Isolation of Salmonellae from Food Samples

III. Dulcitol Lactose Iron Agar, a New Differential Tube Medium for Confirmation

of Microorganisms of the Genus Salmonella

W. I. TAYLOR AND J. H. SILLIKER

Research Laboratories, Swift & Company, Chicago, Illinois

Received for publication March 3, 1958

The many differential media designed for the clinicalmicrobiologist have reflected his interest in the delinea-tion of a wide variety of genera within the entericgroup of organisms. The traditional differential tubemedia, Russell double sugar agar, Krumweide triplesugar agar, Kligler iron agar, and triple sugar iron agar(Hajna, 1945), have been invaluable aids in the promptestablishment of the biochemical patterns which areof generic significance in the identification of Entero-bacteriaceae. These media were created especially forversatility in order to secure the widest applicability.In the laboratory of the food analyst, the greater em-phasis placed specifically on the detection and enumera-tion of members of the genus Salmonella imposes a re-quirement for a medium designed to facilitate thattechnique. In previous papers, the limitations of theenrichment broths (Silliker and Taylor, 1958) and thecapabilities of differential plating media (Taylor et al.,1958) have been demonstrated and discussed. The needfor a simplified test for confirmation of presumptivelypositive colonies of Salmonella on brillant green agarplates became increasingly apparent as these worksprogressed. The tubed differential media used were oflimited value because of the multiplicity of enteric or-ganisms capable of producing reactions identical to thesalmonellae. This, in turn, necessitated the use of manyadditional biochemical tests, such as several carbohy-drate fermentations, indole and urease tests. An in-vestigation to reduce the time and labor involved inbiochemical confirmation of suspect salmonellae cul-minated in the differential medium herein described.

MATERIALS AND METHODS

Dulcitol lactose iron agar (DLI) is a medium designedto differentiate between the genus Salmonella and otherenteric genera. In its most effective form, it is a double-poured medium having slant and butt of different com-

position. It is designed to disclose the fermentation oflactose and dulcitol and the production of hydrogensulfide. The directions given will make 2 L of the com-

pleted medium.

Solution A (slant):

Phenol red agar base (Difcol/BBL2)Lactose, CPFerrous ammonium sulfateSodium thiosulfateWater, distilled

Solution B (butt):

Phenol red agar baseWater, distilled

30 g10 g0.2 g0.2 gI L

30 g1 L

Solutions A and B are brought to a rolling boil inseparate containers, autoclaved for 20 min at 15 lb pres-sure, and cooled to 55 C. A Seitz-filtered 10 per centsolution of dulcitol, CP, is added to solution B in theamount of 20 ml per L (0.2 per cent). If there is a notice-able color change after the addition of the carbohy-drate, restore the pH to about 7.4 with a few drops of1 N NaOH. Upon occasion when a relatively crudedulcitol was used in lieu of CP grade, there was somecolor change, albeit without gas production, after incu-bation with nondulcitol fermenting organisms. This isassumed to result from the fermentation of trace carbo-hydrate impurities which, when coupled with the rangeof the phenol red indicator, is sufficient to produce aslight acid reaction. The same dulcitol in bromeresolpurple sugar broth (BCP) is used to verify the fer-mentation and has always been reliable. Crude dulcitolis now used in BCP broth and CP grade in the DLIslants.

Solution B is dispensed aseptically in 15 by 125 mmscrew-capped tubes in 2 to 3 ml amounts and allowedto solidify. Solution A is added in like amounts andslanted to form a tube half slant and half butt. Sterileseparatory funnels are excellent for dispensing thesolutions.The medium is inoculated by streaking and stabbing,

incubated for 18 to 24 hr at 37 C, and the reactionsread.

I Difco Laboratories, Detroit, Michigan.2 Baltimore Biological Laboratories, Baltimore, Maryland.

335

on June 3, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

W. I. TAYLOR AND J. H. SILLIKER

RESULTS AND DISCUSSION

When Kligler's or triple sugar iron agar (TSI) slantsare used for the presumptive detection of salmonellae,Proteus, Arizona, and Ballerup-Bethesda group or-ganisms produce reactions identical to those of thesalmonellae, that is, alkaline slant, H2S blackening, andacid and gas in the butt. A search was initiated for acarbohydrate fermented exclusively by salmonellaewhich would supplement the primary differentiationafforded by lactose. No such carbohydrate was found,but dulcitol was observed to approach this ideal mostclosely. The obvious advantages secured by the inclu-

TABLE 1Fernmentation reactions of enteric groups

Group* Lactose Dulcitol H2S Ninhydrin

Salmonella ............... -t + () +Arizona .............. X - + +Ballerup-Bethesda ........ (+) +Escherichia ............... + V (-)+Klebsiella .............. + V _ +Providence ............... _Proteus ........ - - .(+) V

* Nomenclature and most fermentationi reactioins shown,from Kauffmann, 1951.

t Key to reactions: - = negative; + = positive in one day;(+) = most strains positive, a few negative; (-) most strainsnegative, a few positive; X = delayed positives or negative;V = variable with serotype or strairns.

sion of dulcitol in the medium are shown in table 1.Since neither the Proteus nor Arizona groups fermentdulcitol, these are easily distinguished from salmonellaewhich ferment it readily in 24 hr with but few excep-tions. Many organisms of the Ballerup-Bethesda groupare able to ferment dulcitol, however, and these con-stitute the only false positives which occur in the useof the DLI medium.

