Growth of Escherichia coli on Short-Chain Fatty Acids - Journal of

JOURNAL OF BACTERIOLOGY, Nov. 1971, p. 885-892Copyright 0 1971 American Society for Microbiology

Vol. 108, No. 2Printed in U.S.A.

Growth of Escherichia coli on Short-Chain FattyAcids: Growth Characteristics of Mutants

JOSEPH P. SALANITRO1 AND WARNER S. WEGENER

Department of Microbiology, School of Medicine, Indiana University, Indianapolis, Indiana 46202

Received for publication 25 May 1971

The parent Escherichia coli K-12 is constitutive for the enzymes of the glyoxy-late bypass and adapts to growth on long-chain fatty acids (C12 to C18). It doesnot utilize medium-chain (C6 to C11) or short-chain (C4, Ca) n-monocarboxylicacids. Several mutants of this strain which grow using short- or medium-chainacids, or both, as the sole carbon source were selected and characterized. One mu-

tant (D,) synthesizes the (-oxidation enzymes constitutively and grows on me-

dium-chain but not on short-chain acids. A second (N3) is partially derepressed forsynthesis of these enzymes and grows both on medium-chain and on short-chainacids. Secondary mutants (N3V-, N3B-, N3OL-) were derived from N3. N3VWgrows on even-chain but not on odd-chain acids and exhibits a lesion in propionateoxidation. N3B- grows on odd-chain but not on even-chain acids and exhibits no

crotonase activity as assayed by hydration of crotonyl-CoA. N30L- grows on ace-

tate and propionate but does not utilize fatty acids C4 to C18; it exhibits multipledeficiencies in the (-oxidation pathway. Growth on acetate of N3, but not of theparent strain, is inhibited by 4-pentenoate. Revertants of N3 which are resistant togrowth inhibition by 4-pentenoate (N3PR) exhibit loss of ability to grow on short-chain acids but retain the ability to grow on medium-chain and long-chain acids.The growth characteristics of these mutants suggest that in order to grow at theexpense of butyrate and valerate, E. coli must be (i) derepressed for synthesis ofthe (3-oxidation enzymes and (ii) derepressed for synthesis of a short-chain fattyacid uptake system.

Escherichia coli grows on acetate (7), propio-nate (26), and long-chain fatty acids (10, 11, 23)by adaptive phenomena. Fatty acids C4 to C11do not support growth of wild-type strains. It hasbeen demonstrated in E. coli K-12 (10) thatgrowth in the presence of long-chain acids resultsin coordinate derepression of synthesis of theenzymes of the (3-oxidation sequence and of afatty acid acyl-CoA synthetase. This synthetaseactivated fatty acids C8 to C18 but did not showactivity with hexanoate or butyrate. Two fattyacid acyl-CoA synthetases were reported (15, 16)in another K-12 strain cultured on oleate. Oneexhibited maximum activity with long-chainacids, whereas the other had maximum activitywith hexanoate as substrate.

Mutants of E. coli K-12 which are perma-nently derepressed for synthesis of the enzymesof (3-oxidation have been described previously(10, 23). Such mutants grow on decanoate butnot on the C4 or C6 fatty acids. The presentstudy provides further evidence that short-chain

' Present address: Biological Sciences, Research Center,Shell Development Co., Modesto, Calif. 95352.

and medium-chain fatty acids, even under condi-tions where they are activated and metabolized,are unable to derepress the A-oxidation operon.For growth to occur on fatty acids with less thantwelve carbons, a mutation is required which al-lows these enzymes to be synthesized constitu-tively. In the K-12 strain studied here, a singlemutation appears to coordinately derepress syn-thesis of the A-oxidation enzymes and of a long-chain and a medium-chain fatty acid-activatingenzyme. Such mutants grow by using fatty acidsCff to C11 as the sole source of carbon but do notgrow on butyrate or valerate. A second mutationwhich confers ability to uptake short-chain acidsis necessary for utilization of the latter sub-strates.

MATERIALS AND METHODS

Materials. All fatty acids were used without furtherpurification. Commercial sources were: butyric, hexan-oic, octanoic, nonanoic, decanoic, dodecanoic, tetradec-anoic, and pentadecanoic (Sigma Chemical Co., St.Louis, Mo.); valeric (Eastman Organic Chemicals,Rochester, N.Y.); heptanoic (Matheson, Coleman andBell); oleic, palmitic, and stearic (Hormel Institute); 4-

885

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

29

Dec

embe

r 20

21 b

y 17

9.10

9.17

2.46

.

SALANITRO AND WEGENER

pentenoic (City Chemical Corp., N.Y.); 2-butenoic, 3-butenoic, 2-decenoic, 2-dodecenoic, and 2-hexadecenoic(Pfaltz and Bauer, Inc.). Radioactive fatty acids were

labeled with 14C in the C-1 position and obtained fromAmersham/Searle Corp.

Substrate terminology. In this communication, n-

monocarboxylic acids having a chain length of C4 (bu-tyrate) or C, (valerate) are designated as short-chainacids. Fatty acids of chain length C6 to C,1 are desig-nated as medium-chain acids and C12 to C18 as long-chain acids. Unsaturated fatty acids are designated-enoic acids (e.g., 4-pentenoic refers to the A 4,5 unsat-urated C, acid).

