J Clin Pathol 1992;45:487-489

Rapid cytochemical demonstration of cytochromeoxidase activity in pathogenic bacteria

M R Barer, P J Marsh

AbstractAims: To develop and assess a cyto-chemical technique for the light micro-scopical detection of oxidase activity inpathogenic bacteria.Methods: Live bacterial cells weredeposited on to aminopropylsilanetreated glass coverslips by centrifugation,dried, then reacted with either 1% (w:v)n,n,n',n'-tetramethyl-p-phenylene diam-ine (TPD) or 5 mM diaminobenzidine(DAB) at 37°C. The preparations weremounted in 50% glycerol and assessed bybrightfield microscopy. An optimisedDAB procedure (5 minutes of drying at37°C and 10 minutes ofreaction time) wasapplied to 44 strains of organisms com-monly associated with infections in manand to two fresh positive blood culturecontaining Gram negative bacilli.Results: TPD gave no discernible local-ised reaction product and was not inves-tigated further. With DAB, oxidasepositive cells (brown) were clearlydifferentiated from oxidase negative cells(colourless) even in mixed preparations.The DAB technique correctly assigned 18oxidase positive isolates (seven genera),26 oxidase negative isolates (eight gen-era), and the organisms present in the twofreshly positive blood cultures to theirappropriate oxidase reactivity as definedby the standard macroscopic TPD tech-nique.Conclusions: The cytochemical reac-tion seems to be a reliable indicator ofthemacroscopic oxidase test. It permitsdetermination of oxidase reactivity at anearly stage in the assessment of clinicalmaterial when infecting organisms can bedemonstrated by microscopy. Furtherdevelopment of this and related cyto-chemical techniques could permit theprovision of microbiological informationthat would be relevant to patient man-agement, well in advance of conventionaltechniques.

The oxidase test is a key step in determining thetaxonomic status and identity of pathogenicbacteria.' 2 The reaction is conventionally doneat the macroscopic level using colonies growingon agar media. Knowledge of the oxidasereactivity of organisms present in clinicalspecimens before they have been fully iden-tified can have a profound effect on patientmanagement. For example, a positive test can

exclude the possibility that an isolate is aSalmonella sp or support a presumptive iden-tification of Pseudomonas sp.A technique capable of demonstrating

oxidase reactivity at a time when the organismsconcerned are only detectable by microscopycould provide clinically valuable informationearlier than conventional methods. It couldalso facilitate taxonomic studies on slow grow-ing organisms. The recent development of amethod for attaching viable bacteria to opticalquality glass without damage and subsequentmicroscopic demonstration of ,B-galactosidasereactivity in the immobilised cells34 led us toexamine the possibility of developing a cyto-chemical oxidase procedure applicable to path-ogenic bacteria within the context of a routinediagnostic laboratory.

Cytochemical demonstration of oxidase hasbeen achieved in bacteria using the NadiReagent.5 Such mixtures of aryl amines andelectron donors have been used extensively forhistological work but the procedures generallyrequire total incubation times in excess of 30minutes,al and there is some doubt about theefficiency of the electron donors used.8 Incontrast, diaminobenzidine (DAB) has beenshown to be valuable for the histochemicaldemonstration9 and quantitation of cyto-chrome oxidase with incubation times under 15minutes. We report results obtained whenDAB and the standard macroscopic reagentn,n,n',n'-tetramethyl-p-phenylene diamine(TPD) were used for the cytochemical demon-stration of oxidase activity in clinical isolates ofpathogenic bacteria.

MethodsStrains examined included laboratory controlsand isolates identified to a level appropriate totheir medical context by standard clinicallaboratory methods.'2 Clinical isolates werestored on nutrient agar slopes and laboratorystrains were maintained by weekly subcultureon blood or chocolate agar (Oxoid). Brothcultures were prepared in Todd Hewitt broth(Oxoid).

