EEt

R*AU

R

aca4fmtaotsdgeascramc

oe

tpc

7

1

Archives of Biochemistry and BiophysicsVol. 368, No. 1, August 1, pp. 156–160, 1999Article ID abbi.1999.1275, available online at http://www.idealibrary.com on

scherichia coli Membrane Fluidity as Detected byxcimerization of Dipyrenylpropane: Sensitivity to

he Bacterial Fatty Acid Profile

icardo Mejıa,* M. Carmen Gomez-Eichelmann,† and Marta S. Fernandez*,1

Department of Biochemistry, Centro de Investigacion y de Estudios Avanzados del Instituto Politecnico Nacional,.P. 14-740, 07000 Mexico D.F.; and †Department of Molecular Biology, Instituto de Investigaciones Biomedicas,niversidad Nacional Autonoma de Mexico, A.P. 70-228, 04510 Mexico D.F.

eceived March 9, 1999, and in revised form April 30, 1999

bIlttcre9oatavmcopumcaTrogpcmmmdmfl

A coordinated study of membrane fluidity and fattycid composition has been carried out in Escherichiaoli W3110. The lipid acyl chain profile of the bacteria,ltered by growing cells in steady state at 30, 37, 42, or5°C, was determined by gas chromatography of theatty acid methyl esters. In parallel experiments, total

embranes obtained from cells of the above-men-ioned cultures were labeled with dipyrenylpropanend their relative fluidity was measured on the basisf the excimer to monomer fluorescence intensity ra-io of the fluorophore. It has been found that, at con-tant assay temperature, fluidity determined withipyrenylpropane decreases gradually with therowth temperature increment, from 30 to 45°C. Inter-stingly, when fatty acid composition is taken intoccount, fluidity increases linearly in the range undertudy, with the proportion of unsaturated fatty acylhains, both variables being highly correlated (0.924 <2 < 0.996). Our results show that dipyrenylpropane is

reliable and quantitative indicator of changes inembrane fluidity, driven by modifications in the acyl

hain composition of bacterial lipids. © 1999 Academic Press

Key Words: Escherichia coli; membrane fluidity; flu-rescence; fatty acid composition; dipyrenylpropanexcimerization.

It is well known that the relative proportions of thehree main phospholipids present in Escherichia coli,hosphatidylethanolamine, phosphatidylglycerol, andardiolipin, are hardly affected, within a given strain,

1

fTo whom correspondence should be addressed. Fax: (525) 747

083. E-mail: msfernandez@mail.cinvestav.mx.

56

y the temperature at which cultures are grown (1–3).n contrast, the fatty acid composition of membraneipids, is strongly modified by manipulating the growthemperature of the bacteria: the lower the tempera-ure, the higher the fraction of unsaturated fatty acylhains (1–6). Changes in fatty acid composition areeflected in modifications of cell surface physical prop-rties (1, 7), such as phase transition temperature (3, 8,) and membrane microviscosity (10, 11). The functionf the bacterial membrane response to thermal alter-tions can be rationalized on the basis of the advan-ages that for an ectothermic organism, arise from thedjustment of lipid bilayer fluidity in the presence ofariable environmental temperature. Such adjustmentay optimize the barrier functions and permeability

haracteristics of the cell envelope (1, 7, 10, 12). It isbvious that modifications in the fatty acid compositionattern are very important for the homeoviscous reg-lation of membranes. Despite this, there is a lack ofeasurements of the precise effects of quantitative

hanges in the proportions of saturated or unsaturatedcyl chains, on the fluidity of bacterial membranes.he aim of the present work is to investigate suchelationships. With this purpose, the fatty acid profilef the bacterial membranes is naturally altered byrowing E. coli cells in steady state at different tem-eratures: 30, 37, 42, and 45°C. The precise fatty acidompositions of bacteria from these cultures are deter-ined by gas chromatography. In coordinated experi-ents, membranes isolated from cells of the above-entioned cultures are labeled with the fluorophore

ipyrenylpropane, and the intramolecular excimer for-ation by this probe (13–15) is used to calculate a

uidity parameter (11). Fluidity is determined at dif-

erent temperatures and analyzed as function of the

0003-9861/99 $30.00Copyright © 1999 by Academic Press

All rights of reproduction in any form reserved.

gs

M

ainoppM

upfls3

cir7deDpsJurLsamImnecmLb

mm1S7caatebaVtgfe4

lptmeaccm(mcatmtmWowom

R

dTlfrtflTwa

I(rsfledsautta

pl

157MEMBRANE FLUIDITY AND BACTERIAL FATTY ACID PROFILE

rowth temperature, as well as of the fraction of un-aturated fatty acyl chains.

