(Received for publication, August 20, 1996, and in revised form, February 10, 1997)
From the The galactose-H+ symport
protein (GalP) of Escherichia coli is very similar to the
human glucose transport protein, GLUT1, and both contain a highly
conserved Asn residue in predicted helix 11 that is different in a
cytochalasin B-resistant member of this sugar transport family (XylE).
The role of the Asn394 residue (which is predicted to be in
putative trans-membrane The D-galactose-H+ symporter
(GalP)1 from Escherichia coli is
homologous to the L-arabinose-H+ symporter
(AraE) and the D-xylose-H+ symporters (XylE) of
E. coli, with 64 and 33% identity, respectively (1, 2).
These transporters are also homologous to a family of mammalian passive
facilitated glucose transporters (GLUT) (2-9). The inclusion of both
passive and active transporters in the homologous transporter family
implies they share common features of both structure and molecular
mechanism. Furthermore, the E. coli and mammalian proteins
are predicted to have a similar membrane topology, comprising 12 membrane spanning
This suggestion is reinforced by the observation that GalP-mediated
sugar transport is inhibited by the antibiotics cytochalasin B and
forskolin (13-17), which are also potent inhibitors of glucose transport mediated by GLUT1, GLUT2, GLUT3, and GLUT4 (18-29). The structures of these antibiotics are shown in Fig. 1.
Both the GLUT1 and GalP proteins can be photolabeled by cytochalasin B and a derivative of forskolin (IAPS-forskolin) in a
substrate-protectable manner (2, 13, 15, 16, 30-34). The binding of
both cytochalasin B and forskolin induces a quench in the intrinsic
protein fluorescence of both GLUT1 and GalP (14, 16, 22, 35-37). This
phenomenon has allowed the kinetics of the binding of cytochalasin B to
GLUT1 and cytochalasin B and forskolin to GalP to be time-resolved
using stopped-flow fluorescence spectroscopy; the binding of
cytochalasin B and forskolin to GalP and of cytochalasin B to GLUT1
occurs by similar mechanisms, with at least three conformational
states of the transport proteins identified (14, 16, 38).
The identification of transporter amino acid residues involved in the
interaction with antibiotics will help elucidate the mechanism of
recognition of both antibiotics and substrates. To this end, we have
exploited the homology of the bacterial and mammalian transporters and
compared their aligned amino acid sequences with their reported
sensitivities to inhibition by cytochalasin B; it is probable that
conserved residues are important in maintaining a common structure and
mechanism and that differences are associated with variations in sugar
specificity and antibiotic binding. In previous studies the covalent
photolabeling of GLUT1 with cytochalasin B has been localized, by
peptide mapping experiments, to a region containing amino acids
Phe389 to Trp412 (30-33, 39), which is
proposed to form part of TM11 and the cytoplasmic loop joining it to
TM10. Site-directed mutagenesis experiments on GLUT1 (40) led to a
model where cytochalasin B can photolabel either Trp388 or
Trp412. These regions of the aligned sequences of GalP,
AraE, XylE, GLUT1, GLUT2, GLUT3, GLUT4, GLUT7, and a high affinity
D-fructose transporter, GLUT5 (41), were compared. The
transporters GalP, AraE, GLUT1, GLUT2, GLUT3, and GLUT4 have a
conserved Asn residue and are all susceptible to inhibition by
cytochalasin B. In contrast, the transporters XylE, GLUT5, and GLUT7 do
not have a conserved Asn residue and show no (XylE22 (41)) or little (GLUT7 (42)) sensitivity to
inhibition by cytochalasin B. We therefore predicted that the conserved
Asn residue would be involved in cytochalasin B binding. To test this
hypothesis Asn394 of GalP, which is predicted in a
three-dimensional model to be in the middle of TM11 (Fig. 1), was
changed by site-directed mutagenesis to Gln. The ability to
over-express wild-type and mutant GalP proteins in high yield in
appropriate E. coli strains has enabled us to achieve a
rigorous characterization, which shows that changing Asn394
to Gln markedly diminishes sensitivity to the antibiotic cytochalasin B, but not to forskolin, and also diminishes binding of sugar to the
inward-facing side of the protein.
E. coli strain JM1100
HfrC his-gnd The E. coli strains JM1100(pPER3) (wild-type GalP (2)) and JM1100(pTPM6)
(GalP with the Asn394 The E. coli cells were disrupted by explosive
decomposition in a French Press at 137.5 MPa, and the inner membranes
were prepared essentially as described by Osborn et al.
(44). This procedure yields predominantly inside-out vesicles (45).
A sample of the
membrane preparation (30 µg of protein) was subjected to
SDS-polyacrylamide gel electrophoresis. After staining with Coomassie
Blue, gels were scanned with a Molecular Dynamics 100A Computing
Densitometer, and the proportion of GalP protein was measured. The
levels of expression of the wild-type and mutant proteins were
determined by comparison of the intensities of the bands corresponding
to the GalP proteins on Coomassie-stained SDS-polyacrylamide gels. The
expression of the Asn394 The concentration of protein in the membrane
preparations was assayed by the method of Schaffner and Weissman
(46).
