(Received for publication, March 28, 1997)
From the Department of Biochemistry and Biophysics, Washington State University, Pullman, Washington 99164-4660
We have shown that the complexes between SecB, a chaperone from Escherichia coli, and two physiological ligands, galactose-binding protein and maltose-binding protein, are in rapid, dynamic equilibrium between the bound and free states. Binding to SecB is readily reversible, and each time the ligand is released it undergoes a kinetic partitioning between folding to its native state and re-binding to SecB. Binding requires that the polypeptide be devoid of tertiary structure; once the protein has folded, it is no longer a ligand. Conditions were established in which folding of the polypeptides was sufficiently slow so that at each cycle of dissociation rebinding was favored over folding and a kinetically stable complex between SecB and each polypeptide ligand was observed. Evidence that the ligand is continually released to the bulk solution and rebound was obtained by altering the conditions to increase the rate of folding of each ligand so that folding of the ligand was faster than reassociation with SecB thereby allowing the system to partition to free SecB and folded polypeptide ligand. We conclude that complexes between the chaperone SecB and ligands are in dynamic, rapid equilibrium with the free states. This mode of binding is simpler than that documented for chaperones that function to facilitate folding such as the Hsp70s and Hsp60s, where hydrolysis of ATP is coupled to the binding and release of ligands. This difference may reflect the fact that SecB does not mediate folding but is specialized to facilitate protein export. Without a requirement for exogenous energy it efficiently performs its sole duty: to keep proteins in a nonnative conformation and thus competent for export.
SecB is a cytosolic chaperone from Escherichia coli
that binds a subset of polypeptides destined to be exported to the
periplasmic space or to the outer membrane. SecB keeps these proteins
in a nonnative conformation, which is competent for export, while
delivering them to SecA, the next component in the export pathway.
There is no specific sequence of amino acids that distinguishes a
polypeptide as a ligand for SecB. The only requirement for binding is
that the ligand be devoid of tertiary structure. The binding frames for
three physiological ligands, maltose-binding protein, galactose-binding protein, and the oligopeptide-binding protein, have been determined (1,
2).1 In all three cases it
was shown that the sequence in contact with SecB is large, minimally
comprising 150 amino acids in a continuous stretch. There is no
similarity in sequence among the binding frames nor is it obvious that
a structural element provides recognition. The large area of contact
between SecB and its ligands allows for interaction at multiple
subsites with each site having low affinity yet together giving tight
binding; the dissociation constant for ligands of SecB is in the range
of 108 M (3-5). Even though the affinity is
high there are several studies that indicate that the complex of SecB
and its ligand is in rapid equilibrium with the unbound forms (3, 4, 6, 7). These previous studies depended on the presence of a competing species and limiting SecB or on the addition of excess SecB to show
that binding is readily reversible. Thus one could not rule out the
possibility that binding of competitors at subsites might in a stepwise
fashion displace the ligand or that transfer between SecB tetramers
might occur directly without release of the ligand to the free state.
Here we show directly by changing only the folding rate of the ligand
that the complex of SecB and two physiological ligands,
galactose-binding protein and maltose-binding protein, is readily
reversible even in the absence of competing ligands. Under conditions
in which nonnative ligands fold slowly, a kinetically stable complex is
observed. Free ligand, present in low abundance, is not readily
detectable. By changing conditions such that only the folding rate of
the ligands is increased, we obtained evidence for the existence of
unbound ligand through the observation of an accumulation of folded,
free ligand.
Ultra-pure guanidinium chloride (GdmHCl)2 was from ICN Biochemicals, Inc. EGTA was from Fluka; calcium chloride and potassium acetate were from J. T. Baker Inc. All other chemicals were from Sigma.
