ant escherichia coli deletion of in reveals its function ... · pdf fileescherichia coli by...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 hy The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 264, No. 34, Issue of December 5, pp. 20297-20302,1989 Printed in U.S.A.

Deletion of ant in Escherichia coli Reveals Its Function in Adaptation to High Salinity and an Alternative Na'/H' Antiporter System(s)*

(Received for publication, July 3, 1989)

Etana Padan$, Noam Maisler8, Daniel Taglichttll, Rachel KarpelS, and Shimon Schuldinert From the $Department of Microbial and Molecular Ecology, Life Science Institute and the §Department of Molecular Biology, Hadassah Medical School, Hebrew Uniuersity, 91010 Jerusalem, Israel

We have deleted the chromosomal ant gene from Escherichia coli by substitution with the kan gene, which encodes kanamycin resistance. The Aunt strains obtained cannot adapt to high sodium concentrations (700 mM, pH 6.8), which do not affect the wild type. The Na+ sensitivity of Aunt is pH dependent, increasing at alkaline pH. Thus at pH 8.5, 100 mM NaCl retard growth of Aunt with no effect on the wild type. The Aunt strains also cannot challenge the toxic effects of Li+ ions, a substrate of the Na+/H+ antiporter system. However, growth of these strains is normal on carbon sources which require Na+ ions for transport and growth. Moreover, antiporter activity, as measured in everted membrane vesicles, is not significantly impaired.

A detailed analysis of the remaining antiporter activity in a Aunt strain reveals kinetic properties which differ from those displayed by the ant protein: (a) K , for transport of Li+ ions is about 15 times higher and ( b ) the activity is practically independent of intracellular pH. Our results demonstrate the presence of an alternative Na+/H+ antiporter(s) in E. coli, additional to ant system.

All growing cells extrude sodium ions actively and maintain a sodium concentration gradient directed inward (1-5). Main- tenance of low intracellular concentration of Na+ seems a necessity in all living cells and is ascribed to inhibitory effect of high Na+ to essential enzymes. The Na' gradient serves as a driving force for transport systems catalyzing sodium/substrate symport (2, 3, 5). Sodium circulation has also been implicated in regulation of intracellular pH (6) and energy homeostasis (7, 8).

In bacteria generation of a Na+ gradient is often attributed to the Na+/H+ antiporter activity which is most widely dis- tributed and is energized by the proton electrochemical gradient (9, 10). Three types of primary pumps extruding Na' have also been described (for review see Ref. 11).

The Na+/H+ antiporter activity has been thoroughly studied in intact cells and isolated membrane vesicles (10). It is electrogenic with an overall stoichiometry of Na+/H+ of above 1 (8, 10, 12), and its activity is regulated by intracellular pH (10). However, although the activity has been reconstituted

* This work was supported by a grant from the United States Israel Binational Foundation (to E. P. and S. S.) and a grant from the Niedersachsiches Ministerium fur Wissenschaft und Kunst (to S. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

tl Supported by a Levy Eshkol fellowship from the Ministry of Science and The National Council for Research and Development.

in proteoliposomes the protein(s) involved has not been purified (13, 14).

A genetic approach has been undertaken to study the Na+/ H+ antiporter system in Escherichia coli. Different mutants have been isolated which lost the Na+/H' antiporter activity as well as the capacity to regulate intracellular pH at alkaline pH, implying the role of the antiporter in pH homeostasis of bacteria (15-17). None of these mutations have been mapped to a structural gene coding for antiporter activity.

Tsuchiya, Wilson and colleagues (18) isolated a mutant in E. coli that shows increased rather than decreased antiporter activity in isolated membranes. These cells also tolerate much higher Li' concentrations (100 mM) as compared with the wild type cells (10 mM). The Li+ resistance has been ascribed to the excretion of the ion by the enhanced antiporter. This mutation designated antup has been mapped at about 0.5 min and the wild type gene ant cloned in plasmid pBR322 (19, 20). When in high copy number (plasmidic + chromosomal) the wild type ant confers Li+ resistance to cells and increases Na+/H+ antiporter activity in membranes (19) i.e. the antup phenotype. The ant gene has been sequenced and found to encode a membrane protein of M , 38,000 (20). The protein has been purified and reconstituted in proteoliposomes in a functional form.'

In the present work we studied the role of ant in wild type cells and inquired whether Na+/H+ antiporter activity(s), other than ant, exists in E. coli. For these purposes we have deleted ant from the chromosome and studied the phenotype of the Aunt strain pertaining to growth tolerance to Na+, Li+, and pH and the Na+/H+ antiporter activity in isolated membrane vesicles.

