JOURNAL OF BACTERIOLOGY, Dec. 1967, p. 1846-1853 Vol. 94, No. 6Copyright @ 1967 American Society for Microbiology Printed in U.S.A

Biosynthesis of Biotin in MicroorganismsV. Control of Vitamer Production

JEROME BIRNBAUM,' CHIK H. PAI,2 AND HERMAN C. LICHSTEINDepartment of Microbiology, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45219

Received for publication 16 September 1967

Use of a yeast-lactobacillus differential microbiological assay permitted investi-gation into the synthesis of biotin vitamers by a variety of bacteria. A major portionof the biotin activity was found extracellularly. The level of total biotin (assayablewith yeast) greatly exceeded the level of true biotin (assayed with lactobacillus).Values for intracellular biotin generally showed good agreement between the assays,suggesting the presence of only true biotin within the cells. Bioautographic analysisof the medium after growth of each organism revealed the presence of large amountsof a vitamer which corresponded to DL-desthiobiotin on the basis of RF value andbiological activity. Biotin, when detected at all, was at very low concentrations.Also, an avidin-uncombinable vitamer was synthesized by a majority of the bacteria.Addition of D-biotin to the growth medium prevented completely the synthesis ofboth vitamers of biotin. D-Biotin-D-sulfoxide had no effect on the synthesis of desthio-biotin or the avidin-uncombinable vitamer. Addition of DL-desthiobiotin did notprevent its own synthesis nor that of the other vitamer. Control of vitamer synthe-sis is therefore highly specific for D-biotin. The avidin-uncombinable vitamer wasproduced only at repressed levels in the presence of high concentrations of both D-biotin and DL-desthiobiotin, which suggested that it is not a degradation product ofthese substances. A possible mechanism for the overproduction of the biosyntheticprecursors of biotin is presented.

- Biotin is overproduced by a variety of bacteriaand fungi (2-5, 14, 15, 17). Moreover, a majorportion of the vitamin activity detected in theculture filtrates of these microorganisms isassociated with desthiobiotin and a vitamer ofbiotin uncombinable with avidin.We recently surveyed several microorganisms

with regard to the biotin vitamers produced andalso inspected the metabolic control of theirsynthesis (16-20).

MATERIALS AND METHODS

Organisms. The bacterial strains used were: Esch-erichia coli (Crookes), E. coli (Davis), E. coli thi- (athiamine requiring strain derived from the Davis wild-type strain), E. coli (Gratia), E. coli (Texas), Aero-bacter aerogenes (Tennessee), A. aerogenes (649),Enterobacter aerogenes subspecies alvei (formerlycalled Bacterium cadaveris), Proteus morganii (pH), P.mirabilis, P. vulgaris, and Bacillus cereus (B-48). Stockcultures were maintained on agar slants containing1% yeast extract (Difco), 1% casitone (Difco), 0.5%

'Present address: Merck Sharp and Dohme Re-search Laboratories, Rahway, N.J. 07067.

'Present address. Department of Microbiology,University of Alberta, Edmonton, Alberta, Canada.

glucose, 0.5% K2HPO4, and 1.5% agar. The orga-nisms were transferred to fresh maintenance media atmonthly intervals, incubated 24 to 48 hr at 30 C, andthen refrigerated.

Growth media and conditions. Wild-type strains ofE. coli and the two strains of A. aerogenes were grownin a chemically defined medium as described previ-ously (16). Growth of E. coli thi- was accomplishedin the same medium, with the addition of 1.5 mg/literof thiamine hydrochloride. For the growth of E.aerogenes, casein hydrolysate was added to the syn-thetic medium at a final concentration of 0.2%.Proteus species were grown in the medium used forthe E. coli wild types, with the following additions to1 liter of medium: 1.5 mg of thiamine hydrochloride,1.5 mg of calcium pantothenate, 0.15 mg of p-amino-benzoic acid, 0.75 mg of folic acid, 1.5 mg of nicotinicacid, 3.0 mg of pyridoxine hydrochloride and 2.0 g ofcasein hydrolysate. However, as described in the dis-cussion of the data of Table 3, some experiments withP. morganii were performed with this medium lackingonly the casein hydrolysate. The growth medium forB. cereus (B-48) contained the above vitamins and thefollowing in grams per liter: glucose, 20; (NH4)2SO4,4.0; KH2PO4, 1.0; K2HPO4, 1.0; NaCl, 1.0;MgSO4.7H20, 0.7; and casein hydrolysate, 2.0. D-Biotin and DL-desthiobiotin were added to the growthmedium as described in the Results section. Growth of