Bacterial suspensions were prepared directlyfrom agar or by dilution of three to five hourbroth cultures in sterile distilled water to adensity not greater than a McFarland No 1standard (about 3 x 108 cells ml/l). Bloodcultures were examined neat and at 1 in 4 and 1in 10 dilutions. Bacteria were immobilised onglass coverslips by the APS-centrifugationtechnique.3 Briefly, 22 x 64 mm glass cover-slips (BDH) which had been treated with 3-

Department ofMicrobiology,University ofNewcastle upon Tyne,Framlington Place,Newcastle upon TyneNE2 4HHM R BarerDepartment ofClinical Microbiology,Royal VictoriaInfirmary, Newcastleupon TyneP J MarshCorrespondence to:Dr M R BarerAccepted for publication17 November 1991

487

on Septem

ber 30, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.6.487 on 1 June 1992. D

ownloaded from

Barer, Marsh

[:

..40. ' 4

............X$4-\

r*s!.

N.w

Figure I Mixtures of oxidase positive and negative organisms staine4reaction. (A) Acinetobacter anitratus (unstained coccobocilli), and r-aeruginosa (stained bacilli). (B) Acinetobacter anitratus and Neisser(stained diplococci). Note that the unstained bacteria are very faintphotomicrographs. They can be clearly discerned by phase contrast mig

aminopropyltriethoxysilaneluwere assembled into Bellco m

bers (Arold Horwell, London'

a 10 chamber (each 2 mm ra

deep) superstructure which pei

ment of different organisms ar

the application of different stair

separate cell monolayers on the

Aliquots of the suspensions (5Cto the chambers and cells depAPS treated surface by subje(assembly to centrifugation at 11minutes. The cell suspensior

removed prior to further proce

Oxidase was detected macro,

TPD. Reactions for microscolwere either done by adding 50 p

mixture to the immobilised

microslide chambers directly a

the suspension medium or after a drying step(37°C). TPD (1% w:v) was prepared in dis-tilled water and DAB (5 mM) in 01 MSorensen's phosphate buffer (pH 7-4). Dryingand reaction times (both at 37°C) were as statedin the text. TPD was made up just before usewhile DAB was prepared at working strengthin advance and stored at - 20°C in 0-5 mlamounts for up to two months before use. Inview of the carcinogenic potential of DAB werecommend use of the Sigma Isopac system forpreparation of stock solutions. Gloves are wornfor the procedure itself and all contaminatedmaterials are discarded into chloros at 2500ppm available chlorine. This has the advantageof providing both a disinfectant action and an

// oxidant for unused DAB. The optimised DABschedule comprised five minutes of dryingfollowed by 10 minutes of reaction. At the endofthe reaction period the chamber was disman-tled, the coverslip rinsed gently in distilledwater, the preparation dried in air at 37°C,mounted in 50% glycerol and sealed with nailvarnish. Mounted preparations were observedby phase contrast and brightfield microscopy;staining was assessed by the latter. Althoughprovisional assessments could be made at amagnification x 600, x 1250 was preferred fordefinitive observations.

.....

ResultsDark purple reactions were sometimesobserved macroscopically when TPD wasapplied to heavily inoculated coverslips, but noreaction product associated with individualcells was recognisable by microscopy. No fur-ther work was done with this reagent. Bycontrast, DAB rapidly produced browncoloured cells with classic oxidase positivegenera. Drying and reaction times of 0-20 and5-20 minutes were examined, respectively.Application of the DAB reagent to undriedpreparations yielded minimal colour even withreaction times up to 20 minutes. Neither dryingbeyond five minutes nor reaction times beyond

d by the DAB 10 minutes produced appreciable enhancement'seudomonas of the colour. The optimised procedure (fiven these minutes drying + 10 minutes reaction) provedcroscopy. equally effective at demonstrating oxidase re-

