Specific assay for endotoxin using immobilized histidine and Limulus amebocyte lysate

ANALYTICAL

BIOCHEMIS’TRY

198,%X!-297

(1991)

Specific Assay for Endotoxin Using Immobilized Histidine and Limulus Amebocyte Lysate Satoshi

Minobe,l

Masahiro

Nawata,

Taizo

Watanabe,

Tadashi

Sato, and Tetsuya

Tosa

Department of Biochemistry ResearchLaboratory of Applied Biochemistry, Tanabe Seiyaku Company, Limited, 16-89, Kashimu-3-chome,Yodogawa-ku, Osaka532, Japan

Received

May

29,199l

The LAL test is inhibited or enhanced by many substances. To overcome these problems, we have developed a specific endotoxin assay method using an ultrafiltration unit, a fluorometric LAL reagent, and immobilized histidine (which is a specific adsorbent for endotoxins). This method is composed of two steps. The first step is the adsorption of endotoxins. Using immobilized histidine, endotoxins are quantitatively adsorbed on the adsorbent, and the adsorbed endotoxins are separated from LAL-inhibiting or -enhancing substances by the ultrafiltration unit. The second step is the reaction of adsorbed endotoxins with the LAL reagent. The endotoxins adsorbed on immobilized histidine are directly reacted with the LAL reagent in a filter cup and show enough activity for assay. The reproducibility and the accuracy of this method are high, and the recovery of endotoxins from a sample solution is more than 96%. The new endotoxin assay method using immobilized histidine can be utilized for the determination of endotoxins in a solution containing LAL-inhibiting or -enhancing substances such as amino acids and antibiotics instead of requiring employment of the more common gel-clot technique. Q lssl Academic press, I~C.

Pharmaceutical products intended for parenteral administration are tested for pyrogenic contamination, which usually takes the form of bacterial endotoxin (lipopolysaccharide; LPS2) from Gram-negative bacteria. There are two methods for determining the presence of r To whom correspondence and reprint requests should be addressed. 2 Abbreviations used: LPS, lipopolysaccharide; LAL, Limulus amebocyte lysate; USP, United States Pharmacopoeia; Boc-Leu-GlyArg-MCA, tert-butyloxycarbonyl-L-leucyl-glycyl-L-ar~nine-4-methylcoumaryl-7-amide; AMC, 7-amino-4-methylcoumarin; NMWL, nominal-molecular-weight limit; EU, endotoxin unit; CV, coefficient of variation.

endotoxins. One is the pyrogen test using rabbits, and the other is the LAL test using Limulus amebocyte lysate (LAL). The LAL test was introduced by Levin and Bang in 1968 (1). The advantages of the LAL test over the rabbit pyrogen test are a higher sensitivity, lower cost, and higher reproducibility. In the LAL test several techniques such as the gel-clot, turbidimetric, chromogenic, and fluorometric are known. In 1980, the gel-clot technique was approved by the USP. In 1985, the turbidimetric technique and the chromogenic technique were also approved by the USP. In 1968, the gel-clot technique was approved by the Japan Pharmacopoeia only for water for injection. The principle of the LAL test for the determination of the presence of endotoxins is based on the endotoxininduced coagulation reaction (2). However, with amebocyte lysate, a @-1,3-D-glucan-induced cascade also exists. This cascade causes a false positive result in the LAL test. Moreover, the LAL test is inhibited or enhanced by many substances such as antibiotics, hormones, heavy metals, amino acids, alkaloids, carbohydrates, plasma proteins, enzymes, and electrolytes in sample solution (3). Many attempts have been made to overcome these problems. A new endotoxin-specific test, from which factor G is removed, has been developed by Obayashi et al. (4). Also, a number of possibilities such as dilution, heating, dialysis, ultrafiltration, and the addition of detergents to eliminate disturbing factors have been described (5). These methods are effective under some conditions. However, they do not fully satisfy the basic requirement for the elimination of interference, especially when endotoxin contaminations are at very low levels. For example, in the dilution method, the sensitivity decreases because the endotoxin is also diluted. The ultrafiltration method cannot be applied to high-molecular-weight interference substances, and the endotoxin might be adsorbed on the ultrafiltration membrane. Thus, to overcome these problems, we

292 All

0003-269’7/91 $3.00 Copyright 0 1991 by Academic Press, Inc. rights of reproduction in any form reserved.