Strains. The parent E. coli is a K-12 derivative(D,H,G7) constitutive for the glyoxylate bypass en-

zymes and auxotrophic for histidine and thiamine (27).For mutant selection, cells were subjected to ultravioletlight treatment, incubated in mineral salts media con-

taining an appropriate fatty acid; and purified on cor-

responding solid media. Mutants Ns and D1 were se-

lected by using nonanoate and decanoate, respectively,as the sole source of carbon. All strains were main-tained on Trypticase soy agar and were stable afterrepeated subculture.

Cultural conditions. The mineral salts medium hasbeen described (24). Fatty acids were neutralized withNaOH, sterilized by filtration, and added as the solesource of carbon at the following concentrations: C2 toC9, 20 mM; C1, to C1, and oleate, 5 mm; and C16 andC18, 2.5 mM.The initial pH of the medium was 7.0.Fatty acids having six or more carbons were solubilizedby using Brij-35 (final concentration in the medium of0.4%). Brij-35 does not alter growth rate on succinateor on fatty acids C2 to C5.To determine growth rate, cells were cultured in 250-

ml nephelometer flasks containing 25 ml of mediumaerated by shaking at 37 C. Media were inoculated toan initial turbidity of approximately 5 Klett units.Washed cells, grown in succinate minimal media to themid-log phase, were used as the inoculum. Turbiditywas monitored at 540 nm with a Klett-Summerson col-orimeter, and the specific growth rate, a, was calcu-lated as defined by:

a = (2.303 log N,/Nl)/(t2 - tl)

where N2 and Nl are the optical densities of the culturein the time interval t2 - t, (6).

Enzyme assays and oxidation studies. For enzymepreparations, cells were suspended in 0.1 M phosphatebuffer (pH 7.0) containing 2 x 10-' M dithiothreitoland disrupted (by a Branson sonifier equipped with9.4-mm disruptor horn) for 90 sec at a power setting of65 w. The preparations were then centrifuged at 16,000x g for 15 min at 5 C.,B-Hydroxyacyl-CoA dehydrogenase (L-3-hydroxy-

acyl-CoA:NAD oxidoreductase, EC 1.1.135) and cro-tonase (L-3-hydroxyacyl-CoA hydro-lyase, EC4.2.1.17) were assayed spectrophotometrically at 35 Cby using acetoacetyl-CoA and crotonyl-CoA, respec-tively, according to the methods of Decker (4) andOverath et al. (11). Initial reaction rates were calcu-lated, and activity was expressed as micromoles of sub-strate utilized per minute per milligram of extract pro-tein. Protein was estimated by the method of Lowry etal. (9).

Oxidation of fatty acids by whole cells was assayedradiorespirometrically by measuring the rate of 14CO2evolution from 1-'4C-labeled fatty acids as describedpreviously (25). The reaction mixture contained: wetcells, 16 mg; labeled substrate (2.5 MCi), 1.0 Amole;and mineral salts, in a volume of 4.5 ml. 14CO2 evolvedwas trapped with hydroxylamine and measured at 5-min intervals. The rate of fatty acid oxidation was ex-

pressed as cumulative percentage of "4C activity, re-

covered as 14CO2 versus time.

RESULTS

Growth behavior on fatty acids. The growthbehavior of the E. coli parent strain and mutantsin mineral salts media which contain n-mono-

carboxylic acids as the sole source of carbon isillustrated in Table 1. The parent strain adaptedto growth on long-chain fatty acids but did notutilize medium-chain or short-chain acids. This isalso the case when cells are derepressed for syn-

thesis of the ,B-oxidation enzymes by precultureon oleate. Both D, and N, grew on medium-chain acids, but only N3 grew on butyrate andvalerate. The pattern of fatty acid utilizationshown by the parent K-12 also is characteristicof strains B and W and various enteropathogenicstrains of E. coli. It was not possible to demon-strate adaptive growth on fatty acids C4 to C,,by incubation at 25, 30, 37, or 42 C in the pres-

ence or absence of limiting amounts of acetate,tryptone, or oleate or by addition of cyclic aden-osine monophosphate to the medium.These studies indicate that, in the parent, the

degradative enzymes required to utilize short-chain and medium-chain fatty acids are not syn-thesized when cells are cultured in the presenceof these substrates. It is possible that the syn-thesis of these degradative enzymes is dere-pressed by other substances present in the nat-ural environment of E. coli. To test this possibil-ity, a number of E. coli isolates from stool speci-mens and direct stool dilutions were inoculatedinto tubes of minimal media containing acetate,oleate, butyrate, valerate, decanoate, or no addedcarbon source. The results of this study indicatethat E. coli cells cultured directly from humanstool grow on acetate and oleate but do not uti-lize short- and medium-chain fatty acids. This isalso the case when E. coli isolates are culturedon fatty acids in the presence of sterile stool fil-trates.