The
codon (AAC) for asparagine 394 of the galP gene was replaced
by that of glutamine (CAG) as follows. The single-stranded DNA template
for mutagenesis was prepared by subcloning a 2.46-kilobase pair
AlwI-AlwI DNA fragment coding for the C-terminal
half of GalP from plasmid pPER3 (which confers elevated expression of GalP (2)) into bacteriophage M13mp18. Mutagenesis was carried out with
the Amersham Corp. in vitro mutagenesis kit using the mutagenic primer 5 The transport of
radioisotope-labeled sugars into intact cells was measured after the
cells were energized with 10 mM glycerol as described by
Henderson et al. (43). All transport measurements were
carried out on cells that had been grown on rich media producing levels
of GalP expression less than 1% total membrane protein. The level of
expression of the mutant GalP protein was very similar to that of wild
type judging by Western blots (data not shown). The velocity of
D-galactose uptake (with D-galactose at 50 µM) without the addition of inhibiting sugar did
not vary greatly between experiments showing little day to day
variation (wild type 24.3 ± 2.8 nmol/mg/min (n = 10) and Asn394 D-Galactose-promoted pH changes
(sugar-H+ symport) were measured using 3.3 mM sugar with energy-depleted anaerobic suspensions of
intact cells in 150 mM KCl, 2 mM glycylglycine,
pH 6.6, as detailed by Henderson and Macpherson (47). Initial
velocities of D-galactose uptake into E. coli
strains expressing wild-type and mutant GalP proteins were measured
after the cells had been aerated in the presence of glycerol (10 mM) and a range of sugar concentrations for 3 min.
Cells were made permeable to cytochalasin B and forskolin
by treatment with Tris-EDTA (48). Cells were grown for uptake experiments as described above. After washing with the growth volume
twice in 150 mM KCl, 5 mM MES, pH 6.5, the
cells were equilibrated in 200 mM Tris/HCl, pH 8.0, before
the addition of an equal volume of 200 mM Tris/HCl, pH 8.0, 1 mM EDTA. After 30 min, 100 volumes of 150 mM
KCl, 5 mM MES, pH 6.5, 10 mM MgSO4
were added and the cells harvested by centrifugation and resuspended in
the same buffer.
Ligands were incorporated into membrane
proteins by irradiation with ultraviolet light (13, 49). Inner membrane
vesicles were preincubated with or without 500 mM
D-galactose at 4 °C in photolabeling buffer (50 mM sodium phosphate, 100 mM NaCl, 1 mM EDTA, pH 7.4) and transferred to quartz cuvettes
containing 0.4 nM 3-125I-APS-forskolin or
0.5 µM 4-[3H]cytochalasin B. The samples
were flushed with argon to reduce free radical production and
irradiated with ultraviolet light for 10 min. After irradiation,
noncovalently bound ligand was removed by diluting the sample in
photolabeling buffer containing 1% mercaptoethanol (for
3-125I-APS-forskolin) or unlabeled cytochalasin B (for
4-[3H]cytochalasin B) and centrifugation (130,000 × g for 2 h at 4 °C). The resulting membrane pellet
was resuspended in 15 mM Tris/HCl, pH 7.5, and assayed for
protein. The proteins were separated by SDS-polyacrylamide gel
electrophoresis. The incorporation of 4-[3H]cytochalasin
B was monitored by fluorography and the incorporation of
3-125I-APS-forskolin by autoradiography.
Fluorescence spectra were measured in
a Perkin-Elmer LS50B spectrophotofluorimeter. The protein was excited
at either 280 or 297 nm and the fluorescence emission monitored between
300 and 400 nm. Rapid reactions were followed using an Applied
Photophysics (London, UK) spectrophotofluorimeter, operated at
20 °C, as described by Walmsley et al. (14, 17, 50).
Fluorescence measurements were carried out in 50 mM
potassium phosphate, 100 mM NaCl, and 1 mM
EDTA, buffered to pH 7.4 at 20 °C. Routinely the membranes were used
at a protein concentration of 200 µg/ml, in the above buffer.
Equilibrium dialysis was carried out
as described by Walmsley et al. (50). The binding of
12-[3H]forskolin and 4-[3H]cytochalasin B
to inner membranes (1.0 mg/ml) containing GalP was measured over the
range of 0.05-80 µM ligand at 4 °C. Ratios of bound
to free ligand were calculated from the equilibrium distribution of the
radiolabeled ligand and used to determine both the
Kd and the number of ligand binding sites by an
unweighted nonlinear least-squares fit of the data to an hyperbola
using the Biosoft program Ultrafit 2.1.
The apparent Km
values, with their standard deviations, for transport of radiolabeled
sugar were obtained by a least-squares fit of the unweighted initial
rate data directly to a hyperbola using Biosoft program Ultrafit 2.1. Generally, rates of the quench in the intrinsic fluorescence of GalP,
when it was mixed with the antibiotics in a stopped-flow apparatus,
were fitted to an exponential function using the non-linear
least-squares regression program supplied with the Applied Photophysics
SF.17MV stopped-flow equipment. The data sets, for the
concentration dependence of the rate of binding of the antibiotics,
were analyzed by fitting to the appropriate equation using the
non-linear least-squares regression program SIGMA PLOT (Jandel
Scientific).
