(Received for publication, January 14, 1997)
From the Department of Biotechnology, Division of
Utilization of Biological Resource, Faculty of Agriculture, Gifu
University, Yanagido, Gifu 501-11, Japan and the § Division
of Biological Science, Graduate School of Science, Nagoya
University, Chikusa-ku, Nagoya 464-01, Japan
The Escherichia coli chemoreceptor Tsr mediates tactic responses to serine, repellents, and changes in temperature. We have previously shown that the serine-sensing ability of Tsr-T156C, which has a unique cysteine in place of threonine at residue 156, is specifically inactivated by thiol-modifying reagents and that L-serine protects the receptor from modification. In this study, we demonstrated the correlation between protective effects of various attractants and their potencies to elicit attractant responses. This indirect binding assay was used to monitor the affinity of the receptor for L-serine under various conditions. It has been demonstrated by in vitro assays that the ligand-binding affinities of Tsr and the related chemoreceptor Tar are unaffected by changes in the methylation state of the receptor. Using the serine protection assay, we re-examined this issue both in vitro and in vivo. The methylation levels of Tsr-T156C did not affect its ligand-binding affinity. We also showed both in vitro and in vivo that the ligand-binding affinity was unaffected by temperature. These results suggest that the structure of the periplasmic domain of the receptor is uncoupled from the signaling states of the cytoplasmic domain. This ligand-binding assay system should be applicable to other receptors.
In many sensory systems, desensitization or adaptation to a persisting stimulus plays a crucial role in highly sensitive detection of stimuli over a comprehensive range. Extracellular signals (e.g. binding of ligands) are received and transduced into intracellular signals (e.g. activation or inactivation of kinases) by cell surface receptors. Binding of a ligand to a receptor extracellular domain induces some changes in structure of a receptor intracellular domain and, hence, in its activity. By contrast, upon adaptation, the signaling activities of the intracellular domains are often down-regulated by intracellular feedback regulatory systems (typically via covalent modification of the receptors). The effects of receptor down-regulation may be limited to the intracellular domain or may induce a global structural change in the receptor.
Four closely related chemoreceptors of Escherichia coli serve as model systems for investigating both transmembrane signaling and down-regulation (for reviews, see Refs. 1-6), Tsr (for serine), Tar (for aspartate and maltose), Trg (for ribose and galactose), and Tap (for dipeptide). These chemoreceptors function as homodimers regardless of ligand occupancy state (7), and the homodimer forms a ternary complex with a homodimer of a cytoplasmic autokinase CheA and two molecules of an adaptor protein CheW (8). Each receptor monomer consists of an N-terminal periplasmic ligand-binding domain, a C-terminal cytoplasmic signaling domain, and two membrane-spanning segments. Binding of a ligand to the interface between the two periplasmic domains triggers some structural change within the receptor dimer. This structural change leads to activation or inactivation of autophosphorylation of CheA and phosphotransfer from CheA to CheY. Phosphorylated CheY binds to the flagellar motor and induce clockwise (CW)1 rotation of the motor (causing tumbling of the cell), which otherwise rotates counterclockwise (CCW) (causing smooth swimming).
Adaptation is achieved via methylation and demethylation of the chemoreceptors (9). Each chemoreceptor contains four to six glutamic acid residues that are reversibly methylated. Methyltransferase CheR catalyzes transfer of a methyl group from S-adenosylmethionine to a glutamic acid side chain, and methylesterase CheB hydrolyzes the methyl ester bond of a methylated glutamate residue. The latter enzyme also serves as deamidase, converting specific glutamine residues of the nascent chemoreceptors to methylatable glutamic acid residues (10).
Recent in vitro studies (11-13) suggest that methylation of a receptor modulates its signaling state but has little effect on its ligand-binding affinity. This means that the structure of the periplasmic ligand-binding domain is uncoupled from the signaling state of the cytoplasmic domain under some conditions. However, contradictory results have been reported for receptor-containing membranes (14). To explain the discrepancy, Lin et al. (13) suggested that association of chemoreceptors with cytoplasmic proteins, such as CheA and CheW, might influence their ligand-binding behavior. Therefore, the effects of methylation and demethylation on ligand-binding affinity of receptors must be investigated under conditions closer to their native setting.
The four chemoreceptors are unique in that they function also as thermosensors (15, 16). Tsr, Tar, and Trg are warm receptors that produce CCW signals upon a temperature increase and CW signals upon a temperature decrease, whereas Tap is a cold receptor with opposite signaling behavior (16). Temperature should affect the signaling states of the chemoreceptors, but it remains unclear how this can be achieved. Does a change in temperature cause a global change in receptor structure or just some local change within the cytoplasmic signaling domain?