Fatty acid mutants. A summary of the deriva-tion of the fatty acid mutants is illustrated inFig. 1. Ni was selected after extended incubationin nonanoate media. N,B-, N3V-, and N,OL-were derived from N3 and selected, by using thepenicillin technique, as secondary mutants unableto grow on butyrate, valerate, and oleate, respec-tively. N3B- grows on C, but not C,; N3V-

886 J. BACTERIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

29

Dec

embe

r 20

21 b

y 17

9.10

9.17

2.46

.

GROWTH OF E. COLI ON SHORT-CHAIN FATTY ACIDS

TABLE 1. Growth pattern of Escherichia coli parentstrain and mutants on fatty acids

Fatty acid chain length Parent Ns D,

C2, 3 ............... + + +C4. 5 ........... .... -a +

C6, 7, 8, 9. 10, 11 * + +

C 12, 14, 13, 16, 18: 1 + + +

a -, No growth in 120 hr.

Parent (B-V-)

P D1 (B-V-)

N3 (B+V+)

N3PR N3B- N3V N,OL-(B-V- ) (B- V+ ) (B+V- ) (B-V- )FIG. 1. Derivation of short-chain fatty acid mu-

tants. B+, B-, V+, and V- refer to growth or nogrowth, on butyrate or alerate, respectively.

grows on C, but not C5; and N3OL- does notgrow on fatty acids C, to C18. N3PR was se-lected from N, by ability to grow in acetatemedia in the presence of 4-pentenoic acid. Itconcomitantly has lost the ability to utilize bu-tyrate and valerate. D1 was selected on deca-noate and is similar to the ,B-oxidation-constitu-tive mutants described previously (11, 23). D1grows on medium-chain but not on short-chainfatty acids.

Table 2 compares the specific growth rate, a(fraction of cells which increase per unit of timeduring exponential growth), of the parent andmutants grown at the expense of straight-chainfatty acids. Several aspects are notable.

(i) Growth of the parent is not supported byfatty acids with less than 12 carbons. Above C12,an increase in chain length is correlated with anincreased a, suggesting a direct relationship be-tween chain length and capacity to derepress the,-oxidation operon. A proportional relationshipbetween chain length and the level of activity ofseveral ,B-oxidation enzymes was reported in theK-12 strain studied by Overath et al. (10). 2-Dodecenoate and 2-hexadecenoate (not shown inTable 2) do not support growth of the parentstrain but are metabolized by the j3-oxidation-constitutive mutants.

(ii) Both N3 and D1 synthesize the ,B-oxidationenzymes constitutively. However, the mutantsdiffer in degree of derepression of enzyme syn-thesis as measured by enzyme activity formedduring growth on acetate. Unlike the parent, D1exhibits a constant specific growth rate (0.21 to0.23) on fatty acids C10 to C18. In contrast to

D1, N3 exhibits a lower a on fatty acids C1O toC12 (0.14 to 0.16) than on fatty acids C14 to C16(0.23 to 0.24). This is in agreement with thefinding that D1 is fully derepressed whereas N3 isonly partially derepressed for synthesis of the /#-oxidation enzymes. In this mutant as in the par-ent, fatty acids with a chain length less than 12to 14 carbons would appear to be unable tocause derepression of the ,B-oxidation operon.

(iii) Both N, and D1 exhibit specific growthrates on fatty acids C, to C, of approximatelyone-half of that shown on the C10 to C12 acids.These data may reflect distinct activation sys-tems for fatty acids C, to C9 and for fatty acidsC10 to C ,. Separate'acyl-CoA synthetase activi-ties for hexanoate and long-chain acids were re-ported in E. coli K-12 by Samuel et al. (15, 16),but this aspect has not been investigated in thestrain employed here.

(iv) N3, but not D1, grows on butyrate anr4valerate. Data presented in the accompanyingpaper (14) suggest that N3 possesses an addi-tional mutation such that cells take up short-chainacids. N,,PR (4-pentenoate-resistant revertant ofN3) is comparable to N3 with respect to growthon long-chain and medium-chain acids. It doesnot grow on C4 or C8, grows slowly on C6, andappears to have lost the ability to take up short-chain acids.

(v) N3B- is similar to N3 with respect togrowth on odd-chain fatty acids but does not uti-lize even-chain acids Co to C12 and grows slowlyon tetradecanoate and oleate. This suggests thepossibility that the ,-oxidation of odd-chain andeven-chain acids in E. coli may involve at leastone step catalyzed by distinct enzymes.

TABLE 2. Comparison ofspecific growth rates onstraight-chain fatty acids

Fatty acid Specific growth rate (a)a of cell typeschainlength Parent N' N,PR N,B- N3V- D,

2 .28 .29 .31 .31 .31 .314 b .23 - - .25 -

5 - .21 - .19 - -

6 - .08 .03 - .06 .067 - .09 .07 .10 - .128 - .08 .06 - .07 .119 - .08 .08 .08 - .1110 - .15 .16 - .17 .2111 - .14 .14 .16 - .2112 .04 .16 .13 - .16 .2214 .11 .24 .25 .03 .20 .2215 .14 .23 .24 .18 .06 .2118:1 .22 .24 .23 .07 .18 .23

a a, Specific growth ratefined in Materials and Methc

b _, No growth in 120 hr.