The transport of
D-galactose by the wild-type GalP protein, under energized
conditions, was characterized by a hyperbolic increase in the initial
rate with increasing concentrations of sugar. A least-squares fit
yielded apparent Km and Vmax values of 42.4-58.9 µM and 59.1-65.4 nmol/mg/min
(Fig. 2 and Table I). The
Km and Vmax values quoted
here and subsequently are apparent as it was not known whether the
co-substrate, H+, was at saturating amounts. The
substitution of Asn394 by Gln led to a 2-fold increase in
the Km for D-galactose and a 2-fold
decrease in the Vmax, compared with the wild
type (Fig. 2 and Table I). The data indicated a relatively small
(4.8-fold) reduction in the specificity (as defined by
Vmax/Km) of the mutant
protein for D-galactose relative to the wild type. The
uptake of 50 µM D-galactose was accumulated
by the mutant to 42-fold that of the external media, compared with
88-fold by the wild type. These data demonstrated that the mutant
protein can catalyze energized transport and that the gross
conformation of the protein is not significantly impaired.
Table I.
The kinetic parameters Km and Vmax for the
transport of D-galactose, D-glucose, and
D-xylose catalyzed by the wild-type and Asn394 Department of Biochemistry and Molecular
Biology,
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION AND CONCLUSIONS
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
-helix 11) in the structure/activity
relationship of the D-galactose-H+ symporter
(GalP) was therefore assessed by measuring the interaction of sugar
substrates and the inhibitory antibiotics, cytochalasin B, and
forskolin with the wild-type and Asn394
Gln mutant
proteins. Steady-state fluorescence quenching experiments show that the
mutant protein binds cytochalasin B with a Kd 37-53-fold higher than the wild type. This low affinity binding was
not detected with equilibrium binding or photolabeling experiments. In
contrast, the mutant protein binds forskolin with a
Kd similar to that of the wild type and is
photolabeled by
3-125I-4-azido-phenethylamido-7-O-succinyl-desacetyl-forskolin.
The mutant protein displays an increased amount of steady-state
fluorescence quenching with the binding of forskolin, suggesting that
the substitution of the Asn residue has altered the environment of a
tryptophan, probably Trp395, in a conformationally active
region of the protein. Time-resolved fluorescence measurements on the
mutant protein provided association and dissociation rate constants
(k2 and k
2),
describing the initial interaction of cytochalasin B to the
inward-facing binding site (Ti), that are decreased
(9-fold) and increased (4.9-fold) compared with the wild type. This
yielded a dissociation constant (K2) for
cytochalasin B to the inward-facing binding site 44-fold higher than
that of the wild type. The binding of forskolin gave values for
k2 and k
2 3.9- and
3.6-fold lower, respectively, yielding a K2
value for Ti similar to that of the wild type. The low
overall affinity (high Kd) of the mutant protein for cytochalasin B is due mainly to a disruption in binding to the
Ti conformation. It is proposed that Asn394
forms either a direct binding interaction with cytochalasin B or is
part of the immediate environment of the binding site and that
Asn394 is in the immediate environment, but not part, of
the forskolin binding site. The ability of the mutant protein to
catalyze energized transport is only mildly impaired with 4.8- and
2.1-fold reduction in
Vmax/Km values for
D-galactose and D-glucose, respectively. In
stark contrast, the overall Kd describing binding of D-galactose and D-glucose to the
inward-facing conformation of the mutant and their subsequent
translocation across the membrane is substantially increased (64-fold
for D-galactose and 163.3-fold for D-glucose).
These data indicate that Asn394 is associated with both the
cytochalasin B and internal sugar binding sites. This conclusion is
also supported by data showing that the sugar specificity of the mutant
protein has been altered for D-xylose. This work powerfully
illustrates how comparisons of the aligned amino acid sequences of
homologous membrane proteins of unknown structure and characterization
of their phenotypes can be used to map substrate and ligand binding
sites.
-helices (12 TM), with helices 6 and 7 connected
by a cytoplasmic domain containing 60-70 amino acids (Fig. 1) (2, 8,
10). The sugar specificities of the E. coli transporters
differ, with GalP primarily transporting hexoses and AraE and XylE
pentoses. However, the sugar specificity of GalP is very similar to
that of the glucose transporters from human erythrocytes (GLUT1) and
rat adipocytes (GLUT4) (11-14) giving rise to the suggestion that GalP
is the bacterial equivalent of the mammalian glucose transporter (9).
Fig. 1.
Two-dimensional model of the
galactose-H+ symport protein. The Asn394
residue of GalP mutated to Gln in this study is indicated. The Asn
residue is predicted to be in the middle of TM11. The structures of the
sugar substrate D-galactose and of the inhibitory
antibiotics cytochalasin B and forskolin are also shown.
[View Larger Version of this Image (67K GIF file)]
Bacterial Strains
thyA galK ptsM galP mglP
ptsF ptsG (43), containing plasmids with wild-type or mutated galP genes, was the host strain for over-production of the
GalP protein. After the mutagenic reaction (see below) the recombinant bacteriophage M13 DNA was transfected into E. coli strain
TG1
(lac-proAB) supE thi
hsd
5 F
[traD36
proAB+ lacIq
lacZ
M15]. The modification (
) and restriction (
)
phenotype of strain TG1 necessitated that the mutant plasmids
constructed from DNA isolated from strain TG1 were transformed into
strain DH1 (modification (+), restriction (
)) before introduction
into E. coli strain JM1100 (modification (+), restriction
(+) phenotype), to prevent restriction of unmodified plasmid DNA.
E. coli strain DH1 has the genotype supE44
hsdR17 recA1 endA1 gyrA96
thi-1 relA1.