In this study, we investigated the effects of methylation/amidation and temperature on the ligand-binding affinity of Tsr. We have previously shown that a Tsr-specific attractant L-serine protects Tsr-T156C, which has a unique cysteine residue at the ligand-binding site, from thiol-modifying reagents such as N-ethylmaleimide (NEM) in a dose-dependent manner (17). This protective effect of L-serine should reflect the affinity of Tsr-T156C for L-serine. If so, this assay can be used to monitor the ligand-binding affinity of the receptor under various conditions. In this report, we demonstrate a correlation between the stimulus strength of various amino acid attractants detected by Tsr-T156C and their abilities to protect Cys-156 from NEM modification. Using this assay system both in vitro and in vivo, we found that the ligand-binding affinity of Tsr is uncoupled from its signaling states modulated by covalent modification and temperature.
All strains used in this
work are derivatives of E. coli K-12. Strain HCB339
[tsr-7021
(tar-tap)5201
trg::Tn10 thr leu his met rpsL136] (18)
lacks all four chemoreceptors, and strain CP553 [trg-100
tsr-7028
(tar-cheB) leu his rpsL lac
xyl ara tonA tsx thi zab::Tn5] (19) lacks
CheB and CheR, as well as all chemoreceptors. Plasmid pGAN1 (17)
carries the promoterless mutant tsr gene encoding the
receptor protein with the single amino acid substitution, Thr-156 to
Cys, placed downstream of the tac promoter. Plasmid pRAB1
(20) carries the methylesterase gene cheB and the
tetracycline-resistant gene tetA (Tcr). A
cheR-overproducing plasmid pKB23 (11) and a
cheB-overproducing plasmid pKB24 (11), both of which carry
the chloramphenicol acetyltransferase gene cat
(Cmr), were provided by M. I. Simon of California Institute
of Technology.
NEM,
isopropyl-1-thio--D-galactopyranoside (IPTG),
phenylmethylsulfonyl fluoride (PMSF), and 1,10-phenanthroline were
purchased from Wako Pure Chemical Industries (Osaka, Japan).
Octyl-
-D-glucopyranoside was purchased from Dojindo
Laboratories (Kumamoto, Japan).
N-[ethyl-1-14C]ethylmaleimide (40 mCi/mmol)
and reagents for the bicinchoninic acid protein assay were products of
Dupont NEN and from Pierce, respectively.
Cells were grown at 30 °C in tryptone-glycerol broth (1% Bacto-tryptone (Difco Laboratories, Detroit, MI), 0.5% NaCl, 0.5% glycerol) supplemented with 50 µg/ml ampicillin (and 25 µg/ml chloramphenicol, when necessary). At late exponential phase, cells were collected by centrifugation, washed twice with motility medium (10 mM potassium phosphate buffer, pH 7.0, 0.1 mM EDTA, 10 mM sodium DL-lactate, 0.1 mM methionine), and resuspended in motility medium.
Temporal stimulation assays for chemoresponse were carried out as described previously (21). The cells suspended in motility medium were pretreated with a repellent, 1 M glycerol, and then stimulated with various concentrations of a Tsr-mediated attractant, L-serine, D-serine, or L-alanine. The change in the smooth swimming fraction after 30 s was measured. Addition of 1 M glycerol reduces the smooth swimming fraction of the cells to nearly zero. Therefore, an increase in the smooth swimming fraction upon addition of an attractant represents the magnitude of an attractant response of the cells.
Temporal stimulation assays for thermoresponse were carried out as described previously (20). Cells were cultured and washed as described above. A drop of cell suspension was placed on a glass slide mounted on a temperature control apparatus (22). When cells swam too smooth or too tumble, glycerol or L-serine was added to the suspension. The temperature was first increased from 20 to 30 °C and then decreased from 30 to 20 °C.
