Vol. 51, No. 2APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Feb. 1986, p. 349-3550099-2240/861020349-07$02.0O/OCopyright © 1986, American Society for Microbiology

Bioluminescence for USP Sterility Testing of PharmaceuticalSuspension Products

DIANE M. BUSSEY AND KIYOSHI TSUJI*Control Analytical Research and Development, The Upjohn Company, Kalamazoo, Michigan 49001

Received 19 June 1985/Accepted 22 November 1985

Bioluminescence measurement significantly improved the accuracy, sensitivity, precision, and reliability ofthe current visual endpoint determination for the USP sterility test and eliminated the day 7 transfer/dilutionstep required for testing suspension products. Thirteen strains of bacteria and fungi (representing potentialcontaminants in sterile products), three pharmaceutical suspensiori products, and four media were used in theexperiment. No interference from suspension products was encountered in the detection of microbial growth bythe bioluminescence measurement. The poor fungal growth encountered was attributed to insufficient diffusionof oxygen into the medium and was circumvented by use of a large tube size (38 by 200 mm) or by vortexingthe medium once during the 2-week incubation period. Bioluminescence measurement would facilitateautomated handling of the sterility test endpoint readout operation. The optimum parameters of biolumines-cence measurement for application in sterility testing were determined.

The firefly luciferin-luciferase bioluminescence measure-ment for detection of ATP has been used for the qualitativeand quantitative determination of microbial content in fluidsand solids (1, 3, 5-6, 7, 9-11, 14, 15). Applications haveincluded clinical screening for bacteriuria in urine and blood(1, 3, 7, 11, 15), antibiotic susceptibility testing (9, 10),microbiological potency assays of antibiotics and vitamins(9, 10), and determination of biomass in environmentalsamples (e.g., water, soil), foods, and beverages (9, 10, 14).The purpose of this paper was not to shorten the incubationtime of the sterility test required by the U.S. Pharmacopeia(16) but to provide objectivity to the visual endpoint readoutoperation currently in use and to eliminate the time-consuming day 7 transfer/dilution operation, which is apotential contamination source, for testing pharmaceuticalsuspension products.

MATERIALS AND METHODSReagents. (i) Bioluminescence. The Firelight ATP test kit

from Analytical Luminescence Laboratories (San Diego,Calif.) was used to perform the bioluminescence measure-ment. The reagents were stored in a refrigerator (4°C) andallowed to reach room temperature before use. The test kitincluded HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid) buffer (pH 7.75), Extralight ATPreleasing reagent, Firelight luciferin-luciferase, and an ATPstandard. Reagents were prepared fresh each day accordingto the manufacturer's instructions. The ATP stock standardsolution was prepared by addition of 1.0 ml of sterilewater-for-injection (WFI; The Upjohn Co., Kalamazoo,Mich.) per ATP standard vial to provide 2 x 10-8 M ATP.ATP working standard solution was freshly prepared eachday by diluting the ATP stock solution. To determineday-to-day assay variation, several vials containing ATPstock solution were combined and frozen until used.

(ii) Media. Four media were used to conduct the simulatedsterility test: Trypticase soy broth (TSB), TSB withazolectin and Tween-80 (TSB-N), thioglycolate (Thio), andThio with azolectin and Tween-80 (Thio-N), all from BBLMicrobiology Systems, Cockeysville, Md. The media were

* Corresponding author.

prepared in Pyrex tubes (25 by 200 mm), 40 ml per tube,capped with Kap-uts (Bellco, Vineland, N.J.). For theoxygen diffusion study, the media (TSB and TSB-N) wereprepared in Pyrex tubes (38 by 200 mm) and autoclaved for15 min at 121°C. For plate counts and streak plates, Trypti-case soy agar, antibiotic medium no. 1, Schaedler agar, andSabouraud dextrose agar (Difco Labs, Detroit, Mich.) wereused.

Test microorganisms. The following 13 strains, represent-ing potential environmental contaminants in sterile products,were inoculated in pharmaceutical suspension products tosimulate positive sterility samples: Aspergilluis niger ATCC16404, Bacillus subtilis ATCC 6633, Candida albicansATCC 10231, Clostridium sporogenes ATCC 11437, Cory-nebacterium sp. strain UC 9165, Escherichia coli UC 3114,Micrococcus sp. strain UC 9168, Penicilliim sp. strain UC7296, pink-pigmented bacterium strain UC 9166, Pseuidomo-nas aeruginosa UC 9170, Pseudomonas cepacia UC 9463,Staphylococcus epidermidis UC 719, and Streptococ cuspyogenes UC 3113. For microcount determinations andstreak plates, the following media were used: antibioticmedium no. 1 for P. aeruginosa and P. cepacia, Schaedleragar for C. sporogenes, Trypticase soy agar for all otherbacteria, and Sabouraud dextrose agar for fungi.

