativehdeterminate sera (group 3-2, sera 248-3 18).

Ingroup3-l(31 sera, sera217-247), 29sera(sera217-245) were consistently positive by gelatin particle agglutination, ELISA using HTLV-I, Western blotting, and the present assay.

gp46(237-262)

to 64-fold.

However, two sera (sera 246 and 247) appeared to be false negative by the present assay.

Group 3-2 (71 sera, sera 248-318) was divided into two groups by the present assay: 63 negative sera (group 3-2- 1, sera 248-310)andeightpositive sera(group3-2-2, sera311-318).

0

o o 8 0 '

10 lo2 lo3 104

Negative Maximal Dilution of Serum with Buffer to Cause Gelatin Particle Agglutination ( -fold )

Comparison of test results for anti-HTLV-I IgG or antibodies by the immune complex transfer enzyme immunoassay using Ala-Cys-env gp46(237-262) (the present assay), gelatin particle agglutination using HTLV-I, and Western blotting using four core proteins (p19, p24, p28, and p53) of HTLV-I. A tentative cut-off value of 7.0 in the present assay is indicated by the dashed line. Open circles indicate groups I and 2; closed triangles, open triangles, and open squares indicate group 3- 1, 3-2- 1, and 3-2-2, respec- tively, as shown in Tables 2 and 3 . Group 3-1 was positive and groups 3-2-1 and 3-2-2 were negativeiindeteminate by Western blotting.

Enzyme lmmunoassay for Anti-HTLV-l IgG 167

TABLE 2. Test Results by Various Methods for Groups 1 (125 sera) and 2 (91 sera)

Optical density

Gelatin at 405 nm Western Fluorescence intensity Inhibition by particle by ELISA blotting for bound P-D-galacto- excess of aggluti- using HTLV-I using sidase activity by the Ala-Cys-env nation with a cut-off four core present assay using gp46 (237-262)

Serum using value of proteins Ala-Cys-env gp46 (237-262) with a cut-off value of 75% with a cut-off value of 7.0 No. HTLV-I 0.62 of HTLV-I

Group I 1-1 15 - 0.12-0.43 N T ( I25 sera) 116, I I7 - 0.69, 0.97 NT

118-124 - 0.12-0.25 NT 125 - 0.46 NT

Group 2 12621 1 + + + b 2.0-14 NT (91 sera) 2 12-2 14 + f + I .9-9.3 NT

215.216 + + + 0.16.0.22 -

0 . 3 4 . 9 1.2,2.1 1 1 4 4 58

NT NT 13-55 93

93-1,000,000 75,8CLIOO 2 . 1 4 . 0 NT 0.9. I . I NT

"NT, not tested. "The maximal dilution of serum samples with the diluent included in the kit to cause gelatin particle agglutination was 128-16,384-fold

In group 3-2-1 (63 sera, sera 248-310), all the sera were negative by Western blotting and the present assay, and two sera (sera 309 and 310) might have been false positive by ELISA using HTLV-I. The results by gelatin particle aggluti- nation for the 63 sera might have been false positive, which was consistent with a previous report that the gelatin particle agglutination kit used in this study often gave false-positive results (25).

In group 3-2-2 (eight sera, sera 311-318), all the sera were positive by the present assay but were positive, negative, or indeterminate by other methods. Notably, Ala-Cys-env gp46(237-262) was highly reactive with

some sera (sera 31 1-3 15), which were indeterminate by Western blotting.

Specificity of the Present Assay Using Ala-Cys- EnV gp46(237-262)

To test the specificity of the present assay using Ala-Cys- env gp46(237-262), sera, which showed fluorescence inten- sities for bound P-D-galactosidase activity over the tentative cut-off value of 7.0, were preincubated with excess of Ala- Cys-env gp46(232-262) and subjected to the present assay (Tables 2 and 3). In most sera, which were positive by both

TABLE 3. Test Results by Various Methods for Group 3 (102 sera)

Optical density

Gelatin at 405 nm Fluorescence intensity Inhibition by particle by ELISA Western blotting aggluti- using HTLV-I using four core sidase activity by the Ala-Cys-env nation with a cut-off proteins of present assay using gp46 (237-262)

