JOURNAL OF BACTrIOLOGY, Feb. 1974, p. 517-526Copyright 0 1974 American Society for Microbiology

Vol. 117, No. 2Printed in U.S.A.

Properties of Mutants in Galactose Taxis and TransportGEORGE W. ORDAL AND JULIUS ADLER

Departments of Biochemistry and Genetics, College of Agricultural and Life Sciences, University ofWiscorain-Madison, Madison, Wisconsin 53706

Received for publication 9 November 1973

,8-Methylgalactoside (mgl) permease mutants of Escherichia coli, which are

defective in three genes, mglA, mglB, and mglC, were assayed for galactose taxisand galactose transport. The mglB product is the galactose-binding protein.Previous evidence, supported by our new findings, shows that the galactose-bind-ing protein is the recognition component for galactose taxis as well as forgalactose transport. Most mutants defective in mglB showed strong effects on

both chemotaxis and transport; however, a couple showed effects chiefly on one

process or the other, thus allowing a separation of chemotaxis and transport. ThemglA and mglC products have not yet been identified, but they must becomponents of the galactose transport machinery since mutants defective inmglA or mglC, or both, showed strongly reduced transport. Although some ofthese mutants showed little chemotaxis, most gave close to wild-type chemotac-tic responses. Thus, transport is not required for galactose taxis. The bacteriadetect changes in the fraction of binding protein associated with galactose, notchanges in the rate of transport.

In the previous paper (17) mutants defectivein the ability to make a ring on galactose swarmplates, i.e., mutants in either galactose taxis ortransport, or both, were characterized geneti-cally. Each was found to lie in at least one ofthree genes, called mglA, mglB, and mglC.Complementation of mutants previously identi-fied as having defective galactose-binding pro-teins showed that mglB is the binding proteingene. Complementation of mutants known tohave intact binding proteins but still mutant inthe ,8-methylgalactoside (mgl) permease, thesystem of which the galactose-binding proteinforms a part, showed that these lie in mgIA,mglC, or both.The galactose-binding protein (5) is needed

for chemotaxis (12) and transport (5, 7) ofgalactose. Anraku (5) has shown that this pro-tein may be released from cells by osmoticshock, a procedure that makes the ,B-methyl-galactoside permease inactive. The purifiedprotein shows conformational change in thepresence of galactose (9, 10, 19), a property thatmight be expected for a protein designed totranslocate substrate.The products corresponding to the new genes,

mglA, and mglC, have not yet been identified,and, up to the present publications, their rolesin chemotaxis were unknown. For transport,additional components besides the galactose-binding protein have been suggested (7). In this

paper, we report assays of chemotaxis andtransport for mutants in mglA, mglB, andmg1C. The results confirm the role of thegalactose-binding protein and the lack of arequirement of transport for chemotaxis (1),and they document the role of the mglA andmglC proteins in chemotaxis and transport.

MATERIALS AND METHODSBacterial strains. The bacterial strains employed

are listed in Tables 1 and 2 and have been furtherdescribed in (17).

Media. Tryptone broth is 1% tryptone, 0.5% NaCl.Vogel-Bonner minimal medium (22) was used asmineral salts base. For growth of bacteria it wassupplemented with any essential amino acids (0.5mM), thiamine (1 gg/ml), and 0.4% D-galactose orglycerol. Eosine methylene blue (EMB) galactoseplates contained 2.75% EMB agar base (Difco) and1.0% galactose.

Radioactivity. D- [1- 14C ]Galactose, at specific ac-tivity of 52.9 mCi/mmol was obtained from NewEngland Nuclear Corp. Under our conditions ofcounting (Packard Tri-Carb liquid scintillationcounter), 1 pmol was found to correspond to 50counts/min.

Preparation of galK- strains. A strain to be madegalK- (defective in galactokinase) was lysogenizedwith X cI857 xis-1 Jam6 (gift from J. S. Parkinson).These lysogens are not viable at high temperaturesince the repressor is thermolabile and the phagecannot be excised (see 11; J. S. Parkinson, personalcommunication). Bacteria were transduced by P1

517

on May 21, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

TABLE 1. MglB mutants

Chemotaxis TransportStraina Class

Peak concn (galactose Peak response K. (galactose V.. (% wildmolarity) (% wild type) molarity) type)

Wild type 10-4to 3.2 x 10-4 100 4 x 10-7 100OW2 I 10-1 20 > lo-OW3 I 10-, 24 > 10-6OW4 I 3.2 x10-2 11 6 x 10-5 47OW5 I 10-2to3.2 x 10-2 52 4 x 10-5 125OW6 I 10-2 to 3.2 x 10-2 49 4 x 10- 6 125OW7 I 10-2 73 2 x 10-5 50OW8 I 10-2 31 10-6 28OW1o I 3.2 x 10- "to 10 x 10- ' 25 6 x 10-' 34OW9 I 3.2 x 10-3 17 6 x 10-' 15OW12 II 10-' 6 4.5 x 10-7 14Owi m 3.2 x 10-"to 10 x 10-' 49 4 x 10-7 9OW13 II 3.2 x 10-6 20 4 x 10-7 8.5

a Strains were analyzed for chemotaxis and transport for galactose by the procedures in Materials andMethods. Wild type is OW1 (16) and showed a Vin.: of 73 pmol of galactose per assay and a peak chemotaxisresponse of 78,000 bacteria.