activity in freshly grown broth or agar cultures(APS) (Sigma) and agar cultures stored for up to seven days atLicroslide cham- + 40C.). These provide To determine whether individual oxidasedius x 12 mm positive and negative cells could be distin-rmits the attach- guished in a single preparation, combinationsid subsequently of appropriate strains were examined.iing processes to Coloured and uncoloured cells were readilysame coverslip. apparent. Moreover, when the two strains were)ul) were added morphologically distinct, colour was onlyosited onto the present in the appropriate cell type (figure).cting the entire The optimised DAB procedure was applied000 x g for two to a series of clinical isolates and laboratoryn medium was strains representing both oxidase positive andssing. negative genera (table). The microscopicscopically using oxidase reaction was consistent with thepic examination macroscopic TPD result in all cases. Very fainttl of the reaction brown deposits were occasionally seen withI cells in the Enterobacteriaceae; these were more noticeableifter removal of in cell clumps.

488

0.1

:W*

* I ...,...J.P.i

on Septem

ber 30, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.6.487 on 1 June 1992. D

ownloaded from

Rapid cytochemical demonstration of cytochrome oxidase activity in pathogenic bacteria

Bacterial strains examined with the DAB oxidaseprocedure

Macroscopic oxidase

Positive No= Negative No=

Pseudomonas aeruginosa 8 Escherichia coli 9Pseudomonas spp 2 Klebsiella spp 6Agrobacterium sp 1 Enterobacter spp 4Neisseria meningitidis 1 Salmonella enteritidis 2Aeromonas hydrophila 2 Proteus spp 2Aeromonas sobria 1 Haemophilus influenzae 1Vibrio cholerae 1 Haemophilus aphrophilus 1Campylobacterjejuni 1 Acinetobacter sp 1Bacillus cereus 1 Xanthomonas maltophilia 3

Streptococcus pyogenes 1Staphylococcus aureus 1

The DAB technique was applied to freshlydetected positive blood cultures (Bactec,radiometric) containing Gram negative bacillion two occasions. Dilution of the blood four or10-fold before centrifugation avoided excessbackground due to blood cell debris. Thereaction was negative in both instances andsubcultures subsequently yielded isolates ofEscherichia coli. In each case examination of asecond preparation in which a control oxidasepositive strain (Pseudomonas aeruginosa) hadbeen added to the aliquot of positive bloodculture facilitated a more confident assignmentof negative reactivity to the test organism.

DiscussionThe cytochemical DAB oxidase reaction des-cribed requires a total processing time ofunder25 minutes, assigns appropriate reactivity toindividual bacterial cells and, for the organismstested to date, produces results in accordancewith the standard macroscopic technique. Theexact nature of the activity demonstrated is notclearly defined but is presumably related tocytochrome C oxidase or its equivalents indifferent bacteria. Although correspondencebetween the DAB and macroscopic TPD tech-niques seems to be excellent, the many differentpatterns of electron transport possessed bybacteria11 indicate that this correspondenceshould not be assumed for all genera.The key factors determining feasibility of

cytochemical demonstrations in the immobil-ised preparations used here are access of thereagents to the appropriate cell compartmentand localisation of the final reaction product.Drying seems to be an essential step for accessof DAB as well as the indoxyl substrate usedfor the fJ-galactosidase technique describedpreviously.3 The reasons for this are not clearbut may be related to the disruption of an outerhydrophilic barrier to reagent penetration as

both reactions are diminished if cells are

rehydrated again before substrate is added(Barer and Marsh, unpublished observations).The TPD reagent evidently suffers from poorlocalisation of the final reaction product as

activity was demonstrated rapidly without dry-ing but no colour was observed in individualcells.Minimal non-specific reactivity was

observed with some Enterobacteriaceae. Thisdid not lead to any difficulties in interpretationand may be related to endogenous hydrogenperoxide formation and catalase activity inthese organisms.78 Benzidine combined withhydrogen peroxide has been used as a reagentfor the macroscopic demonstration of cyto-

chrome systems in bacteria'2 and a DAB basedreaction for catalase has been applied to animaltissues.7 On this basis a cytochemical catalasetest should also be feasible in bacteria, buttechnical realisation ofthis has eluded us so far.