ATERIALS AND METHODS

Chemicals. Fatty acid methyl ester standards and the 14% meth-nolic solution of boron trifluoride, were obtained from Sigma Chem-cal Co. Chloroform and methanol Uvasol grade, were from Merck;-pentane and n-hexane were HPLC grade solvents from Baker. Allther chemicals were analytical reagents. The fluorophore 1,3-di(1-yrenyl)propane (DPyP)2 was purchased from Molecular Probes. Tri-le-distilled water further purified through a Milli-Q System fromillipore, Inc., was employed.Microbiological procedures. Wild-type E. coli K-12 (W3110) was

sed throughout. After diluting an overnight culture 1:50 into 10 mlrewarmed Luria broth (16), cells were grown in 125-ml Erlenmeyerasks with shaking (180 rpm) to mid-log phase (OD550 of 0.4, corre-ponding to approx 1 3 108 cells ml21) at the desired temperature, 30,7, 42, or 45°C.Membrane fluidity determinations. Total membranes of E. coli

ells were obtained as previously described (11). Fluorescence label-ng of E. coli membranes was performed as follows. Membranes wereesuspended in the adequate volume of 10 mM Tris–HCl buffer, pH.8, to give 0.14 mM phospholipid. To 2.5 ml of the membraneispersion subjected to mild vortexing, 8 mL of 31.25 mM DPyP inthanol (Uvasol) was quickly injected. The final concentration ofPyP was 0.1 mM, such that the molar ratio of fluorophore to phos-holipid was 1:1400. After the DPyP injection, membrane disper-ions were incubated in the cold room with magnetic stirring for 24 h.ust before use, samples were mildly sonicated for 1 min in anltrasonic bath at room temperature. Fluorescence of DPyP incorpo-ated into membranes was measured in a computer-controlledS50B Perkin–Elmer spectrofluorometer equipped with a thermo-tated, magnetically stirred four-position sample chamber. Temper-ture was measured directly in the quartz cuvette using an IT18icroprobe attached to a digital thermometer (Bat 8) from Bailey

nstruments. The fluorophore was excited at 329 nm and the mono-er and excimer fluorescence intensities were read at 379 and 480

m, respectively, using a program that automatically changes themission monochromator wavelength in a few seconds. Prior to re-ording a fluorescence value, samples were maintained at least for 5in at the desired temperature, which was approached upward.ight scattering corrections, obtained from readings of adequatelanks, were applied to all fluorescence values.Fatty acid analysis. E. coli lipids were extracted according to theethods of Bligh and Dyer (17) and Ames (18) to which some minorodifications were introduced (19). Cells were centrifuged at 4°C for

0 min at 6000 rpm in a Sorvall SS-3 centrifuge equipped with aS34 rotor. The pellets were washed once with 10 mM Tris–HCl, pH.8, and were employed directly to obtain membrane lipids since noytoplasmic, extracellular, or periplasmic lipid-containing structuresre known in E. coli (2, 20). This is similar to the approach of otheruthors (3) who obtain from such pellets the so-called IO lipids, i.e.,he lipids from total (inner plus outer) membranes. Total lipids werextracted with a mixture of methanol, chloroform, and water (2:1:0.8,y volume). After adding the solvent mixture, the samples, kept onn ice bath, were subjected to vigorous shaking in an adapted Bigortexer (Kraft Apparatus, Inc.) for 30 min. To favor further extrac-

ion, samples were homogenized with a Ten-Broeck Pyrex tissuerinder (10 strokes in approx 2 min). Subsequently, 1 vol of chloro-orm and 1 vol of water were added, with brief homogenization afterach addition. The resultant two-phase system was centrifuged at°C for 10 min at 6000 rpm in the Sorvall centrifuge. The chloroform

s2 Abbreviation used: DPyP, 1,3-di(1-pyrenyl)propane.

ayer was collected and evaporated under a nitrogen stream, and thehospholipid content was calculated from phosphorus determina-ions (21). The fatty acid composition of bacterial lipids was deter-ined by gas chromatography after transesterification in the pres-

nce of boron trifluoride and methanol (22). The gas chromatographicnalysis was carried out in a computer controlled Star 3400 gashromatograph from Varian equipped with a 1077 split/splitlessapillary injector and a flame ionization detector (19). Fatty acidethyl esters were separated in a 30-m fused silica megabore

