Biochi,zica t't Biophysica Acta. I 115 ( 1991) 75-83
75
~) 1~,~91Elsevier Science Publishers B.V. All rights reserved 03114-4t65/ql/$03.5|1
BBAGEN 23613
Continuous measurement of the cytoplasmic pH in Lactococcus lactis with a fluorescent pH indicator D o u w e M o l e n a a r , T j a k k o A b e e a n d W i i N. K o n i n g s D~7~artl~ienl o]' .'~ticrobi~h~.'. Unit't'r.~ilv o1"(ironin.~cn~ t hJrcn ( Th,' Nt'therland.~ )
(Received S July I~J()l )
Key words: Intraccllular pll; pH rcgulatam, Fluorescent probe, t~'arboxyfluorcsccin:Multidrug resistance;~1,. h~ctis)
The cytoplasmic pH of Lactococcus lactis was studied with the fluorescent pH indicator 2',7'-bis-(2-carboxyethylJ-.~ (and-6)-carboxyfluorescein (BCECF). A novel method was applied for loading bacterial cells with BCECF, whicl~ consists of briefly treating a dense cell suspension with acid in the presence of the probe. This results in a pH gradient, which drives accumulation of the probe in the cytoplasm. After neutralization the probe was wet! retained in cells stored on ice. BCECF-Ioaded cells were metabolically active, and were able to genera*.e a pH gradient upon energization. The probe leaks out slowly at elevated temperatures. Effiux is stimulated upon energization of the cells, and is most likely catalyzed by an active transport system, it is a first-order process, and the rate constant could be deduced from the decrease of the fluorescence signal in periods of constant intracellular pH. This allowed a correction of the fluorescence signal for effiux of the probe. After calibration the cytoplasmic pH could be calculated from elllux-corrected fluorescence traces.
Introduction
The cytoplasmic pH is an important factor in regulating enzyme activity and cell function, and the mechanisms involved in the regulation of cytoplasmic pH are studied intensively [1-3]. The only generally applicable methods for measuring intracellular p l t in bacteria are nuclear magnetic resonance techn;ques [4] or analysis of the distribution of (radiolabeleu)weak acid or weak bases (for reviews see Refs. 5 and 6). These methods have a limited time resolution, and can only be used to measure the internal pH under steady-state conditions or when the cytoplasmic pH changes slowly. N M R techniques have the additional disadvantage that high cell densities are required, which can cause diffi.
Abbreviations: Bt'ECF, 2',7'-bis-(2-carboxyethyl)-5(and-6)-carboxyfluorescein; CCCP, carbonyl cyanide m-chlorophcnylhydraztme: pyranine, 8-hydroxy-l,3,6-pyrenetrisulhmicacid: Mes, 2-(N-mt~rpholinok,~thanesulfi)nic acid; Mops, 3-(N-morpht~iint~)propancsulfimic acid; Hepes, N-2 hydrt~xyethylpiperazine-N'-2-ethanesulfi)nic acid: Tricine. N-tris(hydroxyme|hytlmethylglycinc:bislris propane, 1,3bis[tris(hydroxymethyl)methylamint~lpropane. Correspondence: W.N. Konings, Departmenl of Microbi~h~gy,University of Groningen, Kerklaan 3t), 9751 NN II:iren, The Nether~ lands.
culties in controlling the external medium composition (for example pH and oxygen concentration). Fluorescent pH indicators have occasionally been used for bacteria, but either the methods used for loading he cells with the probe were not generally applicable, or the mca,~urcments were difficult to interpret due to high eftlex of the probe [7,8]. The use of a fluorescent probe as a pH indicator has obvious potential advantages. Most indicators used so far have no known serious negative effects on cellular metabolism and measurements can be performed with high time and pH resolution [9], Loading of cells with the fluorescent indicator BCECF can b : achieved by adding the membrane permeable acetoxymethyl estc~ 'ff the ind:. ~tor to cells. The indicator will be trapped intracellulary upon hydrolysis by the action o f esterases. This method finds wide application in the study of pH control and transport systems in eucaryotic cells [9,10]. However, in bacterial cells the method had a narrow perspective for applicability. In H a l o b a c t e r i u m h a l o h i u m the probe BCECF ~as only usable after 4 days incubation of cclls with the acctoxymethyi ester [8]. A carboxyfluorescein ester has been used in Excht, richht coli, but its use is limited since a cumbersome correction has to be madc for the high rate of efflux of carboxyfluoresccin and hydrolysis of ester bonds [7].
76 In the studies described in this article a simple method has been used for loading bacterial cells with BCECF directly. Cells were incubated with the fluorescent probe and 50-1(10 mM HCI for 5 min at room temperature, and then brought to pH 7 with excess buffer. After this treatment, BCECF is located intracellularly. Thc cells are viable and retain BCECF well when stored on ice. Measuremcnt of the fluorescence signal can bc performed with high time resolution. The probe cffluxes at elevated temperatures, but the process can be well characterized, and the fluorescence signal can be casiIy corrected for efflux. Correlation of the signal to cytoplasmic pH can be done by a proper calibration method. In the pK range of BCECF (7-7.5) this method has a high sensitivity towards changes in pH. The method has been applicd to study pH regulation in the Gram-positive fermentative bacterium Lactococcus lactis, it was also applied in E. col(, l.~hodohacter sphaeroides. Arthrot,acter and Acinetohatter species (Molcnaar, D., unpublishcd results and Sikkcma, J., personal communication). Materials and Methods
Strabr and growth condition L. lactis, subspecies lactis MI,3, was grown anaerobically at 2 9 ° C on a modified MRS medium [I I], containing per liter 10 g tryptonc, 10 g 'Lab-Lemco' powder (Oxoid, Basingstokc, U.K.), 5 g yeast extract, 2 g K : H P O 4, 0.2 g MgSO~, 35 mg MnSO4, 0,5 ml Tween8(1, adjusted to pH 6.3 with HCI. The medium was supplemented with separately sterilized 2% lactose, or 2% galactose and 20 mM arginine. Loading of cells with BCECF Ceils were grown in 100 or 200 ml medium and harvested in the logarithmic phase of growth, at an absorbance at 660 nm of approx. 0.6. They were washed twice in 50 mM potassium phosphate, pH 7 (hereafter called "buffer'), and resuspended in 2.5 or 5 ml of this buffer. The protein concentration of this suspension was approx. 5 mg/ml. The suspension was kept on ice until further use. An optimal acid shock requires a distinct quantity of acid to be added to the cells. Enough acid has to be added to get sufficient loading, without damaging the cells. For optimization six tubes were prepared. Aliquots of 200/,1,1 of the suspension were centrifiiged tk~r 15 s in 1 ml tubes in an Eppendorf centrifuge and rcsuspendcd in 2[) tzl buffer. 1 tzl containing 5 m g / m l ( 1(1 mM) BCECF stock solution was ~dded. Droplets of 1.5 to 4 tst (intervals of 0.5 p,l) of 0.5 M HCI were adhercd to the walls of these tubes. After mixing of the droplets with the cell suspension, the tubes were left at room temperature tor 5 min. 1 mt buffer was then added, the tubes wcrc shaken and cells were spun
down in an Eppendorf centrifug:~ for 15 s. The pellet which contained a fair amount ol 3 C E C F (judged by the intensity of its yellow coiour) af er treatment with the smallest amount of acid, was used for further experiments (usually 2 to 3 p.I HCI treatments). These cells contain I to 3 mM BCECF (see Results). The cells were washed four times with 1 ml buffer by vortexing and centrifugation, and finally rcsuspended in 200/.LI buffer and stored on ice.