Gln mutation, this work) were
used for the constitutive over-expression of the wild-type and mutated
GalP proteins. To gain maximum expression of GalP, strains were grown
overnight in minimal media with 15 µg/ml tetracycline, supplemented
with L-histidine (80 µg/ml) and thymine (20 µg/ml). For
sugar transport and sugar-H+ symport experiments, requiring
lower levels of expression, cells were grown on rich media (2TY
supplemented with 20 mM glycerol, 20 µg/ml thymine, and
15 µg/ml tetracycline).
Gln mutant protein was 107%
that of the wild-type protein.
-GGCAATCCACTGGGTGGCAGTG-3
(the mismatches are underlined) complementary to the coding strand of
galP. The base changes were confirmed by single-stranded
dideoxynucleotide sequencing. A double-stranded 1.561-kilobase pair
MluI-NdeI fragment containing the base changes
was isolated from the bacteriophage M13 construct and substituted for
the wild-type DNA in MluI-NdeI cut plasmid pPER3
to give plasmid pTPM6 (Asn394
Gln). The substitution
was confirmed by subcloning the DNA containing the mutation back into
bacteriophage M13mp18 and dideoxynucleotide sequencing. Double-stranded
sequencing of the entire mutant galP gene in the expression
plasmid confirmed that there were no other base changes in the
galP gene sequence.
Gln 6.73 ± 0.88 nmol/mg/min
(n = 10)).
Substitution of Asn394 by Gln Does Not Greatly Impair
the Sugar Transport Activity of GalP
Fig. 2.
Energized transport catalyzed by the
Asn394 Gln mutant protein is mildly impaired. E. coli strain JM1100 expressing the wild-type and mutant proteins
were grown in 2 TY supplemented with 20 mM glycerol until
A680 = 0.6. The cells were harvested and washed
twice in the growth volume of 150 mM KCl, 5 mM
MES, pH 6.6, and finally resuspended to a density of 0.68 mg of dry cell mass/ml. Aerated samples (0.25 ml) were incubated at 25 °C for
3 min with 25 mM glycerol before addition of 6.25 µl of
radiolabeled sugar to give the indicated concentration range. After
15 s, 2 ml was filtered rapidly and washed four times in 2 ml of
the above medium. The initial rates of sugar uptake (nmol/min per dry
mass) were calculated and displayed as a Lineweaver-Burk plot. The
graph is a typical example of the uptake of
1-[3H]D-galactose catalyzed by wild-type and
mutant proteins. The data yielded Km values of
58.9 ± 8.6 and 107.6 ± 34.0 µM and
Vmax values of 65.4 ± 2.3 and 28.5 ± 2.2 nmol/mg/min for wild-type and mutant proteins, respectively.
Similar experiments were carried out on the uptake of
D-glucose and D-xylose, and the results are
shown in Table I.
[View Larger Version of this Image (22K GIF file)]
Gln mutant GalP proteins
Parameter
Wild-type
Asn394
Gln
Km
(D-galactose)
42.4-58.9
µM
107.6 ± 3 µM
Vmax (D-galactose)
59.1-65.4
nmol/mg/min
28.5 ± 2.2 nmol/mg/min
Vmax/Km
(D-galactose)
1.4-1.1
nmol/mg/min/µM
0.26 nmol/mg/min/µM
Km (D-glucose)
10.2 ± 3.5 µM
34.2 ± 3.2 µM
Vmax (D-glucose)
15.6 ± 1.4 nmol/mg/min
23.6 ± 0.9 nmol/mg/min
Vmax/Km
(D-glucose)
1.5 nmol/mg/min/µM
0.7
nmol/mg/min/µM
Km
(D-xylose)
3800 ± 600 µM
16500 ± 7900 µM
Vmax
(D-xylose)
297.8 ± 25.5 nmol/mg/min
241 ± 88.9 nmol/mg/min
Vmax/Km
(D-xylose)
0.0768 nmol/mg/min/µM
0.0146
nmol/mg/min/µM
D-Glucose was transported into the mutant with a Vmax 1.5-fold higher than that of the wild-type protein, although the Km was 3.3-fold higher (Table I). The transport of D-xylose was characterized by a 4.3-fold increase in the Km and a 0.8-fold decrease in the Vmax compared with that of the wild type. The Vmax/Km values were 2.1-fold (D-glucose) and 5.3-fold lower (D-xylose) than that of the wild type.
The Asn394The sugar-H+
proton symport activities of the wild-type and mutant proteins were
assayed by monitoring D-galactose-promoted alkaline pH
changes (data not shown), which are diagnostic of sugar-H+
symport (51, 52). The extent of D-galactose-H+
symport catalyzed by the mutant protein (5.19 ± 2.7 nmol of
H+ per mg (n = 5)) was not significantly
different from that of the wild-type protein (4.87 ± 3.88 nmol of
H+ per mg (n = 8)). However, the rate of
D-galactose H+-symport catalyzed by the mutant
(1.16 ± 0.41 nmol of H+ per mg (n = 5)) was 2.8-fold lower that of the wild type (3.31 ± 1.54 nmol of
H+ per mg (n = 8)). Again this implies that
the gross conformation and transport competence of the mutant protein
are not seriously affected by the Asn394 Gln
mutation.