0.53-mm i.d.) DB-5 capillary column, cross-linked with 5% phenyl-ethylpolysiloxane. Ultra highly pure nitrogen was used both as

arrier (8.5 ml/min) and makeup (30 ml/min) gas. The injection portnd detector temperatures were 250 and 280°C, respectively. Ahree-step temperature program was applied to the column: isother-ic at 180°C for 2 min, followed by an increase at a rate of 11°C/min

o a final temperature of 245°C, which was then maintained for 12in. Data were handled by means of the Star Chromatographyorkstation software. Identification and concentration calculations

f the individual acyl chains in each sample were done by comparisonith retention times and peak areas relative to a standard mixturef fatty acid methyl esters. Fatty acid compositions are reported asol%.

ESULTS

W3110 E. coli cells were grown in steady state atifferent temperatures in the range from 30 to 45°C.otal membranes obtained from these cultures were

abeled with DPyP and their fluidity was determinedrom the excimer to monomer fluorescence intensityatio (Ie/Im) of the fluorophore at a certain absoluteemperature, T. As described previously, the relevantuidity parameter employed is [(1/T) 3 (Ie/Im)] (11).he relative membrane fluidity (F) can be estimatedith respect to the fluidity (Fr) of a reference system,ccording to the following expression:

FFr

5

1T f

S Ie

ImD

1Tr

S Ie

ImD

r

. [1]

n the numerator, on the right side of this equation,Ie/Im) is the excimer to monomer fluorescence intensityatio of dipyrenylpropane for the membrane undertudy measured at Tf, the kelvin temperature of theuidity assay. As for the denominator, (Ie/Im)r is thexcimer to monomer fluorescence intensity ratio ofipyrenylpropane in the reference membrane mea-ured at Tr, the reference kelvin temperature. Ourrbitrary reference system of fluidity (Fr) equal tonity has already been defined (11) and corresponds tootal membranes from E. coli grown at 30°C for whichhe fluidity parameter [(1/T) 3 (Ie/Im)] is also measuredt 30°C (Tr 5 303K).Figure 1 shows that at constant measurement tem-

erature, the higher the growth temperature, theower the membrane fluidity measured by DPyP. A

imilar trend is found at any of the measurement tem-

pGracrpcc(ip(rorofltftpras

attcm(pawe

crctifupflta

FWtƒ

}

(sCcwso(

Fscpgs1mRitrmcc

158 MEJIA, GOMEZ-EICHELMANN, AND FERNANDEZ

eratures (tf) employed, i.e., 30, 37, 42, and 45°C.rowth in steady state at several temperatures in the

ange from 30 to 45°C should naturally alter the lipidcyl chain profile of E. coli and, in this way, inducehanges in the fluidity of membranes. Any hypotheticalole of possible modifications of relative levels of phos-holipid species, in the thermal regulation of fluidity,an be ruled out, since the growth temperature of E.oli does not affect the polar head-group composition1–3). The relative proportions, in exponentially grow-ng cells, of phosphatidylethanolamine (75–80 mol%),hosphatidylglycerol (15–20 mol%) and cardiolipin2–6 mol%), are fairly constant for a particular strain,egardless of growth temperature (1–3, 18, 23). Basedn these considerations, the analysis of our fluidityesults is focused on changes in acyl chain profiles. Inrder to establish a quantitative relationship betweenuidity and bacterial fatty acid pattern, it is necessaryo determine the precise fatty acid composition of lipidsrom W3110 E. coli cells grown under identical condi-ions as the ones used to obtain membranes for thehysicochemical study with dipyrenylpropane. Suchesults are shown in Fig. 2. The figure depicts the fatty

IG. 1. Relative membrane fluidity determined at different con-tant temperatures (tf) as a function of the growth temperature of E.oli cells. Fluidity was estimated from excimerization of dipyrenil-ropane incorporated into E. coli membranes, prepared from cellsrowing in mid-log phase at the temperatures indicated in the ab-cissas. The molar ratio of DPyP to membrane phospholipid was:1400. The probe was excited at 329 nm and the excimer (Ie) andonomer (Im) emissions were read at 480 and 379 nm, respectively.elative membrane fluidity (left-hand scale) was calculated accord-

ng to Eq. [1], by taking as reference the fluidity at the growthemperature of membranes from E. coli cells grown at 30°C. Theight-hand scale represents the fluidity parameter used in the nu-erator on the right-side of Eq. [1]. In each panel, the particular

onstant temperature (tf) at which fluidity was determined is indi-ated.