Bioluminiescence. Portions (0.1 ml) of sample or ATPreference standard were pipetted into plastic cuvettes (cat-alog no. 2000-10, Analytical Luminescence Laboratories).For the negative control, 0.1 ml of WFI or medium wassubstituted for the sample. A 0.1-ml volume each of HEPESbuffer and Extralight reagent were added to samples, and themixture was incubated for 7 min at room temperature torelease/extract ATP from microorganisms. The cuvette wasthen placed in a bioluminometer (Monolight 2001, AnalyticalLuminescence Laboratories), and 0.1 ml of Firelightluciferin-luciferase reagent was automatically pipetted intothe sample. The amount of bioluminescence emitted wasintegrated for 10 s and recorded in relative light units. Allsamples and standard solutions were tested in quadruplicate.The simulated positive samples, microbial suspensions, andATP standards in media were diluted 1:10 in WFI beforeassay. Gilson precision microliter pipettes (Rainin Instru-ment Co., Woburn, Mass.) equipped with sterile disposable

349

on May 1, 2019 by guest

http://aem.asm

.org/D

ownloaded from

350 BUSSEY AND TSUJI

pipette tips were used to prepare dilutions and to transfersamples into the cuvettes.Simulated positive sterility test sample. For each of the 13

organisms 1.0-ml volumes, containing less than 100 cells perml, were pipetted into 12 tubes, each containing 40 ml ofTSB, TSB-N, Thio, or Thio-N. A 1.0-ml volume of eachsuspension product (Depo-Medrol Sterile Aqueous Suspen-sion, Cortef Sterile Aqueous Suspension, and Depo-ProveraSterile Aqueous Suspension) was then aseptically added tothe tubes. Negative controls contained no microorganisms,and positive controls contained no product. The sampleswere then incubated for 2 weeks at room temperature forTSB and TSB-N media and at 35°C for Thio and Thio-Nmedia to simulate sterility test condition. On days 3, 7, and14 of the incubation period, four tubes of the inoculatedsamples and one tube each of the positive and negativecontrols were tested for bioluminescence.For the bioluminescence assay, a sample was recorded as

positive if the relative light units value was greater than orequal to twice the background of the respective negativecontrol. On each of the three testing dates, all biolumines-cence-negative samples were streaked on agar plates toverify the lack of microbial growth or to confirm that thepopulation was below the detection limit of the biolumi-nescence measurement. Microorganisms in positive-biolu-minescence samples were identified to verify growth of theinoculated microorganism. On day 14, microcounts wereperformed to determine the sensitivity and detection limitsof the bioluminescence method.TSB and Thio were used for Depo-Provera Sterile Aque-

ous Suspension and Cortef Sterile Aqueous Suspension,while TSB-N and Thio-N were used to test Depo-MedrolSterile Aqueous Suspension.Oxygen diffusion study. To examine the oxygen content

and diffusion rate, TSB (40 ml per tube) was prepared in twodifferent size containers: a large tube (38 by 200 mm) and asmall tube (25 by 200 mm). The oxygen level of medium wasdetermined with an oxygen probe (model 97-08; OrionResearch, Inc., Cambridge, Mass.) coupled to a pH meter(model 130; Corning Glass Works, Coming, N.Y.). Resultswere expressed in parts per million (microliter per liter) ofoxygen.

RESULTS AND DISCUSSIONOptimization of bioluminescence assay. ATP standard

curves were constructed using the peak and the integrationmodes. Although the peak mode gave better precision (rel-ative standard deviation [RSD], ca. 5 to 8%), the integrationmode (RSD, ca. 4 to 15%) was chosen because of its superiorsensitivity (10-14 M ATP) over the peak mode (10-13 MATP).

(i) Integration time. An integration time of 10 s wasselected from a choice of 10, 30, or 60 s because of lowerbackground bioluminescence, superior linearity (r = 0.9992),sensitivity (10-11 to 10-14 M ATP), and shorter assay time.