Measurenwnt of fluorescence Spectra of intracellular and extraccllular BCECF were recorded. Excitation (emission at 525 nm) and emission (excitation at 500 nm) spectra of BCECF were measured with slit widths of 5 nm. Optima were found at 502 nm in both excitation spectra, and at 521 and 523 nm in the emission spectra of extra- and intracellular BCECF, respectively. The isosbestic (pH invariable) point which can be found at 440 nm in the excitation spectra of free BCECF was not found in the spectra of intracellularly located BCECF. Fluorescence measurements with probe-loaded cells were performed in a 3-mi cuvette containing 3 ml of the desired buffer and 5 to I0 p,I cells, depending upon the concentration of BCECF. The suspension was stirred and thermostated. The excitation and emission monochromator wavelengths were 502 and 525 nm with slit widths of 5 and 15 nm, respectively. The fluorescence signal was averaged over time intervals of 0.5 or ! s. L f F t - assay A L. lactis cell suspension of 100 /zl, loaded with BCECF, was diluted in a tube thermostated at 3 0 ° C containing 2 ml 50 mM potassium phosphate of the desired pH. A time series was made by withdrawing aliquots of 2(~) ~l. Cells were spun down immediately for 3 min in an Eppendorf centrifuge, and 180 ~l of the supernatav, was carefully pipetted off. This 3 min centrifugation is short compared to the time-scale in which leakage occurs, and gross errors are not expected. To obtain a 'total fluorescence' value the suspensions were permeabilized at the end of the experiment by incubation with 0.1% Triton X-100 for an additional 31) rain. The supernatant samples were diluted 60-fold in 50 mM bistris propane (pH 9.1). The fluorescence of this diluted sample was measured with the monochromators set to 502 nm excitation and 525 nm emission, it was assumed that this fluorescence was proportional to the BCECF concentration. The fluorescence of the supernatant at the start was not zero, but approx. 10% of the fluorescence in the 'total BCECF' sample, probably duc to leak during or after washing or to lysis. This value was subtracted as a background from all samples. For calculation of intracellular concentrations of BCECF an intracellular vol-
77 umc of 2.8 p.I/mg protein was used, as determined previously [ 12].
Materials attd bzstrumentation BCECF was obtained from Molecular Probes. Eugene, OR, U.S.A. A stock suspension of 5 m g / m l was made in water, which was brought to approx, pH !1 with KOH, and after a few minutes neutralized with HCi, as was recommended by the supplier. The solution was stored at - 2 1 ) ° C and always kept in the dark. The fluoromcter was a Perkin-Elmcr LS 5(1 with computer controlled data acquisition and storage.
Results Accumulation and retention of BCECF Loading L. lactis with 0.5 mM BCECF cxtraccllularly resulted in an internal concentration of 1 to 3 mM BCECF, depending on the batch of cells. This could bc deduced from the 'total fluorescence' value in cfflux experiments as described below. The 2-6-fold accumulation can be explained if a transient pH gradient is present during the loading procedure. Such a transient pH gradient will lead to the uptake of the protonated species of BCECF which is more membrane permeable than the deprotonated species (BCECF has four carboxyl groups). To get some information about the existence of a transient pH gradient the external pH was recorded with a pH electrode upon the addition of HCI (data not shown). The initial external pH just after the addition of acid was approx. 2, subsequently, the pH increased and stabilized within I h at approx, pH 5. The slow pH increase is probably due to slow proton
!
II
loo %~.
v
influx as a result of the existing pH gradient across the membrane. In the normal loading procedure excess buffer was added 5 rain after the shock. At this time the external pH was approx. 3. The intracellular pH will then have decreased from approx. 7 to a value intermediate between 7-5 (see also discussion). Thus, a large pH gradient of more than 2 pH units, inside alkaline, existed during the loading procedure. The same technique has been applied succcssfully for loading cells with pyraninc, but the levels of fluorescence obtained with this indicator were much lower than for BCECF. The retention of BCECF in L. lactis was studied under different conditions. BCECF was wcli retained in cells storcd on ice. Therc was no detectable efflux in 2.5 h (Fig. IA, triangles). However, slow efflux of the probe occurred from non-energized ceils at 30°C, which were uncoupled with the protonophore CCCP. This efflux was slightly pH-dependent, and more rapid at lower pH. At pH 6 and 8 the tt/2 values (the time nccdcd for cfflux of 50% of the original probe content), were approx. 75 rain and 110 min, respectively (Fig. IA). Energization of the cells with lactose stimulates the cfflux of the probe dramatically (Fig. I B). Rapid efflux was also observed upon the addition of lactose to the cells and rcmaincd rapid when the proton motive force was decreased by the addition of valinomycin and nigericin. The applied valinomycin concentration was sufficient to completely abolish the electrical membrane gradient, as was shown previously by Poolman et al. [12] and was checked with a TPP ~ electrode (data not shown). The addition of nigericin can introduce a
I
?001
~v
|
I
~
"'
It'"
1
t
A
B
80
"6 6o c
"6
h
~
Ld
L=J
T
20
~~
o
lactose 0
"
0
50
""
'
1O0
Time (rain)
150
0
'
"
'
'
'
t0
20
30
40
50
60
70
Time (rain)
Fig, I, Efflux of BCECF, measured with the efflux assay described in Materials and Methods. (A) (.:ells were incubated on ice in 51) mM potassium phosphate pH 7 (v) or at 30 ° C with 51)/~M CCCP in 50 mM potassium phosphate pl[ 6 (o) or pH 8 (o). The vertical axis has a logarithmicscale. (B) Cells were incubated at 311oC in 5(I mM potassium phosphate pl t 6 (el and pH 8 1o1 without additions,or ptl 7 with 2 p.M nigericinand 1/zM valinomycinadded ( v 1. At the time indicatedcells were energizedby the additionof 2.8 mM lactose. The vcrticat axis has a logarithmicscale.
7g smaI1 inverted pH gradient (inside acid) most likely as a result of high internal K + concentration [12] (see also Figs. 4 and 5). The same observation of energy stimulated efflux was made when cells were energized with argininc via the arginine deiminase route [13] (data not shown). Also under energized conditions efflux was slightly dependent on the external pH. Semi-logarithmic plots of the fluorescence decrease yield straight Iines (Fig. I B), indicating that efflux occurred by a first-order process. The t=/z values in the energized state were 5.7 rain at pH 6 and 1 1.2 min at pH 8.
