Vol. 147, No. 1JOURNAL OF BACTERIOLOGY, July 1981, p. 181-1850021-9193/81/070181-05$02.00/0

Metabolism of L-Fucose and L-Rhamnose in Escherichia coli:Differences in Induction of Propanediol Oxidoreductase

ALBERT BORONATt AND JUAN AGUILAR*Department of Biochemistry, School ofPharmacy, University of Barcelona, Barcelona 28, Spain

Received 12 December 1980/Accepted 20 April 1981

Escherichia coli is capable of growing on L-fucose or L-rhamnose as a solesource of carbon and energy. When grown under anaerobic conditions on eithersugar, a nicotinamide adenine dinucleotide-linked L-lactaldehyde:propanedioloxidoreductase activity is induced. The functioning of this enzyme results in theregeneration of oxidized nicotinamide adenine dinucleotide. Conditions of induc-tion of the enzyme activity were studied and were found to display differentcharacteristics on each sugar. In the rhamnose-grown cells, the increase in enzymeactivity detected under inducing conditions was accompanied by the synthesis ofpropanediol oxidoreductase, as measured by the appearance in the extracts of aprotein that reacts with propanediol oxidoreductase antibodies. In contrast, infucose-grown cells, the level of propanediol oxidoreductase as measured byenzyme antibody-reacting material was high under noninducing and inducingconditions. Thus, the increase in enzyme activity detected in going from nonin-ducing to inducing conditions in fucose-grown cells did not depend on theappearance of the specific protein but on the activation of the propanedioloxidoreductase already present in the cells in an inactive form. The propanedioloxidoreductase of both homologous systems should consequently be regulated bydifferent control mechanisms.

L-Fucose and L-rhamnose are metabolized inEscherichia coli through parallel pathways (Fig.1) mediated by the sequential action of a per-mease (10), an isomerase (9, 23), a kinase (13,24), and an aldolase(4, 8). The two homologoussets of inducible proteins are each specific forthe metabolism of its corresponding sugar andare coded by two different gene clusters locatedon the E. coli chromosome at 60 min for fucoseand 87 min for rhamnose (2). The genes of therhamnose system constitute a well-defined op-eron (18), whereas the fucose system seems tomaintain the gene for aldolase under separatecontrol (11).Both pathways converge after the correspond-

ing aldolase action takes place, cleaving the six-carbon derivative of either methyl pentose intothe same products: dihydroxyacetone phosphateand L-lactaldehyde (Fig. 1). Aerobically, L-lac-taldehyde is oxidized in two steps to pyruvateby means of NAD-dependent lactaldehyde de-hydrogenase (22) and a flavin-linked lactate de-hydrogenase (5), thus channeling all of the car-bons from fucose or rhamnose into central met-abolic pathways. Anaerobically, lactaldehyde isreduced to L-1,2-propanediol, which is excreted

t Present address: Department of Biochemistry, Universityof Cambridge, Cambridge CB2 1QW, United Kingdom.

into the medium by an NAD-linked oxidoreduc-tase, regenerating the oxidized coenzyme andallowing the fermentation of fucose or rhamnoseto proceed (3).

Propanediol oxidoreductase has been de-scribed previously as an enzyme inducible byanaerobic growth on fucose (5) or rhamnose (3)which is never found under aerobic conditionseven in the presence of the inducer. Neverthe-less, in this report we present evidence of thepresence of inactive propanediol oxidoreductase,as measured by its reaction with specific anti-bodies in cells grown aerobically on fucose butnot in cells grown aerobically on rhamnose. Asa result, the fucose system, in contrast to therhamnose system, calls for a regulatory mecha-nism other than the control of gene expressionto modulate the oxidoreductase activity.

