Ž .Journal of Microbiological Methods 46 2001 119–129www.elsevier.comrlocaterjmicmeth

Measurements of net fluxes and extracellular changes of Hq,Ca2q, Kq, and NHq in Escherichia coli using4

ion-selective microelectrodes

Lana Shabala a,), Tom Ross a, Ian Newman b, Tom McMeekin a, Sergey Shabala a

a School of Agricultural Science, UniÕersity of Tasmania, GPO Box 252-54, Hobart, Tas 7001, Australiab School of Mathematics and Physics, UniÕersity of Tasmania, GPO Box 252-54, Hobart, Tas 7001, Australia

Received 31 October 2000; received in revised form 21 February 2001; accepted 26 March 2001

Abstract

Ž .This study introduced the use of a non-invasive ion-selective microelectrode MIFE technique to study membrane-trans-port processes in bacteria. Net ion fluxes and changes in the extracellular concentrations of Hq, Ca2q, Kq and NHq in4

Ž .adherent bacteria, isolated from cultures at different growth stages exponential, late exponential, and stationary phases ,2q Ž .were monitored. With the exception of Ca , a significant Ps0.05 difference was found in the magnitude of net fluxes of

the ions measured from bacterial cells at different stages of the population growth curve. The magnitude of the Hq responsewas glucose-dependent with maximum changes occurring at the highest concentration. There was a progressive increase inHq extrusion followed by a gradual return to zero at late stationary phase. Measurements of net ion fluxes crossing thebacterial cytoplasmic membrane, demonstrated here for the first time, may offer insight into underlying mechanisms of iontransport kinetics. Applications of the non-invasive ion-selective microelectrode technique in microbiology are discussed.q 2001 Elsevier Science B.V. All rights reserved.

Keywords: Bacterial growth; Escherichia coli; Hydrogen; Ion fluxes; Microelectrode technique

1. Introduction

Bacterial growth and adaptive responses are cru-cially dependent on the cell’s ability to regulate itsnutrient uptake and to compartmentalize or extrudewastes of metabolism. Membrane transport systemsare central to both of these processes. To understandthe mechanisms involved, a spectrum of methods

) Corresponding author. Tel.: q61-03-6226-2620; fax: q61-03-6226-2722.

Ž .E-mail address: L.Shabala@utas.edu.au L. Shabala .

have been employed. All of them have certain advan-tages and disadvantages.

Of those currently in use, radioactive labelingŽSoupene et al., 1998; Onoda and Oshima, 1988;

.Kashket, 1981 is relatively straightforward and sim-ple. However, the kinetics of only one ion can bemonitored in the same experiment. Also, the timeresolution of this method is rather poor whichseverely limits its application to the study of kineticsof fast adaptive responses in bacteria.

At the other end of the spectrum is the patch-clamptechnique, a sophisticated method to study ion-trans-

Žport processes at the molecular level Petersen, 1992;

0167-7012r01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved.Ž .PII: S0167-7012 01 00270-6

( )L. Shabala et al.rJournal of Microbiological Methods 46 2001 119–129120

Saimi et al., 1992; Kubalski, 1995; Blount et al.,.1996; Levina et al., 1999 . The patch clamp tech-

nique can provide comprehensive information aboutthe kinetics and properties of specific transport pro-teins at the bacterial cytoplasmic membrane. How-ever, this method is not easily applied to all bacteria.To our knowledge, almost all successful patch-clampstudies from bacterial membranes used giant cells or

Žspheroplasts Saimi et al., 1992; Blount et al., 1996;.Cui and Adler, 1996 . These represent a highly

artificial systems.Ž .Fluorescence ratio imaging microscopy FRIM

has recently emerged as a popular tool to studyŽbacterial adaptive responses Gangola and Rosen,

1987; Breeuwer et al., 1996; Budde and Jakobsen,.2000; Siegumfeldt et al., 2000 . The cells are loaded

with fluorescent dyes, and changes in cytosolic ionconcentrations may be monitored with high temporaland spatial resolution. Unfortunately, the applicationof this method is restricted by the number of dyesavailable and only cytosolic Hq and Ca2q can bemeasured.

Finally, membrane-transport processes associatedwith bacterial adaptive responses may be measuredby use of ion-selective electrodes. The use of com-

Žmercially available ion-selective electrodes Am-.mann, 1986 is severely limited by several factors.