250
A
b
200 5 150
g ~oo
5O
Correction of fluorescence for BCECF effltL,; Efflux of the probe complicates the interpretation of changes in the fluorescence signal. The changes measured arc pa)tty due to efflux and partly to changes in the cytoplasmic pH. For analysis of the fluorescence signal the assumption is madc that there exist two homogeneous BCECF populations: a cytoplasmic and an cxtraccltular population. Both populations contribute to the measurcd fluorescence signal. To analyse this total fluorescencc the fl)llowing parameters are introduced: (i) ['~, the fluorcsccncc measured if all BCECF werc located intraccllularly: (it) F0, the fluorescence measured if till BCECF wcrc located extracellularly. This second situation occurs in fitct after a long tin)e, when all the probe has effluxed. In the experiments described bcIow the probe is located both intracellularly and extracellularly. Consequently only a fraction o, corresponding to the fraction of intracellularly located probe, of the hypothetical signal F~ is measured, in addition to a fraction (1 - q ) o f the signal F,,. Together they constitute the total signal F at a certain point m time: [ ' = q'F, +(1 - q).l.:,
(I)
It was found in the experiments described above that efflux of the probe is ~ first-order process. Therefore, the fraction of probe located intracellularly (q) decreases exponentially with time (t): q(t)=qn,
e
(2)
f,t
in which q0 represents the fraction of probe located internally at t = 0, and k is the first-order rate constant for efflux. The measured fluorescence F is a function of time (expressed as F(t)), and so is the fluorescence of cytoplasmic probe F i if the cytoplasmic pH changes in time (Fi(t)). Fo will be constant, since in the experiments described the external pH does not change. SubstitutinL" Eqn. 2 in Eqn. 1 then yields: F(t) = F,,+ q,)'e-~"( F i ( t ) - F~)) From this equation
(3)
it f o l l o w s t h a t u n d e r c o n d i t i o n s o f
Ioc
n)..,
T
6.0
5.5
0
1'
val ni
org l
I
5
10
1't
otg 15
TTnig
T
vol nig
arg
I
I.
I
5
10
15
val
,.5
10
15
Time (rain) Fig. 5. ('ylopl:tsmic pH traces of arginine energized L. hwtis al different external pHs. Conditions, buffers, additions, and concentrations were Ihc same as in Fig. 5. pHs were 6 (A) 7 (B), and 8 (C). At the times indicated 1.3 mM arginine ttarg) was added.
TABLE I
Comparisota of tTtOplaxmic pH of L. lactis umh'r dtfftwcnt conditttm.~ th'termbted b)" the distribution of raditdatwlcd benzoate ocid or methylamine atut b)' B('ECF fluoresccm'e Energy
Arginine External pH:
Cytoplasmic pH determined by distribution " Cytoplasmic pH det~,rmincd by BCECF fluort:s¢cncc
I..actosc
6
7
8
0
7
8
7.1
7.3
7.8
7.3
7.6
7.9
7.35
7.61) 7.85
7.05
7.75
7.85
" Data from Rcf, 16. The buffers and condifion,; wcrc the same as used in Fig. 4.
increase of the pH gradient upon dissipation of the (larger) electrical potential. The mechanism bchind these phenomena is currently under investigation. The steady-state pH values in energized cells were compared to values measured earlier [16] under the same experimental conditions with the distribution of radiolabeled benzoic acid and methylamine (Table !). In general, the values measured by BCECF fluorescence were higher than those measured by the distribution method, at least for pH gradients alkaline internally, which were measured with benzoic acid. Witil both methods a higher pH gradient was found in lactose energized cells compared to arginine energized cells. However, with BCECF fluorescence the difference tl~e differences found in cytoplasmic oH were smallcr at the different external pHs, indicating a more stringent regulation of cytoplasmic pH than found previously. These differences could be due to an overe~,~timation of cytoplasmic pH by BCECF fluorescence, but it could also be the result of an underestimation of pH gradients by benzoic acid accumulation or be due to an inhibition of the build up of the pH gradient by accumulated benzoic acid [1,17]. Discussion
This paper describes a new and simple method for loading bacterial cells with the fluorescent pH indicator BCECF for measuring quantitatively the cytoplasmic pH. It can be applied both in (3ram-positive and Gram-negative organisms. BCECF is accumulated intracellularly during a low pH shock. During this shock the external pH drops to values where all carbo~lic groups of BCECF are likely to be protonated. This fully protonated neutral molecule is probably more membrane permeable than the negatively charged species, and diffuses inward across the membrane. The slow rise in external pH during the acid shock points to the existence of a proton gradient across the membrane. This proton gradient will form a driving force for the accumulation of BCECF, since the equilibrium is shifted to the deprc~tonated species intracellularly.
BCECF i,: we!! retained when ccll:~ arc st~-~..-.d on ice. and at ctevatcd temperatures efflux is slow in the non-energized state. The stimulation of efflux upon energization of the cells makes this process intriguing. especially since this stimulation is also observed in uncoupled cells. If effiux were a passive, non-carriermediated process non-carrier-mediated process (leak), the pH gradient would certainly not stimulate efflux because an increase of the internal pH will lead to a further di.,;~ociation and t h u s to a more negatively charged form of the probe, which will be less membrane permeable. An inverted pH gradient, generated upon addition of nigericin, could possibly stimulate BCECF efflux. However, the observation that a positive pH gradient alone (by energizing cells in the presence of valinomycin) did not drive uptake and did not prevent cfflux argues against this explanation (Molenaar. D.. unpublished data and Fig. 2). An electrical gradient, inside negative, can. in principle, pull the negatively charged probe across the membrane to the outside and can stimulate leak. Yet stimulation of effiux was also observed in cells in the absence of an electrical gradient, and seems to be dependent on the presence of other forms of metabolic energy. There are also differences in efflux rate observed between species. Co~.nebacterium sp. C125 [18] showed a very high effiux rate. whereas Arthrohacter sp. T2 [19] had a very low efflux rate (Sikkema, J., personal communication). Also within species such differences can be observed depending on the mode of energization (Molenaar, D., unpublished results). These results could indicate the presence of an active extrusion system for BCECF driven hy metabolic energy in bacteria. In a forthcoming article we will present more evidence for such a system, and show that it has properties different from the recently cloned protonmotive force driven bacterial multi-drug resistence system [20]. Energy dependent efflux of BCECF is also in line with the recent suggestion for a system that catalyzes this process in epithelia[ ,:ells [21]. BCECF effiux is a first-order process whicla could indicate that this export system is working far below the affinity constant for BCECF. This K, will be much larger than 3 raM, suggesting that the transport system is not very specific for the probe. Although the mechanism of efflux of the probe is still unclear, it is phenomenologieally a simple and well characterized process. Therefore, correction of the fluorescence signal for efflux can be easily performed. However. as can be deduced from Eqn. 4 the error in the corrected signal will be proportional to errors in qo and F,,, and exponentially related to errors in k. The error also increases exponentially in time. This accumulation of errors can introduce large noise and systematic errors in the calculated pH, especially when k is high. A method to correct for probe effiux from suspensions of eukaryotic cells has been developed