SPECIFIC

ASSAY

have investigated a new specific assay method for endotoxins using immobilized histidine. Immobilized histidine is a specific adsorbent for endotoxins that we developed (6). This adsorbent is produced by the covalent binding of histidine on agarose through hexamethylenediamine and can be applied to remove endotoxins in various solutions (6,7). In this paper, we describe the optimum conditions for an operation, the reproducibility and the accuracy of this method, and the assay of endotoxins in various substances by this method.

MATERIALS

AND

METHODS

LPS (Escherichiu coli 0128:B12) was purchased from Difco Labs (Detroit, MI); U.S. standard endotoxin (Lot EC-5), Limulus II single-test Wako (LAL, manufactured by Associates of Cape Cod, Inc.), Limulus amoebocyte lysate HS II (manufactured by Associates of Cape Cod, Inc.), and Curdlan ((3-l,%D-ghCan) were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Boc-Leu-Gly-Arg-MCA and AMC were purchased from Peptide Institute (Osaka, Japan). The Toxicolor test was purchased from Seikagaku Kogyo Co., Ltd. (Tokyo, Japan). Immobilized histidine (sold as PyroSep) was produced from Tanabe Seiyaku Co., Ltd. (Osaka, Japan). Ultrafree CL-THK (a unit composed of a filter cup with an ultrafiltration membrane, 100,000 NMWL, and a tube) was purchased from Nihon Millipore, Ltd. (Tokyo, Japan); Immersible CX-10 ultrafiltration units were purchased from Millipore Corp. (Bedford, MA). The sterile test tube (12 X 75 mm) was purchased from Iwaki Glass (Tokyo, Japan). Crystalline penicillin G potassium was purchased from Meiji Seika Kaisha, Ltd. (Tokyo, Japan); pyrogen-free water, Amizet (amino acidinfusion), L-phenylalanine, L-methionine, L-cysteine hydrochloric acid, and L-a&nine hydrochloric acid were the products of Tanabe Seiyaku Co., Ltd. (Osaka, Japan). All other chemicals were of analytical reagent grade.

Endotoxin Adsorption on Immobilized Histidine

Into a sterile test tube, 30 mg of wet-type immobilized histidine (the dry weight per 1 g of wet-type immobilized histidine is equivalent to 60 mg) and 2 ml of LPS (E. coli 0128:B12) solution (0.27 EU/ml) were added, and the suspension was stirred at 50 rpm by a round mixer (endto-end mixer) at 25°C for 15-120 min. After being stirred, the tube was centrifuged (23OOg, 5 min), and the concentration of endotoxin in the supernatant was measured by the chromogenic technique (8), using the Toxicolor test.

FOR

293

ENDOTOXIN

Depyrogenation

of Filter Cup

A filter cup was soaked in 5% hydrogen peroxide solution and then heated at 70°C for 3 h. After heating, the filter cup was washed thoroughly with pyrogen-free water and dried at 60°C

Preparation of Pyrogen-Free Sample Solution

Removal of pyrogen from various solutions was carried out as follows. In 20 ml of a sample solution, 0.3 g of wet-type immobilized histidine was suspended, and the suspension was stirred at room temperature (25°C) for 4 h. After being stirred, the adsorbent was separated by centrifugation (23OOg, 5 min).