TABiE 2. MgIA, mglC, and mglAC mutantsa

Chemotasis TransportStrain Gene Peak concn Peak response Km (galactose V=.,, (% wild

(galactose molarity) (% wild type) molarity) type)

Wild type 104to 3.2 x 10-4 100 4 x 10-7 100OW14 A 10-4 115 10-7 39OW15 A 10-4 113 3 x 10- 4OW23 C 10-4 110 8 x 10-5 23OW37 AC 10-4 110 2.5 x 10-8 11OW16 A 10-4 110 10-7 19OW24 C 10-4 85 3x10-5 4OW35 AC 10-4 84 1.5 x 10-7 21OW17 A 10-4 83 3x10-5 9OW18 A 3.2 x 10-5 81 2 x 10- 1.6OW25 C 10-4 80 5x10-8 10OW38 AC 3.2x10-6 75 <1OW39 (AC) 10- 70 <1OW26 C 3.2x10-5 66 4 x 10-8 8OW40 AC 3.2x10- 66 3x10- 5OW29 C 3.2x10-"to10xl10- 63 2x10- 4OW27 C 3.2x10-5 59 <1OW19 A 10-4 57 5 x 10- 12OW28 C 10-4 57 2.5x 10-8 19OW30 C 3.2x10- 53 2x10-8 4OW31 C 3.2x10- 44 <1OW41 AC 3.2x10-5 44 <1OW42 (AC) 10-4 40 1.5 x 10-7 10OW21 A 3.2 x 10- 33 2.5 x 10- 4OW43 AC 3.2x10-5 32 <1OW20 A 3.2 x 10- 31 5.5 x 10- 5OW32 C 3.2x10- 29 <1OW33 C 3.2x10-5 25 <1OW44 AC 10-5 <1OW22 A _ 5 <1OW34 C 2 <1

a Wild type is OW1 (16) and showed a V... of 73 pmol of galactose per assay and a peak chemotaxis responaeof 78,000 bacteria. (AC) means mgLA, mglC, or mglAC (see ref. 16). A - means the response was too low todetermine a peak value.

518

on May 21, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

MUTANTS IN GALACTOSE TAXIS AND TRANSPORT

grown on W3092i (23), which is galK-, for ability tosurvive at 42 C (due to transductional loss of Xprophage) on EMB galactose plates (see 17 formethod of transduction). Light colonies were galK-and were about 5% of the total temperature-insensi-tive transductants. Control experiments showed en-hancement of 100- to 1,000-fold in frequency oftemperature-insensitive colonies due to the transduc-tion (compared with absence of transducing phage).

Chemotaxis assays (2). Bacteria were grown withshaking at 35 C overnight in tryptone broth anddiluted 1:50 into minimal medium containing galac-tose for growth. Bacteria were harvested at an opticaldensity of 590 nm (OD,90) between 0.25 and 0.40(OD,90 of 1.0 = 6 x 10' bacteria/ml), centrifuged, andwashed by centrifugation three times at room temper-ature in 10 mM potassium phosphate, pH 7, 0.1 mMEDTA, and finally resuspended in this buffer at OD,,0= 0.003. A capillary containing attractant (D-galac-tose or D, L-a-methylaspartate) was placed into abacterial suspension at 30 C. Bacteria traveled intothe capillary and after 1 h were extruded into a testtube and titered (see 2).The data were handled as follows: A capillary

without attractant was run in this chemotaxis assayand the number of bacteria which entered was takenas the "blank" value. This value was subtracted fromthose obtained using attractants. Finally, to reducethe effect of day to day variations, the ratio of theresponse to 1 mM a-methylaspartate by the wild type(typically 90,000 bacteria) to that by the particularstrain being assayed was used to normalize thegalactose responses. Control experiments were per-formed to be sure that a-methylaspartate taxis is notaffected by mutations in galactose taxis.

Transport assays. Bacteria were grown as forchemotaxis assays except that glycerol and 1 mMD-fucose (10 mM or 100 mM for mglB strains since 1mM was found inadequate in some cases) was usedinstead of galactose to induce for mgl permease andbacteria were washed three or four times by centrifu-gation and resuspended in Vogel-Bonner salts with0.02% glycerol (suspension medium) at OD,,, = 0.03.The assay was carried out at 30 C as follows: 100

,gliters of "4C-galactose was added to 900 1sliters ofbacteria (about 1.5 x 106 bacteria) at time zero andmixed. At 15 s 10 ml of suspension medium was addedand the mixture was filtered on a Millipore filter. Thezero time point was obtained by diluting the bacteriainto the 10 ml of suspension medium before addingthe "4C-galactose. The value of the zero time pointwas subtracted from that of the 15-s point.

