uptake ferrienterochelin escherichia coli: energy ... · energy-dependent process and requires the...

JOURNAL OF BACTERIOLOGY, Apr. 1977, p. 26-36Copyright X) 1977 American Society for Microbiology

Vol. 130, No. 1Printed in U.S.A.

Uptake of Ferrienterochelin by Escherichia coli: Energy-Dependent Stage of Uptake

ANTHONY P. PUGSLEY* AND PETER REEVESDepartment of Microbiology, University of Adelaide, Adelaide, South Australia, 5000

Received for publication 27 October 1976

The uptake of the siderophore-iron complex ferrienterochelin was found to bestrongly dependent upon an energized membrane state, as demonstrated by itssensitivity to dinitrophenol, azide, and cyanide. Ferrienterochelin uptake mayalso be dependent upon phosphate bond energy, as indicated by sensitivity toarsenate and iodoacetic acid. Although the adenosine triphosphatase does notappear to be involved in this energy coupling mechanism, ferrienterochelinuptake was shown to be less dependent upon phosphate bond energy than was

glutamine uptake. Sensitivity of ferrienterochelin uptake to osmotic shock was

shown to be due to the release of a ferrienterochelin binding compound located inthe outer membrane of the cells and probably identical to the colicin B receptorprotein.

The uptake of nutrients, growth factors, andions (referred to herein as transport substrates)by Escherichia coli involves transfer across twolipid bilayers (7). Transfer across the outermembrane is thought, in most cases, to be anentirely passive or nonspecific process, thetransport substrates normally being smallenough to diffuse through pores in the outermembrane (27, 29). Transport against a concen-tration gradient into the cell cytoplasm is anenergy-dependent process and requires thepresence of one or more specific transport sys-tems (6). Although it is not completely clearhow energy is coupled to the uptake process,several workers using different uptake systemshave postulated the existence of three distinctmechanisms (15, 38). The group translocationsystem is the best understood of these threemechanisms but is applicable only to the up-take of a limited number of sugars (15, 38). Theremaining two mechanisms can be distin-guished by such parameters as sensitivity touncouplers of oxidative phosphorylation and in-hibitors of electron transport or glycolysis, bythe effects of loss, through mutation, of cyto-chrome or adenosine triphosphatase (ATPase)activities, by the direct utilization of ferment-able or oxidizable carbon sources to drive activeuptake, and by ability to accumulate the trans-port substrate after anaerobic growth (4, 14, 23,33). On the basis of results obtained in experi-ments of this type, several workers postulatedthat the two energy coupling processes are (i)utilization of an energized membrane state (aproton or charge gradient across the cytoplas-mic membrane) generated either by electron

transport during respiration or by hydrolysis ofadenosine 5'-triphosphate (ATP) by the ATP-ase under anaerobic conditions, and (ii) directutilization of a high-energy phosphate bond(probably ATP) (4, 14, 23).

It has been shown that uptake systems thatutilize phosphate bond energy are highly sensi-tive to osmotic shock, and this sensitivity isassociated with the loss of specific binding pro-teins (4, 23, 35). These proteins are located inthe periplasmic space, the region between theinner and outer membranes of E. coli, and areprobably associated with the outer surface ofthe inner membrane (7). These proteins alsoappear to be involved in chemotactic responsesto certain nutrients (21).We previously reported the results of a series

of experiments on the uptake by E. coli K-12 ofiron complexed with the siderophore (iron che-lator) enterochelin, a cyclic trimer of 2,3-dihy-droxybenzoyl serine (30). Enterochelin (alsocalled enterobactin) is secreted by E. coli (andby other Enterobacteriaceae) under conditionsof iron stress and serves to solubilize iron thatwould otherwise be unavailable to the cells(36). Ferrienterochelin (the ferric ion entero-chelin complex) is transported into the cell viaspecific outer and inner membrane transportsystems and is therefore similar to the uptakeof vitamin B,2 and maltose, both of which alsohave specific outer membrane uptake systems(39, 41). The uptake of ferrienterochelin is in-hibited in a number of mutants of E. coli thatlack outer membrane ferrienterochelin receptoractivity (cbr mutants) (32; unpublished data) orferrienterochelin esterase activity (fes mu-

26

on July 6, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

http://jb.asm.org/

FERRIENTEROCHELIN UPTAKE 27

tants) (24). The nature of the block in ferrien-terochelin uptake in other mutants ofE. coli isless well understood, but some may involvestages in transfer across the outer or innermembranes (fep [8], feuA and feuB [16, 20],exbB and exbC [30], and tonB [12, 30]). Many ofthese mutants are also resistant to a number ofcolicins (20, 30), but the mechanisms throughwhich the colicin action and ferrienterochelinuptake systems interact are not fully under-stood (31).To learn more of the stages involved in the

uptake of ferrienterochelin we have examinedthe energy dependence of the active uptakestage by using some of the parameters de-scribed above. The results presented in thispaper indicate that the uptake of ferrientero-chelin is strongly dependent upon an energizedmembrane state. However, the results also in-dicate that energy for ferrienterochelin uptakeis dependent upon a supply of phosphate bondenergy such as ATP. The implications of theseresults are discussed.

MATERIALS AND METHODSBacterial strains and culture media. Strains ofE.

coli K-12 used in this study are listed in Table 1.Mutants defective in ATPase activity (unc-405 anduncA) were cloned at regular intervals and checkedfor inability to grow with succinate as the sole car-bon source. Phosphate minimal salts medium (me-dium A) was prepared as described previously andwas supplemented with amino acids and growthfactors (Table 1) and with a trace salts solution atthe concentrations used previously (30). Tris(hy-droxymethyl)aminomethane minimal salts solution(medium B) contained (per liter): tris(hydroxy-

methyl)aminomethane base, 12.1 g; NaCl, 4.5 g;KCI, 1.5 g; Ca(NO3)2, 0.01 g; MgSO4-5H2O, 0.2 g;and (NH4)2SO4, 2.0 g (pH 7.3 with HCI). Iron-de-pleted salts solutions and buffer solutions wereprepared by extracting the iron with 8-hydroxy-quinoline (30), and sterile glucose was added to thesterilized growth and uptake media at a final con-centration of 5 g/liter unless stated otherwise.