1846

on July 7, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

CONTROL OF BIOTIN VITAMER BIOSYNTHESIS

all organisms was in 200 ml of medium in 500-mlErlenmeyer flasks. Cultures were incubated on a NewBrunswick gyrotary shaker at 150 strokes per min for16 hr. Temperature of incubation for B. cereus was 30C, and all other bacteria were grown at 37 C. In-oculum cells were prepared from cultures of eachorganism grown for 16 to 18 hr in their respectivemedia. Cells were collected by centrifugation, washedthree times with sterile 0.8% NaCl, and the final cellpellets were resuspended in 0.8% NaCl to give theoriginal cell densities. Each 100 ml of growth mediumwas inoculated with 0.1 ml of these suspensions.

The casein hydrolysate (vitamin-free casaminoacids) used for preparation of the growth media waspurchased from Difco Laboratories, Detroit, Michi-gan. The commercial vitamin-free casamino acidswere purified further by charcoal adsorption to re-move traces of biotin (21). This precaution was nec-essary to reduce the endogenous growth levels causedby the presence of small amounts of biotin as a con-taminant. The possibility that growth stimulationmight be due to the sparing effect of an amino acid,such as aspartic acid, is precluded, since this phenome-non is observed only with organisms which requirebiotin rather than those which synthesize the vitamin.Moreover, the assay organisms used were not affectedby the casein hydrolysate in the absence of addedbiotin as may be seen by the negligible endogenousgrowth levels in Fig. 1 and 2. D-Biotin and DL-desthiobiotin were purchased from California Foun-dation for Biochemical Research, Los Angeles, Calif.Avidin (3,300 units/g) was obtained from NutritionalBiochemicals Corp., Cleveland, Ohio. The avidin wasdissolved in distilled water, sterilized by filtration, andstored in a freezer.

Microbiological assays. Biotin was assayed withLactobacillus plantarum ATCC 8014, formerly calledL. arabinosus, according to the gerneral procedure ofWright and Skeggs (26). Saccharomyces cerevisiaeATCC 9896 was also used for assay according to theprocedure of Hertz (8). Slight modifications in bothmethods were used. These changes, as well as thepreparation of samples for microbiological assay,have been described previously (17). Since L.plantarum is limited in its biotin requirement, theassay values obtained with this organism representtrue biotin. S. cerevisiae utilizes a much greater varietyof biotin vitamers than does the lactobacillus, andtherefore the values obtained with this organism arerepresentative of total biotin. Differences betweentotal and true biotin values were indicative of thepresence of biotin vitamers inactive for the lacto-bacillus but utilizable for the yeast. To observe thepresence of vitamers of biotin uncombinable withavidin, excess avidin (0.02 to 0.04 mg) was addedaseptically to each tube before seeding with the yeast.The amount of avidin-uncombinable vitamer wasexpressed as biotin equivalents for S. cerevisiae.

Bioautography of biotin vitamers. Preparation ofculture filtrates for bioautography was as describedpreviously (19). Descending paper chromatographywas employed with a solvent system composed of n-butanol-water-acetic acid (4:5:1) and Whatman no. 1paper strips. The solvent front was permitted to travel

about 30 cm at room temperature. After air-drying,the chromatograms were subjected to a bioauto-graphic technique as described by Wright and co-workers (25). The paper strips were cut into 1-cmsections and were placed in separate tubes where theywere eluted with distilled water. The eluates wereautoclaved, cooled, and assayed for biotin activitywith S. cerevisiae and L. plantarum. When the tur-bidity was plotted as a function of the distance fromthe origin, a smooth curve connecting the pointspermitted localization of the RF value by interpolationto within 0.03 units.

RESULTS

In agreement with previous results from thislaboratory (17), the medium of E. coli (Crookes)after growth was found to contain high biotinactivity assayable with S. cerevisiae and a muchlesser amount of true biotin active for L. plan-tarum (Table 1). Values for intracellular biotinshowed excellent agreement between the twoassays. This suggested that the medium con-tained a high level of biotin vitamers inactive forthe lactobacillus, whereas the cells contained onlytrue biotin. This same general pattern appearedwith all other bacteria tested, except for the Pro-teus species and B. cereus. Apparently these cellsretain some of the biotin vitamers intracellularly.Further, these organisms excreted by far thegreatest amount of biotin vitamers as measuredwith S. cerevisiae. Overproduction of true biotinby all bacteria tested was slight, since the levelexcreted relative to total biotin ranged from zerofor E. coli thi-, A. aerogenes (Tennessee), E.aerogenes, and B. cereus, to 19.5% for E. coli(Crookes). This same pattern has been reportedfor a large number of different microorganisms,including fungi, bacteria, and streptomyces (14,15).The extent of final growth, whether due to

