(Received for publication, April 1, 1996, and in revised form, October 2, 1996)
From the Centro de Biología Molecular
"Severo Ochoa" (CSIC-UAM), Cantoblanco, 28049 Madrid, Spain and
the ¶ Instituto de Química Física Rocasolano
(CSIC), Serrano 119, 28006 Madrid, Spain
The single-stranded DNA-binding protein (SSB) of
Bacillus subtilis phage 29 is absolutely required for
viral DNA replication in vivo. About ~95% of the
intrinsic tyrosine fluorescence of
29 SSB is quenched upon binding
to ssDNA, making tyrosine residues strong candidates to be directly
involved in complex formation with ssDNA. Thus, we have studied the
spectroscopic properties of the
29 SSB tyrosines (Tyr-50, Tyr-57,
and Tyr-76) using steady-state and time-resolved fluorescence
measurements.
29 SSB tyrosines do not seem to be highly restricted
by strong interactions with neighbor residues, as suggested by (i) the
high value of the average quantum yield of the
29 SSB fluorescence
emission (
F = 0.067 ± 0.010), (ii) the fast
motions of the tyrosine side chains (
short = 0.14 ± 0.06 ns), and (iii) the lack of tyrosinate emission at neutral pH.
Stern-Volmer analysis of the quenching by acrylamide and
I
indicates that
29 SSB tyrosines are surrounded by a
negatively charged environment and located in a relatively exposed
protein domain, accessible to the solvent and, likely, to ssDNA.
Changes in the intrinsic fluorescence upon ssDNA binding allowed us to determine that temperature has an opposite effect on the thermodynamic parameters K (intrinsic binding constant) and
(cooperativity) defining
29 SSB-poly(dT) interaction, the effective
DNA binding constant, Keff = K
,
being largely independent of temperature. Altogether, the fluorescent
properties of
29 SSB tyrosines are consistent with a direct
participation in complex formation with ssDNA.
Interactions of nucleic acids and proteins provide the structural
framework for a wide variety of biological processes. Among them,
complexes formed by the association of single-stranded DNA-binding proteins (SSBs)1 with single-stranded DNA
(ssDNA) are of particular interest. First, they modulate key processes
of DNA metabolism: DNA replication and DNA repair (1, 2), transcription
and translation regulation (3, 4), and recombination (5, 6, 7). Second,
although all SSBs bind ssDNA with high affinity, the functional outcome of this interaction could be radically different. Thus, while most of
them activate DNA replication, e.g. Escherichia coli SSB, T4
gp32, T7 gene 2.5 protein, 29 SSB, RP-A (reviewed in Ref. 1), some
SSBs are inhibitory. Thus, SSBs of filamentous phages M13, fd and f1 or
Pf3 block DNA replication by preventing the conversion of the viral
strand into the replicative form (8). Third, detailed structural
studies of a few SSBs have revealed that virtually each SSB interacts
with DNA using a characteristic and largely specific set of contacts
(9, 10, 11, 12, 13). As a consequence of this diversity, the molecular nature of
the SSB-DNA interactions are less understood than those occurring with
double-stranded DNA-binding proteins.
Bacillus subtilis phage 29 DNA replication requires a
SSB. In vivo, the elongation process of DNA replication is
dependent on the expression of a virally-encoded SSB protein,
29 SSB
(reviewed in Refs. 14 and 15), the cellular SSB being apparently unable to support
29 DNA replication. In vitro,
29 SSB is
essential to obtain full-length
29 DNA when the amount of DNA used
as template is low (16), the closest condition to the initial steps of
viral infection. Even when the amount of DNA template is high,
29
SSB is important for the in vitro system, increasing several
times dNTP incorporation (17, 18). This stimulatory effect is related to binding to the ssDNA produced during