Chemicals. [3Hlproline (677 mCi/mmol), [14C]-glutamine (57.3 mCi/mmol), [3H]arginine (22 mCi/mmol), and 55FeCI3 (11.4 mCi/mmol) were pur-chased from the Amersham/Searle Corp. and werediluted with appropriate carrier before use. Entero-chelin was as used previously (30), and spectinomy-cin was kindly provided by G. Whitfield of UpjohnLtd. All other chemicals were obtained commer-cially and were of the highest purity available.

Protein determinations. Protein concentrationswere determined by the method of Schacterle andPollack (37) by using bovine serum albumin as astandard.Uptake studies. The method used to study entero-

chelin-mediated iron uptake under aerobic condi-tions was as follows unless stated otherwise. Cellswere grown overnight in minimal medium A con-taining an additional 5 ,uM FeCl3. The cells wereharvested by centrifugation and washed twice andsuspended in iron-depleted medium A to an initialoptical density at 660 nm (OD660) of 0.01. The cellswere then grown with good aeration until an OD.0of 0.2 was reached, at which stage the cells werechilled to 4°C and harvested by centrifugation. Thecelis were washed twice and resuspended in freshiron-depleted medium A containing 100 ,uM nitrilo-triacetic acid (NTA) and 0.5% glucose (uptake me-dium; final OD60, 0.8). Cells were stored at 4°C untiluse and were always used within 2 h of harvesting.In studies on the effect of arsenate on uptake, theiron-starved cells were washed and resuspended iniron-depleted medium B containing 100 ,uM NTA,

TABLE 1. E. coli K-12 strains

Relevant genotypea

ilvC7 argH entAilvC7 argH entA unc-405PrototrophicuncA derivative of strain 7his proBuncA derivative of A428aroE spcb derivative of 7aroE spcb derivative of NR70aroE spCb derivative of A428aroE spcb derivative of N144thr leu his proA argE supE aroE spc str thiAs P1552 but exbC51As P1552 but exbB71, metCpabA his arg ilv purE aroE str spcaroE thicbr-7 derivative of AT2472

Source (reference)

F. GibsonF. GibsonR. KadnerR. KadnerR. KadnerR. Kadner

G. C. WoodrowB. Stillman

Strain desig-nationAN248AN2497NR70A428N144P1571P1572P1683P1691P1552P1554P1555AN366AT2472P1798

(9)(9)

(23)(23)(23)(23)

(30)(30)(30)

a Terminology as in Bachmann et al. (2) and in quoted references.b aroE spc strains derived by cotransduction of these markers from strain AN366 with selection for

spectinomycin resistance as described previously (30).

VOL. 130, 1977


.org/D

ownloaded from

http://jb.asm.org/

28 PUGSLEY AND REEVES

0.5% glucose, and either 10 mM sodium arsenate or10 mM sodium phosphate buffer, both at pH 7.3.Uptake was initiated by the addition of the u5Fe(m)-labeled ferrienterochelin [final concentrations, 6,uM 55Fe(m) and 12 ,uM desferrienterochelin] to thecells in. plastic vials which were shaken in an oscil-lating water bath (200 oscillations per min) at 37°C.Samples were removed from the uptake mixture atappropriate time intervals and filtered through aGelman GA-6 membrane filter (0.45 ,um mean porediameter) that had been presoaked in a solution offerric-ethylenediaminetetraacetic acid (24). The fil-ters were rinsed twice with 10 ml of saline and driedand placed in scintillation vials. The filters werecleared by the addition of toluene containing 0.6%butyl-1,3,4-phenylbiphenyloxidiazole (Sigma Chem-ical Co.), and the radioactivity retained by the filterwas counted in a Packard model 3003 scintillationcounter.Amino acid uptake was determined under the

same conditions as ferrienterochelin uptake (i.e., iniron-depleted medium containing 100 ,uM NTA) us-ing the same iron-starved cells except where notedin the text. The final concentration of labeled aminoacid in the uptake mixture was 500 juM. Chloram-phenicol (100 ,ig/ml) was used to inhibit proteinsynthesis during amino acid uptake.

Counts ofradioactivity retained by the membranefilters were corrected for nonspecific retention ofcell-free radioactivity, and the results are expressedas weight of transport substrate per milligram ofcell dry weight.

Inhibitors, where used, were added to the washedcells which were then incubated for 20 min at 37°C.Inhibitors were dissolved in uptake medium or inethanol; control experiments demonstrated that theaddition of ethanol to the cell suspensions did notaffect uptake. In the case of arsenate-treated cells(see above), cells were again incubated at 37°C for 20min before uptake was measured.

Treatment with 2,4-dinitrophenol (DNP) to de-plete cellular energy was performed by the methodof Berger (3) with some modifications. Cells weregrown first in iron-depleted medium A containing 1g of glucose per liter (final OD660, 0.25 to 0.30) andthen washed twice and suspended in unsupple-mented medium A (with no energy source) plus 200,uM NTA and 2 mM DNP. The cells were incubatedwith good aeration at 370C for 10 h (AN248) or 2 h(AN249). After starvation, the cells were harvestedand washed twice and resuspended in uptake me-dium without glucose. Where indicated, energysources were added to the cells which were thenincubated for 10 min at 37°C before uptake wasinitiated.