82 which migh prevent systematic errors by circumventing the need for precise characterization of the efflux process [22]. Thc~c authors used anion exchanger beads to bind external probe in the cuvette outside the light path, so that only the intracellularly located probe is observed. A concentration independent signal can then be obtained by continuous excitation ratioing. We have not yet made attempts in this direction. For relating the fluorescence signal to the cytoplasmic pH it is of utmost importance to have a reliable calibration method. The ionophore nigcricin is used for this purpose in eukaryotic cells, sometimes in combination with valinomycin, It is assumed that during calibration the cytoplasmic pH equals the external pH. Since nigcricin exchanges potassium ions electroneutrally with proton:;, it is not immediately obvious why internal and external pH should be equal. As can be seen in Figs. 5 and 6 (scc also Rcf. 12) in energized cells these pH values arc indeed not equal. However, during the build-up of a pH gradient, the extrusion of positive charge on the proton is usually compensated for by the uptake of pot;,ssium ions. to conserve electroneutrality. Potassium ions are accumulated to concentrations of several hundred millimolars, reflecting the high buffer capacity of the cytoplasm [i,23]. This pool is again freely exchangeable for protons upon the addition of nigericin. The pK upshift tound for intracellular BCECF relative to the free probe exists both at low and high potassium concentrations, suggesting that it is not due to a Donnan potential. The (apparent) pK downshift of BCECF at high potassium concentration relative to low potassium concentration can be caused by differential effects of ionic strength on the activity coefficients of the protonated and deprotonated probe species. Such effects can in principle cause artifacts when the intracetlular ionic strength changes significantly during an experiment. Other potential sources of artifacts are fluorescence quenching mechanisms. Some authors found that carboxyfluorescein derivatives have in vitro a high degree of self-quenching at the high intracellular concentration (1-3 mM) used in the experiments described here [15]. There exists, however, a controversy about this finding [24]. In our experiments we found no evidence for high self-quenching. For example, at the end of calibration experiments, when the external and cytoplasmic pH were high the probe was released from the cytoplasm by addition of Triton X-100. Upon this addition fluorescence did not increase by more than 1020%, although the probe was diluted approx. 1800-fold, to far below the concentration where self-quenching was implied to become of importance, Such an increase might well be due to other quenching mechanisms, e.g., interaction with proteins [t5]. Applicability ot the method for loading BCECF in other bacterial species is mainly dependent on their
ability to resist a transient low pH internally 5 or externally 2-3. The fermentative organisms E. [aeca/is and L, lactis can survice, but not grow, at a relatively low pH. These organisms still showed metabolism even when the acid shock was not optimized to a lowest HCI load. However, E. coli was more sensitive to the acid shock, and after high acid loads it could no longer build up a pH gradient. The fast pH changes shown in Figs. 4 and 5 are probably difficult or impossible to observe by other methods currently used for measuring pH in bacteria. NMR techniques have a practical time resolution limit because of the time needed to record a spectrum, while distribution methods are limited by probe diffusion, selective electrode response or time needed for separation of cells and medium. The method for measuring cytoplasmic pH described here may not only provide further evidence for regulation mechanisms studied by other methods, but can additionally provide new information about proton and proton-linked transport, Because of its high time resolution and sensitivity, it makes the study of transient and minimal phenomena easy and feasible.
Acknowledgement This study was financially supported by the Netherlands Organization for Scientific Research (NWO),
References 1 Booth. I.R. (1985) Microbiol. Rev. 49, 359-378. 2 Krulwich. T.A. and Guffanti, A.A. (1983) Adv. Microb. Physiol. 24, 1015-1022. 3 Poolman, B., Driessen. A.J.M. and Konings. W.N. (1987) Microbiol. Rev, 51,498-508. 4 Nicolay. K.. Lolkerna, J., Helllngwerf. K.J., Kapteln, R. and Konings. W.N. (19811 FEBS Lett. 123, 319-323. 5 RoUenberg, it. (1979) Methods Enzymol. 55, 547-569. 6 Kashket, E.R. (1985) Annu. Rev. Microbiol. 39, 219-242, 7 Sheehter. E, Letellier, L, and Simons. E,R. (1982) FEBS Letl. 139, 121-124. 8 Tsujimoto, K, Semadeni, M., Hufleit, M. and Packer, L. (1988) Biochem, Biophys. Res. Commvn. 155, 123-129. 9 Tsien. R.Y. and Poenie, M. (1986) Trends Biochem. Sci,, II, 450-459. 1(I Rink, T.J., Tsicn, R.Y. and Pozzan, T. (19S2) J. Cell Biol. 95, 189-196.
11 De Man, J.C.. Ragosa, M. and Sharpe, M.E. (1960) J. Appl. Bacteriol. 23, 130-135, 12 Poolman, B., Hellingwerf, KJ. and Konings, W.N. (1987) J. Baeteriol. t69, 2272-2276. 13 Poolman, B., Driessen, A.J.M. and Konings, W.N, (1987) J. Baeteriol. 169, 5597-5604. 14 Kobayashi, H. (1985),I. Biol. Chem. 260, 72-76. 15 Graber, ML., DiLillo, D.C.. Friedman, B.L. and Pastoria-Munoz, E. (1986) Anal. Biochem. 156, 202-212, 16 Poolman, B., Nijssen, R.M.J. and Konings, W.N, (1987) J, Bacteriol, 169. 5373-5378. 17 Salmond, C.C., Kroll, R.G, and Booth, I.R. (1984) J. Gen. Microbiol. 130, 2845-2850.
$3 18 Schraa. G., Belhe, B.M.. van Nccrvcn, A.R.W.. van den Twcel, W.J.J.. vail tier Wende, E. and Zehnder, A.J.B. (1987) A. van Leeuweahoek 53. I59- t70. 19 Sikkema, J. and de Bont, J.A.M. (1901) itiodegradalioil, in r, re~. 20 Neyfakh, A.A., Bidnenko, V.E. and Chcn, LB. (ItJ~l) Prec. N,dl. Acad. Sci, USA 88, 4781-4785. 21 Allen. C.N., llarpur. E.S., Gray, T,J.B., Simmons, N.L. and tlirst, B.H. (1900) Biochem, Biophys. Re.';. (_'ommun, 172, 262-267.
22 Noel. J.. Tt:jedt~r. A., Vinay, P. and Laprade, R. (1089) Renal Ph.~siol. Biochcm. 12, 371-3,..;7. 23 Bakkcr, E.P. (Iqt~fl) FEMS Microbiol. Re,,. 75. 31t~-334. 24 Weinstcin. J.N.. Yoshikami. S., I-tenkarl, P., Blumenthal. R, and t t~gins, W.A. (1~)77~ Science 195.4St)-402.