(ii) Precision sensitivity. Results of ATP standard curvesperformed with 25- and 100-pdl samples indicated that the100-,ul volume provided slightly better sensitivity (2 x 10-14M ATP) than did 25 RI (5 x 10-14 M ATP). The linearity ofthe two standard curves was comparable, and the number ofreplicates (either three or four) had no significant effect (P <0.001) on the linear regression parameters of either standardcurve. However, the sample volume exhibited a profoundinfluence on the RSD of the assay. An RSD of 14.5%obtained with the 25-pd sample size was reduced to 6.5%when the 100-pd sample volume was used.

The reproducibility of the bioluminescence measurementwas evaluated by constructing ATP standard curves on 3consecutive days. All reagents and ATP standards wereprepared fresh each day. No significant difference (P <0.002) in the slope, linearity, or RSD of the standard curvesassayed on 3 separate days was observed. Therefore, theATP bioluminescence assay method can reproducibly (RSD,10 to 12%) detect ATP from 2 x 10-11 M to 2 x 10-14 M (r= 0.998).

(iii) Stability of reagent. The stability of the luciferin-luciferase reagent was examined. The results indicated thatthe reagent is stable for at least 6.5 h at room temperature,since no significant difference (P < 0.002) existed betweenstandard curves constructed at 15 min, 3.5 h, and 6.5 h afterreconstitution. However, when the reconstituted reagentwas stored overnight in a refrigerator, the sensitivity andlinearity of the standard curve were reduced by approxi-mately 50%. When the reagent was stored frozen, thebackground (blank) luminescence increased and the linearrange and sensitivity of the assay were significantly de-creased. Thus, the luciferin-luciferase reagent must befreshly prepared and used each day.

Detection of microorganisms. (i) Extraction/release of ATP.The ATP must be extracted/released from the microorga-nisms for detection by the bioluminescence assay. Aperchloric acid method (2) was initially used to release theATP from the cells. However, this procedure was tediousand time-consuming and resulted in low recovery of ATPfrom microorganisms. An ATP releasing reagent, Extralight,available from Analytical Luminescence Laboratories wasevaluated for each of the 13 strains of microorganisms usedin the study. The Extralight required at least 5 to 10 min ofcontact to achieve maximum extraction of ATP from ap-proximately 75% of the microorganisms (Table 1). For P.cepacia and A. niger, the quantity of ATP did not signifi-cantly increase with increased extraction time; however, thebioluminescence remained stable for 15 min. Based on theseresults, an extraction time of 7 min was chosen for release ofATP from microbial cells to shorten the assay time.

(ii) Detection sensitivity. Various concentrations of S.epidermidis (UC 719), E. coli (UC 3114), C. albicans (UC7275), and A. niger (UC 7048) were tested for biolumines-cence. The linear range covered a 3-log concentration ofmicroorganisms with correlation coefficients of 0.990 to0.999 (Fig. 1). The precision of the assay (RSD, ca. 2.5 to10%) for the majority of the microorganisms was equivalentto or better than that of the ATP standard solution (RSD, ca.10 to 12%). The higher RSD (-16%) for S. epidermidis maybe due to the characteristic clumping of the microorganism.The sensitivity of the bioluminescence assay for the de-

tection of microorganisms is dependent on the ATP contentper CFU and the species, age, and growth conditions of themicroorganisms (e.g., medium, temperature) (5). As little as104 CFU of S. epidermidis and E. coli per ml could bedetected by the bioluminescence method, whereas ca. 102CFU/ml was required to detect C. albicans and A. niger(Fig. 1). In contrast, visual detection of turbidity, the currentendpoint detection operation for the sterility test, requires atleast 106 CFU/ml in a clear solution.

Effect of medium. Although the precision of the assay wasnot significantly affected, the medium increased the back-ground (blank) bioluminescence. When the medium wasdiluted 1:10 and 1:100, the background bioluminescence wasreduced by 72 and 90%, respectively. To ascertain thatincreased background levels were attributable to compo-nents in the medium which absorb at the wavelength (560

APPL. ENVIRON. MICROBIOL.

on May 1, 2019 by guest

http://aem.asm

.org/D

ownloaded from

BIOLUMINESCENCE FOR USP STERILITY TESTING 351

TABLE 1. Contact time required by Extralight for the maximum release of ATP from microorganisms

Relative light units (relative SD) after contact time (min):Microorganism

0.0 2.5 5.0 10.0 15.0 20.0Blank 3,557 3,743 2,315 2,736 3,104 2,044

(47%) (23%) (11%) (23%) (37%) (8%)B. subtilis ATCC 6633 76,400 92,400 81,800 104,700 103,400 116,000