In studies on uptake after anaerobic growth, cellswere grown in iron-depleted medium A containing100 ,IM NTA for 16 h at 370C in anaerobic jars filledwith hydrogen-carbon dioxide gas mixture gener-ated by a GasPak unit (Becton-Dickinson & Co.;final ODwo, 0.15 to 0.20.). Cells were washed twiceand resuspended as above except that no glucosewas added to the medium. The cells were thenpurged continually with nitrogen gas that had beenpassed through two towers ofpyrogallol (5% [wt/vol]

in 14 M KOH), and the glucose was added. Radioac-tive transport substrates were similarly purged for10 min before being added to the cells.The effect of osmotic shock on uptake was studied

by using cells grown under the normal conditions ofiron deprivation for 16 h and subjected to the osmoticshock procedure of Willis et al. (42). All reagentsexcept the MgCl2 solution were extracted with 8-hydroxyquinoline (see above) and contained 200 jIMNTA to prevent iron uptake through the low-affinityuptake system during shock treatment (30).Shocked cells were washed and resuspended in up-take medium and assayed for uptake immediately orafter incubation at 37°C for varying lengths of timein the presence or absence of 100 jig ofchloramphen-icol per ml.

Preparation of periplasmic fluid and membranefractions. Periplasmic fluid was prepared from 5-liter cultures of cells of strain AT2472 grown underconditions of iron stress (in iron-depleted medium A)or in the presence of 20 jiM FeCl3 by the proceduredescribed above. The shock fluid obtained was con-centrated by dialysis against Aquacide II (Calbi-ochem) and then dialyzed extensively against 50mM sodium phosphate buffer (pH 7.3), both at 4°C.Debris (largely lipopolysaccharide) was removed bycentrifugation at 46,000 x g for 60 min at 4°C, andthe supernatant was stored at -20°C.

Outer and inner membranes were prepared fromiron-stressed or iron-supplemented cells (20 jiMFeCl3) in the exponential stage of growth. The cellswere disrupted in a French pressure cell (Aminco)and the cell envelope was pelleted, after removal ofcell debris, by centrifugation at 46,000 x g for 60 minat 4°C. The inner membrane was solubilized with2% Triton X-100 as described previously (30), andthe soluble fraction was precipitated in 70% ethanolat -20°C. The ethanol-precipitated material waspelleted by centrifugation at 46,000 x g for 60 min at4°C and washed twice and resuspended in distilledwater. Outer membrane fractions (Triton-insolublewalls) were prepared as described previously (30),and both membrane fractions were stored at -20°C.

Binding assays. Ferrienterochelin binding assayswere performed using dialysis chambers similar indesign to those used by Furlong and Weiner (13). Allreagents were dissolved or resuspended in a solutioncontaining 50 mM NaCl, 10 mM sodium phosphatebuffer, and 100 ,uM NTA (pH 7.0). Samples contain-ing putative binding proteins and the 55Fe(III)-la-beled ferrienterochelin were separated by a boileddialysis membrane (Visking) and allowed to dialyzeto equilibrium (24 h) at 4°C on a vertical rotatingturntable. Samples were removed from both cham-bers after dialysis and emulsified in a scintillantcocktail containing two parts toluene-butyl-1,3,4-phenylbiphenyloxidiazole (see above) and 1 part Tri-ton X-100 (28) and counted as above. Concentrationsof protein and ferrie4terochelin were adjusted togive 10 to 30% binding of available radioactivity,and bovine serum albumin was used as control.PAGE. Polyacrylamide gel electrophoresis

(PAGE) membrane fractions and periplasmic fluidwere analyzed in the presence of sodium dodecylsulfate by PAGE by the technique of Lugtenberg

J. BACTERIOL.


.org/D

ownloaded from

http://jb.asm.org/


et al. (26) in slab gels. Gels were stained by themethod of Fairbanks, Steck, and Wallach (10).

RESULTSExperimental approach. One ofthe purposes

of this study was to determine the nature of theenergy coupling for the active uptake offerrien-terochelin. Uptake of ferrienterochelin wastherefore compared with known examples oftransport substrates that use either the ener-gized membrane state (proline) (4) or phosphatebond energy (glutamine and arginine (4) foruptake, under the conditions of iron stresswhich are necessary for detectable iron require-ment and for ferrienterochelin uptake. Controlexperiments demonstrated that iron starvationhad no direct effect on uptake of proline orglutamine under aerobic conditions, and re-sults obtained in different experiments maytherefore be compared (having regard for minorfluctuations in uptake rates in different experi-ments). The uptake offerrienterochelin is, how-ever, strongly dependent upon the level of ironstarvation achieved during growth ofthe bacte-ria before assay; hence results obtained in dif-ferent experiments should not be compared. Instudies in which the uptake of ferrienterochelinin different strains of E. coli was compared,care was taken to ensure that the cells weregrown under identical conditions. All experi-ments were performed at least twice.

Effect of loss of ATPase activity on ferrien-terochelin uptake. Previous studies haveshown that uptake of a transport substrate maybe reduced in some mutants that are defectivein ATPase functions (23). The results shown inFig. 1 indicate that this is not the case withferrienterochelin uptake by three mutantslacking ATPase activity. The differences notedin Fig. 1 between ferrienterochelin uptake instrain P1683 and in the isogenic uncA strainP1691 are probably not significant when possi-ble differences in the level ofiron starvation aretaken into consideration. Furthermore, treat-ment of strain AN248 (unc+) with N,N'-dicy-clohexylcarbodiimide, which binds to a specificprotein component of the ATPase complex (1)and thereby inhibits ATPase function, did notaffect ferrienterochelin uptake.