ŽFirst, the large size of the probe typically about.8–10 mm in diameter makes concurrent measure-

ments of several ions impractical. Second, there is ahigh probability that in long-term experiments thesensor might be partially blocked by bacteria adher-ing to the probe surface, resulting in incorrect mea-surements. Due to these and other reasons, there is atendency to use instead liquid-type ion-selective mi-croelectrodes, with a tip diameter in the mm rangeŽ .Lindner and Buck, 2000 . Such electrodes may beprepared routinely and impaled in a biofilm to studyconcentration profiles of a specific ion. The latter

Žtechnique has been used for study of pH Masson et. y Žal., 1994; Lens et al., 1995 , Cl DeBeer et al.,

. Ž1994; Xu et al., 1996 , O Horn 1994; Rasmussen2. yand Lewandowski, 1998; Xu et al., 1998 , and NO3

Ž .DeBeer et al., 1997 profiles in various types ofartificial and natural biofilms. The hydrophobic sur-face of the microelectrode minimizes the probabilityof bacteria adhering to the probe. If they do, theelectrode is simply replaced by another one because

the cost of microelectrodes is considerably less thanthe cost of the standard-size commercial electrodes.

Although concentration measurements by meansof ion-selective microelectrodes have become widelyused in microbiology, interpretation of the results isdifficult at times. The major pitfall is the origin ofobserved concentration changes. As many differentprocesses may contribute to these changes, unam-biguous interpretation of the underlying mechanismsis not always possible, and requires that other tech-niques be used in conjunction with ion-selectivemeasurements. These problems could be overcome ifnet fluxes of ions crossing the bacterial membranewere calculated from the measured concentrationgradients originating directly from the activity ofcytoplasmic membrane transporters. The easiest wayto do that is to move an ion-selective probe slowlybetween two positions above the bacterial cell sur-face. Recently, such a non-invasive ion-selective mi-

Ž .croelectrode probe the MIFEe technique has beenapplied successfully to study various aspects ofmembrane-transport processes in higher plantsŽShabala et al., 1997; Pharmawati et al., 1999; New-

.man, 2001; Shabala, 2000 and protoplasts derivedŽ .from their tissues Shabala et al., 1998 . In this

study, we pioneer the use of the MIFE technique tostudy membrane-transport processes in bacteria.

Such an approach offers several advantages. First,fluxes of specific ions can be measured with high

Ž . Ž .spatial 2–3 mm and temporal ca. 5 s resolution.Second, fluxes of up to four ions can be measured inthe same experiment and in a relatively small vol-

Ž .ume -100 ml . This provides a unique opportunityto monitor membrane-transport processes in real time,as well as to allow novel insights concerned with theionic mechanisms involved.

In this paper, we briefly outline the theoreticalprinciples of non-invasive ion-flux measurements us-ing the MIFE technique and exemplify its applica-tion to bacterial systems using adherent Escherichia

Ž q q qcoli cells. Four major ions H , K , NH , and42q.Ca , important in bacterial cell homeostasis, were

studied. We compared the net ion flux measurementswith concentration measurements in batch cultures atanalogous stages of the population growth cycle andwith different concentrations of glucose. Our resultsdemonstrate advantages of non-invasive ion fluxmeasurements and provide strong evidence that the

( )L. Shabala et al.rJournal of Microbiological Methods 46 2001 119–129 121

MIFE technique promises to be a powerful tool forfuture research in microbial ecology and physiology.

2. Materials and methods

2.1. Bacterial strain and culture conditions

E. coli SB1, a commensal clinical isolate pro-vided by Ms. S. Bettiol, School of Pathology, Uni-versity of Tasmania, was used throughout this study.The inoculum was prepared by incubation for 15–16

Žh in minimal medium Minimal Broth Davis, Difco,. Ž .0756-1-7 supplemented with 0.1% wrv glucose at

Ž .378C with shaking 60 rpm followed by regrowth infresh medium under similar conditions. Growth pa-rameters of E. coli were estimated by measuring

Žabsorbance using a spectrophotometer Spectronic.20, Spectronic Instruments, USA . Cultures were

Žharvested in the exponential phase of growth OD540. Ž .s0.3 , late exponential OD s0.8 , and station-540

Ž .ary phases culture grown for 16 h, OD s1.4 ,540

washed, suspended in a modified minimal mediumŽ .‘MMM’, described below , and kept on ice for upto 2 h until used. Unless otherwise noted, the compo-sition of the MMM used throughout the experimentswas—K HPO : 8.6 mM; NaCl: 8.6 mM; MgSO :2 4 4

Ž .0.83 mM; NH SO : 4.5 mM; CaCl : 0.1 mM.4 2 4 2

This modification of Minimal Broth was required toreduce the buffering capacity and to increase thesignal to noise ratio of the flux measurements. Inother experiments, as indicated, the ionic composi-tion of MMM was altered to improve the sensitivityof measurement of concentration gradients of spe-cific ions. The concentration of glucose varied from

Ž . Ž . Ž . Ž0.1% wrv AhighB to 0.05% wrv Ainter-. Ž . Ž .mediateB , and 0.01% wrv AlowB depending on

design of the experiment.In some experiments, carbonyl cyanide-m-chloro-

Ž . Ž .phenyl hydrazone CCCP C 2759, Sigma , wasincluded in the medium to concentrations in therange 10–200 mM.