(7%) (5%) (21%) (2%) (4%) (18%)S. epidermidis UC 719 69,600 133,000 151,400 135,000 118,000 110,400

(27%) (3%) (8%) (7%) (5%) (9%)E. coli UC 3114 79,900 70,200 73,200 75,400 70,300 69,600

(12%) (9%) (16%) (8%) (25%) (5%)Micrococclus sp. strain UC 9168 27,400 25,000 18,700 17,300 16,600 17,600

(21%) (10%) (2%) (15%) (11%) (19%)Corynebacterium sp. strain UC 9165 855,400 1,278,000 1,575,000 1,317,000 1,295,500 1,416,600

(10%) (8%) (11%) (22%) (9%) (3%)Pink-pigmented bacterium, strain UC 9166 83,100 135,000 121,000 139,00 150,000 147,000

(5%) (7%) (7%) (5%) (5%) (6%)P. aeruginosa UC 9170 508,600 467,000 454,400 466,000 458,000 451,000

(14%) (5%) (8%) (10%) (15%) (8%)P. cepacia UC 9463 215,200 254,000 221,000 199,600 237,000 215,000

(5%) (5%) (20%) (15%) (20%) (2%)S. pyogenes UC 3113 389,000 549,000 662,000 625,400 702,000 657,000

(7%) (14%) (16%) (9%) (10%) (8%)C. sporogenes ATCC 11437 71,700 113,400 62,200 37,000 34,000 31,000

(5%) (10%) (5%) (17%) (2%) (4%)C. albicans ATCC 10231 9,375 528,800 607,400 498,000 425,200 374,200

(20%) (15%) (12%) (18%) (9%1) (11%)A. niger ATCC 16404 41,900 114,700 180,400 132,800 209,400 123,500

(34%) (36%) (26%) (12%) (54%) (18%)Penicillium sp. strain UC 7296 49,800 50,000 48,900 47,500 47,600 49,700

(5%) (7%) (4%) (9%) (12%)

nm) used for the bioluminescence assay, a Cary spectropho-tometer was utilized to measure the absorbance for TSB andunoxidized Thio. The data indicated that the absorbancereadings of TSB and Thio were significantly higher than thatof the WFI blank but were reduced by dilution in WFI. Sinceaeration significantly contributes to the high backgroundreadings for Thio (optical density of 0.10 and 0.26 forunoxidized and oxidized Thio, respectively), resazurin, aredox indicator in Thio, may be the compound causing theinterference in the oxidized Thio.TSB inhibited bioluminescence by about 50 to 60%. To

identify the source of interference in TSB, components inthe medium were individually tested for their inhibitoryeffect on bioluminescence. The cause of interference wastraced to the NaCl (86 mM) present in the medium. NaClconcentrations between 20 and 200 mM have been reportedto markedly reduce light emission from ATP standards (8).Various ions, particularly chloride, have been identified asthe primary source of bioluminescence interference in bac-teriuria assays (8, 12, 13). Dilution of the medium eliminatedthe interference without an equivalent loss in sensitivity ofthe bioluminescence measurement, since the increase inlight output with decreasing inhibitor concentration wasexponential, whereas the reduction in light output withdecreasing ATP concentration was linear (12, 13). Thethioglycolate in Thio caused no effect on the biolumines-cence assay.TSB-N and Thio-N inhibited bioluminescence by 20 to

25%. Azolectin was identified as the cause of interference.Dilution of the medium effectively eliminated the interfer-ence. Tween-80 had no effect on bioluminescence.

Suspension products testing. The effects of the suspensionproducts on bioluminescence were investigated for each ofthe three pharmaceutical products. The bioluminescence

results in the presence of the suspension products were notsignificantly different from those of the ATP reference solu-tion; the relative light units of the samples with Cortef,Medrol, and Provera were 107, 103, and 102%, respectively,of those without the products. Results of the biolumines-cence measurements for 13 strains of bacteria and fungi infour media and three suspension products are shown inTable 2. Since the media tested immediately after inocula-tion with less than 100 CFU of microorganisms per tubeprovided bioluminescence readings equivalent to those ofthe negative controls, any significant bioluminescence de-tected in samples assayed on day 3 and beyond can beattributed to the growth of the microorganism and not to a'carry-over' effect from the inoculum. To confirm growth/no growth in the turbid suspensions, the samples werestreaked onto agar plates to substantiate the biolumines-cence data.