* hr-1, calculated as de-

887VOL. 108, 1971

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

29

Dec

embe

r 20

21 b

y 17

9.10

9.17

2.46

.

SALANITRO AND WEGENER

(vi) N3V- is similar to N. with respect togrowth on even-chain acids but does not grow onodd-chain acids C8 to C,1 and grows slowly onpentadecanoate. Respirometric data (see Table7) indicate that N3V- has a lesion in propionateoxidation.

(vii) Mutant N30L- (not shown in Table 2)grows on acetate and propionate but not on fattyacids C4 to C,8. Enzyme assays (see Table 6)indicate that this mutant lacks activity of mul-tiple ,8-oxidation enzymes.

Utilization of unsaturated and substituted fattyacids. Table 3 compares the parent and N3 andD1 mutants with respect to utilization of or inhi-bition by unsat6rated and substituted fatty acids.N3 grows on 2-butenoate and on 3-butenoate asthe sole source of carbon; 4-pentenoate inhibitsgrowth of this mutant on acetate. In contrast, D1does not utilize either saturated or unsaturatedshort-chain acids, and growth on acetate is notinhibited by 4-pentenoate. Both the N3 and D1mutants, but not the parent strain, grow by usinga,4-unsaturated medium-chain and long-chainacids as the sole source of carbon.

Neither the parent nor the mutant strains uti-lize 2-methyl-substituted short-chain acids.Studies with 14C-2-methylpropionate (isobutyr-ate) indicate that this substrate does not pene-trate N3. Hydroxy- and amino-substituted short-chain fatty acids do not support growth of the

TABLE 3. Effect of unsaturated and substituted fattyacids on growth of the parent and mutant strains

GrowthaFatty acid

Parent N3 Di

2-Butenoate. - -3-Butenoate ......... - +4-Pentenoate ........ -(I)2-Decenoate ......... _ + +2-Dodecenoate ....... - + +2-Hexadecenoate ..... - + +

2-Methylpropionate -. _2-Methylbutyrate ....2-Methylvalerate .....

2-Hydroxybutyrate ... - (I) -(I ) -(I )3-Hydroxybutyrate ... _ _4-Hydroxybutyrate ... _ _

2-Aminobutyrate . - (I - (I) - (I)3-Aminobutyrate .....4-Aminobutyrate... -2-Aminovalerate ..... (I) -(I) -(I)2-Aminohexanoate ... (I) -(I) -(I)

a, No growth in 120 hr; (1), fatty acid (10 mM)inhibits growth on acetate.

parent or mutant strains. In both the mutant andparent strains, the 2-hydroxy- and 2-amino-sub-stituted acids inhibit growth on acetate, indi-cating that such substrates penetrate the cell. Thefact that these substituted acids inhibit growth ofthe parent strain suggests that they are not takenup by the short-chain fatty acid transport system.It cannot be determined from the data whetherthe 3- and 4-substituted acids are noninhibitoryor do not penetrate the cell.

Utilization of medium- and short-chain fattyacids. Medium-chain fatty acids support growthof ,@-oxidation-constitutive mutants but inhibitgrowth of the parent strain on acetate (Table 4).It is suggested that, in the parent, medium-chainacids penetrate the cell but are not activated anddo not derepress the ,-oxidation operon. Degreeof inhibition of growth increases with chainlength up to C9 and then decreases. Fatty acidshaving 12 carbons or more support growth of theparent. It would be predicted that ability to pen-etrate the membrane should increase with chainlength. The fact that growth inhibition by hexan-oate is intermediate between that shown byshort- and medium-chain acids may reflect a lowdegree of penetration by this substrate. In thisrespect, it should be noted that the specificgrowth rate of the D1 mutant on hexanoate wasconsiderably less than on fatty acids C7 to C,(see Table 2). It also may be that the C8 and C,acids are especially toxic to an essential biosyn-thetic or catabolic process. Toxic effects of theseacids have been reported in other bacteria (5, 13)and in animal tissues (22).

Table 5 summarizes the effects of short-chainacids on growth of the parent and mutants. Bu-tyrate, valerate, and 4-pentenoate neither support

TABLE 4. Inhibition ofgrowth of the parent bymedium-chain fatty acids

Fatty acid Inhibition ofadditiona growth (%)b

None 0C4 0C, 0Cs 23C7 81C8 100C, 100C1O 41C1l 12

aThe parent strain was cultured in an acetate-min-eral salts medium with or without addition of fattyacids. Fatty acid concentrations were C4 and C, (10mM) and C, to C,1 (5 mM).

',Growth was assayed as turbidity after 12 hr of in-cubation, and inhibition is expressed as per cent de-crease relative to the control.

888 J. BACTERIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

29

Dec

embe

r 20

21 b

y 17

9.10

9.17

2.46

.