2.2. MIFE theory

Ions crossing the cell surface in solution arecarried to or from that surface by diffusion or by

mass flow. In static or slowly changing conditions,when the convection or water uptake are negligiblysmall, measurements of the net diffusive flux of theion in solution close to the sample indicates the netflux of the ion across the sample surface.

If the ion is taken up by living cells, its concentra-tion in the proximity of the cell surface will be lowerthan that further away. Vice versa, if the ion isextruded across the cytoplasmic membrane, therewill be a pronounced concentration gradient directedaway from the cell surface. The magnitude of thisgradient will be proportional to the rate of ion move-ment across the cytoplasmic membrane, i.e. to thenet flux of the specific ion of interest. Such concen-

Ž .tration gradients and ultimately, net ion fluxes maybe measured if a probe, such as an ion-selectivemicro-electrode, is moved between two positionsabove the cell surface. Thus, the principle of theMIFE is the slow square-wave movement of ion-selective microelectrode probes between two posi-

Ž . Žtions, close to position 1 , and distant from position. Ž .2 the sample surface Shabala et al., 1997 . The

Ž .recorded voltage gradients dV between positions 1Ž .and 2 d x are converted into concentration differ-

ences using the calibrated Nernst slopes of the elec-trodes. Net fluxes of specific ions can then be cal-culated from diffusion profiles from the measuredsurface. Depending on the diffusion geometry, dif-ferent equations are used. For the simplest case of a

Žplanar surface e.g., a confluent layer of adherent. Ž y2 y1.bacterial cells , net flux, J nmol m s of a

Žspecific ion can be calculated as follows Newman,.2001 :

JscuF 58rNernst slope dVrd x ,Ž . Ž .

Ž y3 .where c is average concentration mol m of anion of interest between positions 1 and 2; u is ion

y1 y1.mobility speed per unit force, m s per N mol ;Ž y1 .F is Faraday number 96,500 C mol ; dVrd x is

voltage gradient measured by the ion-selective probebetween positions 1 and 2; and Nernst slope is theslope of the calibration graph for the probe.

For a large unicellular organism, spherical orcylindrical diffusion geometry can be used instead.

ŽSoftware included with the MIFE system Newman,.2001 automatically performs the required calcula-

tions based on the geometry of the measured cell

( )L. Shabala et al.rJournal of Microbiological Methods 46 2001 119–129122

Žcell diameter and the probe distance above the cell.surface and provides the tabulated results of mea-

Ž y2 y1.surements as net ion fluxes nmol m s inASCII format.

2.3. Microelectrode fabrication

The full experimental procedure for ion-selectivemicroelectrode fabrication is given in Shabala et al.Ž . Ž .1997 and Shabala 2000 . Briefly, electrodes were

Žpulled from a borosilicate glass capillaries GC150-.10, Harvard Apparatus, Kent, UK , oven dried, and

Žsilanised with tributylchlorosilane 90796, Fluka.Chemicals . Dried and cooled electrode blanks were

then back-filled with appropriate solutions. Immedi-ately after back-filling, the electrode tips were front-filled with commercially available ionophore cock-tails and calibrated against a range of standards.Specific details about the type of liquid ion ex-

Ž .changer LIX , composition of back-filling solutionsand range of standards used, are given in Table 1.Electrodes with responses less than 50 mV per decadefor monovalent ions and 25 mV per decade forcalcium, and with correlation coefficients less than0.999 were discarded. A reference electrode wasfabricated in a similar way from a glass microcapil-lary and filled with 1 M KCl in 2% agar.