75
~) 1~,~91Elsevier Science Publishers B.V. All rights reserved 03114-4t65/ql/$03.5|1
BBAGEN 23613
Continuous measurement of the cytoplasmic pH in Lactococcus lactis with a fluorescent pH indicator D o u w e M o l e n a a r , T j a k k o A b e e a n d W i i N. K o n i n g s D~7~artl~ienl o]' .'~ticrobi~h~.'. Unit't'r.~ilv o1"(ironin.~cn~ t hJrcn ( Th,' Nt'therland.~ )
(Received S July I~J()l )
Key words: Intraccllular pll; pH rcgulatam, Fluorescent probe, t~'arboxyfluorcsccin:Multidrug resistance;~1,. h~ctis)
The cytoplasmic pH of Lactococcus lactis was studied with the fluorescent pH indicator 2',7'-bis-(2-carboxyethylJ-.~ (and-6)-carboxyfluorescein (BCECF). A novel method was applied for loading bacterial cells with BCECF, whicl~ consists of briefly treating a dense cell suspension with acid in the presence of the probe. This results in a pH gradient, which drives accumulation of the probe in the cytoplasm. After neutralization the probe was wet! retained in cells stored on ice. BCECF-Ioaded cells were metabolically active, and were able to genera*.e a pH gradient upon energization. The probe leaks out slowly at elevated temperatures. Effiux is stimulated upon energization of the cells, and is most likely catalyzed by an active transport system, it is a first-order process, and the rate constant could be deduced from the decrease of the fluorescence signal in periods of constant intracellular pH. This allowed a correction of the fluorescence signal for effiux of the probe. After calibration the cytoplasmic pH could be calculated from elllux-corrected fluorescence traces.
Introduction
The cytoplasmic pH is an important factor in regulating enzyme activity and cell function, and the mechanisms involved in the regulation of cytoplasmic pH are studied intensively [1-3]. The only generally applicable methods for measuring intracellular p l t in bacteria are nuclear magnetic resonance techn;ques [4] or analysis of the distribution of (radiolabeleu)weak acid or weak bases (for reviews see Refs. 5 and 6). These methods have a limited time resolution, and can only be used to measure the internal pH under steady-state conditions or when the cytoplasmic pH changes slowly. N M R techniques have the additional disadvantage that high cell densities are required, which can cause diffi.
Abbreviations: Bt'ECF, 2',7'-bis-(2-carboxyethyl)-5(and-6)-carboxyfluorescein; CCCP, carbonyl cyanide m-chlorophcnylhydraztme: pyranine, 8-hydroxy-l,3,6-pyrenetrisulhmicacid: Mes, 2-(N-mt~rpholinok,~thanesulfi)nic acid; Mops, 3-(N-morpht~iint~)propancsulfimic acid; Hepes, N-2 hydrt~xyethylpiperazine-N'-2-ethanesulfi)nic acid: Tricine. N-tris(hydroxyme|hytlmethylglycinc:bislris propane, 1,3bis[tris(hydroxymethyl)methylamint~lpropane. Correspondence: W.N. Konings, Departmenl of Microbi~h~gy,University of Groningen, Kerklaan 3t), 9751 NN II:iren, The Nether~ lands.
culties in controlling the external medium composition (for example pH and oxygen concentration). Fluorescent pH indicators have occasionally been used for bacteria, but either the methods used for loading he cells with the probe were not generally applicable, or the mca,~urcments were difficult to interpret due to high eftlex of the probe [7,8]. The use of a fluorescent probe as a pH indicator has obvious potential advantages. Most indicators used so far have no known serious negative effects on cellular metabolism and measurements can be performed with high time and pH resolution [9], Loading of cells with the fluorescent indicator BCECF can b : achieved by adding the membrane permeable acetoxymethyl estc~ 'ff the ind:. ~tor to cells. The indicator will be trapped intracellulary upon hydrolysis by the action o f esterases. This method finds wide application in the study of pH control and transport systems in eucaryotic cells [9,10]. However, in bacterial cells the method had a narrow perspective for applicability. In H a l o b a c t e r i u m h a l o h i u m the probe BCECF ~as only usable after 4 days incubation of cclls with the acctoxymethyi ester [8]. A carboxyfluorescein ester has been used in Excht, richht coli, but its use is limited since a cumbersome correction has to be madc for the high rate of efflux of carboxyfluoresccin and hydrolysis of ester bonds [7].
76 In the studies described in this article a simple method has been used for loading bacterial cells with BCECF directly. Cells were incubated with the fluorescent probe and 50-1(10 mM HCI for 5 min at room temperature, and then brought to pH 7 with excess buffer. After this treatment, BCECF is located intracellularly. Thc cells are viable and retain BCECF well when stored on ice. Measuremcnt of the fluorescence signal can bc performed with high time resolution. The probe cffluxes at elevated temperatures, but the process can be well characterized, and the fluorescence signal can be casiIy corrected for efflux. Correlation of the signal to cytoplasmic pH can be done by a proper calibration method. In the pK range of BCECF (7-7.5) this method has a high sensitivity towards changes in pH. The method has been applicd to study pH regulation in the Gram-positive fermentative bacterium Lactococcus lactis, it was also applied in E. col(, l.~hodohacter sphaeroides. Arthrot,acter and Acinetohatter species (Molcnaar, D., unpublishcd results and Sikkcma, J., personal communication). Materials and Methods
Strabr and growth condition L. lactis, subspecies lactis MI,3, was grown anaerobically at 2 9 ° C on a modified MRS medium [I I], containing per liter 10 g tryptonc, 10 g 'Lab-Lemco' powder (Oxoid, Basingstokc, U.K.), 5 g yeast extract, 2 g K : H P O 4, 0.2 g MgSO~, 35 mg MnSO4, 0,5 ml Tween8(1, adjusted to pH 6.3 with HCI. The medium was supplemented with separately sterilized 2% lactose, or 2% galactose and 20 mM arginine. Loading of cells with BCECF Ceils were grown in 100 or 200 ml medium and harvested in the logarithmic phase of growth, at an absorbance at 660 nm of approx. 0.6. They were washed twice in 50 mM potassium phosphate, pH 7 (hereafter called "buffer'), and resuspended in 2.5 or 5 ml of this buffer. The protein concentration of this suspension was approx. 5 mg/ml. The suspension was kept on ice until further use. An optimal acid shock requires a distinct quantity of acid to be added to the cells. Enough acid has to be added to get sufficient loading, without damaging the cells. For optimization six tubes were prepared. Aliquots of 200/,1,1 of the suspension were centrifiiged tk~r 15 s in 1 ml tubes in an Eppendorf centrifuge and rcsuspendcd in 2[) tzl buffer. 1 tzl containing 5 m g / m l ( 1(1 mM) BCECF stock solution was ~dded. Droplets of 1.5 to 4 tst (intervals of 0.5 p,l) of 0.5 M HCI were adhercd to the walls of these tubes. After mixing of the droplets with the cell suspension, the tubes were left at room temperature tor 5 min. 1 mt buffer was then added, the tubes wcrc shaken and cells were spun
down in an Eppendorf centrifug:~ for 15 s. The pellet which contained a fair amount ol 3 C E C F (judged by the intensity of its yellow coiour) af er treatment with the smallest amount of acid, was used for further experiments (usually 2 to 3 p.I HCI treatments). These cells contain I to 3 mM BCECF (see Results). The cells were washed four times with 1 ml buffer by vortexing and centrifugation, and finally rcsuspended in 200/.LI buffer and stored on ice.