The abilities of cytochalasin B or forskolin to
inhibit the energized uptake of 50 µM
D-galactose catalyzed by wild-type and mutant GalP into
E. coli cells were investigated. The cell walls were
rendered permeable to the antibiotics by prior treatment with
Tris-EDTA, pH 8.0 (see "Materials and Methods"). In the case of the
wild type, both cytochalasin B (80 µM) and forskolin (80 µM) inhibited the transporter, cytochalasin B being a
slightly more potent inhibitor (Fig. 3). However, the
transport catalyzed by the mutant protein was not significantly
inhibited by cytochalasin B, although it demonstrated a marked increase
in sensitivity to inhibition by forskolin.
The Asn394
The wild-type GalP
protein was susceptible to photolabeling with
4-[3H]cytochalasin B and
3-125I-APS-forskolin, and the protein was protected against
this photolabeling by the physiological substrate,
D-galactose, but not by L-galactose, which is
not a substrate (Fig. 4, lanes 2 and
3). This protection by substrate showed that both
cytochalasin B and IAPS-forskolin specifically label the GalP protein
(13-15).
The Asn394 Gln mutant GalP protein was photolabeled by
3-125I-APS-forskolin, and D-galactose, but not
L-galactose, protected the protein against the ligand; the
extent of labeling and of protection were at levels similar to those of
the wild-type protein (Fig. 4, B and D).
However, no photolabeling of the mutant protein by cytochalasin B
was detected (Fig. 4, A and C).
The equilibrium binding of the antibiotics cytochalasin B and forskolin to GalP over-expressed in E. coli is characterized by linear Scatchard plots indicating a single species of high affinity binding sites; furthermore, one unlabeled antibiotic displaces the other labeled one in a simple competitive manner (2, 15). When the data were fitted to a single hyperbolic function, by least-squares analysis (53), they revealed dissociation constants (Kd) for cytochalasin B binding of 4.5 ± 1.1 µM (n = 10) and for forskolin binding of 1.36-1.43 µM (2, 15). The binding sites were measured at 12.8 ± 2.9 nmol/mg membrane protein (n = 10) and 7.2-8.1 nmol/mg membrane protein for cytochalasin B and forskolin, respectively (2, 15). These values were compared with the amount of expressed GalP as determined by quantitative densitometry of SDS-polyacrylamide gel electrophoresis gels (2, 54) and indicated a 1:1 ratio for the binding of the antibiotics to the GalP protein.
To determine the concentration of the binding sites and the
Kd values for the binding of cytochalasin B and
forskolin to the mutant protein, equilibrium dialysis measurements were carried out with a range of antibiotic concentrations spanning between
0.05 and 80 µM (the maximum level of solubility). In
separate experiments, the concentration of forskolin binding sites was measured at 8.1 ± 0.66 and 6.04 ± 0.4 nmol/mg membrane
protein and is similar to that of the wild-type protein. The
Kd for forskolin binding was 1.28 ± 0.34 and
2.16 ± 0.69 µM (also similar to that of the
wild-type GalP). In contrast to the binding of forskolin, no
cytochalasin B binding could be detected even when using a specific
activity of 4-[3H]cytochalasin B 40-fold of that usually
used. The horizontal Scatchard plot of the data set
(Fig. 5) show that the small amount of bound
4-[3H]cytochalasin B was not displaced by increasing
amounts of cytochalasin B and is indicative of nonspecific binding.
Measurement of the Steady-state Kinetic Parameters for the Binding of Antibiotics to the Asn394
The binding of the antibiotics cytochalasin B and forskolin to the over-expressed wild-type GalP protein was characterized by a quench in the intrinsic protein fluorescence. This has previously allowed us to monitor the binding of these antibiotics by a combination of steady-state and stopped-flow spectrofluorimetric techniques, providing a rigorous kinetic analysis of the binding process (14, 16).
To assess the importance of Asn394 in the binding of
antibiotic, the intrinsic fluorescence of the mutant protein was
titrated with antibiotic over a range of concentrations up to 80 µM (Figs. 6 and 7). A fit of the data to a hyperbola
yielded an overall dissociation constant (Kd) for
forskolin binding of 1.48 ± 0.5 µM
(n = 3), similar to that of 1.83 ± 0.56 µM determined for the wild type (16). This is consistent
with the equilibrium dialysis data (above), which showed similar
Kd values for the binding of forskolin to both
wild-type and mutant proteins. In contrast to the equilibrium dialysis
and photolabeling experiments described above, some low affinity
cytochalasin B binding to the mutant protein was detected using the
fluorescence method; however, this was shown to be with an affinity 37- and 53-fold lower than the wild-type protein (Kd
41.4 and 58.5 µM) in two separate experiments.
The extent of the quenching of fluorescence by saturating amounts of
forskolin (10 µM) was compared between the wild-type and
Asn394 Gln mutant proteins. Measurements were made with
ex at 280 and 297 (
em max 330-340 nm) to
detect any differences due to light absorption by tyrosine and/or
tryptophan residues and at two protein concentrations (50 and 100 µg/ml) to detect any differences due to the light scattering
properties of the membrane preparations; no differences in the level of
quench was found at either wavelength or between the two protein
concentrations in either preparation. A 1.4-fold increase in the
fluorescence quench of the mutant protein, relative to the wild type
(wild type 8.09 ± 0.32% (n = 7);
Asn394
Gln 11.36 ± 0.88% (n = 6)), was observed with
ex 280 and 1.8-fold increase with
ex 297 (wild type 6.16 ± 0.49% (n = 5); Asn394
Gln 11.06 ± 0.14%
(n = 4)). A similar quantification of the amount of
fluorescence quench concomitant with the binding of cytochalasin B to
the mutant protein could not be made as saturating amounts of ligand
could not be obtained.
