APPLIED MICROBIOLOGY, June 1975, p. 713-716 Copyright 0 1975 American Society for Microbiology

Vol. 29, No. 6 Printed in USA.

Semiautomated Microbiological Method for the Assay of Quaternary Ammonium Compounds T. M. BERG%

Organon International B. V., Oss, The Netherlands Received for publication 10 January 1975

A semiautomated method has been developed for quantitatively assaying the activity of detergent disinfectants. Automation permitted a high level of reproducibility, which in turn allowed a meaningful comparison between the activities of some types of quaternary ammonium compounds. survivors were determined in duplicate, using the pour-plate method (soybean casein digest agar, U.S.P. XVIII). A heat-sterilized peptone (0.1%)saline (0.85%) solution was used as dilution liquid. For all pipetting procedures Eppendorf pipettes (Germany) were used. The most reproducible rates of killing were obtained using inocula that had been stored in liquid nitrogen (4). The stock cultures were prepared from 8-h (log-phase) cultures of the respective test organisms grown in nutrient broth at 37 C. Immediately prior to a test, the ampoules containing approximately 5 x 108 cells/ml in peptone-saline were taken out of the liquid nitrogen and placed in water at 35 C for exactly 1 h. A test culture was prepared by diluting the contents of an ampoule with the peptone-saline solution to achieve a concentration of approximately 107 viable cells/ml. An apparatus was developed (Fig. 1) consisting of a conveyor belt fitted with a sampling needle connected MATERIALS AND METHODS to a dilution unit (for mixing the test culture with the Four types of benzalkoniumchloride were studied: disinfectant) and a fixed needle connected to a (i) BTC 2125 (Millmaster Onyx Corp.), which con- dispensing unit (for the addition of the neutralizing tains 25% n-alkyl (60% C14, 30% C16, 5% C12, 5% C18) liquid). All of these units were electrically linked to dimethyl benzyl ammonium chlorides and 25% n- the control unit, which was provided with an adjustaalkyl (50% C12, 30% C14, 17% CIO, 3% C18) dimethyl ble electromechanical timeclock (Schleicher, W. Berethylbenzyl ammonium chlorides; (ii) Nouryquat Duo lin, Germany) controlling the treatment time. The (Akzo), which has- a composition similar to (i); (iii) conveyor belt, the racks, and the sampling needle BTC 824 (Millmaster Onyx Corp.), which contains were constructed according to a design described 50% n-alkyl (60% C14, 30% C1,i 5% C12, 5% C18) earlier (1, 2). The dilution and dispensing units were dimethyl benzyl ammonium chlorides; (iv) U.S. constructed from Cornwall valves and 1- and 10-ml Pharmacopeia benzalkoniumchloride reference stand- glass syringes (Becton-Dickinson & Co., Rutherford, ard, which contains 0.4% C. 0.7% C1,, 55% C12, 34% N.J.). The apparatus was designed to assay consecuC14, 9.7% C1, and 0.9% C is Test organisms included tive samples automatically. In addition, it was in these studies were Pseudomonas aeruginosa (ATCC provided with press buttons for separate operation 9024), Staphylococcus aureus (ATCC 6538), and of the diluter and dispenser and an alarm system. The treatment procedure is outlined schematically Escherichia coli (ATCC 10536). At the start of the test all liquids were equilibrated in Fig. 2. Tube A is filled with 5 ml of an appropriate at room temperature. For each assay 1 ml of the dilution of disinfectant. One milliliter of this sample disinfectant sample was mixed in a culture tube with is sucked up by the large sample needle of the dilution 4 ml of test culture that was prepared as described in unit. The needle is lifted and tube B is transported to the next paragraph. After a 30-s treatment period, 5 a position under the needle; the needle then delivers ml of a heat-sterilized lecithin-Tween 80 suspension the 1-ml sample together with 4 ml of a test culture. (soya lecithin, 5.3 g; Tween 80, 37.3 g; KHP04, 0.5 g; After 30 s of treatment, 5 ml of neutralizing liquid is and distilled water to 1,080 ml) was added to neutral- added via the small fixed needle of the dispensing ize the benzalkoniumchloride. Viable counts of the unit. 713