MATERIALS AND METHODS

Bacterial Strains, Plnsmids, and Culture Conditions-Bacterial strains used in this study are K12 derivatives. TA15 is melBLidant+ AlacZY (19). JC7623 is recB2lrecC22sbcB15thr (21). Cells were grown in L broth or in L broth of which NaCl was replaced by KC1 (87 mM, LBK) or minimal medium A (22) without sodium citrate, supple- mented with thiamine (1 pg/ml), 0.5% of glycerol or 10 mM melibiose as required. For JC7623 and its derivatives threonine (100 pg/ml) was also added. Growth at pH 8.6 was conducted in MTC medium containing 60 mM CAPS: 60 mM Tricine, 7.5 mM (NH&SO,, 10 mM KzHPO,, 0.08 mM MgSO,, and titrated by KOH. Where indicated Na+ and/or K+ concentrations were changed and Li+ added, For plates 1.5% Difco agar was added. For selection of plasmids 100 pg/ ml ampicillin or 50 pg/ml kanamycin were used.

Construction of plasmids are shown in Fig. 1. Transduction was carried out with Plvir phage as described (23).

' D. Taglicht, E. Padan, and S. Schuldiner, manuscript in prepa- ration.

The abbreviations used are: CAPS, 3-(cyclohexylamino)propane- sulfonic acid; MES, 2-(N-morpholino)ethanesulfonic acid; kb, kilo base pairs; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid.

20297

20298 Multiple Na f /H+ Antiporters in E. coli and Adaptation to Salinity DNA-DNA Hybridization-Chromosomal DNA was prepared by

the method of Gillen et al. (24). Restriction endonuclease digests of the DNA were resolved on horizontal 1% agarose gels and transferred to nitrocellulose membranes by a modification (25) of the Southern procedure (26). The filters were hybridized with probes labeled with [32P]dCTP by the multiprime method (Amersham Corp.).

Everted Membrane Vesicles and Measurement of Na+/H+ Anti- porter Activity-Everted membrane vesicles were prepared (19) es- sentially as described by Rosen and co-workers (14, 27, 28).

Na+/H+ antiporter activity in everted membrane vesicles was estimated based on its ability to collapse a transmembrane pH gradient. Acridine fluorescence was monitored to estimate ApH as previously described (14, 19). Fluorescence of acridine orange was monitored in a Perkin-Elmer fluorimeter (Luminescence Spectrometer, LS-5). Ex- citing light was 430 nm and emission light was measured at 570 nm.

Na+/H+ antiporter activity was also determined by measuring "Na uptake driven by a ApH. A ApH (interior acid) was generated by an ammonium ion gradient according to Nakamura et al. (14) as described (19).

Protein concentration in cells and everted membrane vesicles was determined as described (29).

RESULTS

To elucidate the role of ant gene in E. coli cells, it was essential to disrupt the chromosomal ant and to obtain cells devoid of an active gene. For this purpose we have constructed plasmids pDTK1, pDTK2, and pKR36 (Fig. 1) and produced deletion and insertion mutations in the wild type ant gene of

Mst II I

Hind m w II

&/a

EcoRI

3.7 k b

Eco R I

Mlul , Bgl I1 BomHI or MIuI,MstII or EcoRI Klenow dntp Klenow dntp

Lorge fragment Smoll frgment

HindIII I

FIG. 1. Construction of pDTK1, pDTK2, and pKR36. To obtain pDTKl and pDTK2 (not shown) plasmids, pGM36 (19) was restricted with BglII and MluI. The large BglII-MluI fragment was isolated, end filled, and ligated to end-filled BamHI-BamHI DNA fragment carrying the kanamycin gene, excised from pUC71K (30). Recombinant plasmids carrying the kanamycin gene in the same or reversed orientation, as compared with ant, were identified by restriction site analysis and were designated pDTKl and pDTK2, respectively. (Note that in pDTKl during the ligation a BamHI site was regenerated downstream to the C-terminal of the kanamycin gene.) To obtain pKR36 plasmid pGM36 was restricted with MluI and MstII, the large fragment isolated, end filled, and ligated to end-filled EcoRI-EcoRI DNA fragment carrying the kanamycin gene excised from pUC71K. Black arrow, ant gene; white arrow, blu gene; striped arrow, kan gene; - - -, sequences derived from the E. coli chromosome flanking ant gene (19,20); -, sequences of pBR322.

E. coli chromosome. To obtain pDTKl about 2/3 of the ant sequences of pGM36 (19) have been substituted with the gene encoding for kanamycin resistance. The C- and N-terminal regions of ant and additional chromosomal flanking regions of 0.8 and 1.4 kb, on each side, respectively, were left unmod- ified. pDTK2 is identical to pDTKl except for the orientation of the kanamycin gene which is reversed. Almost all ant sequences were deleted from pGM36 to obtain pKR36 leaving only the 22 base pairs of the N-terminal.