Data were graphed as the reciprocal of the galactoseconcentration on the abscissa and the reciprocal of thegalactose accumulated on the ordinate. A straight linewas drawn through the points. The abscissa interceptis then - 1/Km and the ordinate intercept is l/Vmax.

RESULTSGalactose taxis of parent. Figure 1 com-

pares chemotaxis to galactose by the wild-typestrain and a derivative lacking galactokinase,galK-, the first enzyme for metabolism of

galactose. For the wild type, the greatest num-ber of bacteria entered the capillary whoseinitial concentration of galactose was 3.2 x 10-4M. At concentrations above and below this,fewer bacteria entered the capillary (see 16 forexplanation of optimum). For the galK- deriva-tive, the optimum concentration (peak) wasshifted to 10-4 M. In other experiments, thepeak concentration of the wild type variedbetween 10-4 M and 3.2 x 10-4 M and for thegalK- derivative between 3.2 x 10-5 M and10-4 M. Hence, absence of galactose metabo-lism appears to shift the position of the chemo-taxis peak slightly to lower concentrations.Similar experiments were performed with mglmutant strains, and occasionally but not re-producibly there was a 3.2-fold shift of the peakto lower concentrations in the galK- form.Because the effect was small, the gal+ formswere used since they were the available ones.

1i-6 10-5 10-4 10o-3 1-2 10o-GALACTOSE MOLARITY

FIG. 1. Galactose taxis assays of wild-type (mgl+)in gal+ and galK- forms. See Materials and Methodsfor description of assay and treatment of data. Blankvalues: gal+ 500, galK- 400. Accumulation to 1 mMa-methylaspartate: gal+ 86,000, galK- 83,500. ThegalK- strain was grown on glycerol and 1 mM fucose.Symbols: 0, gal+; A, galK-.

519VOL. 117, 1974

on May 21, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

ORDAL AND ADLER

In analysis of such a chemotaxis experiment,there are two parameters of especial interest:the concentration giving the maximum response(the peak) and the size of the response itself. Forcomparison of strains, to reduce the effect ofday to day variations, the size of the galactoseresponse was standardized by comparison withthat of the response to 1 mM a-methylaspartatewhose taxis is unaffected by mutations in galac-tose taxis (see Materials and Methods).Galactose transport of parent. To compare

parent and mutants for transport of galactoseand determine the range of concentrations overwhich the ,8-methylgalactoside permease is themajor system conveying galactose into cells,wild-type bacteria were assayed for transportfrom 10-'M to 5 x 10- M. Figure 2 shows threedouble reciprocal plots of overlapping concen-tration intervals. Figure 2A, which covers con-centrations from 10-8 M to 10-6 M, shows that

A A

0.8_

0.6-

OA

02

-4 -2 0 2 4 6 8 10

107 (GALACTOSE moLARrTY)-IFIG. 2. Galactose transport assays of wild type

(mgl+) in gal+ and galK~forms. The panels showoverlapping concentration ranges. A, 10-' M to 10-M galactose; B, 10-7 M to 10-5 M galactose; C, 3 x10-6 M to 5 x 15-6 M galactose. Straight linesconnecting the points were drawn for panels A andSB and indicated a single transport system having aVmax Of 73 pmol of galactose per 15 s per 1.5 x 10'bacteria and a Km of 4 x 10- 4M. In panel C the mglpermease is saturated so that the line is roughlyhorizontal. Symbols: 0, gal+; *, galK-.

only one transport system operates for thisrange. Figure 2B, which covers concentrationsfrom 10-7 M to 10-5 M shows only this onesystem. Finally, as shown in Fig. 2C, spanningthe range of 3 x 10-6 M to 5 x 10- M, showsthe same transport system at concentrationswhere it is saturated (i.e., operating at max-imum velocity). It is apparent that only onetransport system, having a Km of 4 x 10-7 Mand a Vmax of 73 pmol/assay (per 15 s per 1.5 x106 bacteria) provides the major route for trans-port of galactose from 10-8 M to 5 x 10i'- Mgalactose. A minor transport system also ap-pears to exist, however, but was not measurablein the wild type; see existence of second trans-port system below.

Bacteria used in the above experiments werecapable of metabolizing galactose. However, itmade little difference whether cells were gal+ orgal- (Fig. 2). Only one system, having a Km of 4x 10-7 M galactose, was detected. It should benoted that at very low concentrations, lessgalactose accumulation than expected was ob-served for the galK- strain (Fig. 2A). Oneexplanation for this observation is that nonradi-oactive galactose might leak into the medium,effectively diluting the isotope, more fromgalK- than from gal+ cells since the latterwould metabolize endogenously produced ga-lactose, resulting in a lower internal concentra-tion (22).