differences in inoculum size, character of themedium employed, or type of organism, did notappear to be of significance. For example, theProteus species gave poorest growth but yieldedhigh levels of biotin vitamers; B. cereus gaveexcellent growth and high vitamer levels; and A.aerogenes (649) grew well but produced low levelsof vitamers. Thus, there was no correlation be-tween amount of growth and magnitude of vita-mer excretion. In fact, it has been established byPai and Lichstein (17) that the maximal rate ofproduction of biotin occurs very early in thegrowth cycle. Further, they showed that the levelof the enzymes which convert desthiobiotin tobiotin are at a maximum early in growth anddecrease thereafter. These enzymes could beelevated later if they were derepressed by theaddition of avidin to the growth medium (20).

1 847VOL. 94, 1967

on July 7, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

BIRNBAUM, PAI, AND LICHSTEIN

TABLE 1. Intracellular and extracellular biotin concentrations of various bacterial cultures-

Biotin assays

Organism Intracellular Extracellular

Saccharomyces Lactobacillus S. cerevisiae L. plantarumcerevisiae plantarrm S

Escherichia coli (Crookes). ... 8.4b 7.4 45 8.8E. coli (Davis).6.2 6.0 56 1.9E. coli (Gratia). 7.7 7.5 48 0.2E. coli (Texas). 8.2 7.7 44 3.3E. coli(thi).6.5 6.2 29 0.0Aerobacter aerogenes (Tennessee). 17.0 17.0 37 0.0A.'aerogenes (649).8.8 9.9 12 0.4Enterobacter aerogenes ................... 25.0 24.0 209 0.0Proteus morganii (pH).22.0 15.0 198 8.8P. mirabilis.26.0 15.0 142 7.5P. vulgaris.21.0 14.0 146 6.8Bacillus cereus (B-48) ................|. 32.0 25.0 485 0.0

a Media and growth conditions as described in Materials and Methods.b All values given as 10-4 ,g/ml of culture.

Bioautography of crystalline D-biotin with thesolvent system used in this study revealed twopeaks of activity for both the lactobacillus andyeast assays. Peak of activity at RF 0.80 to 0.84corresponds to D-biotin, whereas the other peak(RF 0.57 to 0.61) represents D-biotin-D-sulfoxide.Oxidation of D-biotin to the sulfoxide form hasbeen observed by several investigators (1, 19, 25)and is thought to be caused by some componentof the chromatography paper (12). When lowlevels of D-biotin were chromatographed, onlythe peak for D-biotin was detected. When D-biotin was dissolved in a complex medium con-taining sulfhydryl compounds, the sulfoxideform failed to appear even when relatively highlevels of D-biotin were employed (unpublisheddata). Thus, D-biotin-D-sulfoxide, detected on thechromatograms of the culture media used in thisstudy, is not considered a microbial product. DL-Desthiobiotin shows one peak (RF 0.88 to 0.91)active for S. cerevisiae but inactive for L. plan-tarum. Though it was not always possible todistinguish between D-biotin and DL-desthio-biotin purely by RF values, they were differenti-ated easily by microbiological activity of thechromatograms.

Bioautography of the culture medium of E.coli (Crookes) revealed that a majority of thebiotin activity stems from two vitamers which areinactive for L. plantarum (Fig. 1). One of thesematerials corresponds to DL-desthiobiotin on thebasis of RF value (0.89) and microbiologicalactivity. The other compound (RF 0.67 to 0.69)represents an avidin-uncombinable vitamer whichis active for S. cerevisiae. A vitamer with the

same RF value, microbiological activity, and lackof avidin combinability has been detected in theculture medium of a variety of microorganisms(2, 3, 7, 14). A small peak of activity, which doessupport the growth of L. plantarum, is combin-able with avidin, and has an RF value correspond-ing to D-biotin, was also detected.When D-biotin was added to the growth

medium of E. coli (Crookes), the peaks corre-sponding to desthiobiotin and the avidin-uncom-binable vitamer were no longer detectable (Fig.2). In this case, peaks for D-biotin (RF 0.81 to0.83) and D-biotin-D-sulfoxide (RF 0.58 to 0.60)were observed. Both materials produced a growthresponse with the yeast and lactobacillus assays.All of the E. coli strains listed in Table 1 pro-duced almost identical results with those shownfor E. coli (Crookes). The other bacteria werestudied in the same manner and results are sum-marized in Table 2. A. aerogenes (Tennessee) andE. aerogenes excreted only one material duringgrowth which corresponded to DL-desthiobiotin.Neither D-biotin nor avidin-uncombinable vita-mer was detected in the medium of these two or-ganisms, although D-biotin (i.e., a materialassayable with L. plantarum) was found intra-cellularly (Table 1). When exogenous D-biotinwas supplied to these organisms, the accumula-tion of desthiobiotin in the medium ceased, andbioautography revealed the presence of D-biotin(major peak) and D-biotin-D-sulfoxide. The twospecies of Proteus showed patterns identical withthose of the E. coli strains. These bacteria pro-duced D-biotin, a material corresponding todesthiobiotin, and the avidin-uncombinable