Serum using value HTLV-I Ala-Cys-env gp46 (237-262) with a cut-off No. HTLV-I of 0.62 ~ 1 9 024 ~ 2 8 053 with a cut-off value of 7.0 value of 75%

for bound P-D-galacto- excess of

Group 3- I 217-245 +" 1.1-15 + + + + 11-55,000 81-100 (31 sera) 246,247 + 4.7, 1.8 + + + + l.0,l.l NTb

- 0 . k 5 . 5 NT - - 3.2, 1.5 NT

- - Group 3-2 Group 3-2-1 248-308 + 0.1-0.58 ~

(7 1 sera) (63 sera) 309,310 + 1.1,0.82 - -

Group 3-2-2 3 1 I + 5.5 + + * ? 17,000 94 (8 Sera) 312 + 1.3 + + * + 880 87

313 + I .3 + - - - 2,900 95 314 + 0.94 ? % - - 12,000 97 315 + 0.39 + - - - 120 88 316 + 0.33 + - - 7.8 40

- 110 95 317 + 0.42 - 16 89 318 + 0.17

-

- - -

- - -

"The maximal dilution of serum samples with the diluent included in the kit to cause gelatin particle agglutination was eight- to 64-fold for sera 217-31 8 . bNT. not tested.

168 Kohno et al.

gelatin particle agglutination and ELISA using HTLV-I or by all gelatin particle agglutination, ELISA using HTLV-I and Western blotting, the degree of inhibition with the preincu- bation was over 80%. When the cut-off value was tentatively taken to be 75%, seven sera in group I (sera 1 18- 124) and one serum in group 3-2-2 (serum 316) were negative, although these sera were positive on the basis of the fluorescence inten- sity for bound P-D-galactosidase activity by the present assay as described above.

DISCUSSION

The results obtained by the present assay using Ala-Cys- env gp46(237-262) can be summarized as follows. 1) The present assay was more sensitive than other methods. 2) Ala- Cys-env gp46(237-262) was reactive with most positive sera by ELISA using HTLV-I and Western blotting using four core proteins (p19, p24, p28, and p53) of HTLV-I. Notably, the use of Ala-Cys-env gp46(237-262) allowed detection of anti- HTLV-I IgC in seven sera (sera 3 1 1-3 15, 3 17, and 3 18) that were indeterminate or negative by Western blotting. 3) How- ever, Ala-Cys-env gp46(237-262) was not reactive with three sera (sera 212-214) that were strongly positive by gelatin par- ticle agglutination and ELISA using HTLV-I and with two sera (sera 246 and 247) that were positive by both ELISA using HTLV-I and Western blotting. 4) Ala-Cys-env gp46(237- 262) probably gave false-positive results for some sera (sera 1 18- 124) that were negative by gelatin particle agglutination and ELISA using HTLV-I, although this may be corrected by the specificity inhibition test with excess of the peptide as described above.

When test results by various methods are inconsistent, it is difficult to decide which is more reliable. However, this difficulty may be overcome at least in part by developing sen- sitive methods using as many synthetic peptides or recombi- nant proteins as possible to demonstrate separately antibodies against as many different epitopes of HTLV-I as possible and to improve the reliability of serodiagnosis. From such con- sideration, sensitive enzyme immunoassays (immune com- plex transfer enzyme immunoassays) using two more peptides, Cys-gug p 19( 100- 130) and Cys-env gp46( 188-224), have been developed, and these two assays have discriminated most of 132-135 negativeand 128-137positivesera( 13,14). By both assays, two sera in group 1 (sera 116 and 117), which were positive by ELISA using HTLV-I but negative by gelatin par- ticle agglutination and the present assay; 49 sera in group 3-2-1 (sera 248-296), which were positive by gelatin particle agglu- tination but negative by ELISA using HTLV-I, Western blot- ting, and the present assay; and one serum in group 3-2-1 (serum 309), which was positive by gelatin particle aggluti- nation and ELISA using HTLV-I but negative by Western blot- ting and the present assay, were negative. That is, these 52 sera were consistently negative by the assays using three dif- ferent peptides and could be taken as negative with higher