Experiments to show that the assay waslinear during the first 15 s are presented in Fig.3. This result shows that at 10-8 M galactose,the lowest concentration used in any of theassays, about 18% of the label was taken up at15 s (including the zero time value of 9%, whichpresumably includes rapid binding to the galac-

O 2 3 4 5 6 7 8 9 10TIME (MINUTES)

FIG. 3. Kinetics of uptake on 10-6 Mgalactose bywild type (mgl+) in gal+ and galK- forms. Totalradioactivity in assay was 543 counts/min. The assaywas carried out according to procedures in Materialsand Methods, except zero time points were notsubtracted. Symbols: 0, gal+; A, galK-.

520 J. BAcToioL.

on May 21, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

MUTANTS IN GALACTOSE TAXIS AND TRANSPORT

tose-binding protein (18)). Eventually, about65% of the label was taken up before the systemreached equilibrium. Figure 3 shows that pres-

ence of the galK- mutation made little differ-ence in the results. An experiment to show thatthere was nothing unusual about the galactosethat was not taken up was performed by allow-ing gaIK- bacteria to equilibrate with 4 x 10-7M galactose, filtering to obtain the free galac-tose, and incubating the bacteria with thisgalactose. This galactose started at a concentra-tion of about 1.4 x 108 M and about 68% of itwas taken up at equilibrium.Analysis of mglB mutants. Mutants in

mglB were analyzed for galactose chemotaxisand transport and the results are presented inTable 1 and summarized in Table 4. Themutants fall into four classes.

Class I. Most of the newly isolated mutantsfall into class I, whose features include (i) shiftof Km for transport to higher concentrations, (ii)a concomitant shift of peak in chemotaxis assayto higher concentrations, always roughly 500-fold higher than the Km for transport, and (iii)variously depressed chemotaxis response (11 to73% of wild type) and Vmax (15 to 125% of wildtype). Besides strains listed in Table 1, strainAW550 (12) is a member of this class. As anexample of class I mutants, Fig. 4 presents a

4

3 CL.qjB CLASS EI)

2- OWIO (m IB CLASS

cc3eI

0 0W12

GALACTOSE MOLARITYFIG. 4. Galactose taxis assays of mglB mutants.

Blank values: OWIO 300, OWII 700, OW12 400.Accumulation to I mM a-methylaspartate: OW1051,000, OW11 78,000, OW12 71,500. Symbols: A,

OW10; 0, OWII; U, OW12.

chemotaxis assay for OW10, and Fig. 5A pre-sents transport assays for this mutant. StrainEH3039, a mutant shown by an altered sequence

of amino acids to be in the structural gene for thegalactose-binding protein (8), is defective in thetransport of galactose (8) and fails to carry outgalactose taxis (M. Goy, unpublished data); arevertant in that gene shows transport (8) andtaxis (M. Goy, unpublished data).

Class II. Two of the newly isolated mglBmutants fall into class II, whose characteristicsare (i) no shift of Km for transport, (ii) smallshift of chemotaxis peak to lower concentra-tions, and (iii) substantially depressed chemo-taxis (6 to 20% of wild type) and Vmax (8 to 14%).Strains previously identified as having defectivegalactose-binding proteins and members of thisclass include AW543 (12) and AW520 (12, 13); inthese cases chemotaxis was even more severelydepressed (not detectable). As an example ofclass II mutants, Fig. 4 shows a chemotaxisassay for OW12, and Fig. 5B presents transportassays for this strain.Class III. Only one strain has been found in

class III, strain OW11, whose features includegood chemotaxis (49%) but little transport(Vmax = 9%). Its Km is essentially unchangedfrom that of wild type. An assay of chemotaxisfor OW11 is shown in Fig. 4, and transport

.4

A

-3

.2v

f0-4-2 O 2 4 6 8 to 12 14

105 (GALACTOSE MOLARrTY)-I

.6

w OWlI (M1gJpCLASS K)

106 (GALACTOSE MOLARITY)-'FIG. 5. Galactose transport of mglB mutants. See

Materials and Methods for description of assay andtreatment of data. Panel A shows OW10, and panel Bshows OWJZ (a) and OW12 (A). Values of Km andVmax are given in Table 1.

VOL. 117, 1974 521

on May 21, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

ORDAL AND ADLER

assays are shown in Fig. 5B.Class IV. Finally, there is a still different

class, for which no representatives were foundamong the newly isolated mutants, exemplifiedby strain AW551 (12). Its main characteristicsare (i) no chemotaxis but (ii) normal transport.Figure 6 presents an assay of chemotaxis forstrain AW551 and its parent, AW546, and Fig. 7shows transport assays for these two strains.Analysis of mglA, mglC, and mglAC

mutants. Mutants in mgLA or C, or both, wereanalyzed for galactose chemotaxis and trans-port and the results are presented in Table 2.The mutants do not sharply fall into classes;however, at the extremes two classes may bedefined.