Purification of ProteinsSecB, mature maltose-binding
protein, and mature galactose-binding protein were purified as
described previously (2, 8, 9). The protein concentrations were
determined by using extinction coefficients at 280 nm of 47,600 M1 cm
1 for tetramer SecB,
37,700 M
1 cm
1 for native
galactose-binding protein (2), and 1.94 ml mg
1
cm
1 for native maltose-binding protein (10). Precursor
galactose-binding protein was purified from a strain of E. coli, MM52, which carries a temperature-sensitive allele of
secA (11) and harbors plasmid pSF5 (12), which contains the
gene for galactose-binding protein under its natural promoter. A
pre-culture was grown at 30 °C in tryptone broth containing 50 µg/ml ampicillin and 0.2% glucose to repress expression of
galactose-binding protein. These cells were harvested by centrifugation
at 4,000 × g for 5 min, and the cell pellet was
suspended in 3 times the original volume of tryptone broth held at
42 °C, containing 1 mM fucose to induce expression of
galactose-binding protein. The culture was grown for 2 h at 42 °C to express the defect in SecA and thereby cause the
accumulation of precursor galactose-binding protein. The cells were
harvested, washed once with 10 mM Tris-Cl, pH 7.6, 0.3 M NaCl, and the cell pellet was frozen at
70 °C. The
frozen pellet was thawed and suspended in 20 mM Tris-Cl, pH
7.6, 2 mM EDTA, and 100 µg/ml lysozyme at a density of
3 × 1010 cells/ml. The cell suspension was sonicated
to lyse the cells and break the DNA and then centrifuged for 20 min at
27,000 × g. The pellet that contained membrane and
precursor galactose-binding protein in inclusion bodies was suspended
at an equivalent cell density of 3 × 1010 cells/ml in
10 mM Tris-Cl, pH 7.6, 3% (w/v) Triton X-100, and 5 mM EDTA to solubilize the membrane proteins. After
centrifugation for 20 min at 27,000 × g, the pelletted
inclusion bodies were solubilized in a minimal volume (approximately
3 × 1011 cell equivalents/ml) of 6 M
urea, 25 mM bis-Tris, pH 7.1, for chromatofocusing.
Chromatofocusing was performed using a Pharmacia MonoP HR5/20 column,
and the gradient was formed with Polybuffer 74 diluted 1/10 in 6 M urea and adjusted to pH 6.6 with HCl. Multiple runs were
necessary to avoid exceeding the capacity of the column. Fractions
containing the precursor galactose-binding protein as detected by
SDS-gel electrophoresis were pooled and dialyzed against 1 M GdmHCl, 20 mM HEPES-KOH, pH 7.4. Concentration was achieved by multiple rounds of reducing the volume to
1/6 in a SpeedVac (Savant) and dialyzing against 1 M
GdmHCl, 20 mM HEPES-KOH, pH 7.4. The final dialysis was
against 1 M GdmHCl, 10 mM HEPES-KOH, pH 7.6, 150 mM potassium acetate, and the concentrated, denatured protein was stored at
70 °C in this solution. The precursor
galactose-binding protein concentration was determined by absorbance at
280 nm of the unfolded protein using the calculated extinction
coefficient of 37400 M
1 cm
1
(13).
The refolding of the galactose-binding protein or maltose-binding protein was monitored by the increase of the intrinsic fluorescence of tryptophan. Galactose-binding protein contains 5 tryptophans (14, 15) and maltose-binding protein contains 8 (16). In neither case are there tryptophan residues in the leader sequence. The fluorescence intensity was measured using an excitation wavelength of 295 nm (slit width 5 nm) and an emission wavelength of 344 nm (slit width 2 nm) on a Shimadzu RF-540 fluorescence spectrophotometer at 5 °C. Refolding was initiated by manually adding 0.8 nmol of galactose-binding protein unfolded in 1.0 M GdmHCl or 0.14 nmol of maltose-binding protein unfolded in 1.5 M GdmHCl to a chilled, stirred solution containing a chosen concentration of guanidinium chloride, 10 mM HEPES-KOH, pH 7.0, 150 mM potassium acetate, and either 0.5 mM EGTA or 1.5 mM CaCl2. The fluorescence intensity was monitored with time, and the relaxation time to reach equilibrium was calculated. That galactose-binding protein was refolded to the native state was confirmed by comparison of the emission spectra before and after the addition of either galactose or glucose. Binding of galactose causes a blue shift (343-340 nm) and an increase in the fluorescence intensity; glucose binding causes only an increase in the fluorescence intensity. For refolding of all of the proteins studied greater than 80% of the expected fluorescence intensity change was accounted for. For each refolding reaction the GdmHCl concentration was accurately determined by measuring the index of refraction. The relaxation times reported for the refolding of precursor galactose-binding protein and mature galactose-binding protein in the presence of calcium are a weighted average of a slow folding rate (representing 60% of the total change in fluorescence for precursor galactose-binding protein and 50% for mature galactose-binding protein) and a fast folding rate. Relaxation times for refolding of both species of galactose-binding protein were measured at several GdmHCl concentrations and showed a direct dependence on the concentration of the denaturant.