MATERIALS AND METHODSBacteria. The strain used in this study was an E.

coli K-12 strain, also known as E-15 (1) and hereinreferred to as strain 1. Strain 430 is a propanedioloxidoreductase-constitutive derivative of strain 1 pre-viously described as producing high levels of the en-zyme (12). Both strains were kindly provided by E. C.C. Lin, Department of Microbiology and MolecularGenetics, Harvard Medical School, Boston, Mass.Chemicals. L-Fucose, L-rhamnose, and NAD were

181

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

10

Febr

uary

202

2 by

189

.89.

216.

106.

182 BORONAT AND AGUILAR

L-1.24WR0PANEOIOLFIG. 1. Scheme for the anaerobic catabolism of L-fucose and L-rhamnose in E. coli. The enzyme catalyzing

the interconversion ofpropanediol and lactakdehyde is referred to as propanediol oxidoreductase because theactual role of this protein depends upon the strain in which it is found.

obtained from Sigma Chemical Co., St. Louis, Mo. L-Lactaldehyde was prepared and purified in our labo-ratory as described previously (3). The other chemicalswere of the purest grade available from commercialsources.Growth of cells. Carbon sources were added to a

basal inorganic medium (3) at 0.01 M concentrationfor aerobic growth and 0.02 M for anaerobic growth.Aerobic growth was carried out at 37°C in 2-literErlenmeyer flasks partially filled (200 ml each) andvigorously swirled in a rotary shaker. Anaerobicgrowth was carried out at the same temperature inflasks completely filled and gently stirred by a magnet.Growth was monitored at 420 nm with a Spectronic 88spectrophotometer (Bausch & Lomb, Inc., Rochester,N.Y.).

Preparation of cell extracts. Cells were har-vested at the end of logarithmic phase by centrifuga-tion, washed in 10 mM Tris-hydrochloride buffer (pH7.3), and suspended in four times their wet weight ofthe sme buffer. The suspension was sonically dis-rupted in an MSE sonicator set at an amplitude of 18to 24 pm for periods of 30 s/ml of cell suspension in atube chilled at 0°C. The supernatant fraction, aftercentrifugation at 100,000 x g for 60 min at 4°C (toremove NADH oxidase), was used for enzyme assays(12).Enzyme assay. Spectrophotometric assay for pro-

panediol oxidoreductase was performed at 25°C in thedirection of lactaldehyde reduction by following theabsorbance decrease at 340 nm (NADH loss). Theactivity was measured in an assay mixture (1 ml) thatconsisted of 2.5 mM L-lactaldehyde, 100 mM sodiumphosphate buffer at pH 7.0, and 0.125 mM NADH.The three-carbon substrate was omitted from theblank mixture. All reactions were started by additionof the enzyme. One unit ofenzyme activity was defined

as the amount of enzyme that transforms 1 pmol ofsubstrate per min.

Concentration of the protein in cell extracts wasdetermined by the method of Lowry et al. (16), usingbovine serum albumin as standard.Immunological techniques. Antisera against pro-

panediol oxidoreductase were raised in New Zealandwhite rabbits, using as antigen the strain 430 enzymepurified by the method described previously (3). Pro-panediol oxidoreductase (1 mg) in 1 ml of 10mM Tris-hydrochloride-150 mM NaCl (pH 7.3) was emulsifiedwith 1 ml of Freund complete adjuvant, and the mix-ture was injected (intramuscularly) into each rabbit.Booster injections were given 3 weeks (4 mg, subcu-taneously) and 4 weeks (1 mg, intravenously) later.The rabbits were bled 1 week after the last boosterinjection.

Double immunodiffusion and quantitative immu-noelectrophoresis were performed as described byOuchterlony (17) and Laurel (15), respectively.

RESULTSGrowth profile in an aerobic-to-anaero-

bic shift. E. coli cells adapted to anaerobicconditions differently on fucose culture than onrhamnose culture. This different adaptation wasfirst found by following the growth profile in aculture shifted at an early logarithmic phasefrom aerobic to anaerobic conditions. It wasimportant to effect this shift at an early logarith-mic phase so that both a perfect aerobicity ofthe culture before the shift and a significantgrowth yield under anaerobic conditions afterthe shift were guaranteed.