Assay of Endotoxin H&dine

Using Filter Cup and Immobilized

(1) Adsorption. Into a filter cup, lo-100 mg of wettype immobilized histidine and l-2 ml of sample solution were added, and the suspension was stirred at 25°C for 60 min. After being stirred, the filter cup was centrifuged at 2300g for 15 min and the filtrate was discarded. The adsorbent was washed with 25 mM sodium chloride solution and/or water. (2) LAL test. Synthetic substrate (Boc-Leu-GlyArg-MCA) was dissolved in water at a concentration of 0.6 mM, and its contaminating pyrogen was removed by ultrafiltration with Immersible CX-10. The contaminating pyrogen in Tris-HCl buffer (0.4 M, pH 8.0) containing 0.04 M magnesium chloride was also removed by the same method. One vial of Limulus II single-test Wako, which gels with 0.2 ml of Japanese standard endotoxin solution (0.25 EU/ml), was dissolved in 1.5 ml of the above buffer. To the adsorbent on which endotoxins were adsorbed in the filter cup, 100 ~1 of Limulus II single-test Wako solution and 200 ~1 of substrate solution were added, and the suspension was incubated at 37°C for lo-60 min. After incubation, the reaction was stopped by adding 2.3 ml of 12.5% acetic acid and the mixture was centrifuged at 2300g for 30 min. The fluorescence intensity of the filtrate at 460 nm, excited at 380 nm, was measured.

Measurement of Gel-Clotting Activity

The sample solution was diluted with pyrogen-free water. One vial of Limu1u.s amebocyte lysate HS II was dissolved in 5 ml of pyrogen-free water. One hundred microliters of sample solution and one hundred microliters of Limulus amebocyte lysate HS II solution were incubated at 37°C for 60 min. After the incubation, the gel clotting in the tube was examined.

294

MINOBE

0

30 Stirring

60

90 time

120

I 150

(min)

FIG. 1. Time

course of endotoxin adsorption on immobilized histidine. Endotoxin (LPS, E. coli 0128:B12) adsorption on immobilized histidine was carried out by the batchwise method described in the text. 0, In the absence of the adsorbent; 0, in the presence of the adsorbent.

RESULTS

Optimum

Conditions

for an Operation

of Adsorption

The adsorption capacity of immobilized histidine is 0.74 mg of LPS (E. coli 0128:B12) or 0.31 mg of LPS (E!. coli UKT-B) per 1 g of wet-type immobilized histidine. The dissociation constant of immobilized histidine for LPS (E. coli 0128:B12 and E. cob UKT-B) are 1.57 X lo-’ and 7.3 X lo-l3 M, respectively, when the molecular weight of LPS is taken as l,OOO,OOO (6,7). For a quantitative assay of endotoxin using immobilized histidine, it is necessary that endotoxins be adsorbed on immobilized histidine quantitatively and that the endotoxins adsorbed on immobilized histidine show enough activity. First, the optimum conditions for endotoxin adsorption on immobilized histidine were investigated. The time course of endotoxin adsorption on immobilized histidine is shown in Fig. 1. The endotoxin concentration in the supernatant decreased with increasing stirring time. After 60, 90, and 120 min of stirring, almost all endotoxins were adsorbed on the adsorbent. Furthermore, by the additional experiments, the optimum conditions for adsorption were determined as follows: wet-type immobilized histidine, 20-50 mg; pH 6-7; ionic strength, O-0.05 T/2; sample volume, l-3 ml; and more than 60 min of stirring time for adsorption. Optimum

Conditions

for LAL Test

The relationship between reaction time and relative intensity of fluorescence during LAL reaction is shown in Fig. 2. Endotoxins adsorbed on immobilized histidine were incubated with LAL reagent. As the reaction time increased, the relative intensity increased. When the re-

ET

AL.