Effects of inhibitors of electron transportand uncouplers of oxidative phosphorylation.Inhibitors of electron transport (potassium cya-nide) block the production of the energizedmembrane state by interfering with respira-tion. Uncouplers of oxidative phosphorylation(DNP and carbonyl cyanide m-chlorophenylhy-drazone [CCCP]) produce a "leaky" membranewhich cannot maintain an energized state. As

shown in Fig. 2, these reagents preferentiallyinhibit proline uptake, which is dependentupon the energized membrane state, and theireffect was more pronounced in strain AN249(unc-405) in which the defective ATPase couldnot be used to energize the membrane by ATPhydrolysis. The effects of DNP and cyanide onglutamine and proline uptake were clearly de-pendent on the dose of inhibitor used (Table 2),but we were unable to demonstrate preferentialinhibition of proline uptake by the uncouplerCCP (Table 2). This effect has been noted previ-ously (14) and is thought to be due to the abilityof CCP to form thiol groups (22). The uptake ofproline was also strongly inhibited by 10 mMsodium azide, which inhibits oxidative phos-phorylation and cytochrome oxidase, whereasglutamine uptake was largely unaffected (Fig.2). The results in Fig. 2 and in Table 2 indicatethat the active uptake of ferrienterochelin wasalso highly sensitive to DNP, cyanide, andazide, which suggests that this uptake systemis strongly dependent upon the presence of anenergized membrane.

Effects of inhibitors of ATP production. In-hibitors of glycolysis (iodoacetic acid [IAA] andsodium fluoride) block the production of ATPand thus reduce the level of cellular ATP avail-able to drive active uptake. Results in Fig. 3show that uptake of glutamine was stronglyinhibited by 200 uM IAA, whereas proline and

L-

3 50

J 40Lu0

30

4=0 20o

101

00 1 2 3 4 5 6

TIME (min]7

FIG. 1. Effect of defective ATPase activity on fer-rienterochelin uptake in E. coli. Symbols: A, AN248(unc+); A, AN249 (unc-405); *, p1571 (unc+); Ol,P1572 (uncA); *, P1683 (unc+); O, P1691 (uncA) .

A

A/ z

VOL. 130, 1977


.org/D

ownloaded from

http://jb.asm.org/


4 5 6 0 1 2 3 4 5 6TIME (min]

FIG. 2. Effect of uncouplers of oxidative phospho-rylation and inhibitors of electron transport on up-take of glutamine, proline, and ferrienterochelin instrains AN248 (unc+; A, C, and E) and AN249 (unc-405, B, D, and F). Symbols: *, control; *, 500 1MDNP; +, 4 mM KCN; A, 10 mM NaN3.

ferrienterochelin uptakes were less affected.However, sensitivity of ferrienterochelin up-take to 200 ,AM IAA was strongly increased inthe absence of a functional ATPase (strainAN249), which suggests that ATP generated bythe ATPase and by glycolysis may play a role inthe energy-dependent uptake of this compound(Fig. 3F). We were unable to differentiate be-tween the energy coupling involved in the up-take of proline and glutamine with 50 mMsodium fluoride (Fig. 3), but we were able toselectively inhibit glutamine uptake in cellstreated with 10 mM sodium arsenate (Fig. 4),which blocks ATP production by competingwith phosphate (4). Ferrienterochelin uptakewas also inhibited in cells treated with sodiumarsenate in the absence of phosphate (Fig. 4).Ferrienterochelin uptake was completely re-

stored by the addition of 100 mM phosphatebuffer to the arsenate-treated cells, thus rulingout the possibility that iron had entered thecells in association with the arsenate duringpreincubation. The results therefore suggestthat ferrienterochelin uptake is dependentupon the presence of cellular ATP or an analo-gous source of phosphate bond energy.Combined effects of inhibitors of glycolysis

and inhibitors of electron transport or uncou-plers of oxidative phosphorylation. To deter-mine whether the effects of DNP and IAA wereadditive, we pretreated cells of strain AN249(unc-405) with a 100 AM concentration of eachor both reagents. As shown in Fig. 5, thesedoses of IAA or DNP individually have only alimited effect on uptake of ferrienterochelin,but the combined effect of the two inhibitorswas to strongly inhibit ferrienterochelin up-take. Similar results were obtained with com-binations of 10 mM sodium arsenate and 100,uM DNP or 1 mM potassium cyanide.Respiration-dependent uptake. Treatment

of cells with high doses of DNP over long pe-riods of time in the absence of energy sourcescompletely reduces cellular ATP levels (3). Up-take of transport substrates can thus be coupleddirectly with the utilization of exogenouslyadded energy sources. As shown in Fig. 6, up-takes of glutamine and proline can both bedriven quite effectively by exogenous glucoseadded to iron-starved, DNP-treated cells ofstrain AN249 (unc-405). Similar results wereobtained with strain AN248 (unc+). Glucose-driven uptake of ferrienterochelin could, how-ever, only be demonstrated in cells of strainAN249 (unc-405) after DNP treatment. Fur-thermore, glucose-driven proline uptake wassomewhat reduced in iron-starved, DNP-treated cells of both strains tested (cf. Fig. 6with Fig. 2). We also noted that proline uptakecould not be effectively coupled with lactaterespiration in iron-starved cells of either straintested (Fig. 6), even though proline uptake isrespiration dependent and is normally highlyefficient. However, higher rates of proline up-take were obtained with DNP-treated cells ofstrain AN249 (unc-405) which were not sub-jected to the iron-starvation regimen.

Effect of anaerobiosis on uptake. Growth ofE. coli under anaerobic conditions in the ab-sence of an electron acceptor eliminates elec-tron transport and inhibits respiration-depend-ent uptake of transport substrates (14). Controlexperiments demonstrated that glutamine up-take in strain AN248 (unc+) was unaffected byanaerobic growth under iron stress, whereasproline uptake was severely inhibited (20 to.30% of control aerobic uptake). We were unableto demonstrate any active uptake of ferrienter-ochelin after anaerobic growth.