cid profile obtained from gas chromatographic analy-is of lipids from E. coli cells grown in steady state, as

gm

function of the growth temperature. As expected,here is a clear modification of the fatty acid composi-ion with temperature, the trend of variation beingomparable to that reported previously (1–6). Theain changes observed are a decrease in cis-vaccenate

18:1c11) and an increase in the palmitate (16:0) pro-ortions as the growth temperature is raised. There islso a slight reduction in palmitoleic acid (16:1c9),hereas myristoyl (14:0) and stearoyl (18:0) chain lev-ls show only marginal modifications.Our data from gas chromatography are used to cal-

ulate the percentage of unsaturated fatty acids withespect to the total acyl chain content of bacteria fromultures at different temperatures. This percentage isaken as the variable to which relative membrane flu-dity might be expected to respond. As shown in theour panels of Fig. 3, it has been found that in the rangender study and at all measurement temperatures em-loyed, i.e., 30°C (A), 37°C (B), 42°C (C), or 45°C (D),uidity determined with DPyP increases linearly withhe proportion of unsaturates. Changes in the fattycid profile of bacteria may generate significant fluidity

IG. 2. Fatty acid composition of lipids from mid-log phase E. coli3110 cultures as a function of the steady-state growth tempera-

ure. The acyl chains quantified were as follows: Œ, 16:0, palmitoyl;, 16:1c9, palmitoleoyl; F, 18:1c11, cis-vaccenoyl; h, 14:0, myristoyl;, 18:0, stearoyl; E, unidentified. Symbols represent mean values

6S.D.) of 9–17 different determinations from three or four differentamples. Standard deviations smaller than symbols are not shown.ontinuous lines are employed throughout to connect symbols, ex-ept for the datapoints corresponding to unidentified fatty acids, forhich a dotted line is used. For simplicity, results are fitted by

econd-order regression analysis, although such adjustment maynly apply to the range of growth temperatures utilized in this work30 to 45° interval). Data were obtained by temperature-pro-

rammed gas chromatography of the fatty acid methyl esters using aegabore column and a flame ionization detector.

vttgctdcps3

D

rgipsfebii

bttalsm

cbttaattatcpiteopbvfmtggmsmbcwactacifgwmpdbdcd

Ftc(had3ditrpi

159MEMBRANE FLUIDITY AND BACTERIAL FATTY ACID PROFILE

ariations. In Fig. 3A, for instance, it can be observedhat an increase of unsaturated fatty acids from 36.32o 54.53%—obtained by modifying the steady-staterowth temperature from 45 to 30°C—induces an in-rement of the relative membrane fluitidy from 0.657o 1.000 at 30°C. Similar patterns of response are evi-ent in the other panels. In all cases, there is an ex-ellent correlation between fluidity and fatty acid com-osition, as shown by the r2 values of the linear regres-ion analysis of data: 0.9869 in Fig. 3A, 0.9961 in Fig.B, 0.9236 in Fig. 3C, and 0.9245 in Fig. 3D.

ISCUSSION

Membrane fluidity is thought to play an importantole in numerous cellular phenomena of different or-anisms (11, 24–26). In bacteria, there is evidencendicating that changes in membrane fluidity may beart of the control signaling mechanisms of the heathock response (11, 27–29). Intramolecular excimerormation by dipyrenylpropane makes it possible tostimate E. coli membrane fluidity. This procedure haseen employed before in our laboratories to detect thencrease in fluidity induced by the abrupt temperature

IG. 3. Relative membrane fluidity of E. coli at different constantemperatures (tf) as function of the percentage of unsaturated acylhains in bacterial lipids. The values of relative membrane fluidityleft-hand scale) as well as those of the fluidity parameter (right-and scale) were taken from Fig. 1. Fatty acid composition wasltered as shown in Fig. 2, i.e., by growing cells in steady state atifferent temperatures as follows: (h) 45°C, ({) 42°C, (‚) 37°C, (ƒ)0°C. The percentage of unsaturated acyl chains was calculated fromata of the cis-vaccenate and palmitoleate contents of bacterial lip-ds, as represented in Fig. 2. In each panel, the particular constantemperature (tf) at which fluidity was determined is indicated. Cor-elation coefficient (r2) values for the linear regressions used to fitoints are 0.9869 in (A), 0.9961 in (B), 0.9236 in (C), and 0.9245n (D).