Figure 2 displays the growth curves of cells on

J. BACTERIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

10

Febr

uary

202

2 by

189

.89.

216.

106.

REGULATION OF FUCOSE AND RHAMNOSE METABOLISM 183

each sugar and shows, at the time of the shift,an immediate adaptation to the anaerobic con-ditions for the cells growing on fucose (doublingtime, 80 min), reflected by uninterrupted growthat a lower rate (doubling time, 220 min) (Fig.2A). The profile for the cells growing on rham-nose (doubling time, 100 min) presented a com-pletely different behavior. The cells stoppedgrowing as soon as the oxygen in the culture wasexhausted (Fig. 2B) and resumed growth onlyafter a term of approximately 10 h had elapsed.Cells then grew anaerobically on rhamnose to a

yield equal to that obtained with fucose althoughat a slightly lower rate (data not shown).Propanediol oxidoreductase activity in

fucose- and rhamnose-grown cells. The dif-ferences in the adaptation to anaerobiosis infucose or rhamnose indicated by the growthprofile prompted us to analyze the activities ofpropanediol oxidoreductase, the key enzyme forthe fermentation of both sugars. The assay ofthis enzyme in cells grown under anaerobic con-

dition requires special caution because anaero-biosis induces NAD-linked glycerol dehydroge-nase which also attacks propanediol but at C-2(10). To avoid glycerol dehydrogenase activity,propanediol oxidoreductase activity measure-

ments were performed in the direction of thereduction of lactaldehyde which is not a sub-strate of glycerol dehydrogenase (10).Table 1 shows the propanediol oxidoreductase

activities found in crude extracts of cells grown

A FUCOSE 8 RHAMNOSE

0.8

0.6-CE7

CD

-p4CsCs

HOURS

FIG. 2. Growth profile of a culture on fucose orrhamnose shifted from aerobic to anaerobic condi-tions. At the time of the shift, a 10-ml portion of a200-ml culture was transferred to a sterile spectro-photometer tube containing a magnet; the tube wasfilled to the top and tightly capped. The tube wasgently stirred, and the cell mass was monitored (0).The rest ofthe culture was maintained under aerobicconditions as the control, cell mass being monitored(0) at the same time intervals. 0. D., Optical density.

TABLE 1. Activities ofpropanediol oxidoreductasein crude extracts of strain 1 grown on different

mediaEnzyme ac-

Carbon source Growth condition tivity (U/mgof protein)'

Fucose Aerobic 0.27Rhamnose Aerobic 0.02Fucose Anaerobic 1.42Rhamnose Anaerobic 1.15

a One unit of enzyme activity was defined as theamount of enzyme that transforms 1 ,Lmol of substrateper min at 25°C.

under aerobic or anaerobic conditions on bothsugars. The extracts of cells grown aerobicallyon fucose had a low level of activity of 0.27 U/mg of protein. This activity increased fivefoldunder anaerobic conditions, reaching 1.4 U/mgof protein. On the other hand, the basal activityin extracts of cells grown aerobically on rham-nose was found to be lower than that found onfucose, but the growth ofthese cells on rhamnoseunder anaerobic conditions promoted a 50-foldincrease of enzymatic activity to levels close tothose found in cells grown anaerobically on fu-cose.Immunological quantification of pro-

panediol oxidoreductase in fucose- andrhamnose-grown cefls. A more direct indica-tion of the induction ofthe enzyme was achievedby immunological techniques which determinedthe amount of the propanediol oxidoreductaseprotein in the extracts used for enzyme activityassays. This quantification was performed bymeans of an immunoelectrophoresis of crudeextracts against propanediol oxidoreductase an-tiserum obtained with the strain 430 purifiedenzyme. Immunological identity between pro-panediol oxidoreductase, as induced by eitherfucose (aerobically or anaerobically) or rham-nose (anaerobically) in strain 1, and the consti-tutive enzyme of strain 430 was confirmed bydouble immunodiffusion (Fig. 3A).Immunoelectrophoresis of extracts of cells

grown aerobically on fucose, hence with a lowlevel of enzyme activity, surprisingly showed ahigh level of propanediol oxidoreductase. Thisamount ofoxidoreductase protein as representedby the length of the rocket was nearly two-thirdsof that found in extracts of cells whose enzymeactivity had been fully induced under anaerobicconditions (Fig. 3B).The extracts of cells grown aerobically on

rhamnose presented an undetectable level ofpropanediol oxidoreductase, in contrast to theabove-described results for fucose. Nevertheless,under anaerobic growth on rhamnose (inducing