Effects of sodium arsenate and potassiumcyanide on mutants partially defective in fer-rienterochelin uptake. We previously de-scribed two classes of mutants (exbB and exbC)that are partially defective in ferrienterochelinuptake (30). We reexamined ferrienterochelinuptake in these mutants following treatment

J. BACTERIOL.


.org/D

ownloaded from

http://jb.asm.org/


TABLE 2. Effect ofDNP, CCCP, and KCN on active uptake ofglutamine, proline and ferrienterochelin instrains AN248 (unc+) and AN249 (unc-405)a

Inhibition of active uptake (%)

Inhibitor and concen- n i 5'Fe(III)-labeled ferrientero-tration ['4C]glutamine [3H]proline chelin

AN248 AN249 AN248 AN249 AN248 AN249

DNP100 PUM 6 0 25 12 7 31200 ,uM 15 40 52 57 33 77500 ,uM 33 62 76 90 66 80

1,000 ,uM 64 74 89 97 88 851,500 ,M 86 86 96 100 96 982,000,M 100 100 100 100 100 100

CCCP10 ,uM 11 14 13 7 4 1120 ,uM 51 56 62 72 72 6830 uM 85 93 95 96 93 9350 uM 90 92 100 100 98 95

KCN2 mM 0 26 38 94 42 924 mM 3 51 40 97 44 896 mM 23 62 55 98 66 938 mM 62 75 53 97 69 95

10 mM 74 78 68 98 74 9320 mM 72 100 98a Amino acid uptake rates were determined from three samples taken within the first 3 min of uptake.

Ferrienterochelin uptake rates [55Fe(III)] were determined from three samples taken between min 1 and 4 ofuptake to avoid problems caused by initial binding of the ferrienterochelin to the outer membrane receptorprotein. KCN, Potassium cyanide.

with 10 mM sodium arsenate or 4 mM potas-sium cyanide. The results in Fig. 7 indicatethat uptake of ferrienterochelin in the parentstrain (P1552) was inhibited after treatmentwith arsenate or cyanide. Uptake of ferrienter-ochelin was also inhibited in strains P1554(exbC) and P1555 (exbB) after treatment witharsenate or cyanide. We conclude that thesemutants are not defective in ferrienterochelinuptake through inability to couple either phos-phate bond energy or the energized membranestate with the active uptake process.

Effect of osmotic shock on ferrientero-chelin uptake. Sensitivity to osmotic shock isthought to indicate dependence upon a peri-plasmic binding protein and is usually corre-

lated with utilization of phosphate bond energyfor active uptake (4). Results in Fig. 8 showthat arginine, proline, and ferrienterochelinuptake are all strongly inhibited when cells are

assayed immediately after osmotic shock. Pro-line uptake, which involves no periplasmicbinding protein (4), was almost completely re-

stored after 30 min of incubation in the pres-ence of chloramphenicol (Fig. 8B), whereas ar-

ginine and ferrienterochelin uptake were not.Uptake of arginine could, however, be fully

restored in shocked cells incubated for 2 h in theabsence ofchloramphenicol, under which condi-tions fresh periplasmic binding protein(s) couldbe synthesized. Uptake of ferrienterochelincould be partially restored by incubation in theabsence of chloramphenicol after osmoticshock. This restoration was largely achievedthrough an increase in apparent uptake duringthe first minute (approaching normal after 2 hof incubation) rather than an increase in therate of active uptake of ferrienterochelin (nobetter than 50% recovery after 2 h of incubationcompared with over 80% in the case of arginineuptake).

Material released by osmotic shock fromiron-starved or iron-supplemented cells ofstrain AT2472 or P1798 (cbr) was examined forferrienterochelin binding activity. As shown inTable 3, the level of ferrienterochelin bindingactivity in the periplasmic material from iron-starved cells of strain AT2472 was higher thanthat released from the corresponding iron-sup-plement cells. An increase in ferrienterochelinbinding activity was also detected in the outermembranes prepared from iron-starved cells ofthis strain. However, material released by os-motic shock from iron-starved cells of strain

VOL. 130, 1977


.org/D

ownloaded from

http://jb.asm.org/

6 A B complex than the systems involved in the up-Uls - take of other transport substrates such asz 4 amino acids. Apart from the involvement of

-,4X/ /specific systems for transporting the complex33 // ./ across the outer membrane (19, 20, 30, 40),

2 there also appear to be several stages involved0 ,1 / / in transfer across the inner membrane. The

1. , t-Z_2--t *_.. . data presented in this paper indicate that the

> 6: energizedmembrane state isnecessary forthis

6 C D latter, active uptake stage, as demonstrated byw 5z the effects of azide, and particularly of cyanide0 o4 S/ ~ ./ and DNP in an unc- mutant (Fig. 2; Table 2).

2X3-/ Arsenate and IAA were also shown to haveEn:2 $/-*-|- */"<>A- some effect on the uptake of ferrienterochelin

w - (Fig. 3 and 4) although the effects were not as

i!o; . tt pronounced as those of the same reagents upon70 X e | | | §glutamine uptake. The incomplete effects of70 E F IAA and arsenate on ferrienterochelin uptake

6 compared with glutamine uptake may be due to50 / - the fact that residual levels of cellular ATP,*cs./ which remain after treatment with these re-

. / agents, are sufficient for the relatively lowLO . level (in molar terms) of ferrienterochelin up-

'10 ///F//,A ,...0 I. w5A B

0 1 2 3 4 5 6 0 1 2 3 4 5 6z 4

TIME [min]

FIG. 3. Effect of inhibitors ofglycolysis and ATP- 3a 2ase function on uptake of glutamine, proline, and IU 1 * ...- *ferrienterochelin in strains AN248 (unc+; A, C, and .- *E) and AN249 (unc-405; B, D, and F). Symbols: *, TZ 0 _control;*, 50 mMNaF; 0,200 pMIAA; A, 200-M20 DIAA plus 100 MM N,N'-dicyclohexylcarbodiimide.