2.4. MIFE apparatus

A new MIFE experimental setup was designed inthe School of Agricultural Science, University of

Ž .Tasmania Fig. 1 to enable the measurements of net

Table 1Fabrication details of ion-selective microelectrodes used in experi-ments

ŽIon LIX Fluka Back-filling Calibration set. Ž .catalogue no. solution mM

Ž .Hydrogen 95297 15 NaClq 5.1–6.4–7.8 pH40 KH PO2 4

Ž .Potassium 60031 200 KCl 0.1–0.2–0.5 mMŽ .Calcium 21048 500 CaCl 0.1–0.5–1.0 mM2Ž .Ammonium 09882 500 NH Cl 0.1–0.5–1.0 mM4

ion fluxes from bacteria adherent on a flat surface.Microelectrodes were held in E45P-F15PH electrode

Žholders CDR Clinical Technology, Middle Cove,.Australia mounted on a three-dimensional microma-

Žnipulator. An open-type perfusion chamber ap-.proximately 1.2 ml volume inserted on a micro-

scope stand allowed easy access to both sides of theglass coverslip with a dense layer of adherent bacte-

Ž .ria being measured Fig. 1 . The rate of solution flowthrough the perfusion chamber was within the range0.5–4 ml miny1. All measurements were made un-der steady state conditions after solution flowstopped. This was to satisfy the requirements of the

Žtheoretical basis of the MIFE method Newman,.2001 and meant that measurements started 1.5–2

min after solution changes were completed.

2.5. Preparation of bacteria

Net ion fluxes were measured from a confluentlayer of bacteria adhering to a glass surface, andprepared as follows. A glass coverslip was cleanedwith ethanol, thoroughly rinsed with running distilled

Žwater and then dried. Poly-L-lysine 0.1% wrv aque-.ous solution, P 8920; Sigma, St. Louis, USA was

used to attach cells to the glass surface. One dropŽ .f30 ml of the poly-L-lysine solution was appliedto the glass surface and left for about 3 min. Thecoverslip was then thoroughly rinsed again withrunning distilled water, followed by washing withMMM, and placed at the bottom of a perfusionchamber. A 1-ml sample from a batch culture washarvested by centrifugation at 4200=g for 10 minŽ .Biofuge A, Heraeus Sepatech, Germany , the pelletsuspended in 20 ml of MMM, the culture applied toprepared coverslip, and allowed to settle on. Themeasuring chamber was filled with measuring solu-tion, and about 20 ml of measuring solution werepassed through the chamber to flush out the cellswhich had not stuck to the coverslip. This processwas monitored using an inverted system microscopeŽ .IX50, Olympus Optical, Tokyo, Japan with long

Ž .working distance objectives 2.6 mm providing totalmagnification of 500= .

The electrodes were set 10–20 mm above thecoverslip and adherent bacterial layer. The electrodeholders were positioned at an angle of 308 to the

( )L. Shabala et al.rJournal of Microbiological Methods 46 2001 119–129 123

Ž .Fig. 1. An open-type perfusion chamber used in this study. Solution is held by surface tension created between a glass cover slide 4 andŽ . Ž . Ž .the glass base of the chamber 2 allowing safe insertion of microelectrodes 5 and a reference electrode 6 through the lateral meniscus.

Ž . Ž .Solution enters the chamber through a needle 7 and is pumped out via a tube 8 by a peristaltic pump. A glass slide with adherent bacteriaŽ . Ž . Ž . Ž .3 is put on the glass base 2 . The perfusion chamber base 1 is fixed to a microscope stage by screws 9 .

surface of the coverslip and moved horizontally by acomputer-driven hydraulic manipulator providingelectrode movement of 30 mm from the primary

Ž .position at a frequency of 0.05 Hz a 20-s cycle .The recorded voltages were converted into concen-tration differences using the calibrated Nernst slopesof the electrodes. A period of 4 s was allowed aftereach movement for the response to stabilize before

Ž .measurements recommenced Shabala et al., 1997 .Ž .Flux of an ion inward positive was then calculated

by the MIFE software, assuming planar diffusionŽ .geometry see above .

At each site the flux was measured for ca. 3 min.Electrodes were then re-positioned to another regionŽ .typically 100–200 mm distant , and the measure-ments were resumed. On average, for preparationsfrom each growth stage studied, 5–7 sites on theadherent layer were measured.

A series of experiments was undertaken to deter-mine a possible association between ion flux kineticsand the stage of the population growth cycle in

Ž .liquid culture. Samples f3 ml of the culture wereaseptically transferred from the flask into sterile vialsat 15-min intervals, and concentrations of four major

Ž q q 2q q.ions H , K , Ca , and NH were measured for4

34 min with the same microelectrodes used for fluxmeasurements. In a parallel experiment, growth pa-rameters of E. coli batch cultures were estimated bymeasuring absorbance at ls540 nm using aspectrophotometer.