It is not suggested that the DAB test shouldreplace the conventional macroscopic oxidasetest, nor that it should be used in the routineassessment of colonial growth. In the clinicalcontext it is envisaged that this and othercytochemical techniques34 13 could be used forcritical specimens where the early availabilityof information about the organisms present, inaddition to their Gram or acid fast properties,could influence subsequent patient man-agement. Appropriate specimens could includepositive blood cultures before subculture,cerebrospinal fluid, peritoneal dialysis fluid,broncho-alveolar lavage specimens and anyother specimens where the choice betweentherapeutic options might be resolved byapplication of one or more cytochemical tech-niques. The ability of the technique to detectthe correct oxidase reactivity of organisms intwo blood cultures provides preliminaryevidence that the presence of tissue compon-ents will not compromise its accuracy.

In conclusion, we have shown that a cyto-chemical DAB based technique can be used todetermine the oxidase reactivity of pathogenicbacteria. Preliminary results indicate that theprocedure can be applied at a stage in theprocessing of clinical specimens when theorganisms present could only be demonstratedby microscopy in most routine laboratories.Because it does not depend on cell growth orrequire pure cultures, we suggest that furtherdevelopment of this cytochemical approach tocharacterising micro-organisms could provideimportant information that would be relevantto patient management much earlier than con-ventional techniques.

We thank Dr S J Pedler for access to clinical isolates from theRoyal Victoria Infirmary, Newcastle upon Tyne.

1 Cowan ST, Steel KG. Manualfor the identification ofmedicalbacteria. 2nd edn. Cambridge: Cambridge UniversityPress, 1974.

2 MacFaddin JF. Biochemical testsfor identification of medicalbacteria. 2nd edn. Baltimore: The Williams & Wilkins Co.,1982.

3 Barer MR. New possibilities for bacterial cytochemistry:light microscopic demonstration of (J-galactosidase inunfixed immobilized bacteria. Histochem J 1991;23:529-33.

4 Barer MR, Entwistle A. Confocal microscopy of surface andcytoplasmically labelled bacteria immobilized by APS-centrifugation. Lett Appl Microbiol 1991;13: 186-9.

5 Mudd S, Winterscheid LC, DeLamater ED, Henderson HJ.Evidence suggesting that granules of mycobacteria aremitochondria. J Bacteriol 1951;62:459-75.

6 Chayen J, Bitensky L, Butcher R. Practical histochemistry.London: John Wiley, 1972:164-7.

7 Lojda Z, Gossrau R, Scheibler TH. Enzyme histochemistry.Berlin: Springer-Verlag 1979:233-8.

8 Old SL, Johnson MA. Methods of microphotometric assayof succinate dehydrogenase and cytochrome c oxidaseactivities for use on human skeletal muscle. Histochem J

_ 1989;21:545-55.9 Seligman AM, Kamovsky MJ, Wasserkrug HL, Hanker JS.

Non-droplet ultrastructural demonstration ofcytochromeoxidase activity with a polymerizing osmiophilic reagent,diaminobenzidine (DAB). J Cell Biol 1968;38:1-14.

10 Robinson PJ, Dunnill P, Lilly MD. Porous glass as a solidsupport for immobilization or affinity chromatography ofenzymes. Biochim Biophys Acta 1971;242:659-61.

11 Morris JG. The metabolism and growth of bacteria. In:Wilson GS, Dick HM, eds, Topley and Wilson's principalsof bacteriology, virology and immunity. Vol I. 7th edn.London: Edward Arnold. 1984:39-69.

12 Deibel RH, Evans JB. Modified benzidine test for thedetection ofcytochrome-containing respiratory systems inmicroorganisms J Bacteriol 1960;79:356-60.

13 Thom SM, Barer MR. Cytochemical tests of bacterialviability. Abstracts of the 162nd meeting of the Path-ological Society of Great Britain and Ireland. J MedMicrobiol 1991;34:iv.

489

on Septem

ber 30, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.6.487 on 1 June 1992. D

ownloaded from