P1798, which lacks outer membrane receptor a 15Xactivity for ferrienterochelin (Table 3) and for , z/colicins B and D (32), did not contain ferrienter- , 01.oochelin binding activity (Table 3). E X /

arations and the membrane preparations from w 2 / I

strain AT2472 were examined by PAGE in the /presence of sodium dodecyl sulfate by the ,method of Lugtenberg et al. (26). Outer mem- 30 E Fbranes prepared from iron-starved cells and ex-amined by this method showed strong derepres- --20 in..sion of three outer membrane proteins with / .molecular weights ranging from 76,000 to I85,000, as judged by their positions in the cells L, gcompared with proteins of known molecular / /weights (16, 32; A. P. Pugsley and P. Reeves, 0 ._._._Biochem. Biophys. Res. Commun., in press). 0 1 2 3 4 5 6 0 1 2 3 4 5 6No similar derepression of any proteins was TIME [min]detected in either the periplasmic or inner FIG. 4. Effect of10 mM sodium arsenate on gluta-membrane preparations. mine, proline, and ferrienterochelin uptake in strains

AN248 (unc+; A, C, and E) and AN249 (unc-405; B,

DISCUSSION D, and F). Symbols: *, cells incubated in medium BUSSIOMniInLn mMenI&rMinhOnJnhLItIp-AW] inPU_The uptake of iron complexed with the sider-

ophore enterochelin by E. coli is clearly more

bated in medium B containing 10 mM sodium arse-nate.

J. BACTERIOL.32 PUGSLEY AND REEVES


.org/D

ownloaded from

http://jb.asm.org/


70>' 60

_j50w40U

20 /1

0 1 2 3 4 5 6TIME [min]

FIG. 5. Combined effects ofIAA and DNP on up-take offerrienterochelin in strain AN249 (unc-405).Symbols *, control; *, 100 pM IAA; 0, 100 pMDNP; A, 100 pM IAA plus 100 pM DNP.

take. An absolute or partial requirement forcellular ATP or a similar source of phosphatebond energy for ferrienterochelin uptake maytherefore be indicated. In support of this, wehave consistently shown that ferrienterochelinuptake is more sensitive to treatment with ar-senate than is proline uptake, which is depend-ent solely upon an energized membrane state,irrespective of the strain used. On the otherhand, it is also possible that the effects of IAAand arsenate on glucose uptake may also resultin a reduction of energy available to deriveother uptake systems.

If two distinct energy coupling mechanismsare indeed involved in the uptake of ferrienter-ochelin, they do not appear to interact in theway demonstrated previously for phosphate up-take, for which there are two parallel uptakesystems that are genetically and functionallyindependent (33). In the case of ferrientero-chelin, the evidence suggests that both cou-pling mechanisms are required for a single up-take system. The coupling of energy with theuptake system may be further complicated bythe involvement of a requirement for reductionof the metal ion from the ferric to the ferrousform during uptake. This possibility arises (i)by analogy with the postulate of Leong andNeilands (25) that iron in complex with thesiderophore ferrichrome must be reduced to theferrous form before it is transported across thecytoplasmic membrane of E. coli, and (ii) byanalogy with the uptake of ferric sucrose by ratliver mitochondria where ferric irons are re-duced by reducing equivalents supplied by themitochondrial cytochrome system during up-take (11).

No uptake of ferrienterochelin was observedin cells grown anaerobically. One might predictthat a lower level of iron would be required byanaerobically grown E. coli since one of themajor sites of incorporation of iron is the cyto-chrome system, which would not be operating

_

3:

J-

z

.j

0cm

E

w

4a.

/0.

* * * *

0 1 2 3 4 5 6TIME [min]

FIG. 6. Uptake ofglutamine, proline, and ferrien-terochelin in iron-starved cells ofstrain AN249 (unc-405) following treatment with 2 mM DNP for 2 h inthe absence of energy source. Symbols: *, uptake inthe absence of exogenous energy; U, uptake after 10min of incubation with 20 mM DL-lactate; 0, uptakeafter 10 min of incubation with 20 mM glucose.

VOL. 130, 1977


.org/D

ownloaded from

http://jb.asm.org/


-A B

/ -- P b -~~~~~, ,K* D /

.U .- H-,_

IF

U.-

-U--a

aI

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7TIME [min]

FIG. 7. Uptake of ferrienterochelin in strainsP1552, P1554 (exbC51), and P1555 (exbB71) incu-bated for 20 min in the presence of 10 mM sodiumarsenate (A, C, and E) or 4 mMKCN (B, D, and F).Symbols: *, controls; 0, cells treated with inhibi-tors.

under these conditions. Furthermore, sinceiron is soluble in the ferrous form under anaero-

bic conditions, siderophore-dependent systemsfor the uptake of iron would not normally benecessary, and siderophore uptake systemsmay be repressed by anaerobiosis. It has beenshown previously that enterochelin is not hy-persecreted by mutants defective in entero-chelin uptake when the cells are grown anaero-

bically (31). Cells incubated aerobically in thepresence of DNP may also have a reduced re-

quirement for iron, particularly after long pe-

riods of incubation.Some of the data presented in this paper may

also be useful in determining the role of thetonB gene product in iron uptake. The tonBproduct is required for the action of the group Bcolicins (31), for irreversible adsorption of bac-teriophages Ti and 480 and for vitamin B12uptake (C. A. Schnaitman, personal com-

munication). It has recently been shown thatirreversible adsorption (deoxyribonucleic acidinjection) ofbacteriophages Ti and 480 requiresan energized membrane state (17), and thetonB gene product may be involved in coupling

this energization with some stage of Ti or 080infection. If this were so, one might predict thatgroup B colicin action and other chelator-mediated iron uptake systems, as well as fer-rienterochelin uptake (this paper) and vitaminB12 uptake (5), would also require a functionalelectron transport system or energized mem-brane state and would be coupled to it by thetonB gene product. At this stage it is still notpossible to say whether the tonB gene productis located in the inner or outer membranes ofE. coli (12).The sensitivity of ferrienterochelin uptake to

osmotic shock seems to be due to the release of

w

zz

I

CII

01

I

3

-J-J

0)

w

4

6

5

4

32

0

7

6

w

z

OJ 4

I 2

0)

a1

0

140

120

10080

li,60

In 40

0

0 2 4 6 8 10TIME [min]

FIG. 8. Uptake of arginine, proline, and ferrien-terochelin in cells of strain AT2472 before and afterosmotic shock. Cells were tested immediately beforeosmotic shock (-), immediately after osmotie shock(0) or after 30 min of incubation in iron-depletedmedium A containing 100 Mg ofchloramphenicol per

ml, 0.5 mg ofglucose per ml, and 100 MNTA (-).