To lay down precise specifications for a disinfectant (especially one which is chemically poorly defined), quantitative data are required regarding the biological activity of the compound. For this purpose reliable microbiological assay methods are necessary. In the commonly used tests for establishing the efficacy of disinfectants, both time and concentration are used as variables. Usually, however, these tests are not sufficiently reproducible to be used for quantification of the potency. Thus, during the development of an assay method for benzalkoniumchloride, we attempted to devise a sensitive test system that was highly reproducible and which could be standardized.

714

BERG

|-0 X.--0

FIG. 1. Apparatus for disinfectant Xassays. Two racks with tubes are shown on the conve A flask containing the test culture and a magneticc is placed on a magnetic stirrer (right). Or flask containing the neutralizing liquid. The control unit is housed behind the front panel, ccarrying: 1, 2, power light and switch; 3, press button dlilution unit; 4, press button dispensing unit; 5, alarim switch; 6, fixed needle connected to dispensing unit ; 7, sampling needle connected to dilution unit; 8, himeclock for adjusting treatment time. Just visible i s 9, a 10-mI syringe plus a valve from dilution unit.

syorr therleng bsar

4 ml

t1 ml

LI

1 ml

I I

A

B

A

TREATMENT

30

SAMPLE

B SAMPLE

CULTURE

CULTURE

B

A

cluded maximally four samples of unknowns together with one sample of the reference standard BTC 2125. Each sample was represented by triplicate high and low dose levels. One assay thus comprised 30 cultures for viable count determination. The viable counts were used for calculation of the potencies of the unknowns. The potencies were expressed in micrograms of reference standard. The calculation was automated by using a Wang 700 desk computer provided with an IBM electric typewriter. An example of such a calculation for three batches of Nouryquat Duo is given in Fig. 3.

RESULTS Using the semiautomated procedures, we prepared log dose-response curves of BTC 2125 against the three test organisms (Fig. 4). As could be expected, P. aeruginosa was the least sensitive test organism. E. coli yielded the curve and was therefore used as test organism for the microbiological assay. Dose-response curves for BTC 2125 and Nouryquat Duo against E. coli did not show

steepest dose-response

significant differences. The reproducibility and accuracy of the method were demonstrated by determining the potencies of an identical series of samples in seven different assays. Included was a solution of the reference BTC 2125 itself and samples of batches of Nouryquat Duo produced at two

4 l

LI

I SAMPLE DISINF.

5 ml

APPL. MICROBIOL.

+

NEUTR. FIG. 2. Scheme of assay procedSure. Intensive homogenization of the conitents of the tubes is effected by using dispensing neeedles with a small (0.7-mm, inner diameter) bore tip. Apart from the treatment period, the volumeMs of sample, test culture, and neutralizing liquid also can be varied by adjustment of the dilution an d dispensing units. The quantitative method for the deteLrmination of the benzalkoniumchloride was designed according to the two-point parallel line assay (3). 2An assay in-

factories (locations I and II). The most reproducible results (Table 1) were found for the BTC 2125 references tested as sample (coefficient of variation, 3.2%). Only slightly larger coefficients of variation were obtained for Nouryquat Duo produced at locations I and II (4.5 and 5.6%). Potencies calculated for these two samples ranged from 90 to 111%. During a period of 3 years routine testing of batches of Nouryquat Duo (Table 2) was carried out first by the "manual" and later by the semiautomated method. The production methods for Nouryquat Duo had not changed

during this period. Hence, the decrease of the

coefficient of variation from 6.8 to 3.4% is probably due to the improved reproducibility of

the semiautomated assay method. To check whether different types of benzalkonium-chloride produced different slopes, BTC 2125, BTC 824, and U.S.P. benzalkoniumchloride reference standards were tested against E. coli and S. aureus. Against both test organisms BTC 2125 was effective in lower concentrations than BTC 824 and benzalkoniumchlo'ide U.S.P. As the slopes of the parallel portions of the death rate curves did not differ significantly between the various types of benzalkoniumchloride, the potencies (expressed in mi-