1848 J. BACTERIOL.

on July 7, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

CONTROL OF BIOTIN VITAMER BIOSYNTHESIS

10 15 ;DISTANCE FrOM ORIGIN (cm)

FIG. 1. Bioautography of medium after growth ofEstherichia coli (Crookes). (A) Assayed with Sac-charomyces cerevisiae; I Klett unit = 0.0036 mg (dryweight) of cells, with blue filter no. 42, 400 to 450 mu.(B) Assayed with S. cerevisiae in the presence ofavidin.(C) Assayed with Lactobacillus plantarum (I Klettunit = 0.0013 mg (dry weight) of cells, with red filterno. 66, 640 to 700 m,u.

vitamer (RF approximately 0.7). When D-biotinwas supplied, the vitamers no longer accumu-lated in the medium. B. cereus produced the twovitamers of biotin, but no biotin was detected.Addition of exogenous D-biotin to the Bacillusmedium prevented the accumulation of desthio-biotin and the unknown vitamer. The bioauto-graphic data were therefore in good agreementwith those obtained by differential assay of theculture supernatant fluids (Table 1). Thus, in allbiotin-synthesizing organisms studied, addition ofD-biotin prevented accumulation of desthiobiotinand the avidin-uncombinable vitamer in thegrowth medium. The vitamers were not retainedwithin the cells, since values for intracellularbiotin showed excellent agreement between theyeast and lactobacillus assays. Furthermore, thevitamers could not be degradation products ofbiotin, because the addition of high levels of biotinprevented rather than enhanced synthesis of thevitamers.The only organism in which synthesis of the

biotin vitamers did not seem at first to be con-trolled by D-biotin was P. morganii (Table 3).

0 5 10 15 20

DISTANCE FROM ORIGIN (cm)

25 30

FIG. 2. Bioautography of medium with added D-biotin after growth of Escherichia coli (Crookes).Initial concentration ofD-biotin, 100 X 10-4,g/ml. (A)Assayed with Saccharomyces cerevisiae (0), assayedwith S. cerevisiae in the presence of excess avidin (0).(B) Assayed with Lactobacillus plantarum. See legendof FIG. 1 for Klett unit equivalents in dry weight ofcells.

When this organism was grown in a mediumdevoid of biotin and casein hydrolysate, itexcreted biotin vitamers in the same fashion as theother Proteus and E. coli strains. However, apeak for D-biotin-D-sulfoxide had appeared onthe chromatogram and this was not observedwith the other organisms when grown withoutexogenous D-biotin. Moreover, the peak of ac-tivity at RF 0.82 to 0.83 (D-biotin) was very smallrelative to the peak for the sulfoxide form. WhenD-biotin was added to the medium, it still couldnot be detected on the chromatograms. However,there was a large amount of activity associatedwith a peak corresponding to D-biotin-D-sulfoxide.It appeared, therefore, that theaddedD-biotin wasoxidized to the sulfoxide form. The sulfoxidecould not repress synthesis of the biotin vitamers,since RF values corresponding to the avidin-un-combinable vitamer and desthiobiotin still weredetected on the chromatograms. When uninocu-lated media containing D-biotin were subjected tobioautography (after shaking for 16 hr at 37 C),only D-biotin-D-sulfoxide was detected, thusillustrating that the oxidation was not caused bythe organism but rather by some component ofthe medium. When casein hydrolysate was addedto the growth medium, the organism synthesized

1849VOL. 94, 1967

on July 7, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

BIRNBAUM, PAI, AND LICHSTEIN

TABLE 2. Bioautography of media after growthof various bacteria

Organism

Aerobacter aer-ogenes (Tennes-see)

Eniterobacteraerogenes

Proteus vulgaris

Proteus mirabilis

Bacillus cereus(B-48)

Concn ofD-biotinin growthmedium(10-4 /Ag/ml)

0200

0200

0

200

0

200

0

200

Biotin

Saccharo-myces

cerevisiae

RF0.870.590.830.890.620.830.690.830.910.610.830.700.820.900.600.840.710.900.620.84

assays

Lacto-bacillus

plantarum

RFNAb0.600.83NA0.600.84NA0.84NA0.600.84NA0.82NA0.600.83NANA0.620.83

Avidincombin-ability'

+++

++++

a Combinable with avidin (+); avidin uncom-binable (-).