29 DNA replication (19), which, likely, protects it from nuclease degradation and prevents unproductive binding of the
29 DNA polymerase.
29 SSB also
disrupts, by its helix-destabilizing activity, the secondary structure
formed in ssDNA parts of
29 replicative intermediates. This activity has been proposed to account for the stimulation of the DNA elongation velocity by
29 DNA polymerase under conditions in which the DNA opening is impaired (20).
The thermodynamic parameters defining 29 SSB-ssDNA complex formation
have been recently described: it binds to ssDNA with relatively low
affinity (Keff = 105
M
1) and moderate cooperativity (
= 50-70), covering 3-4 nt per monomer, probably as a single array of
protein units along the DNA (21). However, the precise mechanism that
accounts for the requirement of
29 SSB during viral DNA replication
is still unknown. The lack of significant sequence homology with other
SSBs (18) has made difficult the identification of the
29 SSB
domain(s) responsible for DNA binding and protein-protein contacts that lead to a stable protein-DNA interaction. Interestingly, complex formation occurs with significant alterations in the intrinsic tyrosine
fluorescence of
29 SSB. Thus, 95% of the protein fluorescence is
quenched upon binding (21), this value being one of the highest reported for a SSB, and similar to that observed for proteins that bind
DNA through direct interactions with tyrosine residues (22). If the
three tyrosines of
29 SSB (at positions 50, 57, and 76) were in fact
forming part of the domain directly involved in ssDNA complex
organization, they should not be highly restricted by interactions with
other neighboring residues and, likely, placed in an accessible domain
of the protein. To test this possibility, we have characterized the
global spectroscopic properties of the tyrosines of
29 SSB by using
complementary approaches: (i) steady-state and time-resolved
measurements of the total emission and anisotropy of
29 SSB
fluorescence, (ii) analysis of the resistance of
29 SSB tyrosines to
undergo ionization by pH increase, and (iii) selective attenuation of
the tyrosine fluorescence by small diffusible quenchers. In addition,
the fluorescent properties of
29 SSB were also considered to study
ssDNA binding under different temperatures. Our results regarding the
mobility and accessibility of
29 SSB tyrosines strongly suggest a
direct role of these residues in complex formation. In the accompanying
paper (23), some of the fluorescence features of
29 SSB tyrosines
are combined with other spectroscopic and hydrodynamic studies, to
model the conformational properties of
29 SSB and the effect of
complex formation on both
29 SSB and ssDNA structure.
All the solutions were made
with distilled, deionized (MilliQ) and sterilized water, and finally
filtered through 0.45-µm filters using a Millex-HA Filter Unit
(Millipore). The buffer used in time-resolved fluorescence analysis was
5 mM Tris-HCl, pH 7.5 (20 °C) (buffer A), while the
buffer used in absorption and steady-state fluorescence studies was 50 mM Tris-HCl, pH 7.5 (20 °C), 4% glycerol (buffer B).
Poly(dT) was from Pharmacia Biotech Inc. 29 SSB was purified by
differential ammonium sulfate fractionation of
29-infected cellular
extracts followed by phosphocellulose and DEAE-cellulose
chromatography, essentially as described (20). Protein concentration
was estimated by the Lowry (24) method, using bovine serum albumin as a
standard.
Absorption spectra were carried out in a CARY 219 (Varian) at a constant temperature of 22 °C with cells of path-lengths from 2 to 10 mm.
Steady-state Fluorescence SpectroscopyMeasurements of the
intrinsic tyrosine fluorescence of 29 SSB were recorded, unless
otherwise stated, at 22 °C. Fluorescence excitation and emission
spectra, solute quenching experiments, as well as ssDNA binding assays
were performed in a Schoeffel Instrument Corp. RRS100
spectrofluorimeter (slits set to 2 nm), with a 2-mm path-length cell,
and corrected for residual emission of the buffers. Fluorescence
quantum yield and fluorescence emission anisotropy measurements were
carried out in a SLM-8000D photon counting fluorimeter fitted with
Glan-Thompson polarizers. The excitation and emission slits were 4 and
8 nm, respectively. When necessary, corrections for wavelength
dependence of the instrument response were taken into account as
described elsewhere (25).