VOL. 108, 1971 GROWTH OF E. COLI ON SHORT-CHAIN FATTY ACIDS

TABLE 5. Effect of short-chain fatty acids on growthofparent and mutant strdins

Growth on or inhibition bya

Cell type 4-Pent-Butyrate Valerate enoate

Parent ..............Di ................ _N3 ............... + + -(I)N3PR ..............N3B- ............... - (1) + -(I)N3V- ............... + -(I) -(1)N30L- .............. () -(I) -(1)

a_, No growth in 72 hr; (1), fatty acid (10 mM)inhibits growth on acetate as measured by turbidityafter 12 hr of incubation.

nor inhibit growth of the parent or D1 mutantand appear not to penetrate. 4-Pentenoate in-hibits growth of N., N3B-, N3V-, and N3OL-.N3PR is pentenoate-resistant and appears tohave concomitantly lost the ability to uptakebutyrate and valerate. N3B- does not grow onbutyrate but is inhibited by it, indicating thatbutyrate penetrates the mutant. A similar situa-tion exists with respect to penetration of short-chain acids in N3V- and N30L-.

Levels of activity of ,B-oxidation enzymes inparent and mutant cells. Table 6 compares theparent and mutants with respect to activity oftwo enzymes of the ,B-oxidation sequence, cro-tonase and ,B-hydroxyacyl-CoA dehydrogenase(HOADH). The data are expressed as relativespecific activity, considering the parent straingrown at the expense of oleate as 100. The abso-lute specific activities of these enzymes in theparent cultured on oleate were 2.7 and 1.4,umoles per min per mg of protein for crotonaseand HOADH, respectively, and are comparableto values reported earlier (11, 23).

Several aspects are notable: (i) Enzyme syn-thesis is derepressed during growth of the parenton oleate but not on acetate. This is in agree-ment with previous studies. (ii) Enzyme activitiesin N3 cultured on oleate are equivalent to thoseof the parent grown on oleate. N3 differs fromthe parent in that acetate-grown cells exhibit 40to 50% of the fully derepressed level of activity.Moreover, enzyme activities were not signifi-cantly higher when N3 was grown on butyrate,valerate, or decanoate than was the case for cellsgrown on acetate. This indicates that short- andmedium-chain acids, even when activated andmetabolized, do not derepress synthesis of theseenzymes. Activities were only slightly decreasedin glucose-grown cells. (iii) N3B-, which growson odd- but not even-chain acids, has no detect-

able crotonase activity and shows a 50% reduc-tion in HOADH activity. N30L- exhibits nocrotonase and low HOADH activity. N3,PR iscomparable to N3, as would be expected. (iv) D,differs fromn N, in that enzyme activities are sig-nificantly higher than the fully derepressed levelof activity found in the parent grown on oleate.Moreover, enzyme activities were comparablewhether D1 was grown at the expense of acetateor oleate. The fact that synthesis of the ,8-oxida-tion enzymes is derepressed to a greater extent inD1 than in N3 is in agreement with the higherspecific growth rate shown by D1 on mediumchain acids.

Biochemical lesion in N3V-. N3V- grows oneven- but not on odd-chain fatty acids, includingpropionate. In addition, propionate inhibitsgrowth on acetate. Propionate-positive revertants

TABLE 6. Specific activities of crotonase and ,3-hydroxyacyl-CoA dehydrogenase in extracts offatty

acid mutants

Relative specificCelltype Growth activity (%)bCell type substratea

Crotonase HOADH

Parent ........ Oleate 100 100Acetate 2 5

N3 .......... Oleate 110 103Acetate 45 34Butyrate 65 43Valerate 54 34Decanoate 60 40Glucose 32 25

N3B- ......... Oleate 0 52Acetate 0 11Valerate 0 24

N30L- ........ Acetate 0 13Succinate 0 0

N3PR ........ Oleate 103 107Acetate 36 36

DI ........... Oleate 258 170Acetate 294 174

a Cells were cultured on the substrates indicated,harvested at late log phase, and extracts were preparedas described in Materials and Methods.

b Crotonase and f3-hydroxyacyl-CoA dehydrogenase(HOADH) were assayed as described in Materials andMethods. Data for each enzyme are expressed in termsof relative specific activity, considering the activity forthe parent strain grown on oleate as 100. The respectivespecific activities for crotonase and HOADH in theparent cultured on oleate were 2.7 and 1.4 gmoles permin per mg of extract protein.

889

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

29

Dec

embe

r 20

21 b

y 17

9.10

9.17

2.46

.

SALANITRO AND WEGENER

grow on valerate and vice versa. The data inTable 7 provide evidence that the lesion in N3V-is in the pathway of propionate oxidation. Inthis table, the rates of 14CO3 evolution from val-erate-J-'4C and from propionate-l-'4C by N3were each given a value of 100. The rate of oxi-dation of these substrates by N3V- was thenexpressed as a relative percentage. It is apparentthat although N3V- oxidized valerate-1-'4C to14CO2, it did not significantly oxidize propionate.Although addition of unlabeled propionate de-creased the oxidation of valerate in N3, the effectwas much more pronounced in N3V-. Preincuba-tion of N. with unlabeled acetate had little effecton oxidation of propionate-1-54C to 14CO2, butpreincubation with unlabeled valerate markedlydecreased propionate oxidation. This is in accordwith the concept that valerate-1- 4C is oxidizedby p-oxidation to acetyl-CoA-I-14C plus pro-pionyl-CoA.We previously reported (25, 26) that the major