80

60

40

Io 203:c o

'_J 30w

E 201-

,n 10

30

201

10

0~

A U

/// o:+U__~~~

//~~~B , . --

I

.U

A/=t~, . . . . .

J. BACTERIOL.

4-


.org/D

ownloaded from

http://jb.asm.org/


TABLE 3. Ferrienterochelin binding activity ofoutermembranes and periplasmic proteins prepared fromiron-starved cells ofstrains AT2472 and P1 798 (cbr)

ng of mFe(III)-labeled ferrienterochelin/mgof proteina

Cells Outer membrane Periplasmic pro-teins

AT2472 P1798 AT2472 P1798Iron-starved 15.14 0.44 3.32 0.36Iron-supple- 0.17 0.18 0.37 0.37menteda Ferrienterochelin binding to bovine serum al-

bumin occurred at a rate of 0.23 ng of 55Fe(III)-labeled ferrienterochelin/mg of protein. The ferrien-terochelin reagent contained 70 nM 55Fe(III) and140 nM of desferrienterochelin. Protein concentra-tions ranged from 1.0 to 0.02 mg/ml. Each result isthe mean of several independent experiments.

the outer membrane ferrienterochelin receptorby the shock treatment rather than to the lossof a periplasmic binding protein. The evidencefor this is that the initial rate of ferrientero-chelin uptake, which is considered to be largelybinding to the outer membrane receptor (30), isseverely reduced by osmotic shock (Fig. 8), andthat the ferrienterochelin receptor mutant(P1798, cbr-7) did not release ferrienterochelinbinding activity after osmotic shock. We couldnot, however, detect appreciable amounts ofany protein with a similar electrophoretic mo-bility to the outer membrane ferrienterochelin-colicin B receptor protein in any of the osmoticshock fluids examined by PAGE. These obser-vations raise the possibility that other uptakesystems, which are thought to have both outermembrane and periplasmic binding proteins, inreality have only the outer membrane receptor.This possibility has already been considered inthe case of the uptake of vitamin B,2 (5, 41).

LITERATURE CITED1. Altendorf, K., and W. Zitzmann. 1975. Identification of

the DCCD-reactive protein of the energy transducingadenosinetriphosphatase complex from Escherichiacoli. FEBS Lett. 59:268-272.

2. Bachmann, B. J., K. B. Low, and A. L. Taylor. 1976.Recalibrated linkage map of Escherichia coli K-12.Bacteriol. Rev. 40:116-167.

3. Berger, E. A. 1973. Different mechanisms for energycoupling to the active transport of proline and gluta-mine in Escherichia coli. Proc. Natl. Acad. Sci.U.S.A. 70:1514-1518.

4. Berger, E. A., and L. A. Heppel. 1974. Different mecha-nisms for energy coupling to the shock-sensitive andshock-resistant amino acid permeases of Escherichiacoli. J. Biol. Chem. 249:7747-7755.

5. Bradbeer, C., and M. L. Woodrow. 1976. Transport ofvitamin B,, by Escherichia coli: energy dependence.J. Bacteriol. 128:99-104.

6. Boos, W. 1974. Bacterial transport. Annu. Rev. Bio-chem. 43:123-146.

7. Costerton, J. W., J. M. Ingram, and K.-J. Cheng. 1974.Structure and function of the cell envelope of gram-negative bacteria. Bacteriol. Rev. 38:87-110.

8. Cox, G. B., F. Gibson, K. J. Luke, N. A. Newton, I. G.O'Brien, and H. Rosenberg. 1970. Mutations affect-ing iron transport in Escherichia coli. J. Bacteriol.104:219-226.

9. Cox, G. B., F. Gibson, and L. McCann. 1974. Oxidativephosphorylation in Escherichia coli K12. An uncou-pled mutant with altered membrane structure. Bio-chem. J. 138:211-215.

10. Fairbanks, G., T. L. Steck, and D. F. H. Wallach. 1971.Electrophoretic analysis of the major polypeptides ofthe human erythrocyte membrane. Biochemistry10:2606-2617.

11. Flatmark, T., and I. Romslo. 1975. Energy-dependentaccumulation of iron by isolated rat liver mitochon-dria. J. Biol. Chem. 250:64334438.

12. Frost, G. E., and H. Rosenberg. 1975. Relationshipbetween the tonB locus and iron transport in Esche-richia coli. J. Bacteriol. 124:704-712.

13. Furlong, C. E., and J. H. Weiner. 1970. Purification ofa leucine-specific binding protein from Escherichiacoli. Biochem. Biophys. Res. Commun. 38:1076-1083.

14. Gutowski, S. J., and H. Rosenberg. 1976. Energy cou-pling to active transport in anaerobically grown mu-tants of Escherichia coli K-12. Biochem. J. 154:731-734.

15. Hamilton, W. A. 1975. Energy coupling in microbialtransport, p. 1-53. In A. H. Rose and D. Tempest(ed.), Advances in microbial physiology, vol. 12. Aca-demic Press Inc., London.

16. Hancock, R. E. W., and V. Braun. 1976. The colicin Ireceptor ofEscherichia coli K-12 has a role in entero-chelin-mediated iron transport. FEBS Lett. 65:208-210.