Separation of the salmonellae from the Ballerup-Bethesda groups is effected by use of the ninhydrin test.This is the only other biochemical reaction necessary tocomplete the confirmation of salmonellae short of anti-gen analysis. The ninhydrin test (Carlquist, 1956) isused with onie change in technique: no centrifugationis used, 5 to 10 min of standing effecting suitable separa-tion of the chloroform layer. Salmonellae are ninhydrinpositive; Ballerup-Bethesda group organisms are not(table 1).When H2S producing organisms are incubated on TSI

or Kligler's iron agar slants, the fermentation of theglucose in the butt is often obscured by excessiveblackening. This is of relatively small importance be-cause all such organisms are glucose fermenters andthe assumption is that a positive butt has been ob-scured. The significance of dulcitol fermentation as adifferential reaction makes it imperative that the resultbe observed. The omission of the sulfide indicator wouldlose a valuable diagnostic characteristic. On the otherhand, early trials with the medium produced equiv-

Figure 1. A comparison of the biochemical reactions of closely related enteric organisms on Kligler's iron agar and dtulcitol lactoseiron agar slants.

[VOL. 63.36

DIEWITOL JACTOU IROrtA. 41.3,AR

41. 0

on June 3, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

ISOLATION OF SALMONELLAE FROM FOOD

ocal results because many organisms, especially pro-tei, blackened all but the uppermost portion of theslant. The amount of the indicator was reduced withno appreciable improvement in the medium. Differentcompositions for butt and slant portions of the mediumwere used in order to limit the blackening reaction.Since separate pouring would be used, there was noneed to mix the carbohydrates, so amounts of dulcitoland lactose to be used were determined by trial. Thefinal formula evolved is one in which the dulcitol isconfined to the butt in such a concentration thatfermentation reactions are unequivocal. The mediumis stable and it is not normally kept refrigerated but al-lowed to stand at room temperature. Under these con-ditions, there has been no noticeable migration on theindicator or carbohydrates and exceedingly crisp andclear.-cut reactions characterize the medium. For pur-poses of comparison, duplicate cultures of organismswere used to inoculate Kligler's iron agar and DLI. Thetotal blackening of the butt of the Kligler's may be ob-served in contrast to the distinctive separation of fer-mentation reactions shown by the DLI. All of the

Providence Gp.

Pseu domonas

Parocolobactrum

Arizona Gp. So

Proteus

3Imonella

Col iform

E. freundii

KEYBallerup,Bethesdd E

Red Yellow H2S

Figure 2. The characteristic reactions of enteric organismson dulcitol lactose iron agar slants.

organisms except Escherichia freundii were identical inappearance on Kligler's; on DLI, E. freundii differedfrom Salmonella and Ballerup by having a yellow slant,the latter differed from Proteus and Arizona by havingyellow fractured (acid and gas) butts (figure 1). Aschematized presentation of the correlation of typicalreactions to commonly encountered organisms fromenrichment broths is showvn in figure 2. The typicalcoliform and E. freundii reactions should, of course, bescreened out of contention by their appearance on bril-liant green agar plates or other such differential platingmedia but, unfortunately, they will occasionally showas lactose-negative or doubtful colonies when first iso-lated because of initially delayed fermentation reactions.

In use, the medium has proved itself both valid andreliable. In the course of more than 150 isolations ofnaturally occurring salmonellae, more than 95 per centof these organisms have produced the characteristicappearance of red slant, yellow fractured butt, andblack-banded middle. The only exceptions have beenone dulcitol negative S. worthington, and a few1H2S-negative strains of various serotypes.The need for a single medium, DLI, to replace two

media such as Kligler's and dulcitol broth is best under-stood from the viewpoint of the requirements of Sal-monella enumeration by the most probable number(MPN) method, discussed in a previous paper of thisseries (Silliker and Taylor, 1958). By this method, in a9 tube MPN determination, for example, 3 dilutions ofsample would be used to inoculate triplicate enrichmentbroths and thence streaked to brilliant green agarplates. From these 9 plates a statistically valid numberof Salmonella, suspect colonies, usually from 3 to 8 perplate, must be picked for identification as salmonellaein order to determine the population density. Thiscould amount to 72 cultures per sample, and manysamples require more than a 9 tube determination. Ifone were to use the two media mentioned in place ofDLI, it would obviously entail twice the storage spacefor the tubed media, two inoculations per culture,twice the incubator space, two readings of reactions,and twice the amount of glassware to disinfect, handleand wash. Under these circumstances the reductionfrom two media to one has been amply justified in ourexperience.