The steady-state average fluorescence quantum yield
(FST) of the protein was
calculated as in Demas and Crosby (26) using an aqueous solution of
5-methoxy-indole (5-MeOI) as reference (
REF = 0.28 ± 0.01; Ref. 27). Magic angle orientation of the polarizers was
selected to avoid anisotropic effects (28).
The accessibility to solvent and the ionic environment of the
fluorescent tyrosines of 29 SSB were analyzed by monitoring the
fluorescence quenching (
exc = 276 nm,
em = 308 nm) induced by adding increasing amounts of stock solutions (2.5 M) of neutral (acrylamide) or charged quenchers,
I
(as KI) and Cs+ (as CsCl) to the protein
sample (7-11 µM). To prevent iodide oxidation, and avoid
light dispersion and chemical reactions (29), Na2S2O3, at a final concentration
of 200 µM, was added to the KI solution. Corrections for
dilution and inner filter effect due to quencher absorption were made
as described previously (21). Data were analyzed by Stern-Volmer plots
of F0/F versus [Q],
where F0 and F stand for the
fluorescence intensities in the absence and in the presence of
quencher, respectively, and [Q] is the quencher
concentration. Since these plots were linear, the "static" quenching contribution was considered <10% (see "Discussion"), and the Stern-Volmer constant, Ksv, could be
directly obtained from the slope of the line (see Refs. 30 and 31).
This parameter represents the weighted sum of the individual constants
of the emitting tyrosine residues and, therefore, is referred as
Ksv,eff in the text.
Ksv,eff was used to calculate the bimolecular
rate constant of the quenching process, kq,eff,
from the expression Ksv,eff = kq,eff·
, where
is the
weighted average lifetime of
29 SSB fluorescence (see below). The
fractional accessibility of the tyrosine residues to quencher was also
estimated from the y axis intercept of double reciprocal
plots as proposed by Lehrer (Ref. 29; see also Refs. 30, 31, 32). The solid
angle
, under which the quencher can freely approach the tyrosine
residues, was calculated as described (33), taking the values of the
diffusion constants of the acrylamide, the iodide ion and the tyrosyl
residues as 2 × 10
5 cm2
s
1, 1.5 × 10
5 cm2
s
1, and 0, respectively (33, 34).
N-Acetyl-L-tyrosine amide (NAcTyrA) was taken as
a control for a fully accessible residue.
The effect of temperature on complex formation with ssDNA was analyzed
by monitoring the fluorescence intensity changes during direct
titrations of a fixed amount of poly(dT), 30-40 µM (in nt), with increasing amounts of 29 SSB as described previously (Ref.
21 and references therein). The reaction buffer, buffer B, was adjusted
to pH 7.5 at 20 °C. Since in the temperature range analyzed
(10-42 °C) the pH variation (from pH 7.0 to 7.6) does not alter the
global structural properties of
29 SSB (see accompanying paper
(23)), no pH readjustments were made during the binding measurements.
The thermodynamic parameters, nucleotide site size, n,
intrinsic affinity constant, K, and cooperativity,
, were obtained by fitting the experimental data to the theoretical ones given
by the McGhee and von Hippel's (35) equations for binding of proteins
to large lattices with overlapping binding sites (see Ref. 21, for
further details). The enthalpy and the entropy of the binding process
were approximated from the slope and the y axis intercept of
the best fit of van't Hoff plots of lnK
versus 1/T, as
Ho/R
and
So/R (R being the gas constant),
respectively, since the response was linear and therefore
Cpo = 0. Enthalpy and entropy changes were
calculated for the indicated buffer and protein concentrations, and
therefore named as
Hobso
and
Sobso throughout the
text.
To determine the steady-state fluorescence anisotropy,
<r>, of 29 SSB, the corresponding vertically
(I
) and horizontally (I
) polarized
emission intensities elicited by vertically polarized excitation were
corrected for background scattering and included in the described
equations for L-format geometry fluorimeters. Thus, <r>
was obtained from the ratio (I
GI
)/(I
+ 2GI
). The
constant G, accounting for the differential polarization sensitivity of
the photodetection set-up, was estimated as described previously (36).