When pGM36 is transformed to TA15 cells it confers the antup phenotype upon the cells, i.e. they become Li+ (100 mM) resistant and their membranes exhibit enhanced Na+/ H+ antiporter activity (19). TA15 cells transformed with either pDTK1, pDTK2, or pKR36 remained Li+ sensitive like the parent strain and did not show enhanced Na+/H+ antiporter activity in their membranes (not shown). These results verify that the ant gene is inactive in the plasmids bearing the disrupted ant genes.

To obtain cells without a functional ant, the disrupted ant genes were exchanged with the chromosomal wild type ant of E. coli by homologous recombination. Plasmids pDTK1, pDTK2, and pKR36 were linearized by digestion with PuuII and transformed into E. coli strain JC7623, a strain in which propagation of intact plasmids is drastically hampered, while linear DNA is not digested, allowing homologous recombination between the linear DNA and the chromosome (21).

Based on the functions attributed to the Na+/H+ antiporter (see Introduction and Refs. 2-6) it was expected that recombinants carrying the disrupted ant will be sensitive to Na+ and prefer neutral pH and rich medium. Therefore, trans- formants of linearized pDTK1, pDTK2, or pKR36 were plated on LBK, pH 6.8, agar plates containing kanamycin. Recom- binants originated by double crossing over (Fig. 2) were obtained by growth on kanamycin and scoring for sensitivity to ampicillin, the vector's selective marker. Hence, KanRAmpS recombinants were expected to contain, instead of the wild type ant, the deleted ant of pDTK1, pDTK2, or pKR36 with the insertion of the kanamycin gene. One of each type of recombinants designated NM8, NM9, and JCK2, respectively, were used for further study.

To verify that the recombinants contain the deletion and insertion of the ant gene, we looked for the presence of different DNA sequences by hybridization to the DNA of NM8 as compared with the DNA of the wild type. When the BamHI-BamHI DNA fragment of pUC71K containing the kanamycin gene (Fig. 1) was used as a DNA probe, hybridization was observed with the DNA of NM8 but not with the wild type DNA (Fig. 3A). As expected for the distances between the restriction sites of NM8 DNA (Fig. 2) only one BamHI-BamHI fragment of NM8 DNA hybridized, and it was of 2.5 kb. On the other hand two HindIII-Hind111 fragments of 1.4 and 3.3 kb hybridized (Fig. 3A).

When the MluI-BglII DNA fragment of pGM36 which was deleted from ant in NM8 (Figs. 1 and 2) served as a probe there was hybridization only with the wild type DNA while the deletion mutant DNA did not show any homology to the probe (Fig. 3B). As expected after restriction with BamHI or HindIII (Fig. 2), the wild type DNA fragments which hybridized were 1.85 and 5.2 kb or 4.1 kb, respectively.

Utilizing the AuaI-BstXI fragment of pGM36 which bears the entire ant gene (Figs. 1 and 2) as a probe both the wild type DNA and NM8 DNA exhibit hybridization albeit the size of the DNA fragments were different and accorded the changes in the restriction map of the two strains DNA. The BamHI-BamHI and the HindIII-Hind111 fragments of the wild type which hybridized were the same as with the previous

Multiple Na+/H+ Antiporters in E. coli and Adaptation to Salinity 20299

1 kb

FIG. 2. Construction of an E. coli mutant carrying a deletion and insertion in ant. Plasmid pDTKl was linearized with PvuII and transformed in .JCi623 which allows recombination between the linearized plasmid and the chromosome but does not allow propagation of plasmids. (For details see text). A scheme of the double crossing over event leading to the deletion insertion mutant is depicted. This mutant was selected as KanRAmps as opposed to a single crossing over product which is Kan"Amp". Genes and flanking regions are marked as in Fig. 1. Thin arrows describe distances with the number of kb below them.