For eight microorganisms, B. subtilis, S. epidermidis, E.

105 C.albicansAngr E.coll

S.epldermidls

1031102 103 l1, o 106 167Log Of Colonies/ML

FIG. 1. Bioluminescence measurement of ATP extracted from S.epidermidis (r = 0.994), E. coli (r = 0.999), A. niger (r = 0.990), andC. albicans (r = 0.999).

VOL. 51, 1986

on May 1, 2019 by guest

http://aem.asm

.org/D

ownloaded from

352 BUSSEY AND TSUJI

TABLE 2. Bioluminescence measurement in simulated sterility test samplesBioluminescence (no. positive/no. of replicates) in the following media at incubation daya:

Microorganism TSB/TSB-Nb Thio/Thio-Nband product

3 7 14 3 7 14

B. subtilisDepo-ProveraCortefDepo-MedrolPositive controlcVisual

C. sporogenesDepo-ProveraCortefDepo-MedrolPositive controlcVisual

Corynebacterium sp.Depo-ProveraCortefDepo-MedrolPositive controlcVisual

E. coliDepo-ProveraCortefDepo-MedrolPositive controlcVisual

Micrococcus sp.Depo-ProveraCortefDepo-MedrolPositive control'Visual

Pink-pigmentedbacteriumDepo-ProveraCortefDepo-MedrolPositive controlcVisual

P. aeruginosaDepo-ProveraCortefDepo-MedrolPositive controlcVisual

P. cepaciaDepo-ProveraCortefDepo-MedrolPositive controlcVisual

S. epidermidisDepo-ProveraCortefDepo-MedrolPositive controlcVisual

S. pyogenesDepo-ProveraCortefDepo-MedrolPositive controlcVisual

A. nigerDepo-ProveraCortefDepo-MedrolPositive controlcVisual

4/44/44/42/2+

NR:NRNRNRNR

2/44/44/43/3+

4/44/44/42/2+

0/41/42/42/3

2/42/40/41/3

4/44/44/43/3+

1/44/44/43/3

4/43/44/42/2

4/44/44/42/2+

0/40/40/40/2

4/44/44/42/2+

4/44/44/43/3+

4/44/44/42/2+

0/43/41/43/3

4/40/42/42/3

4/44/44/43/3+

1/44/44/43/3+

4/44/44/42/2+

4/44/44/42/2+

1/41/41/40/3

4/44/44/44/4+

4/44/44/44/4+

4/44/44/44/4+

0/44/44/43/3+

4/44/44/44/4

4/44/44/44/4

4/4+

4/44/44/44/4+

4/44/44/44/4+

4/44/44/42/2+

1/40/42/44/5

4/44/44/42/2+

4/44/44/42/2+

1/40/44/43/3

4/44/44/42/2+

0/40/40/43/3

0/40/40/40/3

4/44/43/33/3+

4/44/44/43/3+

4/44/44/42/2+

4/44/44/42/2+

0/40/40/42/2+

4/44/44/42/2+

4/44/44/44/4+

4/44/44/42/2+

4/44/44/42/2+

4/40/44/43/3+

4/41/44/44/4+

4/44/44/42/2+

0/40/40/44/4+

2/40/42/43/3+

3/44/44/44/4+

0/40/44/42/3

4/44/43/33/3+

4/40/44/43/4

4/44/44/44/4+

4/44/44/43/3+

4/44/44/42/2+

4/44/44/42/2+

4/44/44/44/4+

4/44/44/44/4+

4/44/44/42/2+

0/4 1/41/4 4/41/4 2/42/3 3/3+ +

Continued on following page

APPL. ENVIRON. MICROBIOL.

on May 1, 2019 by guest

http://aem.asm

.org/D

ownloaded from

BIOLUMINESCENCE FOR USP STERILITY TESTING 353

TABLE 2-Continuted

Bioluminescence (no. positive/no. of replicates) in the following media at incubation day":Microorganism TSB/TSB-Nb Thio/Thio-Nhand product _______ ___ ____________

3 7 14 3 7 14

C. albicansDepo-Provera 4/4 4/4 4/4 2/4 1/4 1/4Cortef 4/4 2/4 4/4 2/4 2/4 4/4Depo-Medrol 4/4 4/4 4/4 0/4 0/4 3/4Positive control' 2/2 3/3 4/4 2/2 3/3 3/3Visual + /- + + + + +