2.6. Data analysis

w Ž .Standard EXCEL Microsoft tools were used tocalculate correlation coefficients between differentdata series. Significance of the differences between

( )L. Shabala et al.rJournal of Microbiological Methods 46 2001 119–129124

Žmeans was calculated using the t-test ns5–8 in.most cases .

3. Results and discussion

3.1. Feasibility of net ion flux measurements fromthe surface of adherent bacterial cells

Currently, for net ion fluxes from bacterial cellsurfaces to be measured, the bacteria must be immo-bilized on a plane surface. In this study, we used

Ž .0.1% wrv poly-L-lysine to attach cells to a glasssurface. The adhesion was strong enough to allowrapid changes of the measuring solution through theperfusion chamber without displacing adherent cellsŽ .Fig. 1 . Examination by microscope showed thatafter the chamber was flushed with measuring solu-tion and non-adherent cells were removed, the re-maining cells formed a confluent layer.

To ensure that the experimental methods involvedin immobilisation of bacterial cells did not inducefalse ApositiveB ion flux measurements, net Hq fluxmeasurements from the clear glass surface, from theglass surface coated by poly-L-lysine, and from ad-herent E. coli cells prepared from a batch culture at

Ž .the exponential phase of development OD s0.3540

were measured. No fluxes were measured in theabsence of bacteria, while adherent cells exhibited

q Ž .significant net H efflux Fig. 2A indicative of thehigh level of cell metabolic activity.

The variability of flux measurements and theirdependence on duration of measurement were inves-tigated by measurements from the different sites onthe glass surface of a single preparation. Fig. 2Billustrates measurements of net Hq fluxes from threedifferent areas of bacterial cells adherent to thecoverslip and from two different batch cultures ofE. coli SB1 harvested during exponential growthŽ .OD s0.17 . The data are highly reproducible.540

The flux patterns are quite uniform between differentŽspots, but have a tendency to increase with time see

.the difference between samples 1 and 3, or 4 and 6 .This may be explained by increasing activity of cellsin the perfusion chamber and associated changes inthe Hq transport activity.

Fig. 2. Methodological experiment validating the applicability ofthe MIFE technique to measure net ion fluxes from adherent

Ž . q Ž .bacteria. A Net H flux inward positive measurements for theŽglass surface, for the glass surface coated by poly-L-lysine used

.to attach cells to the glass surface , and for the adherent E. coliŽ .cells prepared from an exponential phase culture OD s0.3 A .540

Ž . q Ž .B Net H fluxes from three different sites different symbolsŽof adherent cells from duplicate cultures open and closed sym-

. Ž .bols, respectively of exponential phase OD s0.17 A E.coli540

SB1 cells. The time scale represents a relative time with barsindicating units.

3.2. Ion flux kinetics associated with bacterial growth

ŽConsistent with the literature reports Kashket,.1981; Booth and Kroll, 1983; Luli and Strohl, 1990 ,

there was a significant acidification of the batchculture during the late exponential phase of growthŽ .Fig. 3 . After the onset of a stationary phase, pHchanges were slightly reversed.

Concentration measurements alone do not allowunambiguous interpretation of these results. Severalmechanisms, including changes in respiratory activ-

Ž .ity Harold and Maloney, 1996 , membrane-transportŽprocesses at the cytoplasmic membrane Padan and

Schuldiner, 1987; Benjamin and Datta, 1995; Booth,

( )L. Shabala et al.rJournal of Microbiological Methods 46 2001 119–129 125

Ž . q Ž .Fig. 3. Kinetics of pH changes for the E. coli SB1 cells in batch culture open symbols and associated net H fluxes bars measured fromq Ž .the surface of adherent cells prepared from confluent culture. Note the logarithmic scale for H fluxes. Error bars are S.E. ns6 .

. Ž1999 , formation of acid byproducts Luli and Strohl,. Ž1990; Neijssel et al., 1996 , cell lysis Buurman et

.al., 1991 and some others might be involved. How-ever, when net Hq fluxes were measured, a progres-sive increase in the net Hq extrusion peaking aty150 nmol my2 sy1 for OD s0.3 and followed540

by a gradual return to zero net flux at late stationaryphase was found. The quantitative information is

q Ž .given in Fig. 3, where H flux kinetics bars wereŽ .compared with pH changes open circles . The close

association between these two processes, with themagnitude of the net Hq flux leading the resultantpH changes, offers strong evidence that the observedacidification of the external media during exponen-tial growth of E. coli originates mainly from net Hq

movement across the bacterial membrane. It appearsthat membrane-transport processes are the majorcontributors during the exponential stage of growth.This conclusion was further supported by experi-