' 1 2 3 4

M W T M w r HindIU BamHI

-3.3

-2.5

-1.4

' I 2 3 4

-5.2

.4. I

- 1.85

WT M W T M Barn HI HindIII

C l 2 3 4

0 -:

-5.2

1. I

5.3

- - I

!.5

I .8

- -I 4 M WT M WT Barn HI Hindm

FIG. 3. Southern blot of restriction endonuclease digest of the DNA of the mutant (NM8) and wild type (TA15) hybridized to different probes. Southern blots and DNA-DNA hybridization were obtained as described under "Materials and Methods." The following different probes were used. A, the BarnHI-BarnHI 1.2- kb fragment excised from pUC71K and containing the kanamycin gene; R, the RglII-MluI 1.1-kb fragment of ant obtained from pGM36; C, the AuaI-RstXI 1.38-kb fragment bearing the entire ant derived from pGM36. Chromosomal DNA of TA15 ( W T ) and NM8 (M) were digested with RarnHI or Hind111 as depicted. Fragment sizes are given in kh.

probe (compare Fig. 3, C and B ) . On the other hand the NM8 mutant hybridizing fragments after BamHI or Hind111 restriction were of 2.5 and 5.1 kb or 3.3 and 1.42 kb, respectively. We therefore conclude that NM8 indeed harbors the deletion and the kanamycin insertion in its ant gene. These results also show that ant is present in only one copy in the wild type chromosome.

TA15 is the E. coli strain which we have used for all physiological studies of ant. Therefore, for the study of the effect of mutations in ant on the physiology of the cell we have transferred the deletion/insertion mutations to TA15 by P1 transduction and obtained strain NM81 and NM91 from NM8 and NM9, respectively, and strain RK20 from -JCK2. In accordance with the map position of ant (19), the linkage between thr and the ant mutation as selected by kanamycin was confirmed in all transductions. The hybridization experiments conducted with NM8 (Fig. 3) were repeated with NM81 and RK20 and in all cases the deletion and insertion were confirmed (not shown).

It is anticipated that a deletion/insertion mutation in the ant gene will reveal ant functions. It will grow only under conditions compensating for and therefore reflecting ant ac-

tivities. The growth of NM81 on minimal medium A with no Na' added and with either melibiose or glycerol (not shown) as carbon sources was identical to that of the wild type-TA15 (Fig. 4A). Although not added, this medium contains a con- tamination of 0.5-1 mM Na' (31). As shown before addition of 100 mM Li' but not Na' or K' markedly decreased the growth of TA15 (19). However, growth of NM81 was more sensitive to Li+ with doubling time of 7 h as opposed to 3.9 h of TA15 (Table I). Addition of 10 mM Li' already retarded growth of NM81 but not of TA15 (not shown). This supersen- sitivity of NM81 is due to lack of functional ant since NM81 cells transformed with pGM36, high copy number plasmid bearing ant, become resistant to Li' as previously documented for TA15 cells transformed with this plasmid (19). The transformed cells grow in the presence of Li' (100 mM) as well as without it (Table I). These results confirm our previous sug- gestion that ant confers Li' resistance.

The growth of NM81, NM91, and RK20 on LB broth, pH 6.8, of which Na+ has been replaced by K' (LBK) was identical to that of the wild type (not shown). Whereas the wild type retained identical growth capacity up to at least 0.7 M

A

TAl5

I Na+

1 Li+

d

I rntn

B

NM81

Na+ Li'

FIG. 4. Na+/H+ antiporter activity in everted membrane vesicles derived from NM81, TA15, and TA15/pGM12. Everted membrane vesicles were prepared from strains TA15 (a), NM81 ( b ) , and TA15/pGM12 (c). Cells were grown in LBK. Acridine orange was used to monitor ApH (19). The reaction mixture contained 10 mM Tricine-choline. pH 8.0, 140 mM KCl, 5 mM MgC12. At the onset of the experiment, acridine orange was added to a 0.5 p~ concentration and the fluorescence signal monitored. Thereafter membrane vesicles (130 pg of protein) and Tris-D-lactate (2 mM) were added and the fluorescence quenching (Q) recorded (100%). To measure Na+/H+ antiporter activity, NaCl or LiCl was added to 8 mM concentrations. The dequenching obtained is depicted.

20300 Multiple Na+/H+ Antiporters in E. coli and Adaptation to Salinity

TABLE I Doubling time (h) of NM81 in various ion concentrations of Nu+, K', and Li+

Cells from an overnight inoculum were washed and diluted hundred-folds into the indicated fresh media in which arowth was followed at 600 nm. ND. not determined.

Medium A, melibiose Medium ion conc. No Li' Na+ Na' K' Na' Na' Na'

LB (pH 6.8) LB (pH 7.5)

addition (0.1 M) (0.08 M) (0.7 M) (0.5 M) (0.4 M) (0.5 M) (0.6 M)

Strain TA15 NM81

1.1 3.9 0.7 1.0 0.8 0.9 1.0 1.1 1.1 7.0 1.1 5.0 0.9 1.6 5.8 >24

NM81/pGM36 1.0 1.1 1.0 1.0 ND ND ND ND -

NaCl, increasing NaCl concentration led to decrease in the growth of the mutants, and at 0.7 M NaCl it was practically arrested (Table I). The sensitivity of NM81 to Na' was clearly due to lack of functional ant. It was not caused by the presence of the kanamycin insertion since the mutant NM82, with the reversed kan gene, was sensitive to Na' to the same degree, both in the presence and absence of kanamycin (not shown). Furthermore, NM81 transformed with pGM36, which carries an intact ant gene, regained the wild type resistance to Na+ (Table I).