Measurenwnt of fluorescence Spectra of intracellular and extraccllular BCECF were recorded. Excitation (emission at 525 nm) and emission (excitation at 500 nm) spectra of BCECF were measured with slit widths of 5 nm. Optima were found at 502 nm in both excitation spectra, and at 521 and 523 nm in the emission spectra of extra- and intracellular BCECF, respectively. The isosbestic (pH invariable) point which can be found at 440 nm in the excitation spectra of free BCECF was not found in the spectra of intracellularly located BCECF. Fluorescence measurements with probe-loaded cells were performed in a 3-mi cuvette containing 3 ml of the desired buffer and 5 to I0 p,I cells, depending upon the concentration of BCECF. The suspension was stirred and thermostated. The excitation and emission monochromator wavelengths were 502 and 525 nm with slit widths of 5 and 15 nm, respectively. The fluorescence signal was averaged over time intervals of 0.5 or ! s. L f F t - assay A L. lactis cell suspension of 100 /zl, loaded with BCECF, was diluted in a tube thermostated at 3 0 ° C containing 2 ml 50 mM potassium phosphate of the desired pH. A time series was made by withdrawing aliquots of 2(~) ~l. Cells were spun down immediately for 3 min in an Eppendorf centrifuge, and 180 ~l of the supernatav, was carefully pipetted off. This 3 min centrifugation is short compared to the time-scale in which leakage occurs, and gross errors are not expected. To obtain a 'total fluorescence' value the suspensions were permeabilized at the end of the experiment by incubation with 0.1% Triton X-100 for an additional 31) rain. The supernatant samples were diluted 60-fold in 50 mM bistris propane (pH 9.1). The fluorescence of this diluted sample was measured with the monochromators set to 502 nm excitation and 525 nm emission, it was assumed that this fluorescence was proportional to the BCECF concentration. The fluorescence of the supernatant at the start was not zero, but approx. 10% of the fluorescence in the 'total BCECF' sample, probably duc to leak during or after washing or to lysis. This value was subtracted as a background from all samples. For calculation of intracellular concentrations of BCECF an intracellular vol-
77 umc of 2.8 p.I/mg protein was used, as determined previously [ 12].
Materials attd bzstrumentation BCECF was obtained from Molecular Probes. Eugene, OR, U.S.A. A stock suspension of 5 m g / m l was made in water, which was brought to approx, pH !1 with KOH, and after a few minutes neutralized with HCi, as was recommended by the supplier. The solution was stored at - 2 1 ) ° C and always kept in the dark. The fluoromcter was a Perkin-Elmcr LS 5(1 with computer controlled data acquisition and storage.
Results Accumulation and retention of BCECF Loading L. lactis with 0.5 mM BCECF cxtraccllularly resulted in an internal concentration of 1 to 3 mM BCECF, depending on the batch of cells. This could bc deduced from the 'total fluorescence' value in cfflux experiments as described below. The 2-6-fold accumulation can be explained if a transient pH gradient is present during the loading procedure. Such a transient pH gradient will lead to the uptake of the protonated species of BCECF which is more membrane permeable than the deprotonated species (BCECF has four carboxyl groups). To get some information about the existence of a transient pH gradient the external pH was recorded with a pH electrode upon the addition of HCI (data not shown). The initial external pH just after the addition of acid was approx. 2, subsequently, the pH increased and stabilized within I h at approx, pH 5. The slow pH increase is probably due to slow proton
!
II
loo %~.
v
influx as a result of the existing pH gradient across the membrane. In the normal loading procedure excess buffer was added 5 rain after the shock. At this time the external pH was approx. 3. The intracellular pH will then have decreased from approx. 7 to a value intermediate between 7-5 (see also discussion). Thus, a large pH gradient of more than 2 pH units, inside alkaline, existed during the loading procedure. The same technique has been applied succcssfully for loading cells with pyraninc, but the levels of fluorescence obtained with this indicator were much lower than for BCECF. The retention of BCECF in L. lactis was studied under different conditions. BCECF was wcli retained in cells storcd on ice. Therc was no detectable efflux in 2.5 h (Fig. IA, triangles). However, slow efflux of the probe occurred from non-energized ceils at 30°C, which were uncoupled with the protonophore CCCP. This efflux was slightly pH-dependent, and more rapid at lower pH. At pH 6 and 8 the tt/2 values (the time nccdcd for cfflux of 50% of the original probe content), were approx. 75 rain and 110 min, respectively (Fig. IA). Energization of the cells with lactose stimulates the cfflux of the probe dramatically (Fig. I B). Rapid efflux was also observed upon the addition of lactose to the cells and rcmaincd rapid when the proton motive force was decreased by the addition of valinomycin and nigericin. The applied valinomycin concentration was sufficient to completely abolish the electrical membrane gradient, as was shown previously by Poolman et al. [12] and was checked with a TPP ~ electrode (data not shown). The addition of nigericin can introduce a
I
?001
~v
|
I
~
"'
It'"
1
t
A
B
80
"6 6o c
"6
h
~
Ld
L=J
T
20
~~
o
lactose 0
"
0
50
""
'
1O0
Time (rain)
150
0
'
"
'
'
'
t0
20
30
40
50
60
70
Time (rain)
Fig, I, Efflux of BCECF, measured with the efflux assay described in Materials and Methods. (A) (.:ells were incubated on ice in 51) mM potassium phosphate pH 7 (v) or at 30 ° C with 51)/~M CCCP in 50 mM potassium phosphate pl[ 6 (o) or pH 8 (o). The vertical axis has a logarithmicscale. (B) Cells were incubated at 311oC in 5(I mM potassium phosphate pl t 6 (el and pH 8 1o1 without additions,or ptl 7 with 2 p.M nigericinand 1/zM valinomycinadded ( v 1. At the time indicatedcells were energizedby the additionof 2.8 mM lactose. The vcrticat axis has a logarithmicscale.
7g smaI1 inverted pH gradient (inside acid) most likely as a result of high internal K + concentration [12] (see also Figs. 4 and 5). The same observation of energy stimulated efflux was made when cells were energized with argininc via the arginine deiminase route [13] (data not shown). Also under energized conditions efflux was slightly dependent on the external pH. Semi-logarithmic plots of the fluorescence decrease yield straight Iines (Fig. I B), indicating that efflux occurred by a first-order process. The t=/z values in the energized state were 5.7 rain at pH 6 and 1 1.2 min at pH 8.