Ž .ments with manipulation of substrate glucose con-centration, where a strong dependence of net Hq

fluxes on external glucose concentration was found,and by experiments with CCCP, a known proto-nophore dissipating the proton motive force across

Ž .the cytoplasmic membrane data not shown .Fluxes of other ions also exhibited regular changes

associated with population growth. Fig. 4 showsaverage flux values measured for Hq, Kq, NHq, and4

2q Ž .Ca for middle exponential OD s0.3 , late ex-540Ž . Žponential OD s0.8 ,and stationary phase 16-h-540

( )L. Shabala et al.rJournal of Microbiological Methods 46 2001 119–129126

q q q 2q Ž .Fig. 4. Net fluxes of H , K , NH , and Ca measured from adherent bacterial layers prepared from middle exponential Exp , late4Ž . Ž . q qexponential L-exp , and stationary Stat phase of culture development. Concentrations of K and NH in the measuring solution were 0.24

Ž .and 1 mM, respectively, to reduce the noise during flux measurements. Data are mean"S.E. ns10, 8, 6, and 11, respectively . Note thatthe scale of the Ca2q flux response is 1–2 orders of magnitude less than other ions measured, and is of the same order as the noise level.

.old culture of development. Measurements of netion fluxes from adherent cells, isolated from differ-ent stages of the population growth curve, may be asensitive tool by which to monitor underlying activ-ity of ion-transporting mechanisms associated withbacterial growth and nutrition.

Unlike Hq, Kq, and NHq, measurements of the4

net Ca2q fluxes revealed no significant net Ca2q

Ž .transport across the bacterial membrane Fig. 4 .Concentration measurements in batch cultures, how-

Ž .ever, showed complex biphasic kinetics Fig. 5A ,Ž .highly correlated Rs0.96 with population growth

Ž .data not shown . We hypothesized that such changesoriginate from Ca2q binding to phosphate, known tobe a pH-dependent process, and measured changes inconcentration of free Ca2q in the cell-free medium at

Ž .different pH values. The results data not shownsuggest strongly that the observed Ca2q concentra-tion changes are due to pH-dependent changes inCa2q binding to phosphate due to modulation of the

external pH in the batch culture. Fig. 5A shows thestriking similarity in the measured changes in Ca2q

concentration and those changes calculated from ob-servations in cell-free medium due to binding withphosphate.

These two examples illustrate the power of theMIFE technique and the advantages of net ion fluxmeasurements over traditional concentration mea-surements with ion-selective electrodes. The absence

2q Žof any significant net Ca fluxes measured Figs. 4.and 5B is a strong reason to rule out the contribu-

tion of bacterial membrane-transport systems to ob-Žserved concentration changes based on the absence

2q . 2qof Ca concentration gradients . If only Ca con-centration kinetics were measured, but not net Ca2q

fluxes, such an interpretation of the results would notbe possible. On the other hand, fluxes of other ionsŽ q q.H and NH displayed regular changes and4

showed strong dependence on glucose availabilityŽ .data not shown . That is consistent with the essen-

( )L. Shabala et al.rJournal of Microbiological Methods 46 2001 119–129 127

Ž . 2q ŽFig. 5. A Measured changes in Ca concentration open sym-.bols in batch cultures compared with changes calculated from

observations in pH-modified cell-free medium due to associatedŽ .binding with phosphate closed symbols assuming pH changes as

Ž . 2qshown in Fig. 3. B Net Ca fluxes from adherent bacteriaassociated with different stages of growth of E. coli SB1 batchculture.

tial role of these ions in maintaining homeostasis inŽ .E. coli cells Neijssel et al., 1996 .

4. Conclusions and perspectives

The aim of the present work was to demonstratethe feasibility of non-invasive microelectrode ionflux measurements from a layer of adherent bacterialcells. This is the first such report in the literature. Inthis study, we have demonstrated advantages of the

MIFE system over classical concentration measure-ments in microbiological research. As illustrated bythe example with Ca2q, classical concentration mea-surements do not provide sufficient information forunambiguous data interpretation but measurement ofnet ion fluxes crossing the bacterial cytoplasmicmembrane may offer insight into underlying mecha-nisms.

High temporal resolution of the MIFE techniqueŽ .up to 5 s provides a unique opportunity for real-timein vivo studies of kinetics of bacterial adaptive re-sponses to a large number of environmental stresses.This list might include pH, salt, osmotic, and tem-perature stresses, screening of antibiotic and growthregulators and many others. The ability to measurefluxes of several ions simultaneously and essentiallyat the same spot enables the stoichiometry betweenvarious ions to be studied, ultimately resulting inbetter understanding of ionic mechanisms responsi-ble for bacterial adaptation.