The Na+ sensitivity of the mutants could be due to the nonspecific effect of increased ionic strength and/or increased osmolarity. These possibilities were ruled out since when 0.7 M KC1 rather than NaCl were added to the LBK medium growth of NM81 was identical to that of the wild type or to NM81/pGM36 (not shown). The Na' sensitivity could also stem from lack of enough K+ under the conditions of high Na+. This alternative was also ruled out since addition of up to 10 mM KC1 did not relieve the inhibition caused by NaCl (not shown). Hence, high Na+ has a specific inhibitory effect on the mutant lacking intact ant. Taken together we conclude that functional ant gene is required for the adaptation of E. coli to high salinity as well as for enduring Li+.

Changing the pH of LBK to 7.5 had no effect on growth of NM81 nor on that of TA15 (not shown) even when KC1 concentrations increased to 0.5 M (Table I). However, in marked contrast to TA15, whose growth was not significantly affected by addition of NaCl, the Na+ sensitivity of NM81 increased at pH 7.5 (Table I). Whereas at pH 6.8,0.7 M NaCl were required to inhibit growth of NM81, at pH 7.5, similar inhibitory effects were obtained already by addition of 0.5 M NaC1. To further study the effect of pH, NM81 and TA15 were grown in MTC medium at pH 8.6. Their doubling time was identical (1.5 h) with either glycerol or melibiose as carbon sources. However, addition of only 0.1 M NaC1, but not 0.1 M KC1 (not shown) inhibited growth of NM81 (doubling time of 4.5 h). Growth of TA15 was not affected. Hence, although NM81 is not sensitive to pH its Na+ sensitivity is markedly influenced by pH.

It has previously been shown that ant affects the Na+/H+ antiporter activity of isolated membrane vesicles (19, 20). Cells bearing the mutation antup in the ant locus or transformed with high copy number plasmid bearing ant display increased Na+/H+ antiporter activity and Li+ resistance. These results suggested that the increased antiporter activity is the basis of the resistance. Therefore, we measured the Na+/H+ antiporter activity in membrane vesicles isolated from NM81 and TA15.

Surprisingly, membrane vesicles from NM81, (the Aant strain) displayed Na+/H+ and Li+/H+ antiporter activities that were only 40-50% lower than that displayed by membranes from TA15 (Fig. 4,A and B ) . This finding was observed regardless of whether the cells were grown in minimal medium

or in LBK, although in the latter case all signals were pro- portionally higher (compare Fig. 4 with 5). It is evident, therefore, that in spite of the deleted ant, NM81 possesses a Na+/H+ and a Li+/H+ activity. It also grows in minimal medium with melibiose as a carbon source, and this growth is Na+ dependent (not shown). These results imply that the antiporter activity observed in NM81 is sufficient to maintain a Na+ gradient and support melibiose transport.

A K+/H+ antiporter activity has previously been documented in membrane vesicles isolated from E. coli cells (32, 33). In contrast with the Na+/H+ antiporter activity which is specific to Na+, this system has a low ion specificity and exchanges H+ for both Na+ and K+. In addition the K+/H+ antiporter system is very sensitive to digestion by trypsin whereas the Na+/H+ antiporter is resistant to such treatment (33). Since all measurements summarized in Fig. 4 were performed at K' concentrations (150 mM) saturating the K+/ H+ antiporter (Km, 3-7.5 mM (33)), we assumed that under these conditions the K+/H+ antiporter does not play a major role in the Na+/H+ antiporter activity of any of these membranes. To further verify this assumption, we measured the K+/H+ antiporter activity in membrane vesicles isolated both from TA15 and NM81. When measured in the presence of choline chloride rather than KC1, in both membranes similar K+/H+ antiporter activity was monitored (Fig. 5). Further- more in both membranes treatment by trypsin markedly decreased the K+/H+ antiporter activity with almost no effect on the Na+/H+ antiporter activity (see for example Fig. 5C). These results imply that like the activity of TA15, the Na+/

TA 15 NM81 NM8 I

f KC I

f KC1

a b C

FIG. 5. K+/H+ antiporter activity in everted membrane vesicles derived from TA16 and NM81. Everted membrane vesicles were prepared from TA15 (a) , and NM81 ( b and c ) grown in minimal medium A with glycerol as a carbon source. The experimental system and protocol was that described in Fig. 4 but the reaction mixture contained 10 mM Tris-Hepes, pH 8.0,140 mM choline chloride, 5 mM MgC12, and 100 pg of protein. When indicated KC1 or NaCl were added to 8 mM concentrations. Trypsin treatment was carried out according to Ref. 33. Trypsin (Sigma, 15 pg/ml) was added to the reaction mixture containing the membranes and incubation contin- ued for 3 min. Then the trypsin inhibitor (Sigma) was added (75 pg/ ml), and 10 s later lactate was added to onset the reaction.