Preparation of Membranes and Solubilization of Membrane ProteinsMembranes were prepared by the method of Foster et al. (23). A fresh overnight culture of CP553 cells carrying pGAN1 was inoculated (1:100 dilution) into LB medium (1% Bacto-tryptone, 1% yeast extract, 0.5% NaCl) supplemented with 200 µg/ml ampicillin and grown at 30 °C. If the cells also carried pRAB1 (Tcr) or pKB23 (Cmr), 12.5 µg/ml tetracycline or 25 µg/ml chloramphenicol was added. After 3 h, 1 mM IPTG was added to the culture. After further incubation for 4 h, the cells were harvested by centrifugation, washed with motility medium, and were treated with 100 µM NEM in the presence of 10 mM serine for 60 min at 0 °C. The cells were then collected again, resuspended in lysis buffer (100 mM sodium phosphate, pH 7.2, 5 mM EDTA, 10% glycerol, 5 mM 1,10-phenanthroline, 2 mM PMSF) at 0 °C, and lysed by sonication (Heat Systems-Ultrasonics, Inc., model W-225). Unbroken cells and cell debris were removed by centrifugation for 20 min at 10,000 rpm in a Sakuma 7B rotor at 4 °C. The supernatant was centrifuged for 60 min at 40,000 rpm in an RP 50-2 rotor (Hitachi Koki Co., Ltd.) at 4 °C. The pellet was washed with wash buffer (50 mM sodium phosphate, pH 7.2, 2 M KCl, 10% glycerol, 5 mM EDTA, 5 mM 1,10-phenanthroline, 1 mM PMSF) and then resuspended in a buffer containing 50 mM Tris-HCl, pH 7.0, 10% glycerol, 5 mM 1,10-phenanthroline, and 1 mM PMSF at 0 °C. Protein concentration was determined by the bicinchoninic acid protein assay system using bovine serum albumin as a standard. The suspension was diluted with the same buffer to approximately 10 mg of protein/ml. Octylglucoside was added to 1.25%, and the mixture was incubated for 20 min at 0 °C and centrifuged for 60 min at 40,000 rpm (Hitachi RP65 rotor) at 4 °C. The supernatant was stored at 4 °C.
Modification with Radioactive NEMFor modification of Tsr-T156C in intact cells, N-[ethyl-1-14C]maleimide (10 mCi/mmol; 50 µM final concentration) was added to the cells (A590 = 1.5), and the mixture was incubated at 25 °C for 30 min. For membranes or octylglucoside extracts, the mixture was incubated at 25 °C for 6 min. When necessary, various concentrations of L-serine, D-serine, or L-alanine were added prior to the addition of NEM. The reactions were terminated by the addition of 5% trichloroacetic acid, and the samples were collected by centrifugation. The precipitates were then subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (21) and autoradiography. The labeled proteins was analyzed using Bio Imaging Analyzer (Fuji Photo Film Co., Ltd., BAS-1000).
We first examined the
sensing abilities of Tsr-T156C for various amino acid attractants.
Plasmid pGAN1 carrying the tsr-T156C gene was introduced
into strain HCB339, which is defective in all four chemoreceptor genes.
As shown in Fig. 1, 40 µM
L-serine or 1 mM D-serine was
required for attractant responses in 50% of the cells. Only a small
response was induced by 50 mM or higher concentrations of
L-alanine. These results are essentially consistent with
the published data for wild-type Tsr (24), that the affinity of Tsr for
L-alanine is lower than that for L-serine by
more than two orders of magnitude.
Next, we investigated the ability of various attractants to protect
against NEM modification. As shown in Fig. 2, the
addition of L-serine or D-serine to membrane
proteins extracted with octylglucoside significantly prevented NEM from
reacting with Tsr-T156C, but L-alanine appeared to have no
protective effect. Similarly, glycine and -aminoisobutyric acid, a
non-metabolizable attractant, had no effect (data not shown). The
protective effect of D-serine was smaller than that of
L-serine, consistent with the fact that the concentration
of D-serine for a half-maximal response of cells expressing
Tsr-T156C is about 10-fold higher than that of L-serine. Similarly, the inability of L-alanine to protect
corresponds to the low apparent affinity of Tsr-T156C for
L-alanine. Therefore, we conclude that the relative extent
of protection by an attractant against NEM modification of Tsr-T156C
reflects the affinity of Tsr for the attractant.
Effects of Covalent Modification on Serine Protection in Vitro
We next examined whether methylation of the chemoreceptor
affects its affinity for attractants. Plasmid pGAN1 carrying the tsr-T156C gene was introduced into strain CP553, which lacks
CheB and CheR as well as all four chemoreceptors. To modulate the
methylation level of Tsr-T156C, the resulting strain was further
transformed with plasmids carrying either the methylesterase/deamidase
gene cheB or the methyltransferase gene cheR.