pathway of propionate metabolism in E. coli isa-oxidation to pyruvate. Since N3V- grows onlactate, the lesion appears to be in the formationrather than utilization of pyruvate. Enzyme ex-tracts of N3 and N3V- catalyze the formation ofpropionyl-hydroxamate from propionate, CoA,adenosine triphosphate (ATP), and hydroxyl-amine to the same extent, and it is probable thatthe lesion in N3V- is not in propionate activa-tion. Further studies are required to demonstratethe nature of the lesion. It is suggested thatN3V- is unable to grow on odd-numbered fattyacids because the oxidation of such acids resultsin the formation of propionyl-CoA. In this mu-tant, propionyl-CoA can not be oxidized to pyru-vate, and, as a result, CoA is tied up as a non-

TABLE 7. Comparison of N3 and N3V- with respect tooxidation of valerate and propionate

Unlabeled Relative rate ofI-14C substrate sunstabele oxidation ()

addeda addedb N, N,V

Valerate None 100 55 60Propionate None 100 5-10Valerate Propionate 60-65 2-6Propionate Acetate 90-95Propionate Valerate 30-35

a 1-14C fatty acids (2.5 IACi, 1.0 jLmole) were added, and therate of "4CO, evolution was assayed as described in Materialsand Methods.

bCells were preincubated at 37 C in the presence or absenceof the indicated fatty acids (3 mM) for 10 min before additionof radioactive substrate.

'The rates of "4C0 evolution from valerate-1-14C andfrom propionate-1-14C, respectively, shown by N3 were as-signed a value of 100, and other rates are expressed as a rela-tive percentage.

utilizable acyl-ester resulting in depletion of theCoA pool. Alternatively, it is possible that accu-mulation of propionyl-CoA or a further metabo-lite inhibits key metabolic processes.

DISCUSSIONRelatively few studies have dealt with regula-

tion of growth of bacteria on fatty acids as thesole source of carbon. Serratia marcescens andBacillus brevis oxidize medium-chain fatty acidsafter a lag but not short-chain acids (18, 19). Inthe same study, Pseudomonas aeruginosa andNeisseria catarrhalis oxidized the C9 and C,0fatty acids without a lag and may be constitutivefor the ,3-oxidation enzymes. A number ofaerobic pseudomonads (1, 12) utilize short-,medium-, and long-chain acids as the sole sourceof carbon, as do the Acinetobacter group (3) andthe fungus Cunninghamella echinulata (8). Bu-tyrate and valerate are metabolized by Clostri-dium (2) and by Streptococcus mitis (28), appar-ently by constitutive pathways.

Regulation of metabolism of long- and me-dium-chain fatty acids in E. coli K-12 was studiedby Overath et al. (10, 11) and by Weeks et al.(23). These authors suggest that fatty acids withless than 12 to 14 carbons do not support growthof K-12 because they are unable to derepress the,3-oxidation operon. This is in agreement withthe data presented here. Furthermore, data con-cerning enzyme activities formed by the N3 mu-tant grown on medium- and short-chain acids(Table 6) suggest that these substrates are un-able to derepress the ,8-oxidation operon evenunder conditions where they are activated andfurther metabolized. It is also of interest that 2-dodecenoate and 2-hexadecenoate supportgrowth of the N3 and D1 mutants but not of theparent strain. Since these a,/-unsaturated long-chain acids can be activated and further metabo-lized in ,B-oxidation-constitutive mutants, theirinability to support growth of the parent suggeststhat such substrates do not derepress the ,3-oxi-dation operon. The mechanism by which long-chain acids cause derepression of the operon isnot understood.

Growth on oleate results in a coordinate dere-pression of synthesis of the enzymes of the /3-oxidation sequence along with a long-chain fattyacid acyl-CoA synthetase. The synthetase geneappears to be regulated with, but not linked to,the operon controlling formation of the /3-oxida-tion enzymes. The synthetase is ATP-dependentand activates fatty acids C8 and higher but showsno activity with hexanoate and butyrate (10).Samuel et al. (15, 16) presented evidence for twoATP-dependent acyl-CoA synthetase activities inanother K-12 strain grown on oleate. One ex-

890 J. BACTERIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

29

Dec

embe

r 20

21 b

y 17

9.10

9.17

2.46

.

VOL. 108, 1971 GROWTH OF E. COLI ON SHORT-CHAIN FATTY ACIDS

hibits maximum activity with hexanoate, theother with long-chain acids. There does not ap-pear to be significant activation of butyrate bythe medium-chain synthetase. A guanosine tri-phosphate-dependent synthetase active withshort-chain fatty acids has been found in animaltissues (17). An acetyl-CoA:butyrate transferaseactivity has been described for Clostridia (20)and for E. coli K-12 (21).The specific growth rates of the f,-oxidation-