VOL. 147, 1981

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

10

Febr

uary

202

2 by

189

.89.

216.

106.

184 BORONAT AND AGUILAR

C

I

2

e

jn1[>

A:FIG. 3. (A) Immunological identity betwe

or rhamnose-induced propanediol oxidorestrain I and constitutivepropanediol oxidoof strain 430. Wells in the Ouchterlony infusion plates contained 25 Al of cell extrastrain 1 grown anaerobically on rhamnoseprotein); (b) and (e) strain 430 grown aerocasein hydrolysate (3 pg ofprotein); (c) straiaerobically on fucose (40 pg ofprotein); an

(t) strain I grown anaerobically on fucoseprotei. The center well contained 25 Al ofcHomology between the enzyme of strainanaerobically on rhamnose and strain 1Iobically on fucose was alo checked (data ni(B) Immunoekectrophoretic quantificatioipanediol oxidoreductase. The Laurell rocobtained by applying to a gel containinantiserum, 25 pg ofprotein in 2 Al of cellstrain I grown: (1) aerobically on fucose,cally on rhamnose, (3) anaerobically on ft

(4) anaerobically on rhamnose.

conditions), the oxidoreductase proteira level similar to that obtained by agrowth on fucose (Fig. 3B).

DISCUSSIONThe data obtained show that the e:

and activity of propanediol oxidoredi

fucose- and rhamnose-grown cells is regulated ina different way (Fig. 2).On one hand, the regulation of propanediol

oxidoreductase activity in cells grown on rham-nose seems to be clearly mediated by a mecha-nism of induction of the gene coding for theenzyme. Although the nature of the inducer

d itself is yet to be described, it is clear that theinduction requires both the presence ofthe sugarand the absence of oxygen.On the other hand, the presence of high

amounts of propanediol oxidoreductase in cellsgrown aerobically on fucose, where low levels ofoxidoreductase activity are found, inevitablycalls for a regulatory mechanism of the enzymeactivity other than the control of gene expres-sion. This finding leads us to surmise that thepropanediol oxidoreductase present under aero-bic conditions is inactivated by some biochemi-cal mechanism and that this inactivation will berelieved under the new metabolic situation ofanaerobic conditions.

Several mechanisms may be invoked for sucha regulation, but in any case, a prime candidatefor the modulator would be the state of oxidationof the coenzyme. A number of arguments maybe offered to this end: (i) the process of fermen-tation, in which the enzyme is involved, is initself related to the state of the coenzyme, (ii)the aerobic-to-anaerobic shift will produce inthese cells important changes in the intracellular

en fucose- levels of oxidized and reduced coenzyme (25),ductalle Of and (iii) enzymatic activity may be modulatedreductase by the interaction of the NAD coenzyme withCCts ofd(a oxidoreductases by allosteric interactions (21,et(30og Of 26) or by formation of abortive ternary com-

bically on plexes (6, 7, 14). Thus, propanediol oxidoreduc-in 1 grown tase inhibition in cells grown aerobically on fu-td (d) and cose could be mediated either by the allosteric? (30 pg of inhibition promoted by NAD+ or by the forma-znti8erum. tion of an abortive ternary complex of the NAD-I grown enzyme-lactaldehyde type. As pointed out byrowfn aer- Punich and Fromm (19), the formation of abor-t shown). tive temary complexes would be of special inter-

nketo were est in situations in which two enzymes competeg 1.5% of for a common metabolite. In our case, lactalde-extract of hyde would be in such a situation between pro-(2) aerobi- panediol oxidoreductase and lactaldehyde de-cose, and hydrogenase.