Fig. 3A shows the effects of various
concentrations of L-serine on NEM modification of Tsr-T156C in membrane preparations from CheB+ CheR or
CheB
CheR
cells. In these strains, the five
major methyl-accepting residues of Tsr (collectively designated as
QEQEE) should be modified as follows. In CheB
CheR
cells, all five residues are unmodified (QEQEE); in
CheB
CheR+ cells, the Gln residues are not
deamidated, and the Glu residues are methylated (QEmQEmEm); and in
CheB+ CheR
cells, the Gln residues are
deamidated, and the Glu residues are not methylated (EEEEE). In fact,
Tsr-T156C proteins in those cells showed distinct mobilities on
SDS-PAGE due to methylation or deamidation (25). In each case, the
extent of NEM labeling of Tsr-T156C decreased considerably in the
presence of higher concentrations of L-serine. The
L-serine concentration required for 50% protection is
approximately 500 µM both for CheB
CheR+ cells and for CheB+ CheR
cells (Fig. 3B). Similar results were obtained using
octylglucoside-solubilized Tsr-T156C proteins (data not shown). These
results suggest that covalent modification of Tsr-T156C does not
dramatically affect its affinity for L-serine in the
absence of the cytoplasmic Che proteins, a conclusion consistent with
recent studies on Tar (11, 12) and Tsr (13).
Effects of Covalent Modification on Serine Protection in Vivo
We then examined serine protection of Tsr-T156C from NEM
modification in intact cells. As shown in Fig.
4B, the L-serine concentration required for 50% protection of Tsr-T156C was approximately 1 mM for CheB CheR
(QEQEE),
CheB
CheR+ (QEmQEmEm), and CheB+
CheR
(EEEEE) cells. This was not due to the
ineffectiveness of the method since about 100 mM
D-serine was required for 50% protection of Tsr-T156C
expressed in CheB
CheR
cells (Fig.
4B). Four bands of Tsr-T156C were detected for samples from
CheB
CheR+ cells due to differential
methylation levels (Fig. 4A). So, we further examined serine
protection on each band. The slowest band on SDS-PAGE was considered to
be the unmodified form (QEQEE) of Tsr (Fig. 4C). All bands
were protected from NEM modification by L-serine to almost
the same degree as the slowest band. Taken together, we conclude that
covalent modification of Tsr-T156C does not dramatically change its
affinity to serine even in the presence of the cytoplasmic Che
proteins. However, since Tsr was overproduced, we cannot rule out the
possibility that many of the receptor proteins might not form ternary
complexes with CheW and CheA.
Effects of Temperature on Serine Protection
We next examined
whether temperature influences the ligand-binding affinity of the
chemoreceptor. First of all, we had to examine whether Tsr-T156C has
thermosensing ability in the CheB CheR
background since Tar does not have thermosensing ability in the same
background (20). The results are summarized in Table I. CP553 cells (Tsr
Tar
Trg
Tap
CheB
CheR
) were
transformed with plasmid pJUN11, which carries the tsr-T156C gene. The resulting cells showed tumbling without any stimulus at
25 °C and did not change their swimming behavior upon temperature changes. When tumbling was reduced by the addition of 0.3 mM L-serine, the cells showed thermosensing
abilities. The cells showed smooth swimming as the temperature
increased (from 20 to 30 °C) and then tumbling as the temperature
decreased (from 30 to 20 °C). Similar results were obtained for
CP553 cells expressing wild-type Tsr in the presence of 120 µM L-serine. An analysis by Western blotting with anti-Tsr antiserum detected no change in mobility of either wild-type Tsr or Tsr-T156C during these behavioral responses, indicating that neither spontaneous deamidation nor degradation of the
chemoreceptors had occurred (data not shown). These results suggest
that, unlike Tar, the primary, unmodified translational product of Tsr
can mediate a thermoresponse. That serine does not play an essential
role in thermosensing by Tsr-T156C was verified as follows. To shift
the signaling bias toward smooth swimming (CCW flagellar rotation) by
deamidation, Tsr-T156C was expressed in CP553 cells carrying the
CheB-overproducing plasmid pKB24 (CheB++
CheR
). The resulting cells swam very smoothly even when
temperature was changed. However, in the presence of a repellent
glycerol (10%), the deamidated form of Tsr-T156C mediated a
thermoresponse. We conclude that both the unmodified and the deamidated
forms of Tsr-T156C have thermosensing abilities.
|
Next, we subjected Tsr-T156C expressed in the CheB
CheR
background to NEM modification in the presence of
various concentrations of L-serine. As shown in Fig.
5, both in a membrane preparation and in an
octylglucoside extract, Tsr-T156C was protected by L-serine from NEM modification at 20, 25, and 30 °C to almost the same degree. By contrast, incubation at higher temperature (50 °C) almost
completely abolished the protective effect of serine (data not shown),
indicating (partial) denaturation of the protein. These results suggest
that the ligand-binding affinity of the chemoreceptor is not
significantly affected by a moderate change in temperature which causes
thermotaxis.