The protein concentration (10-15 µM) used was low enough
to prevent turbidity of the solution and, therefore, light scattering
that could result in an artifactual depolarization of the fluorescence
(28).
The decay of total fluorescence intensity, measured at a magic angle orientation of the polarizers, and those of the parallel and perpendicular components were recorded at 22 °C in a time-correlated single-photon counting spectrometer, as described (36), for a protein concentration ranging from 10 to 30 µM. The samples were excited at 290 ± 4 nm with vertically polarized light pulses from a thyratron-gated nanosecond flash lamp (Edimburg Instr. El 199) filled with N2. The emission was isolated with a cut-off 310 nm (Schott KV) and a band-pass U-54 (Schott) filters, and detected with a Phillips XP2020Q photomultiplier.
The kinetic parameters (lifetimes i, and amplitudes,
i) of the decay of the fluorescence intensity were determined by iterative convolution of the signal recorded at magic
angle, I54(t). The fitting routines were based
on non-linear least-squares techniques (37, 38), using standard
weighting factors (39). Wavelength-dependent timing effects
in the photodetection were corrected as indicated (36). The weighted
average lifetime,
, was computed from,
![]() |
(Eq. 1) |
The fluorescence anisotropy decay parameters (rotational correlation
times, i, and amplitudes,
i) were determined by
simultaneously fitting the measured intensities I
,
I
to a sum of exponentials using a non-linear
least-squares global analysis method as detailed elsewhere (36). The
experimental decay of the tyrosine fluorescence anisotropy of
29 SSB
was best fitted to a biexponential function,
![]() |
(Eq. 2) |
The decay of the fluorescence intensity and anisotropy contains
subnanosecond components that were not well resolved in the flash lamp
spectrometer. To have a better definition of these parameters, a set of
both kinds of measurements were also carried out at the Laboratoire of
Biochimie Moleculaire et Cellulaire (Orsay, Paris) with a Ti:sapphire
(Spectra Physics) ps laser spectrometer, tuned at the third harmonic
(287 nm) and with an instrument response function of 75 ps; the
tyrosine emission was detected at 308 and 315 nm ( = 4 nm).
To measure the
tendency of the tyrosine phenolic hydroxyl groups of 29 SSB to
undergo ionization, small amounts of a stock solution of 2.5 M NaOH were added to a solution containing the indicated
concentrations of protein in buffer B, the change in volume at the end
of the titration being <2.5%. The pH increase was recorded
simultaneously with the decrease in the fluorescence at 308 nm. Free
tyrosine was taken as a model since its fluorescent properties
depending on the ionic environment in response to pH increase are well
documented (41, 42, 43) and no equivalent data were, to our knowledge,
available for NAcTyrA. Corrections for dilution and filter effects were
made as described (21).
The amino acid sequence of 29 SSB (Fig.
1A) was analyzed with currently available
computer programs to assess its basic theoretical structural features.
According to Chou-Fasman secondary structure predictions (44), the
region around tyrosines 50 and 57 (see Fig. 1A) showed a
high probability of being in the protein surface. Moreover, analysis of
the sequence-related hydropathy (45) indicated that this same region
(from residues 49 to 69), based on the negative values of the
Kyte-Doolittle index (Fig. 1B), could not be highly hydrophobic.
In order to have a more reliable picture on both, the accessibility of
the tyrosines of 29 SSB to the solvent, and potentially to DNA, as
well as their ionic environment, we analyzed the parameters defining
the steady-state and the time-resolved intrinsic fluorescence properties of the
29 SSB. The results obtained are summarized in
Table I.
|
The fluorescence spectra of 29 SSB free in solution, pH
7.5, and complexed with ssDNA are depicted in Fig. 2.