I NA = no activity.

biotin, desthiobiotin, and the avidin-uncombin-able vitamer (Table 3). In this case, true biotinactivity was somewhat higher than in the mediumlacking casein hydrolysate, and the sulfoxideform was not detectable. Thus, casein hydrolysateappears to prevent oxidation of D-biotin. Addi-tion of D-biotin to the medium containing caseinhydrolysate prevented synthesis of the biotinvitamers in the same fashion as with the otherorganisms (Table 2). In this case, activitiescorresponding to D-biotin and D-biotin-D-sul-foxide were detected; the sulfoxide peak was verysmall relative to the size of the peak of activityfor D-biotin. Thus D-biotin-D-sulfoxide is a poorrepressor of biotin vitamer formation, and thecontrol of vitamer synthesis appears to behighly specific for D-biotin. Other vitamers ofbiotin have been shown to be unable to represssynthesis of the biotin-synthesizing enzymes in E.coli (17, 19).The quantity of each vitamer excreted by

several microorganisms grown in the presenceand absence of exogenous DL-desthiobiotin wasstudied (Table 4). Levels of true biotin (D-biotin)were determined by direct assay of the mediumwith L. plantarum. Avidin-uncombinable vitamer

TABLE 3. Bioautography of media after growth ofProteus morganii (pH)

Biotin assays

Addition to basal Avidingrowth medium Saccear Lacobacilus combin-

cerevisiae plantarum

RF RFNone 0.59 0.60 +

0.70 NAb -0.82 0.83 +0.88 NA +

D-BiOtin (200 X 0.61 0.60 +10-4 mg/ml) 0.69 NA

0.89 NA +Casein hydrolysate 0.68 NA

(0.2%) 0.83 0.84 +0.87 NA +

D-BiOtin + casein 0.60 0.61 +hydrolysate 0.84 0.84 +

a Combinable with avidin (+);binable (-).bNA = no activity.

avidin-uncom-

was measured with the yeast assay by the additionof excess avidin to the assay tubes suspected ofcontaining this material. Quantitation of desthio-biotin was accomplished by subtracting both thevalues for true biotin and avidin-uncombinablevitamer from the S. cerevisiae assay values fortotal biotin. E. coli and Proteus strains accumu-lated approximately equal amounts of desthio-biotin and the avidin-uncombinable vitamer inthe growth medium (Table 4). Concentration oftrue biotin was again very low relative to thelevel of either vitamer. B. cereus produced largeamounts of desthiobiotin, a lesser concentrationof avidin-uncombinable vitamer, and true biotinwas not detectable. In the presence of exogenousdesthiobiotin, more of the same was excreted.Biotin and the avidin-uncombinable vitamer wereproduced in either the same concentration as inthe absence of desthiobiotin or at increasedlevels. Once again, B. cereus failed to excretetrue biotin assayable with L. plantarum. Theseresults suggest that DL-desthiobiotin, as well asD-biotin-D-sulfoxide, cannot suppress the syn-thesis of biotin vitamers by these bacteria. D-Biotin apparently is very specific in this regard.

It was noticed that the avidin-uncombinablevitamer always appeared in medium coincidentwith high concentrations of desthiobiotin. Also,addition of DL-desthiobiotin to the medium ofP. vulgaris and B. cereus caused an increasedproduction of the vitamer. Furthermore, Asper-gillus oryzae has been shown to degrade desthio-biotin to a biologically active vitamer which is

1850 J. BACTERIOL.

on July 7, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

CONTROL OF BIOTIN VITAMER BIOSYNTHESIS

TABLE 4. Concentration of biotin vitamers in media after growth of various bacteria

Concn of biotin vitamersa

Organism Desthiobiotin Biotin Uncharacterized vitamers8'-Des- +Des- -Des- +Des- -Des- +Des-

thiobiotin thiobiotin' thiobiotin thiobiotin thiobiotin thiobiotin

Escherichia coli (Crookes) 50 143 7.5 5.2 59 47E. coli (Davis). 29 219 1.9 2.0 25 19Proteus morganii (pH).132 328 5.0 6.3 133 146P. vulgaris.76 176 2.1 2.5 89 150Bacillus cereus (B-48) ......... 423 670 0 0 77 250

a All values given as 10-4 Mg/ml of medium.b Expressed as D-biotin equivalents for S. cerevisiae.c DL-Desthiobiotin added to media at a concentration of 200 X