Class I mutants have the following character-istics: (i) shift of Km for transport to lowerconcentrations, (ii) substantial to extreme de-creases of Vlmax, (iii) nearly normal chemotacticresponse, and (iv) possibly small shift of chemo-taxis to lower concentrations. It should be notedthat the small shift of the chemotaxis peak doesnot follow the decreases of Km and, as explained

i-0-7 106 Co4 10-4 10-3 -o-ICGALACTOSE MOLARITY

FIG. 6. Galactose taxis assay ofA W551, mutant inmglB, and its parent, A W546. Blank values: A W5519,100; AW546, 14,100. Accumulations to 1 mM a-

methylaspartate: A W551 260,900; A W546, 269,000.These assays were carried out at a higher bacterialconcentration (ODf60 = 0.03 rather than 0.003) thanused in all other experiments; hence the higher blanksand higher responses. Symbols: A, A W551; 0,

A W546.

D75- AW46 (PARENT)

wf AW551 CMq 1B CLASS Z)

OI 1-2 0 2 4 6 8 10

106 CGALACTOSE MOLARITY)-'FIG. 7. Galactose transport assays of AW;55, mu-

tant in mglB, and its parent A W546. The graph showsthat both strains have a Km of 5 x 10-7M galactoseand a Vmax of 90 pmol of galactose per 158 per 1.5 x106 bacteria. Symbols: 0, AW551; A, AW546.

in the analysis of the galK~wild-type strain,may be understood as due to a lower rate ofgalactose utilization, as a result of the defect ingalactose transport. Hence, gradients of galac-tose at low concentrations are preserved longerduring the chemotaxis assay. An example ofclass I is 0W29, mutant in mglC. Figure 8 showsa chemotaxis assay of this strain and Fig. 9 atransport assay.At the other extreme, in class II, are mutants

severely reduced in both chemotaxis and trans-port. Figure 8 shows a chemotaxis assay of0W34, mutant in mglC, an example of class fl.

Finally, at the border between classes I and IIare strains having significantly decreased chem-otaxis, usually accompanied by very low trans-port. Figure 8 shows a chemotaxis assay of0W20, mutant in mglA and an example of aborderline strain. Figure 9 shows a transportassay for this strain.Existence of second transport system. Sev-

eral reports indicate that E. coli possess agalactose transport system with a Km in therange 10-6 to 10-5 M. Rotman and Radojkovic(20) found that strain ML32400 had Km'S forgalactose uptake at 4.9 x 10-7 M and at 4 x10-6 M. Kerwar, Gordon, and Kaback (14)found that strain ML3 showed Km's for galac-tose uptake at 7.4 x 10-7 M and at 9.1 x 10-6M. In our wild-type strain, we could not detect asecond, lower-affinity transport system, and wetherefore assumed that if such a system ispresent, its Vmax must be low relative to that ofthe ,8-methylgalactoside permease.

In our mutant strains with defective mgltransport, we could test directly for presence ofa second galactose transport system. Table 3shows results of tests of several strains with lowmgl permease activity for presence of this sec-

522 J. BATRioL.

on May 21, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

MUTANTS IN GALACTOSE TAXIS AND TRANSPORT

FIG. 8. Galactose taxis assays of mglA and mglCmutants. Blank values: 0W20, 200; 0W29, 800;0W34, 800. Accumulations to 1 mM a-methylaspar-tate: OW20, 74,600; OW29, 66,700; OW34, 125,000.Symbols: A, OW20; *, OW29; *, OW34.

105 (GALACTOSE MOLARITY)-IFIG. 9. Galactose transport assays of 0W20 and

0W29. Values of K. and Vmax are given in Table 2.Symbols: *, OW20; A, OW29.

ond system. It can be seen that mglA, mglC,and mglAC mutants appear to possess a secondsystem. It has not been determined whethervariations in Km and Vmax among these strainsare significant. An example is OW34, a mglCmutant (Fig. 10) showing a Km of 7 x 10-6 Mgalactose and a Vmax of 22 pmol of galactose perassay (compared with a Vmax of 73 pmol perassay of the wild-type mgl permease). Thissecond system does not mediate taxis towardgalactose since OW34 is highly defective inchemotaxis over a very broad range of concen-trations (Fig. 8). OW34 was too low in mglpermease to allow an mgl Km to be determined,but OW21, OW24, OW29, and OW40 showed

both an mgl Km (Table 2) and a second Km(Table 3); thus single strains showed propertiesof two different systems.

However, several mglB mutants do not showa second transport system despite the fact thattheir mgl transport systems are only 8 to 14%that of wild type. Thus, the second systemdepends in some way upon the galactose-bind-ing protein. Figure 10 shows that OW13, a mglBmutant, has only the mgl permease, whose Kmis about 4 x 10-7 M galactose and whose Vmax is8 pmol per assay.