Penicillium sp.Depo-Provera 0/4 0/4 0/4 0/4 0/4 3/4Cortef 0/4 0/4 0/4 0/4 0/4 1/4Depo-Medrol 0/4 1/4 2/4 0/4 0/4 0/4Positive control' 0/3 1/3 1/5 1/3 1/3 2/4Visual - - - + /- +/- +" +, Growth; -, no growth. NR, Not run.b TSB and Thio were used to test Depo-Provera and Cortef. Depo-Medrol was tested in TSB-N and Thio-N." No drug added.

coli, Corynebacteriium sp., P. aeriuginosa, P. cepacia, S.pyogenes, and C. sporogenes, the bioluminescence mea-surement had no problem in detecting the growth of themicroorganisms in the presence of the suspension products(Table 2). Bioluminescence was detected in 100% (four offour) of the tubes tested on each of the three test dates (days3, 7, and 14) during the 2 weeks of incubation. An exceptionwas E. coli in Thio or Thio-N: although significantbioluminescence readings were recorded on days 3 and 7,the culture lysed by day 14 and no bioluminescence wasdetected. Streak plates verified the loss of viable cells in thesamples. In contrast, E. coli in TSB or TSB-N did not lyseand was detected from day 3 to the end of the 2-weekincubation period (Table 2). Thus, this finding does notindicate a potential false-negative problem in thebioluminescence measurement, but rather underscores thereason for using the two types of compendially requiredmedia in sterility testing.Based on the visual examination of positive controls

(without product), pink-pigmented bacteria and Micro-c'occuls sp. proved to be slow growers, especially in TSB(Table 2). Although the visual indication of growth was still

questionable by day 14, bioluminescence measurement de-tected microbial growth in over 60% of the positive controlson day 7 and in 75 to 100% of the tubes on day 14.Some products did inhibit the growth of certain microor-

ganisms in some media: Micrococcus sp. in TSB withDepo-Provera Sterile Aqueous Suspension; Corynebacte-rium sp. and pink-pigmented bacterium in Thio with CortefSterile Aqueous Suspension; C. albicans in Thio with Depo-Provera Sterile Aqueous Suspension; and C. albicans inThio-N with Depo-Medrol Sterile Aqueous Suspension (Ta-ble 2). The lack of growth was verified by the streak plates.In all cases, the growth of these five microorganisms wasfacilitated by one of the other types of medium (TSB or Thio)in the presence of the products (Table 2).The results of the study described above indicate that the

bioluminescence measurement is sensitive for the detectionof microorganisms without interference from suspensionproducts and provides objectivity to the visual endpointreadout operation currently in use for the sterility test.Growth of fungi and mold. It was noted from visual

7.500

TABLE 3. Effect of larger size container (38 by 200 mm) ongrowth of fungi"

Bioluminescence (no. positive/no. ofMicroorganism replicates) at day of incubation:and product

3 7 14

A. nigerDepo-Provera 4/4 4/4 4/4Cortef 4/4 4/4 4/4Depo-Medrol 4/4 4/4 4/4

C. albicansDepo-provera 4/4 4/4 4/4Cortef 4/4 4/4 4/4Depo-Medrol 4/4 4/4 4/4

Penicillium sp.Depo-Provera 2/4 4/4 2/4Cortef 3/4 4/4 2/4Depo-Medrol 0/4 4/4 0/4

aTSB and Thio were used to test Depo-Provera and Cortef. Depo-Medrolwas tested in TSB-N and Thio-N.

0.

ECL

x0

0 4.2 8.4 12.6 16.8 21.0

Time (Days)

FIG. 2. Effect of container sizes on the oxygen content inmedium. Small tubes, 25 by 200 mm; large tubes, 38 by 200 mm.

Voi- 51, 1986

on May 1, 2019 by guest

http://aem.asm

.org/D

ownloaded from

354 BUSSEY AND TSUJI

TABLE 4. Effect of day 7 transfer/dilution on growth of fungi

Bioluminescence (no. positive/no. of replicates) in thefollowing medium at incubation day:

Microorganismand product TSB/TSB-N Thio/Thio-N

3 7 14 7-Ta 3 7 14 7-T'

A. nigerDepo-Provera 4/4 4/4 4/4 2/4 3/4 4/4 4/4 4/4Cortef 4/4 4/4 3/4 3/4 2/4 3/4 4/4 4/4Depo-Medrol 1/4 3/4 0/4 0/4 2/4 2/4 1/4 2/4