10-4 ug/ml of D-biotin equivalents for S. cerevisiae.

not on the pathway of biotin synthesis (10). Thesefindings suggest that the vitamer detected heremight be a degradation product of desthiobiotin.As shown previously, the synthesis of desthio-biotin was prevented by the addition of D-biotinto the media. Under these conditions, the avidin-uncombinable vitamer also was not produced. Itwas thought possible that the lack of productionof desthiobiotin, rather than the repressive effectof D-biotin, was preventing the synthesis of the un-known vitamer. To test this, we studied the accu-mulation of the avidin-uncombinable vitamer inmedia supplemented with D-biotin or DL-desthio-biotin, or both (Table 5). Synthesis of the vitamerwas almost completely prevented by the additionof D-biotin, whereas supplementation of themedia with DL-desthiobiotin had little effect onthe excretion of the vitamer. More important,the avidin-uncombinable vitamer was producedonly at the repressed level in the media contain-ing both D-biotin and DL-desthiobiotin (columnD). This suggested that the vitamer was not adegradation product of desthiobiotin, sinceample desthiobiotin was supplied. Apparently,synthesis of the vitamer as well as desthiobiotinwas under metabolic control by D-biotin.

DIscussioNDesthiobiotin has been considered an inter-

mediate in the biosynthetic pathway leading tobiotin, since it can be converted to the vitamin bybiotin-requiring microorganisms (11, 22, 23).This has been inferred also from studies of amutant strain of Penicillium chrysogenum whichcould synthesize the vitamer but could not con-vert it to biotin (23). The original strain syn-thesized biotin and also converted desthiobiotinto biotin. Addition of pimelic acid, a knownintermediate in biotin biosynthesis (6), to the

10 4g/ml which represents 100 X

TABLE 5. Concentration of the avidin-uncombinablevitamer in media after growth of

various bacteria

Additions to growth mediaOrganism

Aa B C D

Escherichia coli(Crookes).59b 4.5 47 6.8

E. coli (Davis).25 0.9 19 0.9Proteus morganii (pH) 150 2.4 150 2.7P. vulgaris.89 1.4 92 1.5Bacillus cereus (B-48). 77 3.5 250 4.3

a A = none; B = 100 X 104 Ag/ml of D-biotin;C = 200 X 10 4,ug/ml of DL-desthiobiotin; D =1oo X 104 ,ug/ml of D-biotin and 200 X 104 /Ag/ml of DL-desthiobiotin.

b All values are 104 ug/ml (D-biotin equiva-lents for Saccharomyces cerevisiae).

growth media of several bacteria, fungi, andstreptomyces increased the level of desthiobiotinexcreted by these organisms (14, 15). Directevidence for the conversion of '4C-desthiobiotinto biotin has been obtained by useof growingcultures of Aspergillus niger (24). Reports fromthis laboratory have shown that D-biotin re-presses formation of the enzyme system respon-sible for conversion of desthiobiotin to biotin (19).Derepression of this enzyme system followedkinetics typical of other biosynthetic enzymes(20). We observed that desthiobiotin could notprevent its own synthesis, but, rather, that thisfunction was highly specific for biotin.

This finding provides additional evidence tosupport the role of desthiobiotin as an inter-mediate in biotin biosynthesis.

In the synthesis of purines, pyrimidines, oramino acids, the intermediates of these bio-

VOL. 94, 1967 1851

on July 7, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

BIRNBAUM, PAI, AND LICHSTEIN

synthetic pathways are rarely excreted into thesurrounding medium. However, all bacteriaemployed in this study, as well as the largenumber of different organisms used by Ogataet al. (14), excreted relatively high concentrationsof desthiobiotin. This excretion of a biosyntheticintermediate can be explained on the basis ofprevious work from this laboratory. Pai andLichstein (18) have shown that the regulation ofbiotin synthesis in E. coli (Crookes) occurs viarepression of enzyme synthesis rather than feed-back inhibition of enzyme activity. Furthermore,they reported that the concentration of exogenousbiotin required to inhibit synthesis of total biotinis higher than that required to repress synthesis oftrue biotin. This suggested that at least two sitesexist in the biotin pathway of E. coli which arecontrolled by repression. One of the sites islocated before desthiobiotin and the other be-tween desthiobiotin and biotin (17, 19). Sincetrue biotin synthesis was repressed to a greaterextent by low levels of exogenous biotin than wastotal biotin synthesis, it is believed that the sitelocated between desthiobiotin and biotin is moresensitive to the repressive effect of biotin. In thepresent study it was demonstrated that little or notrue biotin was excreted by most microorganisms(Table 1). Therefore, the enzymes convertingdesthiobiotin to biotin, which according to ourhypothesis are more sensitive to repression,would be repressed more rapidly than thosewhich come before desthiobiotin. The result ofthis situation is the excretion of high levels ofdesthiobiotin when the cells are grown withoutexogenous biotin (Table 4, column 1). It has beenwell established by Moyed (13) that repression ofenzymes alone cannot provide adequate controlin a biosynthetic pathway. Thus, the lack of feed-back inhibition and the presence of two sites ofpossibly different sensitivities to repression bybiotin could successfully explain the accumula-tion of intermediates of the biotin pathway.Of all the organisms studied, E. coli (Crookes)