The binding of either forskolin or cytochalasin B to the GalP protein was previously analyzed by stopped-flow fluorescence spectroscopy in terms of the following three-step Model A (14, 16, 38).
![]() |
![]() |
For such a model, the dissociation constant of the transporter-forskolin complex is given by Equation 1.
![]() |
(Eq. 1) |
The Asn394 Gln GalP mutant protein was characterized by
a concentration-dependent linear increase in the rate of
the fast phase of forskolin binding to the Ti conformation
of the protein (Fig. 6), indicating association
(k2) and dissociation rate constants (k
2) of 1.51 ± 0.1 µM
1 s
1 and 2.37 ± 0.96 s
1, respectively; the dissociation constant
(K2
(k
2/k2)) was calculated
to be 1.57 ± 0.64 (Table II). The rates of
association and dissociation of forskolin with the mutant were 4- and
3.6-fold slower, respectively, than for the wild type. Thus, the
dissociation constant (K2) of the mutant is
almost identical to the wild type, demonstrating that the affinity of
the Ti or putative inward-facing transporter of the mutant
protein for forskolin is unchanged. As K2 and
overall Kd values are similar for both wild-type and
mutant proteins there is no evidence to suggest changes to either
K1 or K3.
|
Cytochalasin B also bound to the mutant protein in a linear,
concentration-dependent manner (Fig. 7),
indicating an association rate constant of 0.67 ± 0.28 µM1 s
1. In previous work with
the wild-type GalP protein the dissociation rate constant for
cytochalasin B was measured by directly displacing ligand, bound to the
transporter, by adding sugar substrate (49). The transport protein was
previously equilibrated with 80 µM ligand and mixed with
500 mM sugar in the stopped-flow apparatus to displace the
bound ligand. The increase in protein fluorescence resulting from
displacement of the bound cytochalasin B by D-galactose
could be fitted to a single exponential function revealing a
dissociation rate constant of 12.18 ± 2.78 s
1. The
dissociation constant K2
(k
2/k2) for the
Ti conformation of the mutant protein was calculated as
18.1 ± 8.78 µM. This value is 44-fold higher than
that of the wild type and is similar to the 37-53-fold decrease in the
overall affinity (increase in overall Kd) for
cytochalasin B.
8-Anilino-1-naphthalene sulfonate (ANS)
can be used as a fluorescent probe to monitor conformational changes in
the GalP protein induced by the binding of sugar (17). Transported
sugars such as D-galactose and D-glucose cause
an enhancement in the ANS fluorescence; non-transported sugar analogues
and antibiotics produce only a slight quench in the ANS fluorescence,
but they can reverse the enhancement in fluorescence induced by
transported sugars. The increase in ANS fluorescence is attributable to
the sugar-induced reorientation of the transporter from an inward to an
outward-facing conformation. Non-transported ligands are thought to
reverse the fluorescence enhancement by recruiting the transporter to
an inward-facing conformation. Before interpreting the effects of the
Asn394 Gln mutation, we first considered the kinetic
model for the wild-type protein, which was derived from measurements of
ANS fluorescence (17).
The rate of enhancement in ANS fluorescence produced by D-galactose was measured by stopped-flow spectrofluorimetry (17) and was found to increase in a biphasic manner as [sugar] increased. The fast phase was attributed to the initial binding of the sugar to the inward-facing conformation of the protein and the slow phase to a subsequent conformational change. The rate of the fast phase increased hyperbolically with increasing D-galactose concentration indicative of the rapid binding of the transporter that was rate-limited by a slower conformational change at near saturating concentrations (see Equation 2).
![]() |
(Eq. 2) |
The change in fast phase was fitted to the following hyperbolic function as shown in Equation 3.
![]() |
(Eq. 3) |
The overall Kd for D-galactose binding was calculated at 1.8 mM from the measured values for KD-gal and K2 by the relationship shown in Equation 4.
![]() |
(Eq. 4) |
![]() |
(Eq. 5) |
Hence the Kd would be as shown in Equation 6.
![]() |
(Eq. 6) |
Substrate Specificity of the Asn394
To determine the substrate specificity of the
Asn394 Gln mutant protein, an extensive kinetic
analysis was carried out in which increasing concentrations of the
unlabeled sugar under investigation were used to inhibit labeled
1-[3H]D-galactose transport over a range of
concentrations of both sugars.