C. albicansDepo-Provera 4/4 4/4 4/4 4/4 3/4 3/4 4/4 3/4Cortef 4/4 4/4 4/4 2/4 4/4 4/4 4/4 4/4Depo-Medrol 3/4 4/4 4/4 3/4 1/4 4/4 4/4 4/4

Penicillium sp.Depo-Provera 0/4 1/4 1/4 1/4 0/4 0/4 1/4 0/4Cortef 0/4 1/4 1/4 1/4 0/4 0/4 1/4 0/4Depo-Medrol 0/4 0/4 1/4 0/4 0/4 0/4 0/4 0/4a Day 7 transfer samples, incubated for an additional 7 days.

observations of the positive controls that the cultural condi-tions for the growth of Penicillium sp. and A. niger in TSBmight not be optimum (Table 2). Since TSB was preparedand sterilized in tubes 25 by 200 mm in size, the surfacearea-to-volume ratio may have been unfavorable, causinginsufficient diffusion of oxygen into the medium and therebylimiting the growth of fungi. When TSB was prepared inlarger tubes (38 by 200 mm), the growth of C. albicans andA. niger was detected in 100% of the samples tested on alltesting days (3, 7, and 14 days), and no product interferencewas noted (Table 3). Similarly, Penicillium sp. was able togrow and be detected, without product interference, in allsamples tested on day 7 (Table 3).The results described above clearly indicated that the

growth of fungi in TSB and TSB-N was enhanced bypreparing the medium (40 ml per tube) in a larger tube (38 by200 mm) to increase the surface area-to-volume ratio,thereby enhancing the diffusion of oxygen into the medium.The oxygen content in TSB prepared in the larger tube (38 by200 mm) exhibited a higher level on day 1 and rapidlyreached the maximum, equilibrium concentration of 6.8 to

7.0 ppm by day 5 to 7 (Fig. 2). In contrast, the oxygen

content in the medium placed in the smaller container (25 by200 mm) slowly increased and reached the maximum oxygen

concentration of 7.0 ppm after 13 to 15 days. These resultscompared favorably with the theoretical oxygen level calcu-lated in each of two container sizes using the followingequation (4):

C=1-2 ,

n = 0

(-n) -e(ttl + 1/2)Trr(at/b2) cos (it + L/2)Try/b](n + 1/2)Xr

n ( 1) exp-tt, + 1/2)-1rr(a/b2)t (24)(3,600) cos (n + 1/2)iTyIb]IT (n + I/2)

0 (63662 ) exp-[(n + 1/2)-t 0.07310818] cos [(n + 1/2)2.3561(n + I/2)

where C is oxygen concentration achieved, fraction of theequilibrium; I is depth of medium (5.0 cm for large tube and10.8 cm for small tube); t is time after autoclaving (days); n

is the point of 02 measurement, yll (1/4 from the top of themedium); t is dimensionless time, at/I2; and a is diffusivity of

oxygen in aqueous solution (X l0-5 cm2/s). Use of an aged,oxygen-saturated medium in the smaller tube (14-day-oldmedium gave 14 positives out of 20) or vortexing once duringthe 2-week incubation period to increase dissolved oxygenlevel (10 positives out of 10) also improved fungal growth.

Elimination of day 7 transfer/dilution requirement. Theprimary purpose of the day 7 transfer, used for sterilitytesting of suspension products, is to minimize by dilution theturbidity created by suspension products or medium andthereby permit the visual endpoint determination of micro-bial growth. However, the transfer operation might alsoserve to promote microbial growth by minimizing productinterference via dilution. However, samples which werenegative by bioluminescence on day 7 failed to grow upontransfer into fresh medium (Table 4). In addition, approxi-mately 25 to 50% of the samples transferred from the tubeswith positive bioluminescence results on day 7 failed to showgrowth (visually and by bioluminescence) after incubationfor 7 additional days. This may be due to nonhomogeneousdistribution of microorganisms in the medium on day 7, anda 0.5-ml sample size used for transfer may not have repre-sented the sample. Thus, the transfer/dilution operationwould not alleviate the problem, but would rather increasethe potential for false-negative and false-positive (laboratorycontamination) results. Since the bioluminescence measure-ment is sensitive enough to detect 103 to 104 CFU/ml inturbid suspension samples, this measurement effectivelyeliminates the need of the day 7 transfer/dilution operationand thereby minimizes potential assay problems, providesmanpower savings, and facilitates automated handling of theendpoint readout operation.