The area under the excitation and emission spectra is reduced by 90%
upon binding, although the same excitation and emission maxima,
exc = 276 nm and
em = 308 nm,
respectively, were observed in both cases. Moreover, the emission
spectra did not show any shoulder around 340 nm and, therefore,
ionization or strong hydrogen bonding of the phenol hydroxyl group of
tyrosine does not seem to take place (28, 46) either for free or
complexed
29 SSB.
As expected for A-type proteins, that is, proteins lacking tryptophan
in their primary structure (28, 47), no changes in either the shape of
the excitation or the emission spectra were observed for the denatured
29 SSB (4 M guanidinium chloride) (not shown).
The
anisotropy of 29 SSB fluorescence monitored by continuous excitation
with polarized light gave a value of 0.173 ± 0.010 (at 22 °C),
much smaller than the maximum (ro = 0.32) corresponding to immobilized tyrosines (48, 49). Different extrinsic
mechanisms (once discarded light scattering and other artifactual
effects, see "Materials and Methods"), like radiationless energy
transfer among the residues or rotational mobility of the fluorophore
and/or the protein, could account for this depolarization. The former
is very unlikely since it would require spectral overlap among the
excitation and emission spectra that is not observed (Fig. 2) and a
much higher concentration of the residues (28). To distinguish between
localized motions of the tyrosine residues (that would be indicative of
their rotational mobility) and global motions of the protein, the
time-resolved fluorescence anisotropy of
29 SSB was analyzed.
The decay of the anisotropy was best fitted (2 = 1.17)
to biexponential functions (see Equation 2 and Fig. 3)
with a ro value (0.28 ± 0.01) close to the
maximum expected for the tyrosine chromophore (see above), and two well
separated correlation times. The fast one,
short = 0.14 ± 0.06 ns, accounts for approximately one-fourth of the
depolarization. The remaining anisotropy decays almost completely
through the slow rotational component,
long = 7.0 ± 0.5 ns. Since
short
long the
decay of the total anisotropy can be expressed as a product of two
depolarizing processes, one due to fast movements of the protein
segment containing the tyrosine residues (r
(t)),
and another related to the global rotational motion of the whole
protein (for reviews, see Refs. 50 and 51), as,
![]() |
(Eq. 3) |
![]() |
(Eq. 4) |
The efficiency of the tyrosine emission of 29 SSB was
assessed by measuring the quantum yield of its global fluorescence emission in time-resolved (
FTR)
and steady-state (
FST)
measurements.
The experimental decay of the total protein fluorescence was best
fitted (2 = 1.1) to triexponential functions that gave
the lifetimes
1 = 0.4 ± 0.1 ns,
2 = 1.5 ± 0.1 ns, and
3 = 3.8 ± 0.1 ns, the pre-exponential terms (amplitudes) being
1 = 0.39 ± 0.03,
2 = 0.27 ± 0.03, and
3 = 0.34 ± 0.03, respectively. Since the fluorescence decay kinetics
of polypeptides and proteins, even with a single tyrosyl residue, is
usually complex (22, 46, 49), the results presented here should be
taken as a numerical approximation to a more complex kinetics. However,
the averaged lifetime of the protein fluorescence (
= 1.8 ± 0.1 ns) is a valid parameter to estimate the global quantum yield of
the fluorescence emission of
29 SSB. Thus, the ratio between and the
radiative lifetime of the tyrosine
R (see "Materials
and Methods") gave a (
FTR) of
0.067 ± 0.010. A similar value (see Table I) was obtained when
the quantum yield was estimated in steady-state measurements (
FST) when the area of the
emission spectra of
29 SSB was compared to that of the reference
emitter, the well characterized 5-MeOI (not shown). Therefore, the
29 SSB tyrosine-dependent fluorescence is not being
quenched by ground-state mechanisms. In contrast to the situation found
in most proteins, the value of
FST did not increase upon
denaturation (results not shown).