ncrement from 30 to 45°C, which we used to elicit the m

acterial heat-shock response. We have also been ableo measure the reduction in fluidity as the response tohermal stress evolves (11). Such a reduction can bettributed to remodeling of the bacterial membraneipids (19), although it has also been suggested thatome heat-shock proteins could be responsible for theembrane fluidity control (27).In the present study, we determined the fluidity of E.

oli membranes that had been physiologically alteredy growing cells in steady state at different tempera-ures. Our dipyrenylpropane measurements showedhat membrane fluidity decreases with growth temper-ture (Fig. 1). These results could have been intuitivelyttributed to changes in fatty acid composition, al-hough a more quantitative analysis was desirable. Inrying to gather data from the literature on the fattycid content of bacteria grown at various tempera-ures, on one hand we corroborated that there is strongonsistency in the tendency of change induced by tem-erature, i.e., in the reduction of palmitate and thencrease in cis-vaccenate proportion, as the growthemperature decreases (1–6). On the other hand, how-ver, we found large differences in the absolute valuesf fatty acid composition at several fixed growth tem-eratures. The discrepancies arise from the peculiaracterial strain employed, the growth medium, theariable aeration, and agitation of cultures, as well asrom the analytical procedure employed for the deter-

ination of the acyl chain content. In addition, not onlyhe growth phase, but also the stage within a certainrowth phase at which cells are harvested, may havereat influence on the results. With these caveats inind, and in order to establish a quantitative relation-

hip between fluidity and fatty acid composition, it wasandatory to carry out a coordinated investigation of

oth variables using the same strain as well as identi-al conditions of growth. The analytical data of Fig. 2ere obtained taking into account these considerationsnd are in agreement with the general tendency ofhange in fatty acid composition reported by other au-hors (1–6). However, these results have the uniquedvantage of coming from the same strain and similarulture conditions as those we employed for the phys-cochemical study of the bacterial membranes. There-ore, these composition patterns for cultures at severalrowth temperatures, could be analyzed in conjunctionith the fluidity results. In order to relate our fluidityeasurements to the acyl chain contents, we used the

roportion of unsaturates as the significant indepen-ent variable. The rational choice of this variable wasased on the inverse effects induced in membrane or-er by an increase of unsaturated or saturated acylhain proportions. In this way, it has been clearlyemonstrated (Fig. 3) that the fluidity of total E. coli

embranes measured with DPyP shows a positive,

lf

mataNpitcqmcuqccpc

A

Taa

R

11

1

1

1

1

1

1

11

2

22

22

2

2

2

2

160 MEJIA, GOMEZ-EICHELMANN, AND FERNANDEZ

inear correlation with the percentage of unsaturatedatty acids in the bacterial lipids.

Previously, we have found that dipyrenylpropane exci-erization detects the disordering of bilayers brought

bout by the abrupt, sudden temperature incrementshat give rise to the heat-shock response and that induce

hyperfluid state of the bacterial membrane (11, 27).ow, we show that the probe also responds to the ex-ected reduction in order brought about by an incrementn the proportion of unsaturates and/or a decrement inhe saturated fatty acyl chain fraction. It can be con-luded that measurements with dipyrenylpropane areuantitatively sensitive to the fatty acid composition ofembranes and that the fluidity variations detected are

onsistent with the predictable opposite effects of unsat-rates and saturates on membrane order. As a conse-uence, this excimer-forming fluorescent probe can beonsidered a reliable and quantitative indicator ofhanges in membrane fluidity induced not only by tem-erature changes but also by modifications in the acylhain composition of lipids.

CKNOWLEDGMENTS

The partial financial support of Consejo Nacional de Ciencia yecnologa de Mexico (CONACyT) through research grants to M.S.F.nd M.C.G.E., as well as a doctoral fellowship to R.M., is gratefullycknowledged.