Multiple Na+/H+ Antiporters in E. coli and Adaptation to Salinity 20301

H+ antiporter activity observed in NM81 is not related to the K+/H+ antiporter system and therefore represents an as yet unknown system(s), additional to ant system.

Since ant is disrupted in NM81, the activity observed in NM81 membranes clearly represent a second system(s) other than ant. A mixture of the two types, most probably, com- prises the activity observed in membrane vesicles of TA15, the wild type strain. However, TA15 transformed with pGM12, a high copy number plasmid bearing ant, may yield membranes representing mainly the ant system. In these membranes minimal interference is expected from the second system since it is most probably present in a single copy (see above) as compared with the multicopy ant. We, therefore, compared the properties of the antiporter activity in membrane vesicles derived from NM81, and TA15/pGM12 (Table I1 and Fig. 4). It is evident that both types of membrane vesicles exhibit not very different apparent K, for Na' (0.25 and 0.68 mM, respectively) with NM81 having slightly better apparent affinity. Since the acridine orange technique yields only a relative and a qualitative estimate of the antiporter

TABLE I1 Ion specificity of the Na+/H+ antiporter activity in everted membrane

vesicles derived from NM81 and TAlSIpGM12 Na+/H+ antiporter activity was measured at different concentra-

tions (8-0.1 mM) of NaCl or LiCl as described in Fig. 4. The activity of Na+/H+ antiporter was estimated from the ratio between the dequenching of fluorescence after addition of NaCl or LiCl and the fluorescence quenched after the addition of D-laCtate. The K, was calculated according to Eadie-Hofstee methods (34) where the Na+/ H+ antiporter activity was considered as V according to Ref. 8.

NM81 TA15/pGM12 Na+ 0.25 0.68

Li+ 0.86 0.06 Apparent K , (mM)

100 -

80 -

60 -

40 -

20 -

6 7 8

PH

FIG. 6. Effect of medium pH on the Na+/H+ antiporter activity of everted membrane vesicles of NMSl and TAlB/pGM12. Activity of Na+/H+ antiporter was estimated as described in Fig. 4 at various pH values. Tricine, 10 mM, titrated with choline hydroxide was used to obtain pH 7.0-8.5 while MES-choline was used for pH 6.5. In addition to the buffers, the reaction mixture contained 140 mM KC1 and 5 mM MgCl,. The activity of the Na+/H+ antiporter at each pH was estimated as described in Table 11. Percent of dequenching due to addition of NaCl (8 mM) is depicted versus medium pH.

activity we measured directly initial rates of **Na transport as described (19). Identical results for the K,,, values were obtained (not shown).

However, whereas NM81 vesicles exhibit an apparent K,,, for Li+ of 0.86 mM, TA15/pGM12 vesicles show more than a 10-fold lower apparent K,,, for the ion (60 PM). These results suggest that the ant system has a much higher apparent affinity to Li+ compared with the other system.

Working with right side out membrane vesicles derived from another wild type strain of E. coli (ML308225) Leblanc and his co-workers (10, 35) previously showed that the Na+/ H+ antiporter activity is regulated by pH at the cytoplasmic side of the membrane. Since the cytoplasmic side of everted membrane vesicles is exposed to the medium pH we compared the effect of medium pH on the Na+/H+ antiporter activity of everted membrane vesicles derived from NM81 and TA15/ pGM12 (Fig. 6).

The Na+/H+ antiporter activity of TA15/pGM12 show marked dependence on pH very similar in pattern to that previously observed for a wild type strain (35). Thus, at a low pH of 6.5, the activity is low, increasing with pH about 4-fold at pH 8.5. On the other hand, in marked contrast to this behavior the activity in NM81 vesicles is not affected at all over pH range of 6.5 to 8.5

Taken together the ant-dependent antiporter system differs from the additional system revealed in NM81 both in ion specificity and pH sensitivity.

DISCUSSION

In this work we describe the disruption of the ant gene in the E. coli chromosome and its effect on growth tolerance to Na+, Li+, and pH and on the Na+/H+ antiporter activity in everted membrane vesicles.