The intensity of 29 SSB fluorescence was directly dependent on the
temperature of the solution, this dependence being biphasic with a
clear slope change at 45 °C (Fig. 4). This transition
could reflect conformational changes in the environment of the
fluorescent tyrosines, rather than on the global conformation of
29
SSB, which seems to be stable up to 50 °C (23).
All Tyrosine Residues of
If there
is no energy transfer among the tyrosines (see below), it is possible
to estimate whether or not a tyrosine(s) contributes appreciably to the
observed emission by combining the estimated global F
and of the protein fluorescence. This is based on the fact that the
pre-exponential terms of the function that fits the experimental
fluorescence decay data are related to the fractional concentration of
the emitting species. Thus, the number of emitting fluorophores
(NE) can be related to that of the excited residues by the expression NE = NT
FST/(
/
R),
where NT is the total number of fluorophores (25). In our case, NE was calculated to be ~3.
Therefore, all tyrosines contribute to
29 SSB, although they may not
be necessarily mechanistically identical.
It is well known (41, 42) that positive charges in the surroundings of a tyrosine facilitate ionization, allowing tyrosinate to appear below pH ~ 10, the pKa of the free tyrosine. On the contrary, in the presence of a negative environment or for buried tyrosines, a higher pH is required for effective ionization.
Ionization of 29 SSB tyrosines was followed by the extent of
fluorescence quenching at 308 nm as a function of pH increase, since
the quantum yield of tyrosine emission is highly reduced upon
ionization (47). The results, compared with those of the free tyrosine
are shown in Fig. 5. From pH 7.5 (the standard during the functional assays of
29 SSB) to pH 9, the quenching pattern of
29 SSB closely resembles that of the exposed tyrosine in solution. However, both ionization patterns became different at higher pH. The
initial fluorescence was reduced by 50% at pH 10.2 for free tyrosine,
as expected, and at pH 10.9 for
29 SSB. Therefore, either (i) the
29 SSB tyrosines are located in heterogeneous regions and they are
differentially ionized at different pH or (ii) they are exposed in the
protein surface, and as a consequence of denaturing effects became
protected from ionization (by being included within the protein bulk
and/or by being surrounded by negatively charged residues). More
precise information could not be obtained by this technique, since the
protein structure, as indicated by CD spectra, was significantly
affected at pH > 9 (23). Further insight on the tyrosine
microenvironment, in terms of solvent accessibility and neighbor ionic
residues, was obtained by measuring the quenching of the protein
fluorescence by small and highly diffusible quencher agents:
acrylamide, iodide, and cesium ions.
Acrylamide, a highly efficient quencher of the tyrosine fluorescence
(52), which does not interact with proteins and is not influenced by
nearby charged residues (20, 31), was used to analyze the exposure of
the emitting residues of 29 SSB to solvent. Stern-Volmer plots for
the acrylamide quenching of
29 SSB fluorescence (Fig.
6A) were linearly dependent on the quencher concentration, the fractional accessibility being 1 (Fig.
6B). The rate constant of the global quenching process,
kq,eff (acryl) was (3.8 ± 0.5) × 109 M
1 s
1, 75% of
the value measured for the fully accessible NAcTyrA under the same
conditions. This difference may be related to the difference in the
diffusion coefficients of the tyrosyl residues of the protein relative
to free NAcTyrA, since the solid angle
under which the acrylamide
can approach these fluorophores was estimated as 4
for both, the set
of tyrosyl residues of
29 SSB and the NAcTyrA. Therefore, these
results are consistent with a situation in which the fluorescent
tyrosines of
29 SSB are equally accessible to the solvent, and,
apparently, located in non-buried regions of the protein.