constitutive mutants described in the presentstudy (see Table 2) are consistent with the possi-bility that there are three physiologically distinctentry and/or activation systems for straight-chain fatty acids: one for fatty acids C,, and C5,another for fatty acids C. to C, and a third forfatty acids C10 to C18. It should be emphasizedthat this hypothesis is based only on growth ratedata. The possibility that a single protein acti-vates both long- and medium-chain acids but atdifferent rates cannot be excluded. While the ac-tivation system(s) for long-chain and medium-chain fatty acids is coordinately derepressed withthe $-oxidation enzymes, the entry or activation,or both, of short-chain acids requires an addi-tional mutation. The finding that, in the D1 ,B-oxidation-constitutive m utant, the specificgrowth rate on fatty acids C, to C9 is signifi-cantly lower than that exhibited on the C10 toC18 substrates suggests that the rate of fatty acidentry or activation or both may be a rate-lim-iting step in the oxidation pathway.The behavior of the N3B- mutant is of interest

and requires further investigation. This mutantgrows on odd-chain fatty acids at the same rateas does the N3 parent but does not grow on even-chain acids C4 to C,1 and grows slowly onoleate. Revertants which grow on butyrate regainthe ability to grow quickly on oleate and viceversa. No crotonase activity was detected in cellsgrown at the expense of oleate or valerate as as-sayed by hydration of crotonyl-CoA. The be-havior of this mutant suggests the possibility thatA-oxidation of odd- and even-chain fatty acidsmay involve at least one step which is catalyzedby distinct enzymes, one specific for odd-, theother for even-chain acids. It would be of interestto compare the relative activity of various /3-oxi-dation enzymes in N3 and N,3B- by using, assubstrates, the appropriate acyl-CoA esters ofboth the C4 and C5 acids. N3V- grows on even-but not odd-chain acids because of a lesion in thepropionate oxidation pathway. This indicatesthat odd-chain fatty acids are metabolized via a,B-oxidation and not an a-oxidation pathway.The data presented suggest that short-chain

acids do not penetrate the parent K-12 strain atpH 7. Medium-chain acids penetrate but appear

not to be activated by the parent. For growth tooccur on butyrate and valerate, two mutationsare required: (i) a mutation which allows theenzymes of /3-oxidation to be synthesized consti-tutively and (ii) a mutation which allows foruptake of butyrate and valerate. The latter as-pect is dealt with in the accompanying report(14).The N3 mutant grows on 2-butenoate and 3-

butenoate, and growth is inhibited by 4-pent-enoate. N3PR, which is pentenoate-resistant,has concurrently lost the ability to utilize buty-rate and valerate. These data are consistent withthe possibility that in N3 the saturated C4 and C5acids and the unsaturated short-chain acids maybe taken up by a common system. It also is pos-sible that these substrates are concentrated bydistinct uptake systems which are regulated coor-dinately. The mutation characterizing N3 doesnot permit uptake of the 2-methyl-substitutedshort-chain acids. The 2-hydroxy- and 2-amino-sustituted short-chain acids penetrate the parentand hence do not enter via the short-chain fattyacid uptake system. The fact that 3-butenoate isutilized suggests that the double bond can beshifted from the fly to the a/3 position. Overathet al. (11) reported that an isomerase (i) and anepimerase (ii) enzyme are derepressed with the,8-oxidation operon.

dodecen-3 cis-oyl-CoA xdodecen-2-trans-oyl-CoA (i)

D-3-hydroxylauryl-CoAL-3-hydroxylauryl-CoA (ii)

The inhibitory effect of 4-pentenoate ongrowth of N, can be partially overcome by addi-tion of CoA precursors to the medium. Panto-thenate, pantethine, and f3-alanine at a concen-tration of 1.25 mm relieved 4-pentenoate (1 mM)inhibition of growth on acetate by 82, 62, and60%, respectively. Data are similar for reversalof butyrate inhibition of growth of N30L- onacetate by precursors of CoA. It is suggestedthat 4-pentenoate inhibits growth by depletingthe free CoA pool of the cell. In the N3 cell types,this unsaturated short-chain acid penetrates andappears to be activated to the acyl-CoA ester butnot to be metabolized with regeneration of freeCoA. Stimulation of CoA synthesis by additionof CoA precursors to the medium partially over-comes this inhibition. This hypothesis is basedon growth data, and the accumulation of 4-pentenoyl-CoA has not been demonstrated.

ACKNOWLEDGMENTS

This investigation was supported by grant GB 8146 from theNational Science Foundation.

891

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

29

Dec

embe

r 20

21 b

y 17

9.10

9.17

2.46

.

SALANITRO AND WEGENER

LITERATURE CITED

1. Ballard, R. W., N. J. Palleroni, M. Doudoroff, R. Y.Stanier, and M. Mandel. 1970. Taxonomy of the aerobicpseudomonads: Pseudomonas cepacia, P. marginata, P.allicola, and P. caryophylli. J. Gen. Microbiol. 60:199-214.

2. Barker, H. A. 1956. Bacterial fermentations. John Wileyand Sons, Inc., New York.

3. Baumann, P., M. Doudoroff, and R. Y. Stanier. 1968. Astudy of the Moraxelia group. II. Oxidative-negative spe-

cies (Genus Acinetobacter). J. Bacteriol. 95:1520-1541.4. Decker, K. 1965. Crotonyl coenzyme A, p. 437-444. In H.