Analysis of Complexes between SecB and the Binding ProteinsA complex between SecB and galactose-binding protein was formed by diluting unfolded galactose-binding protein to a solution held on ice containing SecB to give final concentrations of 0.09 M GdmHCl, 10 mM HEPES-KOH, pH 7.0, 150 mM potassium acetate, 0.5 mM EGTA, 5.5 µM SecB tetramer, and 3.7 µM galactose-binding protein. The sample was divided into two portions, and one part was analyzed by HPLC size exclusion chromatography using a TSKgel 3000.SW column (TosoHaas) equilibrated in 10 mM HEPES-KOH, pH 7.0, 150 mM potassium acetate, and 0.5 mM EGTA at 5 °C. To the second portion, CaCl2 was added to give 6 mM and the mixture analyzed on the TSKgel 3000.SW size exclusion column equilibrated in 10 mM HEPES-KOH, pH 7.0, 150 mM potassium acetate, and 1.5 mM CaCl2. Complexes between SecB and maltose-binding protein were formed exactly as for galactose-binding protein above. In all cases, the A280 was monitored, and 1-ml fractions were collected. Carrier protein (~1 µg) was added to each fraction to facilitate precipitation of the proteins of interest. The fractions were treated with 8% trichloroacetic acid, and the precipitated proteins were collected by centrifugation, washed with acetone, and suspended in sample buffer for 14% polyacrylamide gel electrophoresis. One quarter of the protein in each fraction and the equivalent of 1/20 of the sample applied to the column were loaded on the gel unless indicated otherwise.
Our working model describing the role of SecB in facilitating
protein export involves a kinetic partitioning of nonnative precursor
proteins between folding to the native conformation in the wrong
compartment, the cytosol, and binding to SecB which in turn leads to
entry into the export pathway. In this model the ligand is in dynamic,
rapid equilibrium between the free and bound states, and thus the
proportion of the polypeptides that are properly localized is a
function of the rate constant of association relative to the rate
constant of folding. To demonstrate that the ligand is constantly
sampling the free state thereby allowing a kinetic partitioning to
occur, we have made use of two ligands for which we can easily
manipulate the rate of folding within the range that is crucial to
poise the partitioning to favor either complex formation or folding.
Both ligands are sugar binding-proteins, which are exported in E. coli to the aqueous periplasmic space between the cytoplasmic and
outer membranes. The periplasmic binding proteins are monomeric
comprising two /
domains with the ligand binding site between the
domains. The native structure of the galactose-binding protein contains
one Ca2+ ion which stabilizes the structure (17).
Maltose-binding protein has a similar tertiary fold but contains no
metal ions (18).
The reversible folding of both maltose-binding protein and of
galactose-binding protein was monitored by fluorescence spectroscopy. The refolding of both proteins, unfolded by incubation with guanidinium chloride, results in a progressive increase in amplitude of the fluorescence signal that is attributed to the sequestering of tryptophanyl side chains in the interior of the native proteins. The
relaxation time to reach equilibrium in a refolding reaction carried
out at 5 °C in which mature galactose-binding protein was rapidly
subjected to dilution of the denaturant from 1.0 to 0.09 N
was 115 s in the presence of Ca2+, whereas the
relaxation time increased to 6300 s in the presence of EGTA (see
"Experimental Procedures" for details). The refolding of precursor
galactose-binding protein was also very sensitive to the presence of
calcium. In the presence of calcium precursor galactose-binding protein
folded with a relaxation time of 200 s (5 °C, final
concentration of guanidinium chloride, 0.09 M). However, in
the absence of calcium refolding of precursor galactose-binding protein
was so slow it was not measurable. Attempts to measure the folding rate
at 25 °C were also fruitless. The presence of the leader peptide and
the absence of Ca2+ apparently have a synergistic effect on
the refolding reaction. Even though the precursor galactose-binding
protein did not refold, it was not irreversibly aggregated since
addition of 6 mM CaCl2 to the cuvette
containing the precursor galactose-binding protein and EGTA caused a
rapid increase in fluorescence to the expected amplitude. Furthermore,
the unfolded precursor galactose-binding protein in solutions
containing EGTA eluted as a single, symmetrical peak with the expected
amplitude when analyzed by size exclusion chromatography (Fig.
2A, dotted line).
Maltose-binding protein does not contain bound calcium in its native structure, and as expected calcium had no effect on the rate of folding when the refolding of maltose-binding protein was investigated in a similar series of experiments (data not shown).
A complex between SecB and either mature galactose-binding protein or
precursor galactose-binding protein can be formed in the absence of
calcium since both species fold very slowly under these conditions.
These complexes were demonstrated by co-elution of SecB and mature or
precursor galactose-binding protein on a size exclusion column (Figs.
1A and
2A, dashed lines, Figs.
1B and 2B). The complex eluted before free SecB
or either refolded or unfolded galactose-binding protein (Figs.