probability. Six sera in group 1 (sera 118-123), which were positive by the present assay but negative by gelatin particle agglutination and ELISA using HTLV-I, were negative by the two assays. Two sera in group 2 (sera 2 15 and 2 16), which were positive by gelatin particle agglutination but negative by ELISA using HTLV-I, Western blotting, and the present assay; and 12 seraingroup3-2-1 (sera297-307 and310), whichwere positive by gelatin particle agglutination, negative/positive by ELISA using HTLV-I, and negative by Western blotting and the present assay, were negative by one of the two assays, That is, these 20 sera were negative by at least two of the assays using three different peptides and could be taken to be probably negative. Three serain group 2 (sera212-214), which were positive by gelatin particle agglutination and ELISA using HTLV-I but negative by the present assay; and two sera in group 3-1 (sera 246 and 247), which were positive by gelatin particle agglutination, ELISA using HTLV-I and Western blot- ting but negative by the present assay, were positive by the two assays. That is, these five sera were positive by at least two of the assays using three different peptides and could be taken to be probably positive. Five sera in group 3-2-2 (sera 3 1 1-315), which were positive by gelatin particle agglutina- tion and the present assay, positive or negative by ELISA using HTLV-I and indeterminatehegative by Western blotting, were positive by the two assays. That is, these sera were consis- tently positive by the assays using three peptides and could be taken to be positive with higher probability. (These results will be described in more detail elsewhere.) Thus Ala-Cys- env gp46(237-262) was suggested to be useful, in conjunc- tion with other peptides and recombinant proteins, for improving the reliability of serodiagnosis of HTLV-I infec- tion by separately demonstrating antibodies against different epitopes of HTLV-I.

REFERENCES

1.

2.

3.

4.

5 .

Poiesz BJ, Ruscetti FW, Gazdar AF, Bunn PA, Minna JD, Gallo RC: Detection and isolation of type C retrovirus particles from fresh and cultured lymphocytes of a patient with cutaneous T-cell lymphoma. Pruc NatlAcadSci USA 77:7415-7419, 1980. Hinuma Y, Nagata K, Hanaoka M, Nakai M, Matsumoto T, Kinoshita K , Shirakawa S , Miyoshi I: Adult T-cell leukemia: Antigen in an ATL cell line and detection of antibodies to the antigen in human sera. Proc Narl AcadSci USA 78:6476-6480, 198 I . Gallo RC, Kalyanaraman VS, Samgadharan MG, Sliski A, Vonderheid EC, Maeda M, Nakao Y, Yamada K, Ito Y, Gutensohn N, Murphy S, Bunn PA Jr, Catovsky D, Greaves MF, Blayney DW, Blattner W, Jarrett WFH, zur Hausen H, Seligmann M, Brouet JC, Haynes BF, Jegasothy BV, Jaffe E, Cossman J , Broder S. Fisher R1, Golde DW, Robert-Guroff M: Association of the human type C retrovirus with a subset of adult T-cell cancers. Cancer Res 43:3892-3899, 1983. Okochi K, Sato H, Hinuma Y: A retrospective study on transmission of adult T cell leukemia virus by blood transfusion: Seroconversion in recipients. Vux Sang 46245-253, 1984. Ando Y, Saito K, Nakano S , Kamimoto K, Furuki K, Tanigawa T, Hashimoto H, Moriyama I , Ichijo M, Toyama T: Bottle-feeding can prevent transmission of HTLV-1 from mothers to their babies. J lnfecr 19:25-29, 1989.

Enzyme lrnrnunoassay for Anti-HTLV-I IgG 169

virus infection by using a synthetic peptide-based enzyme immunoas- say. JfnfecrDis 163:41-46, 1991.

16. Ishikawa E, Imagawa M, Hashida S, Yoshitake S, Hamaguchi Y, Ueno T: Enzyme-labeling of antibodies and their fragments for enzyme immu- noassay and immunobistochemical staining. Jlmmunoassay4:209-327, 1983.

17. Hashida S , Ishikawa E: Detection of one milliattomole of ferritin by novel and ultrasensitive enzyme immunoassay. J Biochem 108:960-964, 1990.

18. Kohno T, Kitamura T, Tsuda K, Ishikawa E: Novel and sensitive enzyme immunoassay (immune-complex-transfer enzyme immunoassay) for alpha- fetoprotein from hepatocellular carcinoma. J Clin Lab Anal 4: 183- 188, 1990.

19. Kohno T, Ishikawa E: A highly sensitive enzyme immunoassay of anti- insulin antibodies in guinea pig serum. J Biochem 100: 1247- I25 I , 1986.