We used sulfhydryl modification to investigate the ligand-binding affinity of Tsr-T156C under various conditions. Whereas an attractant binds reversibly to Tsr, NEM covalently attaches to the cysteine residue. Therefore, the amount of NEM-labeled to Tsr-T156C increases even in the presence of the competitor during incubation. In this study, we could not determine absolute values of the dissociation constants. More detailed analyses would be required for obtaining such values as demonstrated for dissociation constants between an enzyme and its substrate or inhibitors (26). However, when Tsr-T156C was incubated with NEM for a fixed length of time in the presence of L-serine, D-serine, or L-alanine, the magnitudes of protection were consistent with the apparent affinities of Tsr-T156C for those attractants. Therefore, we could compare relative affinities of Tsr for L-serine under various conditions simply by monitoring the protective effect of L-serine on NEM modification during a fixed length of time.
Using this assay system, we found that the serine-binding affinity of Tsr is uncoupled from its methylation-modulated signaling states both in vivo and in vitro. This is consistent with recent in vitro studies on Tar (11, 12) and Tsr (13). These results suggest that "reverse transmembrane signaling" from the inside to the outside of the cell membrane does not happen in adaptation. However, a previous report (14) demonstrated that methylation of Tsr and Tar substantially decreased the ligand affinity of receptors in the membrane preparations. The discrepancy might be explained by the fact that the latter study did not monitor ligand binding to the receptors directly or by differences in the amount of the receptors in the membrane preparation. Alternatively, the cytoplasmic Che proteins, especially CheA and CheW, which form a ternary complex with the chemoreceptor, might contribute to reduction of the ligand-binding affinity upon methylation, which cannot be observed for the bare receptor, as has been discussed before (13). We cannot rule out this possibility since Tsr-T156C was overproduced in our experiments, and some would not be in ternary complexes with CheA and CheW.
More surprisingly, we found that the serine-binding affinity of Tsr is also uncoupled from its temperature-modulated signaling states in vivo and in vitro. Thus, a moderate change in temperature in the range between 20 and 30 °C may not cause a dramatic change in the receptor structure, such as monomer-dimer transition or a profound conformational change, but may cause some effects limited to the cytoplasmic domain. These local structural changes should be the nature of thermosensing by the chemoreceptor. The importance of the cytoplasmic domain in thermosensing by Tar has been suggested by a previous study (20), which showed that the primary translational product of tar does not have thermosensing ability and the unmethylated and the heavily methylated forms of Tar function as warm and cold receptors, respectively. In this regard, Tsr behaves differently. The unmodified form functions as a warm receptor (Table I), and it seems to lose thermosensing ability upon methylation (27, 28). These differences between Tsr and Tar might be related to differences in the number of methylation sites (25, 29). In any case, covalent modification in the cytoplasmic domain dramatically affects the thermosensing property of the receptor, a fact which is consistent with the notion that the structure of the cytoplasmic domain is altered by temperature changes.
Our results provide insight into the molecular architecture of the
receptor. That the ligand-binding domain may be fairly stable in its
overall organization is suggested by crystallographic analyses
(30-32). There is only a small difference between the ligand-occupied
and -free structures of a periplasmic fragment of Tar. Recent studies
suggest that the first transmembrane helices (TM1 and TM1) pair
stably, and binding of the ligand causes some movement of the second
transmembrane helix (TM2) relative to TM1-TM1
(33-36). However, it
remains unclear what happens to the cytoplasmic domain. Recent in
vitro studies suggest that dimerization (or oligomerization) of
the cytoplasmic domain of Tar plays a critical role (37, 38). On the
other hand, the assays for genetic complementation between two mutant
receptors suggest that only one intact cytoplasmic domain per receptor
dimer is sufficient for signaling (39, 40). Therefore, the signaling
mechanism may involve interactions between dimers, conformational
changes within a single cytoplasmic domain, or both. In any case,
binding of an attractant followed by a relatively small change in the
periplasmic domain triggers a dramatic inactivation of CheA kinase.
Therefore, it is reasonable to assume that the two different signaling
states (CW and CCW; kinase on and off) of the receptor are caused by a
relatively large structural change of the cytoplasmic domain. If so,
covalent modification and temperature changes should cause similar
structural changes in the cytoplasmic domain. However, under such
conditions, the structure of the periplasmic domain remains unaffected,
suggesting structural uncoupling between the two domains across the
membrane.
We thank Dr. M. I. Simon for generously providing plasmids, and S. Nishiyama and I. Tatsuno for collaborating in preparation of antiserum. We especially thank Dr. J. S. Parkinson for critically reading the manuscript.