Initial velocities of D-galactose uptake catalyzed by the wild-type and mutant proteins into E. coli were measured. The 1/v against [I] plots (Dixon plots) for D-galactose were linear, as expected for simple hyperbolic binding, in the presence of unlabeled D-galactose, 2-deoxy-D-glucose, D-talose, 6-deoxy-D-glucose, 2-deoxy-D-galactose and D-mannose. At higher concentrations of D-xylose, 6-deoxy-D-galactose (wild type only), D-glucose, and L-arabinose the Dixon plots were non-linear and deviated systematically from hyperbolic binding (data not shown). However, the data points at the lower concentrations of these inhibiting substrates did display linearity, and extrapolated estimates of Ki were determined. It was reported previously (13) that the kinetics of transport of D-glucose, 2-deoxy-D-glucose, and 6-deoxy-D-galactose transport by GalP were non-hyperbolic. In order that changes in substrate specificity of the mutant protein could be readily detected, a ratio of mutant Ki to wild-type Ki was calculated, which was then normalized by dividing with the ratio for D-galactose (Table III). The recognition by the mutant protein of most of the sugar substrates was only altered from that of the wild type between 0.5-2.0-fold, indicating little change in sugar selectivity. Surprisingly, the specificity for D-xylose was seven times lower with the mutant protein compared with the wild type. There is no obvious, simplistic interpretation of this effect such as an altered substrate recognition in terms of a single H-bonded sugar-protein interaction. Clearly the mutation of the Asn394 residue to Gln has, in terms of D-xylose, selectively affected the substrate specificity of the GalP protein.
|
An alignment of the amino acid sequences of those proteins that are homologous to GalP, together with a knowledge of the ability of these transporters to bind cytochalasin B, suggested that the Asn394 residue might be involved in the recognition of cytochalasin B as a part of the cytochalasin B binding site or pocket. For the purposes of this work we will define the binding site as comprising those amino acids that bind directly with the antibiotic, and the binding pocket is defined as amino acids in the immediate environment of the antibiotic that do not necessarily bind directly to it. To investigate the involvement of Asn394, the residue was altered by mutagenesis to Gln.
Inward D-galactose transport catalyzed by the mutant protein, as defined by the parameters Km and Vmax, was relatively unimpaired, implying that the gross conformation of the protein was intact and the transport mechanism conserved. Similarly, the mutant protein was able to catalyze sugar proton symport, to an extent similar to wild type albeit at a 2.8-fold lower rate, indicating that coupling between proton and sugar transport was maintained.
Previous equilibrium binding experiments on the wild-type protein
revealed mutual competition between cytochalasin B and forskolin for
the antibiotic binding site (16). It was therefore of interest to
investigate the effects of the Asn394 Gln substitution
on the interaction of GalP with both antibiotics. The mutant protein
was photolabeled by IAPS-forskolin at levels similar to that of the
wild-type protein, and the reaction was protected by
D-galactose. These data showed that Asn394 is
not essential for photolabeling by forskolin. However, it cannot be
ruled out that Asn394 of wild-type GalP is the photolabeled
amino acid residue and that Gln394 is being photolabeled in
the mutant protein. Positive identification must wait for the direct
sequencing of protein fragments generated from photolabeled
transporter. A Scatchard analysis demonstrated that the overall
affinity for forskolin binding to the mutant protein and the number of
forskolin binding sites were almost identical to those of the wild-type
protein; this unchanged overall affinity of the mutant protein for
forskolin was confirmed by the titration of the protein's
fluorescence. D-Galactose transport catalyzed by the mutant
protein demonstrated a marked increase in sensitivity to inhibition by
forskolin compared with the wild type (wild type 58.0 ± 2.9% and
Asn394
Gln 7.3 ± 2.9% compared with the ethanol
control). This increase probably reflects a decrease in the level of
competition by D-galactose. The maximal fluorescence quench
induced by forskolin binding was 1.8-fold greater with the mutant
protein than the wild type (
ex 297;
em
max 330-340 nm), although the amount of GalP was the same. This
indicates that the substitution of Asn394 by Gln altered
the environment of one or more tryptophan residues. Characterization of
a Trp395
Phe mutant GalP protein indicated that
Trp395 is in a conformationally active region of the
protein and that this residue is the main reporter of forskolin binding
in fluorescence quenching experiments (38). The increase in the maximal
fluorescence quench due to the binding of forskolin to the
Asn394
Gln mutant can be rationalized by suggesting
that the presence of Gln394 causes a localized
conformational change altering the environment of the neighboring
Trp395 residue. Although the overall affinity for forskolin
binding to the mutant protein is unaffected, transient fluorescence
measurements showed that the rates of association
(k2) and dissociation
(k
2) to the inward-facing conformation of the
protein (Ti) are reduced (3.9- and 3.6-fold for
k2 and k
2,
respectively). The affinity for the initial interaction with forskolin
to the inward-facing conformation of the protein (Ti), as
defined by the dissociation constant K2
(k
2/k2), was unchanged.
As K2 and overall Kd values
are similar for both wild-type and mutant proteins, there is no
evidence to suggest changes to either K1 or
K3.
In stark contrast to the binding of forskolin, cytochalasin B was shown, in fluorescence quenching experiments, to bind to the mutant protein with an overall affinity 37-53-fold lower than that of the wild-type protein. The displacement of cytochalasin B from the mutant protein by sugar substrate provided corroborating evidence of ligand binding. This low affinity cytochalasin B binding was not detectable at all in photolabeling or equilibrium dialysis experiments, most likely due to an insufficient concentration or specific activity of the radiolabeled cytochalasin B.
Transient fluorescence measurements showed that the affinity for the
initial interaction with cytochalasin B to the inward-facing conformation of the protein (K2, as defined by
k2/k2) was also much
lower than wild type (40-fold lower). The reduction in overall affinity
for cytochalasin B cannot be attributed to stabilization of the
To (outward-facing) conformation of the mutant transporter,
otherwise a similar effect should be observed with forskolin binding;
thus, it is most likely to be entirely due to a reduction in affinity
of the Ti conformation. Cytochalasin B was shown to have a
rate of association to the mutant protein (k2)
lower than that of the wild type (9-fold) and as such is similar to
forskolin. However, in direct contrast with the rate of forskolin
dissociation, that of cytochalasin B from the mutant protein is higher
(4.9-fold) than that of the wild type.