DISCUSSIONThe results presented in this and the preced-

ing paper show that three genes play roles intaxis towards and transport of galactose: mglB,

TABLE 3. Second transport systema

Strain Gene Km(galactose Vd.. (pmol of % Wild-typeStraiGene molarity) galactose per mgl V...

OW21 A 5 x 10-6 16 22OW34 C 7 x 10-6 22 30OW24 C 4 x 10-6 9.5 13OW29 C 2 x 10-6 7.5 11OW40 AC 2x10-6 6.5 9OWli B _- -

OW12 B _ - -OW13 B

a Strains were analyzed for Km in the range of 10-6M to 10-' M galactose. All mglA and/or C mutantsexcept OW34 have a second Ki,, in the range of 10-8M to 10-7 M galactose, as shown in Table 2. A -means no system was detected in addition to the mglpermease (see Table 1).

2.0

1 .6-

Ut.. 0~~~~~~~W20(MgAj)

|00.8_

Lo0.4 - A OW29 omlC CLASS 1)

-e /o I 0-6 -4 -d 0 2 4 6 8 t

107 (GALACTOSE MOLARrry)-'

FIG. 10. Galactose transport assays of OW13 and0 W34. Values ofKmwere 7 x 10-6M (OW34) and 4 x10-7M (OW13) and of Vmax were 22pmol of galactoseper 15 s per 1.5 x 106 bacteria (OW34) and 8 pmol ofgalactose per 15 s per 1.5 x 106 bacteria (OW13).Symbols: A, OW13; *, OW34.

VOL. 117, 1974 523

on May 21, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

ORDAL AND ADLER

which is the structural gene for the galactose-binding protein; and two genes, mglA andmglC, whose corresponding polypeptides arepresently unidentified.The following paragraph is a summary of our

interpretations, which will be discussed in moredetail below. Chemotaxis requires a binder, thecomponent that recognizes attractant, in thisinstance the galactose binding protein, whichrecognizes galactose. In addition, chemotaxisrequires a "signaller," the component that de-tects changes in the binder and sends signals tothe flagella to bring about appropriate swim-ming behavior. "Chemoreceptor" refers to thecomplex of binder and signaller. Our studiessuggest that the signaller responds to changes inthe amount of binder-attractant complex. Thebinder is closely associated with the transportmachinery of the cell; indeed it is also therecognition component for transport. Theseideas are diagrammed in Fig. 11.

Interpretation of mglB mutants. The galac-tose-binding protein (the mglB product) is thebinder for galactose taxis based on evidencepresented previously (12), and in addition it isknown to be that component of the ,B-methyl-galactoside permease that recognizes the galac-tose (5, 7).The mglB mutants are basically of four

classes, as summarized in Table 4. The fact thatmembers of each class showed significantly

7'ransport

OtheTransportComponents

Message to flagella

Signaller

Binder

FIG. 11. Model relating transport and chemotaxis. 4.

altered taxis toward galactose-either with re-duced taxis or with a change in the concentra-tions at which it occurs-confirms that thegalactose-binding protein is pivotal for galac-tose taxis.

Class I mglB mutants show parallel shifts oftransport and taxis to higher concentrations.We suggest that galactose does not bind aseasily to the mutant binding protein and so

must be present in higher concentrations toattain the same degree of binding of galactosethat is accomplished in the wild type at lowerconcentrations. This finding is expected of aprotein that is at the same time the bindingcomponent for a transport system and for taxis.

Class II mglB mutants, highly defective inboth galactose taxis and transport, show therequirement for the galactose-binding protein inboth processes.The class III mglB mutant shows that the

galactose-binding protein can be altered in sucha way that good taxis (adequate interactionwith the signaller) is still possible when trans-port is strongly inhibited. Thus, normal trans-port of galactose into the cell is not required forgalactose taxis.

Since the galactose-binding protein is thebinder, it must interact with the signaller. Infact, AW551, the class IV mglB mutant, may beinterpreted as being defective only in the in-teraction of binding protein with signaller. Thefact that AW551 is normal for transport isevidence that some feature of the binding pro-tein itself, probably changes in the fraction ofthe binding protein associated with galactosewith time, are monitored by the signallers,rather than transport of galactose into the cell.

Interpretation of the mglA, mglC, andmglAC mutants. At the extreme, these mu-tants are of two types, as summarized in Table

TABLE 4. Summary of mutant typesa

Gene Classn Chemotaxis Transport

B I Peak shifted to higher concentrations; Km shifted to higher concentrations;response ranges from good to poor Vmax ranges from high to low

B II Peak not shifted or shifted to somewhat Km not shifted; Vmax lowlower concentrations; response low orabolished

B III Peak not shifted; response good (1/2 that Km not shifted; Vma, lowof wild type)

B IV Abolished Wild typeA, C I Peak shifted to somewhat lower concen- Km shifted to lower concentrations;

trations; response close to that of Vmax lowwild type

A, C II Virtually abolished None detectable

a Summary was taken from the data of Tables 1 and 2.