20. Ishikawa E, Kato K: Ultrasensitive enzyme immunoassay. Scand J fmrnunol I(Suppl7):43-55, 1978.

21. Inoue S, Hashida S, Kohno T, Tanaka K, Ishikawa E: A micro-scale method for the conjugation of affinity-purified Fab’ to P-D-galactosidase from Escherichin col i . J Biochem 98: 1387-1394, 1985.

22. Imagawa M , Hashida S, Ohta Y, Ishikawa E: Evaluation of P-D-galac- tosidase from Escherichia coli and horseradish peroxidase as labels by sandwich enzyme immunoassay technique. Ann Clin Biochem 21 :3 10- 317, 1984.

23. Hashida S , Imagawa M, lnoue S, Ruan K-h, Ishikawa E: More useful maleimide compounds for the conjugation of Fab’ to horseradish per- oxidase through thiol groups in the hinge. J Appl Biochem 656-63, 1984.

24. Imagawa M, Hashida S, Ishikawa E, Mori H, Nakai C , Ichioka Y, Nakajima K: A highly sensitive sandwich enzyme immunoassay for insu- lin in human serum developed using capybara anti-insulin Fab‘-horseradish peroxidase conjugate. Anal Letr 16:1509-1523, 1983.

25. Maeda Y, Imai J , Kiyokawa H, Kanamura M, Hino S: Performance certification of gelatin particle agglutination assay for anti-HTLV-1 anti- body: Inconclusive positive results. Jpn J Cancer Res 80:91S-9 19, 1989.

6. lkeda M. Fujino R , Matsui T, Yoshida T, Komoda H, Imai I: A new agglutination tebt for serum antibodies to adult T-cell leukemia virus. Gann 75345-848. 1984.

7. Taguchi H, Sawada T, Fujishita M, Morimoto T, Niiya K, Miyoshi I: Enzyme-linked immunosorbent assay of antibodiei to adult T-cell leukemia-associated antigens. Gann 74: 185-187, 1983.

8. Gallo D, Diggs JL, Hanson CV: Comparison of Western immunoblot antigens and interpretive criteria for detection of antibody to human T-lymphotropic virus types I and 11. J Clin Microbiol 28:2045-2050, 1990.

9. Kannagi M, Sugamura K, Sato H , Okochi K, Uchino H, Hinuma Y: Establishment of human cytotoxic T cell lines specific for human adult T cell leukemia virus-bearing cells. Jlmmunol 130:2942-2946, 1983.

10. Miyoshi I , Kubonishi I , Yoshimoto S , Akagi T, Ohtsuki Y, Shiraishi Y, Nagata K, Hinuma Y: Type C virus particles in a cord T-cell line derived by co-cultivating normal human cord leukocytes and human leukaemic Tcells. Narure294:770-771, 1981.

1 I . Kohno H, Kohno T, Sakoda I , Ishikawa E: Novel and sensitive enzyme inmunoassay (immune-complex-transfer enzyme immunoassay) for anti- human T cell leukemia virus type 1 IgG in human serum using recom- binant gag-env hybrid protein as antigen. J Clin Lab .4nal 4:355-362, 1990.

12. Kohno T, Sakoda I, Ishikawa E: Immune complex transfer enzyme immu- noassay for (anti-human T-cell leukemia virus type I) IpG in serum using a synthetic peptide, env gp46(188-209), as antigen. J Clin Lab Anal 5 2 - 3 7 , 1991.

13. Kohno T, Sakoda I , Suzuki M, Izumi A, Ishikawa E: Immune complex transfer enzyme immunoassay for (anti-human T-cell leukemia virus type I) IgG in serum using a synthetic peptide, Cys-gag p19(100-130~, as antigen. JClin LabAnal5:307-316, 1991.

14. Kohno T, Sakoda I, Ishikawa E: Sensitive enzyme immunoassay (immune complex transfer enzyme immunoassay) for (anti-human T-cell leuke- mia virus type I ) IgG in serum using a synthetic peptide, Cys-env gp46( 188-224), as antigen. J Clin Lub Anal (in press).

15. La1 RB, Rudolph DL, Lairmore MD, Khabbaz RF, Garfield M, Coligan JE, Folks TM: Serologic discrimination of human T cell lymphotropic