17. Hancock, R. E. W., and V. Braun. 1976. Nature of theenergy requirement for the irreversible adsorption ofbacteriophages Ti and 480 to Escherichia coli. J.Bacteriol. 125:409-415.

18. Hancock, R. E. W., and P. Reeves. 1975. Bacteriophageresistance in Escherichia coli K-12: general pattern ofresistance. J. Bacteriol. 121:983-993.

19. Hantke, K., and V. Braun. 1975. Membrane receptordependent iron transport in Escherichia coli. FEBSLett. 49:301-305.

20. Hantke, K., and V. Braun. 1975. A function common toiron-enterochelin transport and the action of colicinsB, I, V in Escherichia coli. FEBS Lett. 59:277-281.

21. Hazelbauer, G. L. 1975. The binding of maltose to "vir-gin" maltose binding protein is biphasic. Eur. J. Bio-chem. 60:445-449.

22. Kaback, H. R., J. P. Reeves, S. A. Short, and F. J.Lombardi. 1974. Mechanisms of active transport inisolated bacterial membrane vesicles. XVIII. Themechanism of action of carbonylcyanide m-chloro-phenylhydrazone. Arch. Biochem. Biophys. 160:215-222.

23. Kadner, R. J., and H. H. Winkler. 1975. Energy cou-pling to methionine transport in Escherichia coli. J.Bacteriol. 123:985-991.

23. Langman, L., I. G. Young, G. E. Frost, H. Rosenberg,and F. Gibson. 1972. Enterochelin system of irontransport in Escherichia coli: mutations affecting fer-ric-enterochelin esterase. J. Bacteriol. 112:1142-1149.

25. Leong, J., and J. B. Neilands. 1976. Mechanisms ofsiderophore iron transport in enteric bacteria. J. Bac-teriol. 126:823-830.

26. Lugtenberg, B., J. Meijers, R. Peters, P. van der Hoekand L. Van Alphen. 1975. Electrophoretic resolutionof the 'major outer membrane protein' ofEscherichia

VOL. 130, 1977


.org/D

ownloaded from

http://jb.asm.org/


coli K12 into four bands. FEBS Lett. 58:254-258.

27. Nakae, T. 1976. Outer membrane of Salmonella. Isola-tion ofprotein complex that produces transmembranechannels. J. Biol. Chem. 251:2176-2178.

28. Patterson, M. S., and P. C. Greene. 1965. Measurementof low energy beta emitters in aqueous solutions byliquid scintillation counting of emulsions. Anal.Chem. 37:854-857.

29. Payne, J. W., and C. Gilvarg. 1968. Size restriction on

peptide utilisation byEscherichia coli. J. Biol. Chem.254:6291-6299.

30. Pusley, A. P., and P. Reeves. 1976. Iron uptake incolicin B-resistant mutants of Escherichia coli K-12.J. Bacteriol. 126:1052-1062.

31. Pugsley, A. P., and P. Reeves. 1976. Characterizationof group B colicin-resistant mutants of Escherichiacoli K-12: colicin resistance and the role of entero-chelin. J. Bacteriol. 127:218-228.

32. Pugsley, A. P., and P. Reeves. 1976. Increased produc-tion ofthe outer membrane receptors for colicins B, Dand M by Escherichia coli under iron starvation. Bio-chem. Biophys. Res. Commun. 70:846-853.

33. Rae, A. S., and K. P. Stickland. 1976. Studies on phos-phate transport in Eacherichia coli. II. Effect of meta-bolic inhibitors and divalent cations. Biochim. Bio-phys. Acta 433:564-582.

34. Rae, A. S., K. P. Stickland, N. Medveczky, and H.Rosenberg. 1976. Studies on phosphate transport inEscherichia coli. I. Re-examination of the effect ofosmotic and cold shock and some attempts to restoreuptake with phosphate binding protein. Biochim.Biophys. Acta 433:555-563.

35. Rosen, B. P. 1973. Basic amino acid transport inEsche-

richia coli. II. Purification and properties of an argi-nine-specific binding protein. J. Biol. Chem.248:1211-1218.

36. Rosenberg, H., and I. G. Young. 1974. Iron transport inthe enteric bacteria, p. 67-82. In J. B. Neilands (ed.),Microbial iron metabolism. Academic Press Inc.,New York.

37. Schacterle, G. E., and R. L. Pollack. 1973. A simplifiedmethod for quantitative assay of small amounts ofprotein in biologic material. Anal. Biochem. 51:664-665.

38. Simoni, R. D., and P. W. Postma. 1975. The energeticsof bacterial active transport. Annu. Rev. Biochem.44:523-554.

39. Szmelcman, S., M. Schwartz, T. J. Silhavy, and W.Boos. 1976. Maltose transport in Escherichia coliK12. A comparison of transport kinetics in wild typeand X-resistannt mutants with dissociation constantsof the maltose-binding protein as measured by fluo-rescence quenching. Eur. J. Biochem. 65:13-19.

40. Wayne, R., and J. B. Neilands. 1975. Evidence for com-mon binding sites for ferrichrome compounds andbacteriophage 4680 in the cell envelope ofEscherichiacoli. J. Bacteriol. 121:497-503.

41. White, J. C., P. M. DiGirolamo, M. L. Fu, Y. A. Pres-ton, and C. Bradbeer. 1973. Transport of vitamin B12in Escherichia coli. Location and properties of theinitial B12 binding site. J. Biol. Chem. 248:3978-3986.

42. Willis, R. C., R. G. Morris, C. Cirakogliu, G. D. Schel-lenberg, N. H. Gerber, and C. E. Furlong. 1974.Preparation of the periplasmic binding proteins fromSalmonella typhimurium and Escherichia coli. Arch.Biochem. Biophys. 161:64-75.

J. BACTERIOL.


.org/D

ownloaded from

http://jb.asm.org/

uptake ferrienterochelin escherichia coli: energy ... · energy-dependent process and requires the...

Documents