Even by itself, this technique offers wide prospectsto study various aspects of bacterial physiology.Application of the MIFE technique to natural andartificial biofilms is, perhaps, one of the mostpromising areas. Even greater progress might beachieved by combining the MIFE technique withother recently developed approaches to study thephysiology of bacteria. From this perspective, ionflux measurements from genetically modified materi-

Ž .als mutant strains may be very promising. It is alsopossible to combine non-invasive ion-selective fluxmeasurements with monitoring of cytosolic pH orCa2q changes by means of fluorescence techniquessuch as fluorescence ratio-imaging microscopy. Incombination, these two techniques may provide newinsights into the underlying mechanisms of pH andion homeostasis in bacterial cells.

Acknowledgements

The authors wish to acknowledge the financialsupport of Meat and Livestock Australia and Schoolof Agricultural Science for funding and stipend sup-port of L. Shabala.

( )L. Shabala et al.rJournal of Microbiological Methods 46 2001 119–129128

References

Ammann, D., 1986. Ion-Selective Microelectrodes. Springer-Verlag, Berlin.

Benjamin, M., Datta, A.R., 1995. Acid tolerance of enterohemor-rhagic Escherichia coli. Appl. Environ. Microbiol. 61, 1669–1672.

Blount, P., Sukharev, S.I., Moe, P.C., Nagle, S.K., Kung, C.,1996. Towards an understanding of the structural and func-tional properties of MscL, a mechanosensitive channel inbacteria. Biol. Cell. 87, 1–8.

Booth, I.R., 1999. The regulation of intracellular pH in bacteria.Novartis Found. Symp. 221, 19–37.

Booth, I.R., Kroll, R.G., 1983. Regulation of cytoplasmic pHŽ .pH in bacteria and its relationship to metabolism. Biochem.i

Soc. Trans. 601, 70–72.Breeuwer, P., Drocourt, J.L., Rombouts, F.M., Abee, T., 1996. A

novel method for continuous determination of the intracellularpH in bacteria with the internally conjugated fluorescent probeŽ .5 and 6- -carboxyfluorescein succinimidyl ester. Appl. Envi-

ron. Microbiol. 62, 178–183.Budde, B.B., Jakobsen, M., 2000. Real-time measurements of the

interaction between single cells of Listeria monocytogenesand nisin on a solid surface. Appl. Environ. Microbiol. 66,3568–3591.

Buurman, E.T., Teixeira de Mattos, M.J., Neijssel, O.M., 1991.Futile cycling of ammonium ions via the high affinity potas-

Ž .sium uptake system Kdp of Escherichia coli. Arch. Micro-biol. 155, 391–395.

Cui, C., Adler, J., 1996. Effect of potassium-efflux system, KefA,on mechanosensitive channels in the cytoplasmic membrane ofEscherichia coli. J. Membr. Biol. 150, 143–152.

DeBeer, D., Srinivasan, R., Stewart, P.S., 1994. Direct measure-ment of chlorine penetration into biofilms during disinfection.Appl. Environ. Microbiol. 60, 4339–4344.

DeBeer, D., Schramm, A., Santegoeds, C.M., Kuhl, M., 1997. Anitrite microsensor for profiling environmental biofilms. Appl.Environ. Microbiol. 63, 973–977.

Gangola, P., Rosen, B.P., 1987. Maintenance of intracellularcalcium in Escherichia coli. J. Biol. Chem. 262, 12570–12574.

Harold, F.M., Maloney, P.C., 1996. Energy transduction by ionŽ .currents. In: Neidhardt, F.C. Ed. , Escherichia coli and

Salmonella. Cellular and Molecular Biology, vol. 1, ASMPress, Washington, pp. 283–306.

Horn, H., 1994. Dynamics of a nitrifying bacteria population in abiofilm controlled by an oxygen microelectrode. Water Sci.Technol. 29, 69–76.

Kashket, E.R., 1981. Effects of aerobiosis and nitrogen source onthe proton motive force in growing Escherichia coli andKlebsiella pneumoniae cells. J. Bacteriol. 146, 377–384.

Kubalski, A., 1995. Generation of giant protoplasts of Es-cherichia coli and an inner-membrane anion selective conduc-tance. Biochim. Biophys. Acta—Biomembr. 1238, 177–182.