250
A
b
200 5 150
g ~oo
5O
Correction of fluorescence for BCECF effltL,; Efflux of the probe complicates the interpretation of changes in the fluorescence signal. The changes measured arc pa)tty due to efflux and partly to changes in the cytoplasmic pH. For analysis of the fluorescence signal the assumption is madc that there exist two homogeneous BCECF populations: a cytoplasmic and an cxtraccltular population. Both populations contribute to the measurcd fluorescence signal. To analyse this total fluorescencc the fl)llowing parameters are introduced: (i) ['~, the fluorcsccncc measured if all BCECF werc located intraccllularly: (it) F0, the fluorescence measured if till BCECF wcrc located extracellularly. This second situation occurs in fitct after a long tin)e, when all the probe has effluxed. In the experiments described bcIow the probe is located both intracellularly and extracellularly. Consequently only a fraction o, corresponding to the fraction of intracellularly located probe, of the hypothetical signal F~ is measured, in addition to a fraction (1 - q ) o f the signal F,,. Together they constitute the total signal F at a certain point m time: [ ' = q'F, +(1 - q).l.:,
(I)
It was found in the experiments described above that efflux of the probe is ~ first-order process. Therefore, the fraction of probe located intracellularly (q) decreases exponentially with time (t): q(t)=qn,
e
(2)
f,t
in which q0 represents the fraction of probe located internally at t = 0, and k is the first-order rate constant for efflux. The measured fluorescence F is a function of time (expressed as F(t)), and so is the fluorescence of cytoplasmic probe F i if the cytoplasmic pH changes in time (Fi(t)). Fo will be constant, since in the experiments described the external pH does not change. SubstitutinL" Eqn. 2 in Eqn. 1 then yields: F(t) = F,,+ q,)'e-~"( F i ( t ) - F~)) From this equation
(3)
it f o l l o w s t h a t u n d e r c o n d i t i o n s o f
Ioc
n)..,
T
6.0
5.5
0
1'
val ni
org l
I
5
10
1't
otg 15
TTnig
T
vol nig
arg
I
I.
I
5
10
15
val
,.5
10
15
Time (rain) Fig. 5. ('ylopl:tsmic pH traces of arginine energized L. hwtis al different external pHs. Conditions, buffers, additions, and concentrations were Ihc same as in Fig. 5. pHs were 6 (A) 7 (B), and 8 (C). At the times indicated 1.3 mM arginine ttarg) was added.
TABLE I
Comparisota of tTtOplaxmic pH of L. lactis umh'r dtfftwcnt conditttm.~ th'termbted b)" the distribution of raditdatwlcd benzoate ocid or methylamine atut b)' B('ECF fluoresccm'e Energy
Arginine External pH:
Cytoplasmic pH determined by distribution " Cytoplasmic pH det~,rmincd by BCECF fluort:s¢cncc
I..actosc
6
7
8
0
7
8
7.1
7.3
7.8
7.3
7.6
7.9
7.35
7.61) 7.85
7.05
7.75
7.85
" Data from Rcf, 16. The buffers and condifion,; wcrc the same as used in Fig. 4.
increase of the pH gradient upon dissipation of the (larger) electrical potential. The mechanism bchind these phenomena is currently under investigation. The steady-state pH values in energized cells were compared to values measured earlier [16] under the same experimental conditions with the distribution of radiolabeled benzoic acid and methylamine (Table !). In general, the values measured by BCECF fluorescence were higher than those measured by the distribution method, at least for pH gradients alkaline internally, which were measured with benzoic acid. Witil both methods a higher pH gradient was found in lactose energized cells compared to arginine energized cells. However, with BCECF fluorescence the difference tl~e differences found in cytoplasmic oH were smallcr at the different external pHs, indicating a more stringent regulation of cytoplasmic pH than found previously. These differences could be due to an overe~,~timation of cytoplasmic pH by BCECF fluorescence, but it could also be the result of an underestimation of pH gradients by benzoic acid accumulation or be due to an inhibition of the build up of the pH gradient by accumulated benzoic acid [1,17]. Discussion
This paper describes a new and simple method for loading bacterial cells with the fluorescent pH indicator BCECF for measuring quantitatively the cytoplasmic pH. It can be applied both in (3ram-positive and Gram-negative organisms. BCECF is accumulated intracellularly during a low pH shock. During this shock the external pH drops to values where all carbo~lic groups of BCECF are likely to be protonated. This fully protonated neutral molecule is probably more membrane permeable than the negatively charged species, and diffuses inward across the membrane. The slow rise in external pH during the acid shock points to the existence of a proton gradient across the membrane. This proton gradient will form a driving force for the accumulation of BCECF, since the equilibrium is shifted to the deprc~tonated species intracellularly.
BCECF i,: we!! retained when ccll:~ arc st~-~..-.d on ice. and at ctevatcd temperatures efflux is slow in the non-energized state. The stimulation of efflux upon energization of the cells makes this process intriguing. especially since this stimulation is also observed in uncoupled cells. If effiux were a passive, non-carriermediated process non-carrier-mediated process (leak), the pH gradient would certainly not stimulate efflux because an increase of the internal pH will lead to a further di.,;~ociation and t h u s to a more negatively charged form of the probe, which will be less membrane permeable. An inverted pH gradient, generated upon addition of nigericin, could possibly stimulate BCECF efflux. However, the observation that a positive pH gradient alone (by energizing cells in the presence of valinomycin) did not drive uptake and did not prevent cfflux argues against this explanation (Molenaar. D.. unpublished data and Fig. 2). An electrical gradient, inside negative, can. in principle, pull the negatively charged probe across the membrane to the outside and can stimulate leak. Yet stimulation of effiux was also observed in cells in the absence of an electrical gradient, and seems to be dependent on the presence of other forms of metabolic energy. There are also differences in efflux rate observed between species. Co~.nebacterium sp. C125 [18] showed a very high effiux rate. whereas Arthrohacter sp. T2 [19] had a very low efflux rate (Sikkema, J., personal communication). Also within species such differences can be observed depending on the mode of energization (Molenaar, D., unpublished results). These results could indicate the presence of an active extrusion system for BCECF driven hy metabolic energy in bacteria. In a forthcoming article we will present more evidence for such a system, and show that it has properties different from the recently cloned protonmotive force driven bacterial multi-drug resistence system [20]. Energy dependent efflux of BCECF is also in line with the recent suggestion for a system that catalyzes this process in epithelia[ ,:ells [21]. BCECF effiux is a first-order process whicla could indicate that this export system is working far below the affinity constant for BCECF. This K, will be much larger than 3 raM, suggesting that the transport system is not very specific for the probe. Although the mechanism of efflux of the probe is still unclear, it is phenomenologieally a simple and well characterized process. Therefore, correction of the fluorescence signal for efflux can be easily performed. However. as can be deduced from Eqn. 4 the error in the corrected signal will be proportional to errors in qo and F,,, and exponentially related to errors in k. The error also increases exponentially in time. This accumulation of errors can introduce large noise and systematic errors in the calculated pH, especially when k is high. A method to correct for probe effiux from suspensions of eukaryotic cells has been developed