The ionic environment of the 29 SSB tyrosine residues was assessed
by using I
(as KI) and Cs+ (as CsCl) as
quenchers. Fig. 6C shows the results obtained in titrations
of
29 SSB and NAcTyrA with KI. Again, the plots were linear, with
all of the fluorophores being accessible (Fig. 6D). However,
the quenching was less efficient, kq,eff
(I
) being (2.3 ± 0.2) × 109
M
1 s
1, ~46% of NAcTyrA, and
the angle
= 3.3
(see Table I). Thus, it seems that although all
of the fluorescent tyrosines of
29 SSB are homogeneously quenched by
I
, they seem to be somehow shielded by negative charges
that would make the interaction with the quencher difficult. We also
used a positively charged quencher (Cs+), but the results
were erratic. Due to its low quenching efficiency (33), high amounts of
salt (more than 0.6 M) had to be added to significantly
quench the
29 SSB fluorescence, giving rise to important changes in
the ionic strength of the solution and, most likely, in the protein
organization. Similar undesired effects of CsCl have been previously
reported for other A-type proteins, e.g. ribosomal proteins
S8 and S15 (33) or calmodulin (52).
Changes in the intensity of
29 SSB intrinsic fluorescence induced upon ssDNA interaction at
different temperatures were studied to determine the relative
contribution of the different thermodynamic parameters on
29
SSB-ssDNA complex formation. The experimental temperature range was
10-42 °C, where the protein conformation is stable (23) and the
intensity of the fluorescence emission is linearly dependent on
temperature (Fig. 4). Data from direct titrations of poly(dT) with
increasing amounts of
29 SSB were processed according to the
"unlimited" cooperative model of McGhee and von Hippel (35) as
outlined under "Materials and Methods" (see Ref. 21, for further
details). Representative examples of the experimental assays at 15 and
42 °C and the fitting procedure are shown in Fig. 7,
A and B, respectively. The value of the
nucleotide site size, n, the intrinsic affinity
K, and the cooperativity,
, are indicated in Table
II. Qmax and n were
found to be relatively independent on the temperature of the binding
solution. On the contrary, K was highly affected by
temperature increase, diminishing from 7100 M
1 (10 °C) to 700 M
1 (42 °C). Interestingly, the
cooperativity parameter
increased with temperature, being ~4
times higher at 42 °C (
= 220) than at 10 °C (
= 56). The
result of these compensating effects was a slight (2.5-fold) reduction
of the effective DNA binding constant K
. The van't Hoff
plot of K
versus 1/T was linear
within the temperature range used (Fig. 8), indicating
that
Cpo = 0 and, therefore, that the
enthalpy (
Ho) is independent of temperature
(53). Approximated values of the enthalpy
(
Hobso) and the entropy
(
Sobso) of the
29
SSB-ssDNA complex formation were obtained from the y axis
intercept and the slope of the plot, respectively. Thus,
Hobso ~
5 kcal
mol
1 and
Sobso
~ 8 cal
K
1 mol
1.
|
SSBs share two main functional properties: several orders of
magnitude more efficient binding to ssDNA than to dsDNA, and critical
roles in the maintenance and expression of the genetic information.
B. subtilis phage 29 encodes its own SSB, which is
absolutely required for viral DNA replication in vivo.
29 SSB plays its role(s) by binding to the ssDNA regions of
the replicative intermediates produced during strand-displacement
29
DNA replication (18, 19, 20). Therefore, the molecular basis of the
contacts responsible for the formation and stability of this complex
must be determined to properly understand the functional significance of
29 SSB during phage
29 infective cycle.