is the only one in which the control mechanismsof biotin biosynthesis have been elucidated (18).However, there exist several similarities amongall of the bacteria, including E. coli (Crookes):(i) all exhibit the same pattern of excretion ofdesthiobiotin; (ii) the level of true biotin isalways small relative to total biotin; (iii) synthesisof biotin vitamers in all the organisms is pre-vented by exogenous D-biotin; and (iv) controlby D-biotin is very specific in all the bacteria sinceDL-desthiobiotin and D-biotin-D-sulfoxide do notinhibit vitamer synthesis. These similarities suggestthat the control mechanism in biotin biosynthesismay be the same for most if not all organisms.

Recently, Eisenberg and Maseda (5) demon-strated the inhibitory effect of D-biotin on thesynthesis of the two vitamers in P. chrysogenum.An avidin-uncombinable vitamer, similar in

RF and biological activity to that reported here,has been observed by several investigators (2, 37, 14). Eisenberg (4) has shown that this vitamerlacks both the urea structure and the sulfur atom.Iwahara and co-workers (9) recently identifiedthis material as 7-keto, 8 amino, pelargonic acid.They showed that this material became labeledwith radioactivity when growing cells of aBacillus species were incubated with '4C-pimelicacid. When this '4C-labeled vitamer was incu-bated with resting cells of Bacillus sphaericus,it was converted to 14C-labeled desthiobiotin. Inthis study, it was seen that the synthesis of thevitamer, as well as desthiobiotin, was meta-bolically controlled by D-biotin. Other evidence(Table 5) illustrates that the vitamer is not adegradation product of desthiobiotin even thoughthe two materials appear coincidentally. Thus,evidence obtained by this and other laboratoriessuggests strongly that the vitamer as well asdesthiobiotin are biosynthetic precursors of bio-tin. The unavailability of the vitamer charac-terized by Iwahara et al. (9) prevented a directcomparison with the vitamer detected in ourstudies.

ACKNOWLEDGME&rS

These studies were supported by the Office ofNaval Research, Department of the Navy (NR 103-555), and the National Science Foundation (GB-2595).The excellent technical assistance of T. Troffkin is

gratefully acknowledged.LITERATURE CITE

1. BIRNBAUM, J., AND H. C. LIcHSTEIN. 1965. Con-version of D-biotin to biotin vitamers by Lacto-bacillus arabinosus. J. Bacteriol. 89:1035-1040.

2. DHYSE, F. G., AND R. HERTZ. 1958. The effect ofoctithiazic acid on egg white induced biotindeficiency and upon the microbial formation ofbiotin vitamers in the rat. Arch. Biochem. Bio-phys. 74:7-16.

3. EISENBERG, M. A. 1963. Biotin biosynthesis. I.Biotin yields and biotin vitamers in cultures ofPhycomyces blakesleeanus. J. Bacteriol. 86:673-680.

4. E1SENBERG, M. A. 1965. An early intermediate inthe biosynthesis of biotin: incorporation studieswith (6S) sulphate. Biochem. J. 98:15-17.

5. EISENBERG, M. A., AND R. MASEDA. 1966. Anearly intermediate in the biosynthesis of biotin:incorporation studies with (1 ,7-14C2) pimelicacid. Biochem. J. 101:601-606.

6. ELFORD, H. L., AND L. D. WRIGHT. 1963. The in-corporation of pimelic acid as a unit in the bio-

1852 J. BACTERIOL.

on July 7, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

CONTROL OF BIOTIN VITAMER BIOSYNTHESIS

synthesis of biotin. Biochem. Biophys. Res.Commun. 10:373-378.