524 J. BACTRIUOL.

on May 21, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

MUTANTS IN GALACTOSE TAXIS AND TRANSPORT

The claim (1) that transport of galactose intothe cell is itself not required for galactose taxisis confirmed by properties of the class I mglA orC mutants, or both. Many of these strains arehighly defective in transport of galactose intothe cell but little or not defective in chemotaxis,indicating that chemotaxis does not depend ontransport. This result confirms the finding thatstrain 20SOK-, an mglAC mutant, showedexcellent chemotaxis but extremely low trans-port (1).What might be the nature of the mglA and

mglC proteins? The fact that there appears tobe no difference in kinds of mutants found ineach indicates that both proteins play similarroles. In this connection, as pointed out inthe preceding paper (17), it is possible thatmglA and mglC mutations lie in the same geneand complement intracistronically. It seemslikely that the A and C proteins reside in thecytoplasmic membrane, since galactose musttraverse this to get into the cytoplasm. The factthat for most mutants chemotaxis is nearlynormal but transport of galactose into the cell isseverely damaged indicates that these proteinsare primarily involved in transport and onlyperipherally in chemotaxis.The type II mglA or C mutants, or both,

however, showed severely reduced chemotaxis.Apparently, sufficiently abnormal mglA ormglC polypeptides can interfere with interac-tion of the binding protein and the signallersince all these proteins are probably near eachother. Thus the mglA and C products appear tobe peripherally involved in galactose taxis,although centrally in transport of galactose intothe cell.

Signaller for galactose taxis. Many exam-ples of a new type of mutant (G. W. Ordal andJ. Adler, unpublished data) were found whosetaxis to galactose and ribose is defective withoutmuch effect on other taxes or on transport ofgalactose. It is noteworthy that ribose taxis isalso mediated by an osmotically shockablebinding protein (4, 12). These mutants, calledtrg (for taxis to ribose and galactose), map atabout 30 min on the E. coli map (G. W. Ordal,unpublished data), as compared with about 41min for the mgl mutants (ref. 21; E. N. Kort,R. W. Reader, and J. Adler, in preparation).These trg mutants may be defective in proteinsthat are candidates for being chemotactic sig-nallers.Two examples of a still different type of

mutant, derived from AW546 (the parent ofAW551) and not found among the mutantsisolated in this study, exist. This type is defec-

tive in galactose taxis but normal for other taxestested, including ribose, and is normal in galac-tose transport. The mutations appear to mapwhere the trg mutants map and because of theirrarity may represent a specialized mutation of atrg allele. These mutants may also be defectivein proteins that are candidates for being signal-lers.The results reported in this investigation

indicate that transport of galactose into the cellis not required for galactose chemotaxis.Rather, change in some feature of the galactose-binding protein itself is what is monitored. Mostlikely, this feature is the fraction of the bindingprotein that is associated with galactose, basedon the following evidence. First, the size of thechemotactic response depends on the change infraction of binding protein occupied; near thedissociation constant of the binding protein thechemotactic response and the change in fractionof protein occupied are greatest (16). Second,the peak concentration for various attractantsbound by this protein parallels their dissocia-tion constants (16). Finally, in the mglB class Imutants the peak concentration for galactosetaxis parallels the Km for transport and presum-ably the dissociation constant of the bindingprotein. Some component, here called signaller,must monitor the change in the binding protein.One may ask how the signallers can detect

whether galactose is complexed with bindingprotein. Boos et al. (9, 10) and Rotman and Ellis(19) have found evidence for two conformationsof the galactose-binding protein. If binding thegalactose causes a conformational change in thebinding protein (as these authors propose), thenpossibly this second conformation might beresponsible for activating chemotactic signal-lers.How does the swimming behavior of bacteria

result in chemotaxis? Berg and Brown (6 andmanuscript in press) and Macnab and Kosh-land (15) have shown that when concentrationof attractant increases, bacteria swim more andtumble less; when it decreases, the bacteriatumble more and swim less, compared to therelative frequencies in isotropic medium. Thusbacteria headed up the concentration gradienttend to keep going, and bacteria headed thewrong way, down the gradient, tend to tumbleso that they may turn around. Hence, in someway, information from the chemoreceptors issent to the flagella to determine the frequencyof tumbling. Macnab and Koshland (15) pro-posed the existence of an unidentified metabo-lite whose level, influenced by enzymes acti-vated by attractant-binder complex, controls

VOL. 117, 1974 525

on May 21, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

ORDAL AND ADLER

coordination of flagella. Further work will beneeded to elucidate the mechanism by whichchemoreception.results in signals to the flagella.