82 which migh prevent systematic errors by circumventing the need for precise characterization of the efflux process [22]. Thc~c authors used anion exchanger beads to bind external probe in the cuvette outside the light path, so that only the intracellularly located probe is observed. A concentration independent signal can then be obtained by continuous excitation ratioing. We have not yet made attempts in this direction. For relating the fluorescence signal to the cytoplasmic pH it is of utmost importance to have a reliable calibration method. The ionophore nigcricin is used for this purpose in eukaryotic cells, sometimes in combination with valinomycin, It is assumed that during calibration the cytoplasmic pH equals the external pH. Since nigcricin exchanges potassium ions electroneutrally with proton:;, it is not immediately obvious why internal and external pH should be equal. As can be seen in Figs. 5 and 6 (scc also Rcf. 12) in energized cells these pH values arc indeed not equal. However, during the build-up of a pH gradient, the extrusion of positive charge on the proton is usually compensated for by the uptake of pot;,ssium ions. to conserve electroneutrality. Potassium ions are accumulated to concentrations of several hundred millimolars, reflecting the high buffer capacity of the cytoplasm [i,23]. This pool is again freely exchangeable for protons upon the addition of nigericin. The pK upshift tound for intracellular BCECF relative to the free probe exists both at low and high potassium concentrations, suggesting that it is not due to a Donnan potential. The (apparent) pK downshift of BCECF at high potassium concentration relative to low potassium concentration can be caused by differential effects of ionic strength on the activity coefficients of the protonated and deprotonated probe species. Such effects can in principle cause artifacts when the intracetlular ionic strength changes significantly during an experiment. Other potential sources of artifacts are fluorescence quenching mechanisms. Some authors found that carboxyfluorescein derivatives have in vitro a high degree of self-quenching at the high intracellular concentration (1-3 mM) used in the experiments described here [15]. There exists, however, a controversy about this finding [24]. In our experiments we found no evidence for high self-quenching. For example, at the end of calibration experiments, when the external and cytoplasmic pH were high the probe was released from the cytoplasm by addition of Triton X-100. Upon this addition fluorescence did not increase by more than 1020%, although the probe was diluted approx. 1800-fold, to far below the concentration where self-quenching was implied to become of importance, Such an increase might well be due to other quenching mechanisms, e.g., interaction with proteins [t5]. Applicability ot the method for loading BCECF in other bacterial species is mainly dependent on their
ability to resist a transient low pH internally 5 or externally 2-3. The fermentative organisms E. [aeca/is and L, lactis can survice, but not grow, at a relatively low pH. These organisms still showed metabolism even when the acid shock was not optimized to a lowest HCI load. However, E. coli was more sensitive to the acid shock, and after high acid loads it could no longer build up a pH gradient. The fast pH changes shown in Figs. 4 and 5 are probably difficult or impossible to observe by other methods currently used for measuring pH in bacteria. NMR techniques have a practical time resolution limit because of the time needed to record a spectrum, while distribution methods are limited by probe diffusion, selective electrode response or time needed for separation of cells and medium. The method for measuring cytoplasmic pH described here may not only provide further evidence for regulation mechanisms studied by other methods, but can additionally provide new information about proton and proton-linked transport, Because of its high time resolution and sensitivity, it makes the study of transient and minimal phenomena easy and feasible.
Acknowledgement This study was financially supported by the Netherlands Organization for Scientific Research (NWO),
References 1 Booth. I.R. (1985) Microbiol. Rev. 49, 359-378. 2 Krulwich. T.A. and Guffanti, A.A. (1983) Adv. Microb. Physiol. 24, 1015-1022. 3 Poolman, B., Driessen. A.J.M. and Konings. W.N. (1987) Microbiol. Rev, 51,498-508. 4 Nicolay. K.. Lolkerna, J., Helllngwerf. K.J., Kapteln, R. and Konings. W.N. (19811 FEBS Lett. 123, 319-323. 5 RoUenberg, it. (1979) Methods Enzymol. 55, 547-569. 6 Kashket, E.R. (1985) Annu. Rev. Microbiol. 39, 219-242, 7 Sheehter. E, Letellier, L, and Simons. E,R. (1982) FEBS Letl. 139, 121-124. 8 Tsujimoto, K, Semadeni, M., Hufleit, M. and Packer, L. (1988) Biochem, Biophys. Res. Commvn. 155, 123-129. 9 Tsien. R.Y. and Poenie, M. (1986) Trends Biochem. Sci,, II, 450-459. 1(I Rink, T.J., Tsicn, R.Y. and Pozzan, T. (19S2) J. Cell Biol. 95, 189-196.
11 De Man, J.C.. Ragosa, M. and Sharpe, M.E. (1960) J. Appl. Bacteriol. 23, 130-135, 12 Poolman, B., Hellingwerf, KJ. and Konings, W.N. (1987) J. Baeteriol. t69, 2272-2276. 13 Poolman, B., Driessen, A.J.M. and Konings, W.N, (1987) J. Baeteriol. 169, 5597-5604. 14 Kobayashi, H. (1985),I. Biol. Chem. 260, 72-76. 15 Graber, ML., DiLillo, D.C.. Friedman, B.L. and Pastoria-Munoz, E. (1986) Anal. Biochem. 156, 202-212, 16 Poolman, B., Nijssen, R.M.J. and Konings, W.N, (1987) J, Bacteriol, 169. 5373-5378. 17 Salmond, C.C., Kroll, R.G, and Booth, I.R. (1984) J. Gen. Microbiol. 130, 2845-2850.
$3 18 Schraa. G., Belhe, B.M.. van Nccrvcn, A.R.W.. van den Twcel, W.J.J.. vail tier Wende, E. and Zehnder, A.J.B. (1987) A. van Leeuweahoek 53. I59- t70. 19 Sikkema, J. and de Bont, J.A.M. (1901) itiodegradalioil, in r, re~. 20 Neyfakh, A.A., Bidnenko, V.E. and Chcn, LB. (ItJ~l) Prec. N,dl. Acad. Sci, USA 88, 4781-4785. 21 Allen. C.N., llarpur. E.S., Gray, T,J.B., Simmons, N.L. and tlirst, B.H. (1900) Biochem, Biophys. Re.';. (_'ommun, 172, 262-267.
22 Noel. J.. Tt:jedt~r. A., Vinay, P. and Laprade, R. (1089) Renal Ph.~siol. Biochcm. 12, 371-3,..;7. 23 Bakkcr, E.P. (Iqt~fl) FEMS Microbiol. Re,,. 75. 31t~-334. 24 Weinstcin. J.N.. Yoshikami. S., I-tenkarl, P., Blumenthal. R, and t t~gins, W.A. (1~)77~ Science 195.4St)-402.