Lens, P., DeBeer, D., Cronenberg, C., Ottengraf, S., Verstraete,W., 1995. The use of microsensors to determine populationdistributions in UASB aggregates. Water Sci. Technol. 31,273–280.

Levina, N., Totemeyer, S., Stokes, N.R., Louis, P., Jones, M.A.,Booth, I.R., 1999. Protection of Escherichia coli cells againstextreme turgor by activation of MscS and MscL mechanosen-sitive channels: identification of genes required for MscSactivity. EMBO J. 18, 1730–1737.

Lindner, E., Buck, R.P., 2000. Microfabricated potentiometricelectrodes and their in vivo applications. Anal. Chem. 72,336A–345A.

Luli, G.W., Strohl, W.R., 1990. Comparison of growth, acetateproduction, and acetate inhibition of Escherichia coli strainsin batch and fed-batch fermentations. Appl. Environ. Micro-biol. 56, 1004–1011.

Masson, F., Lacroix, C., Paquin, C., 1994. Direct measurement ofpH profiles in gel beads immobilizing Lactobacillus helÕeti-cus using a pH sensitive microelectrode. Biotechnol. Tech. 8,551–556.

Neijssel, O.M., Teixeira de Matos, M.J., Tempest, D.M., 1996.Growth yield and energy distribution. In: Neidhardt, F.C.Ž .Ed. , Escherichia coli and Salmonella. Cellular and Molecu-lar Biology, vol. 2, ASM Press, Washington, pp. 1683–1692.

Newman, I.A., 2001. Ion transport in roots: measurement offluxes using ion-selective microelectrodes to characterisetransporter function. Plant Cell Environ. 24, 1–14.

Onoda, T., Oshima, A., 1988. Effects of Ca2q and protonophoreon growth of an Escherichia coli L-form. J. Gen. Microbiol.134, 3071–3077.

Padan, E., Schuldiner, S., 1987. Intracellular pH and membranepotential as regulators in the prokaryotic cell. J. Membr. Biol.95, 189–198.

Petersen, O.H., 1992. Ion channels. Ten years of patch-clampstudies. Biochem. Pharmacol. 43, 1–2.

Pharmawati, M., Shabala, S.N., Newman, I.A., Gehring, C.A.,1999. Natriuretic peptides and cGMP modulate Kq, Naq, andHq fluxes in Zea mays roots. Mol. Cell Biol. Res. Commun.2, 53–57.

Rasmussen, K., Lewandowski, Z., 1998. The accuracy of oxygenflux measurements using microelectrodes. Water Res. 32,3747–3755.

Saimi, Y., Martinac, B., Delcour, A.H., Minorsky, P.V., Gustin,M.C., Culbertson, M.R., Adler, J., Kung, C., 1992. Patchclamp studies of microbial ion channels. Methods Enzymol.207, 681–691.

Shabala, S.N., 2000. Ionic and osmotic components of salt stressspecifically modulate net ion fluxes from bean leaf mesophyll.Plant Cell Environ. 23, 825–837.

Shabala, S.N., Newman, I.A., Morris, J., 1997. Oscillations in Hq

and Ca2q ion fluxes around the elongation of corn roots andeffects of external pH. Plant Physiol. 113, 111–118.

Shabala, S., Newman, I., Whittington, J., Juswono, U., 1998.Protoplast ion fluxes: their measurement and variation withtime, position and osmoticum. Planta 204, 146–152.

Siegumfeldt, H., Rechinger, K.B., Jakobsen, M., 2000. Dynamicchanges of intracellular pH in individual Lactic acid bacteriumcells in response to a rapid drop in extracellular pH. Appl.Environ. Microbiol. 66, 2330–2335.

Soupene, E., He, L., Yan, D., Kustu, S., 1998. Ammonia acquisi-tion in enteric bacteria: physiological role of the

( )L. Shabala et al.rJournal of Microbiological Methods 46 2001 119–129 129

Ž .ammoniumrmethylammonium transport B AmtB protein.Proc. Natl. Acad. Sci. U. S. A. 95, 7030–7034.

Xu, K.D., Stewart, P.S., Chen, X., 1996. Transport limitation ofchlorine disinfection of Pseudomonas aeruginosa entrapped inalginate beads. Biotechnol. Bioeng. 49, 93–100.

Xu, K.D., Stewart, P.S., Xia, F., Huang, C.T., McFeters, G.A.,1998. Spatial physiological heterogeneity in Pseudomonasaeruginosa biofilm is determined by oxygen availability. Appl.Environ. Microbiol. 64, 4035–4039.