1A and 2A, dotted lines). Since galactose-binding
protein folds so slowly in the absence of calcium, each time it
dissociated from SecB it remained a ligand and rebound, thus we
observed a complex.
If calcium was added to the pre-formed complexes comprising SecB and either mature or precursor galactose-binding protein and then the samples were analyzed on the size exclusion column, a complex was not observed: SecB and galactose-binding protein eluted independently (Figs. 1A and 2A, solid lines, Figs. 1C and 2C). Since the calcium increases the rate of folding of galactose-binding protein, it also increases the probability that when galactose-binding protein is released from SecB it will fold before it rebinds. Furthermore, the complex remained readily reversible even 25 h after formation; a complex was held on ice 25 h in EGTA and following addition of calcium analysis by size exclusion HPLC showed no complex present, whereas without the addition of calcium the galactose-binding protein was observed in complex with SecB (data not shown). This shows that even after 25 h the same principles continue to govern the equilibrium.
We eliminated the possibility that calcium interacts directly with SecB
to decrease its affinity for its ligands by demonstrating that calcium
had no effect on a complex formed between SecB and wild-type mature
maltose-binding protein, a ligand that folds at a rate unaffected by
the presence or absence of Ca2+. Thus, maltose-binding
protein was recovered in complex with SecB whether the size exclusion
column was run in calcium or EGTA and whether the complex was initially
formed in calcium or EGTA (Fig. 3,
A-C).
The rate of folding of maltose-binding protein can be increased by
raising the temperature, and we have previously shown that a complex
can be formed between SecB and maltose-binding protein at 5 °C where
the folding is sufficiently slow but not at 25 °C (9). As further
proof that SecB is continuously binding and releasing its ligand, a
complex between SecB and maltose-binding protein was formed at 5 °C
(Fig. 3A), and a portion of the sample was warmed to
25 °C. The difference in temperature affects only the folding rate
of maltose-binding protein and not the ability of the SecB to bind
ligand as has been demonstrated previously by the formation of
complexes between SecB and mutationally altered species of
maltose-binding protein that fold slowly (9, 19). As was the case for
galactose-binding protein, when the folding rate of the maltose-binding
protein ligand increased the complex was disrupted (Fig.
4).
SecB is classified as a protein chaperone based on its ability to selectively bind polypeptides that are in a nonnative state, but it is distinguished from other members of the family by at least one distinctive property. Whereas the chaperones such as GroEL and the Hsp70s, which facilitate folding, reversibly interact with their ligands in a cycle that is coupled to the binding and hydrolysis of ATP (20), SecB does not utilize exogenous energy to carry out its function. As we have shown here the complex between SecB and its ligand is in rapid, dynamic equilibrium with the uncomplexed proteins. Thus the ligand is continuously sampling the free state. It seems likely that this fundamental difference in the modes of binding is related to the fact that SecB is dedicated to protein export, whereas the chaperones that hydrolyze ATP are facilitators of folding of polypeptides. It has been proposed that GroE assists folding by imparting energy to the ligands via the hydrolysis of ATP to let them escape from kinetic traps that exist along the pathway to the low energy, native state (21). In this model the ligand is completely released from GroE. An alternative hypothesis is that GroE sequesters the folding polypeptide in a chamber within the GroEL·S chaperone complex, and the cycle of binding and hydrolysis of ATP is associated with release into the chamber. In this model the residence time of the ligand associated with the chaperone is regulated to allow folding (22). As chaperones, both GroEL and SecB bind unfolded proteins, but the function of GroEL is to facilitate folding of its ligand, whereas the function of SecB is just the opposite, to keep its ligands from folding. To play its role in export it simply must capture the polypeptide before it acquires structure so that the ligand will remain in a conformational state compatible with transfer through the membrane. Perhaps the binding mechanism used by SecB provides the least costly way to accomplish this. There is no need to impart energy to the ligand or to increase the residence time of association. In fact, since SecB appears to have no specificity other than for nonnative structure, it is likely that, particularly while nascent, many proteins that are not destined for export might associate with SecB. The fact that binding is readily reversible, rapid, and without need for exogenous energy would minimize the occupancy of SecB by the wrong ligands and maximize efficiency. Proteins that fold rapidly will partition to the native state, and those destined for export, which fold more slowly because of the presence of leader peptide or of other intrinsic folding properties, will rebind with high probability. Yet when SecB binds SecA, the next component of the export pathway, the ligand can be passively passed on provided that SecA has a higher affinity for ligand than does SecB.
We thank Vijaya J. Khisty for purifying the mature galactose-binding protein.