U. Bergmeyer (ed.), Methods in enzymatic analysis.Academic Press Inc., New York.

5. Ferdinandus, J., and J. B. Clark, 1969. Selective inhibitionof bacterial enzymes by free fatty acids. J. Bacteriol. 98:1109-1113.

6. Herbert, D., R. Elsworth, and R. C. Telling. 1956. Thecontinuous culture of bacteria; a theoretical and experi-mental study. J. Gen. Microbiol. 14:601-622.

7. Kornberg, H. L. 1966. The role and control of the glyoxy-late cycle in Escherichia coli. Biochem. J. 99:1- I 1.

8. Lewis, H. L., and G. T. Johnson. 1966. Intermediates offatty acid metabolism by Cunninghamella echinulata.Mycologia 58:136-147.

9. 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.

10. Overath, P., G. Pauli, and H. U. Schairer. 1969. Fatty aciddegradation in Escherichia coli. An inducible acyl-CoAsynthetase, the mapping of old-mutations and the isola-tion of regulatory mutants. Eur. J. Biochem. 7:559-574.

11. Overath, P., E. M. Raufuss, W. Stoffel, and W. Ecker.1967. The induction of the enzymes of fatty acid degra-dation in Escherichia coli. Biochem. Biophys. Res.Commun. 29:28-33.

12. Palleroni, N. J., M. Doudoroff, R. Y. Stanier, R. E. So-lanes, and M. Mandel. 1970. Taxonomy of the aerobicpseudomonads: the properties of the Pseudomonasstutzeri group. J. Gen. Microbiol. 60:215-231.

13. Rosenfeld, W. D. 1948. Fatty acid transformations by an-aerobic bacteria. Arch. Biochem. 16:263-273.

14. Salanitro, J. P. and W. S. Wegener. 1971. Growth of Es-cherichia coli on short-chain fatty acids: nature of thetransport system. J. Bacteriol. 108:893-901.

15. Samuel, D., and G. Ailhaud. 1969. Comparative aspects offatty acid activation in Escherichia coli and Clostridiumbutyricum. FEBS Letters 2:213-216.

16. Samuel, D., J. Estroumza, and G. Ailhaud. 1970. Partialpurification and properties of acyl-CoA synthetase ofEscherichia coli. Eur. J. Biochem. 12:576-582.

17. Sartorelli, L., L. Galzigna, C. R. Rossi, and D. M.Gibson. 1967. Influence of lecithin on the activity of theGTP-dependent acyl-CoA synthetase. Biochem. Biophys.Res. Commun. 26:90-94.

18. Silliker, J. H., and S. C. Rittenberg. 1951. Studies on theaerobic oxidation of fatty acids by bacteria. 1. The na-

ture of the enzymes, constitutive or adaptive. J. Bac-teriol. 61:653-659.

19. Silliker, J. H., and S. C. Rittenberg. 1951. Studies on theaerobic oxidation of fatty acids by bacteria. 11. Applica-tion of the technique of simultaneous adaptation to thestudy of the mechanism of fatty acid oxidation in Ser-ratia marcescens. J. Bacteriol. 61:661-673.

20. Stadtman, E. R. 1953. The coenzyme A transpherasesystem in Clostridium kluyveri. J. Biol. Chem. 203:501 -

512.21. Vanderwinkel, E., P. Furmanski, H. C. Reeves, and S. J.

Ajl. 1968. Growth of Escherichia coli on fatty acids:requirement for coenzyme A transferase activity.Biochem. Biophys. Res. Commun. 33:902-908.

22. Weber, G., M. A. Lea, and N. B. Stamm. 1969. Regula-tion of hepatic carbohydrate metabolism by free fattyacids and acetyl-CoA: sequential feedback mechanism.Lipids 4:388-396.

23. Weeks, G., M. Shapiro, R. 0. Burns, and S. J. Wakil.1969. Control of fatty acid metabolism. 1. Induction ofthe enzymes of fatty acid oxidation in Escherichia coli.J. Bacteriol. 97:827-836.

24. Wegener, W. S., P. Furmanski, and S. J. Ajl. 1967. Selec-tion of mutants constitutive for glyoxylate condensingenzymes during growth on valeric acid. Biochim. Bio-phys. Acta 144:34-60.

25. Wegener, W. S., H. C. Reeves, and S. J. Ajl. 1967. Propi-onate oxidation in Escherichia coli. Arch. Biochem. Bio-phys. 121:440-442.

26. Wegener, W. S., H. C. Reeves, R. Rabin, and S. J. Ajl.1968. Alternate pathways of metabolism of short-chainfatty acids. Bacteriol. Rev. 32:1-26.

27. Wegener, W. S., E. Vanderwinkel, H. C. Reeves, and S. J.Ajl. 1969. Propionate metabolism. V. The physiologicsignificance of isocitrate lyase during growth of E. colion propionate. Arch. Biochem. Biophys. 129:545-553.

28. Wolin, M. J., J. B. Evans, and C. F. Niven, Jr. 1952. Theoxidation of butyric acid by Streptococcus mitis. J. Bac-teriol. 64:531-535.

892 J. BACTERIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

29

Dec

embe

r 20

21 b

y 17

9.10

9.17

2.46

.