action time was more than 30 min, the blank value was also high. On the other hand, when the reaction time was less than 20 min, the sensitivity was low. Therefore, the optimum reaction time is 20-30 min. The optimum conditions of other factors for LAL test were determined by additional experiments to be as follows: the enzyme used was Limulus II single-test Wako, the substrate used was 0.34 mM Boc-Leu-Gly-Arg-MCA, the reaction volume was 300-350 ~1, and the measurement was carried out using the fluorescent technique. From these results, the standard operation of this method was arranged as follows. Into a filter cup, 100 ~1 of 30% immobilized histidine suspension and 2 ml of sample solution were added, and the suspension was mixed at 50 rpm and at 20-25°C for 60 min in a round mixer. After being mixed, the filter cup was centrifuged at 2300g for 15 min and the filtrate was discarded. For washing the adsorbent, 0.5 ml of 25 mM sodium chloride solution was added and mixed by a vortex mixer. The filter cup was centrifuged at 2300g for 10 min and the filtrate was discarded. For further washing, 2 ml of water was added and mixed by a vortex mixer. The filter cup was centrifuged at 2300g for 25 min and the filtrate was discarded. For LAL reaction, 100 ~1 of LAL solution was added at 4°C and mixed gently, and then 200 ~1 of substrate solution was added at 70°C and mixed gently. The suspension was incubated at 37°C for 25-30 min. After incubation, the reaction was stopped by adding 2.3 ml of 12.5% acetic acid solution and the mixture was centrifuged at 2300g for 30 min. The fluorescence intensity of the filtrate at 460 nm, excited at 380 nm, was measured. In practice, the sensitivity of the endotoxin assay could be changed by changing the incubation time with

*

.=e

760-

2 .c 8

wo-

5:

400-

g 3 E

wo-

e ‘G 2 d

WO-

zoo‘fJoO-

0 FIG. 2.

10

20 30 40 50 Reaction time (min)

60

70

LAL reaction and endotoxins adsorbed on immobilized histidine. Endotoxin adsorption was carried out as described in the text using 30 mg of wet-type immobilized histidine and 2 ml of endotoxin solution (LPS, E. coli 0128:B12: n , 0.4 EU/ml, 0, 0.02 EU/ml, 0, blank) containing 25 mu sodium chloride, and LAL reaction was carried out as described in the text at various incubation times. The relative fluorescence intensity of 2 PM AMC was 150.

SPECIFIC TABLE Reproducibility

of the

ASSAY

Relative

intensity

Mean value SD cv (%)

295

ENDOTOXIN

-

200

1 Adsorption

P

Method

P E

Endotoxin added (EU/ml) Items

FOR

150

E

0

0.25

H

100

15 17 14 11 16 14 16 14 14 15 14.6

176 181 168 168 188 178 161 175 185 170 175

5 F 2al

50

1.6 10.7

a 2 pM AMC showed 150. Note. Endotoxin (EC-5) solutions containing 25 mM ride were assayed by the standard operation (n = 10).

3 d 0 0.00

Endotoxin

chlo-

the LAL reagent. With an unknown sample, many dilutions of the sample would be used. As regards washing the adsorbent, washing with water only is sufficient if the sample is not adsorbed to immobilized histidine; otherwise, washing with sodium chloride solution is necessary. The effect of washing increases with increasing wash volume.

0.30

concentration (Eltlml)

The reproducibility of this assay method is shown in Table 1. When endotoxin was not added, the mean value of the relative intensity, blank value, of 10 times assay was 14.6, and the CV was 10.7%. When endotoxin was added at a concentration of 0.25 EU/ml, the mean value of the relative intensity of 10 times assay was 175, and the CV was 4.5%. Curve

The relationship between endotoxin concentration and relative fluorescence intensity was investigated. As shown in Fig. 3, the calibration curve obtained between 0 and 0.3 EU/ml was well correlated. Recovery of Endotoxins in Several Substances by the Adsorption Method The recovery of endotoxins added to the solutions containing several substances was investigated. To endotoxin-free solutions, 0.2 EU/ml of EC-5 was added, and then endotoxin concentrations in these solutions

(ECby the

were determined by this assay method. As shown in Table 2, in every solution tested except 0.1 M sodium chloride solution, the recovery of endotoxin was high. In 0.1 M sodium chloride solution, the recovery was low. This was due to a low adsorption of endotoxin on immobilized histidine because of the high ionic strength of the solution containing endotoxin (6,7). Assays of Endotoxins in Various Substances by the Adsorption Method Endotoxins in various substances were assayed by the adsorption method. Table 3 shows a comparison of this TABLE

Reproducibility

Standard

0.20

FIG. 3. Standard curve of the adsorption method. Endotoxin 5) solution containing 25 mM sodium chloride was assayed standard operation method.