We found that ant is not an essential gene. Growth of NM81, the Aant strain, is normal both in minimal salt medium and in LB medium. Increasing the Na+ concentration uncovers an impaired adaptation in the Aunt strain. Thus, unlike ant+ strain, Aunt does not grow in LBK, pH 6.8, medium, used for the strain isolation, when 0.7 M NaCl is added. This Na+ sensitivity is markedly dependent on pH, increasing upon alkalinization. At pH 7.5 in LBK medium, 0.5 M NaCl and at pH 8.6 in MTC medium, 0.1 M NaC1, are sufficient to inhibit growth of Aunt. That this sensitivity is specific to Na+ ions is demonstrated by the fact that growth of Aant strains is normal when KC1 is added instead of NaCl at all pH values. Sensitivity to the effects of Li' is also increased in the Aunt strain as compared with ant+. Whereas 100 mM LiCl inhibit growth of wild type cells in minimal medium A, 10 mM LiCl are already deleterious to Aunt cells.

The phenotype of Aunt seems to be due solely to the deletion of ant since we have not disrupted other open reading frames in its vicinity. The results are independent of whether the disrupting gene, coding for kanamycin resistance, was in- serted in a reading frame with or against the direction of ant. In addition, Aunt phenotype is not strain-specific since JC7623 Aunt exhibited Na+ and Li+ sensitivity similar to NM81. Finally, the defective phenotype can be complemented by transformation with a plasmid bearing the ant gene.

Growth of Aunt strains is supported not only by carbon sources such as glycerol but also by ones such as melibiose in a Na+-dependent manner. These findings suggested that even though the Aunt strains have an impaired capacity to adapt to high salinity and to challenge Li+ toxicity it can still generate a Na+ gradient large enough to allow for growth on a carbon source whose transport requires Na+ (18).

Indeed, from studies in everted membrane vesicles prepared

20302 Multiple Na+/H+ Antiporters in E. coli and Adaptation to Salinity

from the Aunt strains a residual significant antiporter activity other than ant became evident. Because of this multiplicity of antiporter systems we shall call ant, antA. The residual activity detected in Aunt strains will be tentatively called antB even though it is not yet established to be a product of a single gene.

The antB protein(s) displays kinetic properties different from those of antA, namely its activity in everted vesicles is practically independent of extravesicular pH while that of antA product is enhanced with increasing pH (10, 35). Since the measurements were conducted in everted membrane vesicles it is implied that antB is insensitive to intracellular pH while antA is very sensitive to this pH (10, 35). In addition, the apparent affinity for Li' of antB at the cytoplasmic face of the membrane is about 15 times lower than that of antA.

The lack of antA and the properties of antB explain the phenotype of the Aunt strain. The poor adaptation of the AantA strains to high Na' is most likely due to lack of adequate V,,,, which in the wild type, is provided by combined activity of both antA and antB. The enhanced Li' toxicity in the AantA strain can be explained in a similar manner. In addition, the lower affinity of the antB system for Li+ ions is in line with the impaired capacity of AantA strains to challenge Li+ toxicity.

The increased sensitivity to Na+ with alkalinization sug- gests the possibility that antA is mainly active at alkaline intracellular pH. This is corroborated by the finding that the antA antiporter activity in membrane vesicles is drastically dependent on intracellular pH, increasing about 4-fold upon alkalinization. It is suggested that a Na' leak into the cell, whether carrier mediated or not, increases with increasing pH. Therefore, at a given Na' concentration increasing the pH increases the Na+ load, which in the absence of antA floods the antB system and brings about a collapse of cell homeostasis.

Growth of Aunt strains at alkaline pH is normal provided Na' ion concentration is kept low. These results suggest that, under these conditions, systems other than antA are sufficient to carry on homeostasis of cellular pH. Further studies are required to dissect in detail the mechanisms for pH regulation.

Multiplicity of transport systems for one substrate is not uncommon in bacteria. Thus, for example, this is the case for ions such as K' (36) and Pi (37) and for several amino acids (38). The experimental demonstration of multiple Na+/H+ antiporters in E. coli is a novel finding. These data are in line with suggestions based on theoretical grounds (39) and raise many questions as to the nature of the alternative systemb), their interaction at the activity level, and their regulation a t the gene level.

REFERENCES

1. Harold, F. M., and Altendorf, K. (1974) Curr. Top. Membr. Trans.

2. Lanyi, J. K. (1979) Biochim. Biophys. Acta 559, 377-397 3. Krulwich, T. A. (1983) Biochim. Biophys. Acta 726,245-264

5,1-50

4. Booth, I. R. (1985) Microbiol. Reu. 4 9 , 359-378 5. Rosen, B. P. (1986) Annu. Reu. Microbiol. 40, 263-286 6. Padan, E., Zilberstein, D., and Schuldiner, S. (1981) Biochim.