EFERENCES

1. Haest, C. W. M., De Gier, J., and van Deenen, L. L. M. (1969)Chem. Phys. Lipids 3, 413–417.

2. Cronan, J. E., Jr., and Rock, C. O. (1996) in Escherichia coli andSalmonella: Cellular and Molecular Biology (Neidhardt, F. C.,Curtiss, R., Ingraham, J. L., Lin, E. C. C., Low, K. B., Magasanik,B., Reznikoff, W. S., Riley, M., Schaechter, M., and Umbarger,H. E., Eds.), 2nd ed. Vol. I, pp. 612–636, American Society ofMicrobiology, Washington, DC.

3. Morein, S., Andersson, A. S., Rilfors, L., and Lindblom, G. (1996)J. Biol. Chem. 271, 6801–6809.

4. Marr, A. G., and Ingraham, J. L. (1962) J. Bacteriol. 84, 1260–1267.

5. Lugtenberg, E. J. J., and Peters, R. (1976) Biochim. Biophys.Acta 441, 38–47.

6. Magnuson, K., Jackowski, S., Rock, C. O., and Cronan, J. E., Jr.(1993) Microbiol. Rev. 57, 522–542.

2

7. Davis, M. T. B., and Silbert, D. F. (1974) Biochim. Biophys. Acta373, 224–241.

8. Baldassare, J. J., Rhinehart, D. B., and Silbert, D. F. (1976)Biochemistry 15, 2986–2994.

9. Jackson, M. B., and Cronan, J. E., Jr. (1978) Biochim. Biophys.Acta 512, 472–479.

0. Sinensky, M. (1974) Proc. Natl. Acad. Sci. USA 71, 522–525.1. Mejıa, R., Gomez-Eichelmann, M. C., and Fernandez, M. S.

(1995) Biochim. Biophys. Acta 1239, 195–200.2. Cossins, A. R., and Sinensky, M. (1984) in Physiology of Mem-

brane Fluidity (Shinitzky, M., Ed.), pp. 1–20, CRC Press, BocaRaton, FL.

3. Almeida, L. M., Vaz, W. L. C., Zachariasse, K. A., and Madeira,V. M. C. (1982) Biochemistry 21, 5972–5977.

4. Almeida, L. M., Vaz, W. L. C., Zachariasse, K. A., and Madeira,V. M. C. (1984) Biochemistry 23, 4714–4720.

5. Jurado, A. S., Almeida, L. M., and Madeira, V. M. (1991) Bio-chem. Biophys. Res. Commun. 176, 356–363.

6. Miller, J. H. (1972) in Experiments in Molecular Genetics, ColdSpring Harbor Laboratory, Cold Spring Harbor, NY.

7. Bligh, E. G., and Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37,911–917.

8. Ames, G. F. (1968) J. Bacteriol. 95, 833–843.9. Mejıa, R., Gomez-Eichelmann, M. C., and Fernandez, M. S.

(1999) Biochem. Mol. Biol. Int. 47, 835–844.0. Rock, C. O., Tsay, J. T., Heath, R., and Jackowski, S. (1996) J.

Bacteriol. 178, 5382–5387.1. Bartlett, G. R. (1959) J. Biol. Chem. 234, 466–468.2. Morrison, W. R., and Smith, L. M. (1964) J. Lipid Res. 5, 600–

608.3. Goldfine, H. (1982) Curr. Topics Membr. Transp. 17, 1–43.4. Gyorfy, Z., Horvath, I., Balogh, G., Domonkos, A., Duda, E.,

Maresca, B., and Vigh, L. (1997) Biochem. Biophys. Res. Com-mun. 237, 362–366.

5. Nadiv, O., Shinitzky, M., Manu, H., Hecht, D., Roberts, C. T., Jr.,LeRoith, D., and Zick, Y. (1994) Biochem. J. 298, 443–450.

6. Tardy, F., and Havaux, M. (1997) Biochim. Biophys. Acta 1330,179–193.

7. Torok, Z., Horvath, I., Goloubinoff, P., Kovacs, E., Glatz, A.,Balogh, G., and Vıgh, L. (1997) Proc. Natl. Acad. Sci. USA 94,2192–2197.

8. Horvath, I., Glatz, A., Varvasovszki, V., Torok, Z., Pali, T.,Balogh, G., Kovacs, E., Nadasdi, L., Benko, S., Joo, F., and Vıgh,L. (1998) Proc. Natl. Acad. Sci. USA 95, 3513–3518.

9. Vigh, L., Maresca, B., and Harwood, J. L. (1998) Trends Bio-chem. Sci. 23, 369–374.