SUMMARYA new differential tube medium, dulcitol lactose iron

agar, which facilitates the biochemical separation ofgenus Salmonella from other commonly encounteredenteric organisms, is presented.The marked superiority of dulcitol lactose iron agar

medium over Kligler iron agar and triple sugar ironagar slants is demonstrated to result from use of a car-bohydrate less frequently fermented by nonsalmonellaethan glucose, and from thc method of construction in

1958] 337

\ N

on June 3, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

BRAZIS, LESLIE, KABLER, AND WOODWARD

which the butt reaction is never obscured by excessiveH2S formation.The only nonsalmonellae which produce the char-

acteristic salmonella reactions on dulcitol lactose ironagar are Ballerup-Bethesda paracolons which may bedifferentiated by the ninhydrin test.

REFERENCES

CARLQUIST, P. R. 1956 A biochemical test for separatingparacolon grouips. J. Bacteriol., 71, 339-341.

HA.NA, A. A. 1945 Triple-sugar iron agar medium for the

identification of the intestinal group of bacteria. J.Bacteriol., 49, 516-517.

KAUFFMANN, F. 1951 Enterobacteriaceae. Ejnar Munksga-ard Ltd., Copenhagen, Denmark.

SILLIKER, J. H. AND TAYLOR, W. I. 1958 Isolation of sal-monellae from food samples. II. The effect of addedfood samples upon the performance of enrichment broths.Appl. Microbiol., 6, 228-232.

TAYLOR, W. I., SILLIKER, J. H., AND ANDREWS, H. P. 1958Isolation of salmonellae from food samples. I. Factorsaffecting the choice of media for the detection and enu-meration of Salmonella. Appl. Microbiol., 6, 189-193.

The Inactivation of Spores of Bacillus globigii and Bacillusanthracis by Free Available Chlorine1

A. R. BRAZIS, J. E. LESLIE, P. W. KABLER, AND R. L. WOODWARD

Robert A. Taft Sanitary Engineering Center, Bureau of State Services, Public Health Service, U. S. Departmentof Health, Education, and Welfare, Cincinnati, Ohio

Received for publication March 4, 1958

The germicidal action of chlorine and chlorine-yieldingcompounds has been studied by a number of investi-gators. However, two important variables were notconsidered: the effect of pH changes and the distinctionbetween free available, combined available, and com-bined chlorine.

Studies conducted by Tilley and Chapin (1930) onanthrax spores indicated that nitrogen trichloride,chlorine in neutral solutions, and chlorine plus N/50HCl were effective in 15 min with available chlorineconcentrations of 10 mg per L or more. In a more re-cent study (Fair et al., 1947), titratable free chlorineresiduals of 1 and 3 mg per L produced a 99.95 per centkill of Bacillus anthracis spores at pH 5.0 and 25 C in60 and 30 minm, respectively. Based on the amount ofhypochlorous acid present, Moore (1951) stated that3 mg per L produced 100 per cent kill of B. anthracisspores in 30 min at 25 C.

This investigation was undertaken to determine theeffects of free available chlorine, hereinafter referredto as FAC, upon the spores of Bacillus globigii (Bacillussubtilis morphotype globigii) and Bacillus anthracisunder the following conditions: (1) hydrogen-ion con-centrations of 6.2, 7.2, 8.6, and 10.5; (2) exposure tem-peratures of 4 and 22 C; and (3) exposure periods of 2and 48 hr. The results obtained should provide practicalinformation on water treatment plant operation intimes of disaster.

' This work was performed under contract with the D)epart-ment of the Navy, Bureau of Yards and Docks.

MIATERIALS AND METHODSTest cultures. Bacillus anthracis no. 811 and Bacillus

globigii no. 102 were obtained from The Ohio StateUniversity.

Development of media. In initial studies with spore-forming microorganisms, media which would inducerapid germination during incubation at 35 C weretested. Nutrient agar plus 0.5 per cent yeast extractand 0.3 per cent sodium citrate seemed to be mostsuitable for B. globigii spores. Tryptose agar plus 0.2per cent liver extract proved to be adequate for B.anthracis spores.

Studies were also made to determine the mediumwhich would attain maximum sporulation of the testorganisms under specified conditions. A mineral mediumcontaining 0.1 per cent Casamino acids, 0.25 per centglucose, 0.5 per cent yeast extract, 0.01 per cent man-ganese sulfate, 0.0001 per cent ferrous sulfate, and 3per cent agar, adjusted to pH 6.8 with sodium hy-droxide, provided the greatest number of matureviable spores.

Vegetative cell autolysis. The presence of vegetativecells in spore suspensions might lead to endogeneousrespiration, resulting in the germination of spores duringFAC evaluations. Germination of the spore suspensionsunder the conditions of the investigation could result inspurious observations. Autolysis obtained by a com-bination of heat shock plus incubation at 45 C, for 24hr, proved to be the most efficient method for destroy-ing vegetative cells.

[VOL. 6338

on June 3, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from