Hitherto, only one gene coding for propanedioloxidoreductase (linked to the fucose system

n reached genes) has been described in the E. coli chro-maerobic mosome (5). Nevertheless, two different meta-

bolic systems (fucose and rhamnose) seem touse the product of a propanediol oxidoreductasestructural gene (3). Two possibilities arise undersuch circumstances: both systems may be using

xpression the same gene linked to the fuc locus althoughuctase in regulating its expression in a different way or

A

a

B

J. BACTERIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

10

Febr

uary

202

2 by

189

.89.

216.

106.

REGULATION OF FUCOSE AND RHAMNOSE METABOLISM 185

each system may be using a different gene, onelinked to fucose and the other presumably linkedto rhamnose. Under the second alternative, itwould clearly be a question of two equal orhighly similar copies of the propanediol oxido-reductase gene since their products, induced byfucose or rhamnose, are indistinguishable (3).The difference, as reported in this article, be-

tween regulation of the synthesis of the pro-panediol oxidoreductase in fucose- or rhamnose-grown cells seems to point toward the two genealternative. This could be reinforced by the factthat fucose and rhamnose genetic systems mayhave a common evolutionary origin. In thissense, it may be pointed out that evolution byduplication and differentiation of one geneticsystem from the other as established by Rileyand Anilionis (20) may be also applicable in thiscase. On the one hand, both systems lie approx-imately 900 apart on the circular map of the E.coli chromosome, thus indicating a possible re-peated duplication of the whole chromosome(27). On the other hand, the substrates, exceptlactaldehyde, of each homologous enzyme inboth pathways are different (Fig. 1), thus exer-cising a pressure for the differentiation of ho-mologous genes derived by duplication. Theidentity in the substrate of propanediol oxido-reductase in both systems would explain theabsence of gene differentiation, although it mustbe remembered that the presence of two identi-cal copies of the same gene may lead to the lossor silencing ofone ofthem by evolutive pressure.

Analysis ofpropanediol oxidoreductase induc-tion in mutants with the fucose genetic systemcompletely deleted is currently under way andwill probably show evidence of the existence ofa second propanediol oxidoreductase structuralgene.

LUTERATURE CITED1. Bachmann, B. J. 1972. Pedigrees of some mutant strains

of Escherichia coli K-12. Bacteriol. Rev. 36:525-557.2. Bachmann, B. J., and K. B. Low. 1980. Linkage map of

Escherichia coli K-12, edition 6. Microbiol. Rev. 44:1-56.

3. Boronat, A., and J. Aguilar. 1979. Rhamnose-inducedpropanediol oxidoreductase in Escherichia coli: purifi-cation, properties, and comparison with the fucose-in-duced enzyme. J. Bacteriol. 140:320-326.

4. Chiu, T. H., and D. S. Feingold. 1969. L-rhamnulose-1-phosphate aldolase from Escherichia coli. Crystalliza-tion and properties. Biochemistry 8:98-108.

5. Cocks, G. T., J. Aguilar, and E. C. C. Lin. 1974.Evoluation of L-1,2-propanediol catabolism in Esche-richia coli by recruitment of enzymes for L-fucose andL-lactate metabolism. J. Bacteriol. 118:83-88.

6. Everse, J., R. E. Barnett, C. H. J. R. Thorne, and N.0. Kaplan. 1971. The formation of ternary complexesby diphosphopyrydine nucleotide-dependent dehydrog-

enase. Arch. Biochem. Biophys. 143:444 460.7. Fromm, H. J., and D. R. Nelson. 1962. Ribitol dehy-

drogenase. III. Kinetic studies with product inhibition.J. Biol. Chem. 237:215-220.

8. Ghalambor, M. A., and E. C. Heath 1962. The metab-olism of L-fucose. H. The enzymatic cleavage of L-fuculose-1-phosphate. J. Biol. Chem. 237:2427-2433.