El 4.5

sodium

0.10

Recovery

2

of Endotoxins in Several by the Adsorption Method

Substances

Substance Name NaCl NaCl NaCl Glucose Glucose Phenylalanine Methionine A&nine-HCl Cysteine-HCl Penicillin G potassium Streptomycin sulfate Amino acid infusion Amino acid infusion @-1,3-D-Glucan

Concentration 0.025 M 0.05 M 0.1 M 5% 20% 2% 5% 1% 0.25% 50,000 U/ml 1% Original 4X dilution 2 rig/ml

Recovery 60) 98 95 46 117 127 98 98 123 109 116 105 125 96 120

Note. To endotoxin-free solutions containing various substances, 0.2 EU/ml of EC-5 was added, and then the endotoxin concentration in these solutions was assayed by the adsorption method as described in the text.

296

MINOBE TABLE Assays of Endotoxins

in Various

ET AL. 3

Substances

by the Adsorption

Method

Substance Adsorption Name W I-b.0 NaCl Phenylalanine Methionine Cysteine-HCl Arginine-HCl Amino acid infusion Penicillin G potassium /3-1,3-D-Glucan

method

Gel-clotting

activity”

Concentration

Endotoxin added (EU/ml)

Dilution

EU/ml

Dilution

Clotting

-

0.2 0.2

-

-

6x 12x

(+) t-1

0.5 M 2% 5% 1.5% 1% Original 50,000 U/ml 2 rig/ml

0.5 0.1 0.1 0.5 0.05 0.2 0.2 0.2

10x 1x 1x 10x 1x 4x 4x 1x

0.48 0.11 0.11 0.54 0.05 0.20 0.22 0.23

15x 3x 3x 15x 1x 6x 6X 10x

(+) (-Jb (-lb (-lb (+) (-jb (+I (+I"

0 Lysate sensitivity, 0.03 EU/ml. b False negative. ’ False positive. Note. EC-5 was added to endotoxin-free solutions containing adsorption method and the gel-clot technique were applied.

various substances, and then after dilution

adsorption method with the gel-clot technique for endotoxin assay. In the gel-clot technique, phenylalanine, methionine, cysteine hydrochloric acid, and amino acid infusion showed false negative results because these substances inhibit the LAL procedure. On the other hand, p-1,3-D-glucan showed a false positive result because it activates the LAL procedure. With the adsorption method, the concentration of endotoxin could accurately be assayed in every case. This indicates that the new adsorption method using immobilized histidine is a favorable assay method for endotoxins in solutions containing LAL-inhibiting or -enhancing substances. DISCUSSION

The LAL test is widely used to assay for bacterial endotoxins because of its advantages. However, the LAL test has some shortcomings, such as false negative results in the presence of LAL-inhibiting substances and false positive results in the presence of p-l,S-D-glucan or LAL-enhancing substances. Although there have been many attempts to eliminate this interference, no attempt using a specific adsorbent for endotoxin has been made. For several years, we have carried out studies on the preparation of adsorbents for the removal of pyrogen and their applications. In so doing, it was found that immobilized histidine has a high affinity for endotoxin at a low ionic strength and over a wide pH range (6,7). We considered that if this adsorbent could adsorb endotoxin quantitatively and the endotoxins adsorbed on the adsorbent could activate the LAL procedure, this adsorbent could be used for a specific endotoxin assay.