Completely

-~.Jw '~ ~.'

randomi ed (2

* 2 + 2 + ..)-point parallel line assay with equal dose ratios for each dose of each Unknown.

inU,

Ztandard dose and

Experimental

unit:

colooy £f1i

LA (mn of tunvishle c os -m

Response variable y' =

H/L,

with

PnS experimental units for Date:

Particulars:

aiiU A

(0 means y = y'; 1 means y = A log(y'+B) + D, with A =

'Number of Unknowns r = i

= 6

nS

nU

(am mple)

home ref.

(Assumed) potency Step

1

Step

2

Dose ratio HA = d = 1*6D

Standard S= U1grC 2125 U= e2= STC 2125 hme ref. Iouryquat Louryquat

Preparation p

U.

U4

nouryqust

Duo B

Duo A

, D =

B =

i = log d

U7

U6

Duo C

_ .-

_

Step 3

-1 0 -miL 40.0000

Volume for dose H

Dose H Treatment t

--lO-aL-.

--1.0-mi--

-.1.0-wl--

SL

SH

.6.

5.0

40 0000 40.0000 40.0000 Pst------ _ 1H ?L 2H 3L 3H 1L 6.1 '5S.0. 6 .3' 5..0 6.5 S.D

6.0

4-9 h.5.5.1 _6A3.

6,3

50

60p

4,9

l---i--

--L.

40 0000 ---_

--

4.9

6..3

144--1974...

)

Program 7 (7). Transformation code number= 0 = 1

each

6,3

4L

4H

60

1.0

5L

6-,

5H

6H

7L

7H

Q..........

IS 63.

6.4 4.9 6,0 3,Q

4,9

y'-values4 -------

observed

w

yr-totala T

TT T H,p L,p

b,p

bp

=

2T

18.9 :14.9 18.3 14.8 19.1 15.0

Tt

In i b,p P

J

number of substituted y-values

V

=

V4F(n5s+nu)var / nSnUi2

n

e

Unknowns with

=

rnU+nS-2r-2

lying outside

bU

20

+ V

bS

'15.1 18.6 14.9 -6.0

-6.0

-6.6

-S 7

-6.5

=

18.8

[ZY2 -( t/n ) / 9

var

4.02 and -5.03

=

are

F

0

F1,;0.05

435

rejected for significant difference in

regression with Standard. For their potencies no confidence limits are calculated. For the

non-rejected preparations holds: and nmber of Unknowns a r' Tb = £T b,p 4' HN nU1rn5 b

N

=

=

l

(*L~g-(C1

2Tb / N'i

iN'(nsTu

{SC

n.TS)

=

dose SN / dos

as =

(is

0.95

the

ame

Te- / (Te

-

FN'var)

r

=

7178

7174

.024

.024

.024

1.044

.981

.993

1.018

1.0000

1.0000

1.0000

1.0000

.024 N

numerical value

antilog

/ bse00ef.

.0203

7174

7154

=

{

=

.008

F/(IL)2

=

being S 0, the assay is not rejected for

regression and the calculations are continued

-.002

-.008

w

C not

1.0058

A= s/IbI

.018

i

=

.1264

/ 2nSTb

__

I

UK

units potency

=

non-significant comon

-6.2

=

R

K

C

units)

wv ^

mvp R x K

potency1

=

[antilog(CH-IL)]

x K

potencyh

=

[antilog(CNtIL) ]

x

j1 11

wvpg 9

1.04

1.02

1.04.93

.94

.96

1.04

1.05

1.08

confidence

interval

Calculatd by

and

K

1.11 Date

FIG. 3. Example of calculation form for the evaluation of a BTC 2125-Nouryquat Duo assay, including four unknowns. Steps 1, 2, and 3, dilution steps necessary to obtain dose H; dose H, high dose per preparation; dose L, low dose per preparation; THP, total of the y-values of the high dose per preparation; TL.P, total of the y-values of the low dose per preparation; bp, regression coefficient per preparation; 4'e, degrees of freedom for error; var, estimated error variance; F1 4e;O05O F value with 1 and 4e degrees of freedom for numerator and denominator, respectively, and 0.05 probability of exceedance; b, mean regression coefficient for the standard preparation and the nonrejected unknown preparations; L, width of 95% confidence interval for the true log potency; w, weighing factor for combining a log potency found in this assay with log potencies in similar assays of the same preparation. 715

716

BERG

APPL. MICROBIOL.