ACKNOWLEDGMENTSWe wish to acknowledge Analytical Luminescence Laboratories

for the use of the Monolight 2001 luminometer. We thank J. M. Priceand S. M. Davenport for technical assistance on the oxygendependence of C. albicans and P. A. Martin for the theoreticaloxygen content in a medium. P. S. Satoh is thanked for his technicaladvice.

LITERATURE CITED1. Alexander, D. N., G. M. Ederer, and J. M. Matsen. 1976.

Evaluation of an adenosine 5'-triphosphate assay as a screeningmethod to detect significant bacteriuria. J. Clin. Microbiol.3:42-46.

2. Bagnora, A. S., and L. R. Finch. 1972. Quantitative extractionand estimation of intercellular nucleotide triphosphates of Esch-erichia coli. Anal. Biochem. 45:24-34.

3. Beckers, B., and H. R. M. Lang. 1982. A simplified method forthe early detection of bacterial growth in blood cultures usingbioluminescent measurement of ATP, p. 1073-1077. In Proceed-ings of XI International Congress of Clinical Chemistry. Walterde Gruyter, Inc., Berlin, N.Y.

4. Bird, R. B., W. E. Stewart, and E. N. Lightfoot. 1976. Transportphenomena, p. 354-356. John Wiley & Sons, Inc., New York.

5. Bush, V. N., G. L. Picciolo, and E. W. Chappelle. 1975. Theeffect of growth phase and medium on the use of the fireflyadenosine triphosphate (ATP) assay for the quantitation ofbacteria, p. 35-42. In E. W. Chappelle and G. L. Picciolo (ed.),Analytical applications of bioluminescence and chemilumines-cence. National Aeronautics and Space Administration, Wash-ington, D.C.

6. Chappelle, E. W., G. L. Picciolo, and J. W. Deming. 1978.Determination of bacterial content in fluids. Methods Enzymol.57:65-72.

7. Lazaroni, J. A., and D. L. Henry. 1981. Bioluminescent screen-ing for bacteriuria, p. 443-452. In M. A. DeLuca and W. D.McElroy (ed.), Bioluminescence and chemiluminescence. Aca-demic Press, Inc., New York.

APPL. ENVIRON. MICROBIOL.

on May 1, 2019 by guest

http://aem.asm

.org/D

ownloaded from

BIOLUMINESCENCE FOR USP STERILITY TESTING

8. Leach, F. R. 1981. ATP determination with firefly luciferase. J.Appi. Biochem. 3:473-517.

9. Lundin, A. 1981. Applications of firefly luminescence, p.187-196. In M. A. DeLuca apd W. D. McElroy (ed.),Bioluminescence and chemiluminescence. Academic Press,Inc., New York.

10. Lundin, A. 1982. Analytical applications of bioluminescence:the firefly system, p. 43-74. In L. J. Kricka and T. J. N. Carter(ed.), Clinical and biochemical luminescence. Marcel Dekker,Inc., New York.

11. Mackett, D., S. Kessock-Philip, S. Bascomb, and C. S. F.

Easmon. 1982. Evaluation of the Lumac kit for the detection ofbacteriuria by bioluminescence. J. Clin. Pathol. 35:107-110.

12. Nicholas, W. W., G. D. W. Curtis, and H. H. Johnston. 1981.The identity and properties of firefly luciferase inhibitors in

urine. Anal. Biochem. 114:433-440.13. Nicholas, W. W., G. D. W. Curtis, and H. H. Johnston. 1981.

Inhibitors of firefly luciferase in clinical urine specimens, p.485-490. In M. A. DeLuca and W. D. McElroy (ed.),Bioluminescence and chemiluminescence. Academic Press,Inc., New York.

14. Stannard, C. J., and J. M. Wood. 1983. The rapid estimation ofmicrobial contamination of raw meat by measurement of aden-osine triphosphate (ATP). J. Appi. I3acteriol. 55:429-438.

15. Thore, A., A. Lundin, and S. Ansehn. 1983. Firefly luciferaseATP assay as a screening method for bacteriuria. J. Clin.Microbiol. 17:218-224.

16. U.S. Pharmacopeial Convention, Inc. 1985. The United StatesPharmacopeia, 21st ed., p. 1156-1160. U.S. Pharmacopeia,Rockville, Md.

VOL. 51, 1986 355

on May 1, 2019 by guest

http://aem.asm

.org/D

ownloaded from