ACKNOWLEDGMENTSWe thank W. W. Cleland for valuable help in analysis of

the data and J. Straughn for performing assays of galactosetaxis and transport. G. W. 0. held postdoctoral fellowshipsfrom the American Cancer Society (PF-656) and the NationalInstitutes of Health (I F02 NS55125-01). The work wassupported by Public Health Service grant AI-08746 from theNational Institute of Allergy and Infectious Diseases.

LITERATURE CITED1. Adler, J. 1969. Chemoreceptors in bacteria. Science

166:1588-1597.2. Adler, J. 1973. A method for measuring chemotaxis and

use of the method to determine optimum conditions forchemotaxis by Escherichia coli. J. Gen. Microbiol.74:77-91.

3. Adler, J., and B. Templeton. 1967. The effect of environ-mental conditions on the motility of Escherichia coli. J.Gen. Microbiol. 46:175-184.

4. Aksamit, R., and D. E. Koshland, Jr. 1972. A ribosebinding protein of Salmonella typhimurium. Biochem.Biophys. Res. Comm. 48:1348-1353.

5. Anraku, Y. 1968. Transport of sugars and amino acids inbacteria. Im. Studies on the restoration of activetransport. J. Biol. Chem. 243:3128-3135.

6. Berg, H. C., and D. Brown. 1972. Chemotaxis in E. colianalyzed by three-dimensional tracking. Nature (Lon-don) 239:500-504.

7. Boos, W. 1969. The galactose binding protein and itsrelationship to the ,@-methylgalactoside permease fromEscherichia coli. Eur. J. Biochem. 10:66-73.

8. Boos, W. 1972. Structurally defective galactose-bindingprotein isolated from a mutant negative in the ,-methylgalactoside transport system of Escherichia coli.J. Biol. Chem. 247:5414-5424.

9. Boos, W., and A. S. Gordon. 1971. Transport properties ofthe galactose binding protein of Escherichia coli. Oc-currence of two conformational states. J. Biol. Chem.246:621-628.

10. Boos, W., A. S. Gordon, R. E. Hall, and H. D. Price. 1972.Transport properties of the galactose-binding protein ofEscherichia coli. Substrate-induced conformationalchange. J. Biol. Chem. 247:917-924.

11. Gottesman, M. E., and R. A. Weisberg. 1971. Prophageinsertion and excision, p. 113-138. In A. D. Hershey(ed.), The bacteriophage lambda. Cold Spring HarborLaboratory, Cold Spring Harbor, New York.

12. Hazelbauer, G. L., and J. Adler. 1971. Role of thegalactose binding protein in chemotaxis of Escherichiacoli toward galactose. Nature N. Biol. 230:101-104.

13. Hazelbauer, G. L., R. E. Mesibov, and J. Adler. 1969.Escherichia coli mutants defective in chemotaxis to-ward specific chemicals. Proc. Nat. Acad. Sci. U.S.A.64:1300-1307.

14. Kerwar, G. K., A. S. Gordon, and H. R. Kaback. 1972.Mechanisms of active transport in isolated membranevesicles. IV. Galactose transport by isolated membranevesicles from Escherichia coli. J. Biol. Chem.247:291-297.

15. Macnab, R. M., and D. E. Koshland, Jr. 1972. Thegradient-sensing mechanism in bacterial chemotaxis.Proc. Nat. Acad. Sci. U.S.A. 69:2509-2512.

16. Mesibov, R., G. W. Ordal, and J. Adler. 1973. The rangeof attractant concentrations for bacterial chemotaxisand the threshold and size of response over this range:Weber law and related phenomena. J. Gen. Physiol.62:203-223.

17. Ordal, G., and J. Adler. 1973. Isolation and complemen-tation of mutants in galactose taxis and transport. J.Bacteriol. 117:509-516.

18. Parnes, J. R., and W. Boos. 1973. Energy coupling of the,6-methylgalactoside transport system of Escherichiacoli. J. Biol. Chem. 248:4429-4435.

19. Rotman, B., and J. H. Ellis, Jr. 1972. Antibody-mediatedmodification of the binding properties of a proteinrelated to galactose transport. J. Bacteriol.111:791-796.

20. Rotman, B., and J. Radojkovic. 1964. Galactose transportin Escherichia coli. The mechanism underlying theretention of intracellular galactose. J. Biol. Chem.239:3153-3156.

21. Taylor, A. L, and C. D. Trotter. 1972. Linkae map ofEscherichia coli strain K-12. Bacteriol. Rev. 36:504-524.

22. Vogel, H. J., and D. M. Bonner. 1956. Acetylornlthinaseof Escherichia coli: partial purification and someproperties. J. Biol. Chem. 218:97-106.

23. Wu, H. C. P. 1967. Role of the galactose transport systemin the establishment of endogenous induction of thegalactose operon in Escherichia coli. J. Mol. Biol.24:213-223.

526 J. BACTFOL.

on May 21, 2020 by guest

http://jb.asm.org/

Dow

nloaded from