9. Green, M., and S. S. Cohen. 1956. The enzymatic con-version of L-fucose to L-fuculose. J. Biol. Chem. 219:557-568.

10. Hacking, A. J., and E. C. C. Lin. 1976. Disruption ofthe fucose pathway as a consequence of genetic adap-tation to propanediol as a carbon source in Escherichiacoli. J. Bacteriol. 126:1166-1172.

11. Hacking, A. J., and E. C. C. Lin. 1977. Regulatorychanges in the fucose system associated with the evo-lution of a catabolic pathway for propanediol in Esch-erichia coli. J. Bacteriol. 130:832-838.

12. Hacking, A. J., J. Aguilar, and E. C. C. Lin. 1978.Evolution of propanediol utilization in Escherichia coli:mutant with improved substrate-scavenging power. J.Bacteriol. 136:522-530.

13. Heath, E. C., and M. A. Ghalambor. 1962. The metab-olism of L-fucose. I. The purification and properties ofL-fuculose kinase. J. Biol. Chem. 237:2423-2426.

14. Kaplan, N. O., and M. M. Ciotti. 1954. Direct evidencefor a diphosphopyridine nucleotide-hydroxylaminecomplex with horse liver alcohol dehydrogenase. J. Biol.Chem. 211:431-445.

15. Laurell, C. B. 1966. Quantitative estimation of proteinsby electrophoresis in agarose gel containing antibodies.Anal. Biochem. 15:45-52.

16. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.Randall. 1951. Protein measurement with the Folinphenol reagent. J. Biol. Chem. 193:265-275.

17. Ouchterlony, 0. 1953. Antigen-antibodies reactions ingels. IV. Types of reactions in coordinated systems ofdiffusion. Acta Pathol. Microbiol. Scand. 22:231-240.

18. Power, J. 1967. The L-rhamnose genetic system in Esch-erichia coli K-12. Genetics 35:557-568.

19. Purich, D. L., and H. J. Fromm. 1972. A possible rolefor kinetic reaction mechanism dependent substrateand product effects in enzyme regulation. Curr. Top.Cell. Regul. 6:131-167.

20. Riley, M., and A. Anilionis. 1978. Evolution of thebacterial genome. Annu. Rev. Microbiol. 32:519-560.

21. Sanwal, B. D. 1969. Regulatory mechanism involvingnicotinamide adenine nucleotides as allosteric effectors.I. Control characteristics of malate dehydrogenase. J.Biol. Chem. 244:1831-1837.

22. Sridhara, S., and T. T. Wu. 1969. Purification andproperties of lactaldehyde dehydrogenase from Esche-richia coli. J. Biol. Chem. 244:5233-5238.

23. Takagi, Y., and H. Sawada. 1964. The metabolism of L-rhamnose in Escherichia coli. I. L-rhamnose isomerase.Biochim. Biophys. Acta. 92:10-17.

24. Takagi, Y., and H. Sawada. 1964. The metabolism of L-rhamnose in Escherichia coli. II. L-rhamnulose kinase.Biochim. Biophys. Acta. 92:18-25.

25. Wimpenny, J. W. T., and A. Firth. 1972. Levels ofnicotinamide adenine dinucleotide and reduced nicotin-amide adenine dinucleotide in facultative bacteria andthe effect of oxygen. J. Bacteriol. 111:24-32.

26. Wright, J. A., and B. D. Sanwal. 1969. Regulatorymechanisms involving nicotinamide adenine nucleo-tides as allosteric effectors. II. Control of phosphoenol-pyruvate carboxykinase. J. Biol. Chem. 244:1838-1845.

27. Zipkas, D., and M. Riley. 1975. Proposal concemingmechanism of evolution of the genome of Escherichiacoli. Proc. Natl. Acad. Sci. U.S.A. 72:1354-1358.

VOL. 147, 1981

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

10

Febr

uary

202

2 by

189

.89.

216.

106.