Possible explanations for these data are that the amino acid substitution has 1) altered the shape of the forskolin binding pocket making it more difficult for the antibiotic to bind, but once bound it is held within the pocket for longer, and (2) disrupted a single H-bonded cytochalasin B-protein interaction resulting in a reduction of overall affinity for the ligand. The reduction in affinity for cytochalasin B and alteration of forskolin binding provides evidence that the cytochalasin B and forskolin binding sites are closely associated. We speculate that Asn394 forms part of the cytochalasin B binding site and part of the forskolin binding pocket. However, the possibility that the disruption in cytochalasin B binding is due to steric hindrance by the larger Gln residue cannot be discounted.
In contrast to the energized transport of sugar, which is relatively unimpaired, ANS fluorescence measurements show that the overall Kd describing the binding of D-galactose and D-glucose to the inward-facing conformation of the mutant protein and their subsequent translocation across the membrane were drastically higher than that for the wild-type protein (64.1-fold for D-galactose and 163.3-fold for D-glucose). The fact that energized transport catalyzed by the mutant transport is relatively unimpaired suggests that the translocation mechanism is intact. It is therefore most likely that the decrease in overall affinity is largely due to loss of affinity for these sugars at the inward-facing binding site.
The substitution of the Asn394 residue by Gln has, at least, at the primary structural level made the GalP protein more like XylE. Such single residue differences may be the means by which proteins are able to discriminate between different substrates, e.g. between hexoses and pentoses. The sugar specificity of the wild-type GalP protein has been extensively investigated (2, 13, 15, 17), which showed the importance of the C-3 hydroxyl and C-6 on the pyranose ring for hexose recognition by GalP and the lack of importance of the C-2 and C-4 hydroxyl groups. These observations parallel the requirements for sugar recognition by the passive glucose transporters of human erythrocytes (GLUT1) and rat adipocytes (GLUT4) (11, 12), confirming the functional similarity of the bacterial and mammalian transporters. The substrate specificity study carried out for energized uptake of sugars in this work did not reveal a general role for the Asn394 residue in the recognition of external sugar substrates, but, surprisingly, the specificity for D-xylose was significantly reduced (7.6-fold compared with the wild type); there is no obvious, simplistic interpretation of this effect, such as an altered substrate recognition in terms of a single H-bonded sugar-protein interaction. Overall, these observations support the other data that indicate Asn394 is part of the internal binding site for sugars.
Cytochalasin B has been shown by kinetic analysis to compete with
D-glucose for binding to the inward-facing conformation of
the GLUT1 protein (55). Our data, which demonstrate that the disruption
of sugar binding to the inward-facing conformation of GalP (as opposed
to the outward-facing conformation) is coupled to the disruption of
cytochalasin B binding by the Asn394 Gln mutation,
provide further evidence of the functional and structural similarity of
the GLUT1 and GalP proteins. These data contribute to the ongoing
debate as to whether cytochalasin B and sugar bind to the same site of
the inhibitor susceptible proteins and suggest, at least with GalP,
that this is the case.
An Asn415 Asp mutant of the GLUT1 transporter has been
isolated (56). Asn415 is conserved in all the GalP, AraE,
XylE, and GLUT transporters and is the 4th residue after
Asn394 in GalP. Characterization of the Asn415
Asp mutant GLUT1 protein showed 40% less cytochalasin B
photolabeling although the inhibition of labeling by ethylene glucose,
which preferentially binds to the outer glucose binding site (57), inhibited cytochalasin B similar to the wild type. This led to the
suggestion that the Asn415 residue is likely to reside
close to the inner glucose and cytochalasin B binding sites of GLUT1.
It is interesting to note that to make a complete turn of an
-helix,
the polypeptide backbone must traverse 3.6 amino acids. Thus amino
acids three or four amino acids apart (as in the case of
Asn412 (aligned with Asn394 of GalP) and
Asn415 of GLUT1) would be expected to project from the same
side of an
-helix. It has also been shown (38) that the affinity for the initial interaction of cytochalasin B to a Trp395
Phe mutant as described by the apparent dissociation constant (K2) is 43-fold lower than that for the wild
type while the K2 for forskolin binding to the
mutant is similar to wild type. These data suggest that the residue
adjacent to Asn394 (Trp395) is also associated
with the cytochalasin B but not the forskolin binding site.
Interestingly, the affinity for D-galactose of the inward-facing binding site of the Trp395
Phe mutant is
similar to wild type and is therefore unlikely to be involved in
D-galactose binding (38). It is by the collation of such
data from many mutants that structural and mechanistic models for these
related sugar transporters will begin to emerge, and the binding sites
of sugar and antibiotic will be determined.
We are also grateful to J. O'Reilly for isolating the inner membrane preparations from E. coli and D. Ashworth for oligonucleotide synthesis and DNA sequencing. We are indebted to Dr. M. F. Shanahan for the kind gift of 3-125I-APS-forskolin.
HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
All ASBMB Journals | Molecular and Cellular Proteomics |
Journal of Lipid Research | Biochemistry and Molecular Biology Education |