with endotoxin-free

water, the

Therefore, we first investigated the possibility of the quantitative adsorption of endotoxin on immobilized histidine, and, as shown in Fig. 1, it became clear that endotoxins were adsorbed on immobilized histidine quantitatively. Next, we investigated the activity of the endotoxins adsorbed on the adsorbent, and, as shown in Fig. 2, it became clear that the endotoxins adsorbed on immobilized histidine showed sufficient activity. We believe that the endotoxins adsorbed on the adsorbent were liberated from the adsorbent in the LAL reaction solution because the pH and ionic strength of the reaction solution were high. When polymyxin-Sepharose (9) was used instead of immobilized histidine, the endotoxins adsorbed on the adsorbent could not activate the LAL procedure. In this paper, we used E. coli 0128:B12, LPS, and EC-5 as endotoxins. Moreover, it is believed that many kinds of endotoxins can be determined by this assay method because immobilized histidine adsorbs endotoxins originating from various Gram-negative bacteria (7). As regards the specificity of adsorption, many substances other than endotoxin are not adsorbed on immobilized histidine at low ionic strength and around neutral pH, and most substances do not interfere with the adsorption of endotoxin. However, some acidic substances are adsorbed on the adsorbent under these conditions. These substances are only slightly adsorbed on the adsorbent at high ionic strength (0.05-0.1 I’/2) (7). Therefore, it may be necessary to choose the proper adsorption conditions for some samples. We used an ultrafiltration membrane unit as a test tube because the separation of the adsorbent from a

SPECIFIC

ASSAY

solution was easy. It is possible to retain liquids in the unit by using this ultrafiltration membrane unit and, if necessary, to remove them by centrifugation. This new endotoxin assay method showed good recovery, reproducibility, and accuracy. Furthermore, when endotoxins were added to the solutions containing LAL-inhibiting or -enhancing substances, such as amino acids and P-1,3-D-glucan and were assayed by this method, the concentration of endotoxin could be accurately assayed. On the other hand, when the gelclot technique was used, many solutions showed false negative or false positive results (Table 3). Thus, this new assay method using immobilized histidine could be used widely as a means of assaying endotoxins in solutions containing LAL-inhibiting or -enhancing substances. However, this method also has some shortcomings: the operation time is long, the operation is complicated, the chromogenic technique cannot be applied because the ultrafiltration membrane adsorbs substrate, and high-molecular-weight substances cannot be applied because the pore size of the membrane is small. Further studies on the assay method using immobilized histidine with the goal of overcoming these shortcomings are in progress.

FOR

297

ENDOTOXIN

ACKNOWLEDGMENT We thank

M. Hase

for her technical

assistance.

REFERENCES 1. Levin, J., and Bang, 186-197. 2. Morita, T., Tanaka,

F. B. (1968) S., Nakamura,

Thromb. Diuth. Haemorrh. S., and

Iwanaga,

19,

S. (1981)

FEBS Lett. 129,318321. 3. Pfeiffer,

4.

M., and Weiss, A. R. (1987) in Detection of Bacterial Endotoxins with the Limulus Amebocyte Lysate Test (Stanley, W. W., Levin, J., and Novitsky, T., Eds.), pp. 251-262, Liss, New York. Obayashi, T., Tamura, H., Tanaka, S., Ohki, M., Takahashi, S., Arai, M., Masuda, M., and Kawai, T. (1985) Clin. Chim. Actu

149,55-65. 5. Zimmermann, G. (1985) Phnrmucol. Ind. 47,203-207. 6. Minobe, S., Watanabe, T., Sato, T., and Tosa, T. (1988) Biotechnol. Appl. Biochem. 10,143-153. 7. Matsumae, H., Minobe, S., Kindan, K., Watanabe, T., Sato, T., and Tosa, T. (1990) Biotechnol. Appi. Biochem. 12,129-140. 8. Iwanaga, S., Morita, T., Harada, T., Nakamura, S., Niwa, M., Takada, K., Kimura, T., and Sakakibara, S. (1978) Haemostusis ‘7,183-188.

9. Umeda, M., Asakawa, Organs 12.275-278.

M.,

and

Terada,

T. (1983)

Jpn. J. Artif.