TABLE 2. Potencies of 51 batches of Nouryquat Duo produced at location I and assayed by using manual and semiautomated methods Method

C~~~~~

so

6¢6-

Manual Semiautomated

%

0)

.,

"a\.\%

K

5

No. of of samples

No.le

Mean potency0

oMeany

Coefficient of variation (%)

28 23

0.94 1.01

6.8 3.4

4)

a Expressed in micrograms of home reference BTC 2125.

3,

10

100

50

Rg/mI

Concentration of BTC 2125

FIG. 4. Log dose-response curves of BTC 2125 prepared with three test organisms. Symbols: P. aeruginosa ATCC 9024; -----, S. aureus ATCC 6538; --, E. coli ATCC 10536. TABLE 1. Potenciesa of one batch of benzalkoniumchloride BTC 2125 and two batches of Nouryquat Duo determined by seven assays performed on different days Day

BTC 2125

Nouryquat

Nouryquat

home

Duol

Duo II

1.05 1.03 1.02 1.02 1.00 0.97 1.07

0.99 0.96 0.98 0.98 1.11 0.95 1.03

0.96 0.90 0.94 0.96 1.04 0.94 0.99

1.02

1.00

0.96

3.2%

4.5%

5.6%

reference

1 2

3 4 5 6 7 Mean

Coefficient of variation

a Expressed in micrograms of home reference BTC 2125.

example, benzalkoniumchloride. The method is based on the relationship between concentration and the death rate of a sensitive bacterial population that is treated for a short, fixed time period. E. coli was a sensitive test organism, yielding reliable results, especially when preparation of the inocula was standardized. The relatively short duration (20 min) of one complete assay guarded against eventual changes in sensitivity of the test culture. The relation between death rate and concentration was characteristic for each type of benzalkoniumchloride. Nevertheless, the log doseresponse curves were parallel, allowing relative potencies to be calculated. Preliminary experiments with another disinfectant for which no suitable neutralizers exist indicated that, in this semiautomated procedure, dilution can also be used to effect neutralization. It is therefore suggested that the rapid, reproducible, and accurate microbiological assay methods described here can be modified for the evaluation of other, nonquaternary types of disinfectants. ACKNOWLEDGMENTS I wish to thank A. van der Veen for technical assistance and J. A. M. van Oorschot for assistance in the design and construction of the equipment. I also am grateful to F. J. Verbon for the statistical design of the assay. LITERATURE CITED

of home reference BTC 2125) could be calculated. Surprisingly, identical potencies were found for the U.S.P. reference benzalkoniumchloride and BTC 824 against both test organisms, E. coli and S. aureus, i.e., 0.9 and 0.75 jig, respectively. crograms

DISCUSSION An accurate and reproducible assay method has been developed that is suitable for disinfectants which possess detergent properties, for

1. Berg, T. M. and H. A. Behagel. 1972. Semiautomated method for microbiological vitamin assays. Appl. Micro-

biol. 23:531-542. 2. Berg, T. M., J. M. den Burger, and H. A. Behagel. 1975. Semiautomated method for microbiological vitamin and antibiotic assays, p. 1-22. In R. G. Board and D. W. Lovelock (ed.), Some methods for microbiological assay. Society for Applied Bacteriology technical series no. 8. Academic Press Inc., London. 3. Finney, D. J., 1952. Statistical method in biological assay. Griffin, London. 4. Sokolsky, W. T., E. M. Stapert, and E. B. Ferrer. 1964. Liquid nitrogen freezing in microbiological assay systems. Appl. Microbiol. 12:327-329.