7. GENGHOF, D. S. 1956. Production of vitamers ofbiotin by Corynebacterium xerose. Arch. Bio-chem. Biophys. 62:63-69.

8. HERTZ, R. 1943. Modification of the yeast-growthassay method for biotin. Proc. Soc. Exptl.Biol. Med. 52:15-17.

9. IWAHARA, S., M. KIKUCHI, T. ToCHIKURA, ANDK. OGATA. 1966. Some properties of avidin-uncombinable unknown biotin-vitamer pro-duced by Bacillus sp. and its role in biosynthesisof desthiobiotin. Agr. Biol. Chem. (Tokyo) 30:304-306.

10. IWAHARA, S., S. TAKASAWA, T. TOCHIKURA, ANDK. OGATA. 1966. Studies on biosynthesis ofbiotin by microorganisms. V. Degradation ofdesthiobiotin by molds. Agr. Biol. Chem.(Tokyo) 30:1069-1075.

11. LEONIAN, L. H., AND V. G. LILLY. 1945. Conver-sion of desthiobiotin into biotin or biotinlikesubstances by some microorganisms. J. Bac-teriol. 49:291-297.

12. MELVILLE, D. B., D. S. GENGHOF, AND J. M. LEE.1954. Biological properties of biotin-d- and 1-sulfoxides. J. Biol. Chem. 208:503-512.

13. MOYED, H. S. 1961. Interference with the feedbackcontrol of histidine biosynthesis. J. Biol. Chem.236:2261-2267.

14. OGATA, K., T. ToCHIKURA, S. IWAHARA, K.IKUSHIMA, S. TAKASAWA, M. KIKUCHI, AND A.NISHIMURA. 1965. Studies on biosynthesis ofbiotin by microorganisms. I. Accumulation ofbiotin-vitamers by various microorganisms.Agr. Biol. Chem. (Tokyo) 29:889-894.

15. OGATA, K., T. TOCHIKURA, S. IWAHARA, S.TAKASAWA, K. IKUSHIMA, A. NISHIMURA, ANDM. KIKUCHI. 1965. Studies on biosynthesis ofbiotin by microorganisms. II. Identification ofbiotin-vitamers accumulated by various micro-organisms. Agr. Biol. Chem. (Tokyo) 29:895-901.

16. PAI, C. H., AND H. C. LICHSTEIN. 1962. Controlof biotin synthesis in Escherichia coli by exoge-nous biotin. Biochim. Biophys. Acta 65:159-160.

17. PAI, C. H., AND H. C. LICHSTEIN. 1965. The bio-synthesis of biotin in microorganisms. I. The.physiology of biotin synthesis in Escherichiacoli. Biochim. Biophys. Acta 100:28-35.

18. PAI, C. H., AND H. C. LICHSTEIN. 1965. The bio-synthesis of biotin in microorganisms. II.Mechanism of the regulation of biotin synthesisin Escherichia coli. Biochim. Biophys. Acta 100:36-42.

19. PAI, C. H., AND H. C. LICHSTEIN. 1965. The bio-synthesis of biotin in microorganisms. III. Thebiosynthesis of (+)-biotin from desthiobiotinand its control in Escherichia coli. Biochim.Biophys. Acta 100:43-49.

20. PAI, C. H., AND H. C. LICHSTEIN. 1966. Biosyn-thesis of biotin in microorganisms. IV. Repres-sion and derepression of (+)-biotin synthesisfrom (i)-desthiobiotin. Arch. Biochem. Bio-phys. 114:138-144.

21. SNELL, E. E. 1950. Vitamin methods, vol. 1, p.327. Academic Press, Inc., N.Y.

22. SToKES, J. L., AND M. GUNNESS. 1945. The micro-biological activity of synthetic biotin, its opticalisomers and related compounds. J. Biol. Chem.157:121-126.

23. TATUM, E. L. 1945. Desthiobiotin in the biosyn-thesis of biotin. J. Biol. Chem. 160:455-459.

24. TEPPER, J. P., MCCORMICK, D. B., AND L. D.WRIGHT. 1966. Direct evidence for the conver-sion of desthiobiotin to biotin in Aspergillusniger. J. Biol. Chem. 241:5734-5735.

25. WRIGHT, L. D., E. L. CRESSON, AND C. A.DRISCOLL. 1954. Bioautography of biotin andcertain related compounds. Proc. Soc. Exptl.Biol. Med. 86:480-483.

26. WRIGHT, L. D., AND H. R. SKEGGS. 1944. Deter-mination of biotin with Lactobacillus arab-inosus. Proc. Soc. Exptl. Biol. Med. 56:95-98.

1853VOL. 94, 1967

on July 7, 2020 by guesthttp://jb.asm

.org/D

ownloaded from