Biophys. Acta 650 , 151-166 7. Osterhelt, D., Hartmann, R., Michel, H., and Wagner, G. (1978)

in Energy Conservation in Biological Membranes (Schager, G., and Klingenberg, M., eds) Springer Verlag, Heidelberg

8. Schuldiner, S., and Fishkes, H. (1978) Biochemistry 17, 706-711 9. West, I. C., and Mitchell, P. (1974) Biochem. J. 144,87-90

10. Leblanc, G., Bassilana, M., and Damiano, E. (1988) in Na+/H' Exchange (Grinstein, S., ed) CRC Press, Boca Raton, in press

11. Skulachev, V. P. (1986) in Zon Transport in Prokaryotes (Rosen,

York B. P., and Silver, S., eds) pp. 131-164, Academic Press, New

12. Castle, A. M., Macnab, R. M., and Schulman, R. G. (1986) J. Biol. Chem. 261,3288-3294

13. Tsuchiya, T., Misawa, A., Miyake, Y., Yamasaki, K., and Niiya, S. (1982) FEBS Lett. 142 , 231-234

14. Nakamura, T., Hsu, C. M., and Rosen, B. P. (1986) J. Biol. Chem.

15. Krulwich, T. A. (1986) J. Membr. Biol. 89,113-125 16. Ishikawa, T., Hama, H., Tsuda, M., and Tsuchiya, T. (1987) J.

Biol. Chem. 262,7443-7446 17. Zilberstein, D., Padan, E., and Schuldiner, S. (1980) FEBS Lett.

116,177-180 18. Niiya, S., Yamasaki, K., Wilson, T. H., and Tsuchiya, T. (1982)

J. Biol. Chem. 257,8902-8906 19. Goldberg, E. B., Arbel, T., Chen, J., Karpel, R., Mackie, G. A.,

Schuldiner, S., and Padan, E. (1987) Proc. Natl. Acad. Sci.

20. Karpel, R., Olami, Y., Taglicht, D., Schuldiner, S., and Padan, E.

21. Winans, S. C., Elledge, S. J., Krueger, J. H., and Walker, G. C.

22. Davies, B. D., and Mingioli, E. S. (1950) J. Bacteriol. 6 0 , 17-28 23. Miller, J. (1972) Experiments in Molecular Genetics, Cold Spring

24. Gillen, J. R., Willis, D. K., and Clark, A. J. (1981) J. Bacteriol.

25. Smith, E., and Summers, M. D. (1980) Anal. Biochem. 109,123-

26. Southern, E. M. (1975) J. Mol. Biol. 98,503-517 27. Ambudkar, S. V., Zlotnick, G. W., and Rosen, B. P. (1984) J .

28. Rosen, B. P. (1986) Methods Enzyrnol. 125,328-336 29. Bradford, M. M. (1976) Anal. Biochem. 7 2 , 248-254 30. Ruther, U. (1980) Mol. Gen. Genet. 178,475-477 31. Mandel, K. G., Guffanti, A. A., and Krulwich, T. A. (1980) J.

32. Brey, R. N., Beck, S. C., and Rosen, B. P. (1978) Biochem.

33. Brey, R. N., Rosen, B. P., and Sorensen, E. N. (1980) J. Biol.

34. Hofstee, B. H. J. (1954) Enzymologia 17,273-278 35. Bassilana, M., Damiano, E., and Leblanc, G. (1984) Biochemistry

36. Walderhaug, M. O., Dosch, D. C., and Epstein, W. (1987) in Zon Transport in Prokaryotes (Rosen, B. P., and Silver, S., eds) pp. 85-130, Academic Press, New York

37. Rosenberg, H. (1987) in Zon Transport in Prokaryotes (Rosen, B.

38. Ames, G. F. L. (1986) Annu. Reu. Biochem. 65,397-425 P., and Silver, S., eds) pp. 205-248, Academic Press, New York

39. Macnab, R. M., and Castle, A. M. (1987) Biophys. J. 52, 637-

261,678-683

U. S. A. 8 4 , 2615-2619

(1988) J. Biol. Chem. 263,10408-10414

(1985) J. Bacteriol. 161 , 1219-1221

Harbor Laboratory, Cold Spring Harbor, NY

145,521-532

129

Biol. Chem. 259,6142-6146

Biol. Chem. 255,7391-7396

Biophys. Res. Commun. 8 3 , 1588-1594

Chem. 255,39-44

23,5288-5294

647

ant escherichia coli deletion of in reveals its function ... · pdf fileescherichia coli by...

Documents