Steady-state fluorescence studies have served to
determine that the interaction between 29 SSB and ssDNA is
characterized by an unlimited cooperative DNA binding mode (
= 50-80 at 25 °C), a relatively low binding affinity
(Keff = 1-3 × 105
M
1 at 25 °C), a small nucleotide binding
size (n = 3-4 nt), and a net release of a very low
number of cations (21). The lack of a clear sequence homology or
defined patterns in protein motifs involved in ssDNA binding made
difficult the identification of the
29 SSB residues interacting with
ssDNA. One feature of the
29 SSB-DNA interaction that could give a
hint on a potential DNA binding domain is the high reduction of the
intrinsic tyrosine fluorescence of
29 SSB upon complex formation,
the maximal quenching (Qmax > 95%) being one
of the highest values reported for a SSB-ssDNA interaction (21). Among
the several mechanisms described to account for tyrosine quenching in
proteins that bind to DNA (22, 46, 54), those leading to the highest
reduction in fluorescence intensity have been related to direct
interactions between tyrosines and DNA (22, 46). It should be kept in
mind that since
29 SSB binding is cooperative, the possibility that
quenching occurs as a consequence of protein-protein interactions must
be also taken into account. In any case, a detailed knowledge of the
tyrosine environment and accessibility in the
29 SSB molecule is
necessary to understand the molecular nature of the interactions that
contribute to
29 SSB-DNA complex formation.
The average quantum yield F of the intrinsinc
29 SSB
fluorescence emission, analyzed by steady-state and time-resolved
measurements, gave a value of 0.065 ± 0.010. The fact that both
techniques gave the same
F value can be considered as
indicating that none of the three possible emitters has been eliminated
by ground-state deactivating processes. In other words, the three
tyrosines of
29 SSB contribute to its intrinsic fluorescence. This
F value is unusually high for A-type proteins, namely
proteins that lack tryptophan in their primary sequence. Thus, values
of 0.037, 0.016, and 0.001 have been reported for insulin, octopus
calmodulin, and ovomucoid protein, respectively (47, 55). The main
mechanism leading to tyrosine quenching in A-type proteins is related
to peptide bond formation, and additionally, to interactions with the
side chains of neighboring residues, which can act as partners in
hydrogen bond formation (46, 56).
F of
29 SSB is
~70% of that of N-acetyl-L-tyrosine amide,
the closest analogous of tyrosines included in polypeptidic bonds.
Therefore,
29 SSB tyrosines do not appear to maintain strong
interactions with other residues of the protein since this would lead
to a much reduced value of the efficiency of fluorescence emission
(56). In addition, the excitation and emission maxima of
29 SSB
fluorescence at neutral pH (see Table I) are indicative of a lack of
surrounding interactions leading to tyrosinate formation. An opposite
situation has been described for the SSBs of the filamentous phages
Pf1, Pf3 (57).
The Stern-Volmer analysis of the quenching by acrylamide and KI allowed
us to propose that the fluorescent tyrosines are equally accessible to
small quenchers and, likely, they are placed in an accessible domain of
the protein. The kq,eff(acryl) for 29 SSB is
(3.8 ± 0.5) × 109 M
1
s
1, similar to those reported for accessible tryptophan
residues of E. coli SSB (58) and T4 gp32 (59). Moreover, the
value of
, the solid angle related to the accessibility of tyrosines to acrylamide, together with the fractional fluorescence accessible to
the quencher, 4
and 1 for
29 SSB, respectively, correspond to
those expected for totally accessible residues. The differences in the
accessibility to a negatively charged quencher, iodide ion,
kq,eff(IK) (2.3 ± 0.2) × 109
M
1 s
1, indicate that
fluorescent tyrosines should have a net negatively charged
microenvironment, in agreement with the pH dependence of tyrosine
fluorescence of
29 SSB. The absence of upward curvature for both,
acrylamide and iodide, could be interpreted as due to a dynamic
quenching process (31), the contribution of a static quenching being
negligible, at least for the range of quencher concentrations used in
this work. This linear behavior is characteristic of a relatively
homogeneous population of the fluorescence emitters (30). Although the
fluorescent tyrosines could be located in different exposed regions of
the protein, it would be rather unlikely that all of these
domains have a similar ionic environment. Therefore, it is tempting to
speculate that
29 SSB tyrosines could be located within the same
solvent-accessible protein region.
Further information regarding the environment of the 29 SSB
tyrosines was obtained from the analysis of the rotational
depolarization of the intrinsic protein fluorescence. As shown above,
time-resolved anisotropy revealed two depolarizing motions with clearly
different correlation times. In the absence of resonance energy
transfer, the short correlation time,
short = 0.14 ± 0.06 ns, can be assigned to fast localized motions of the tyrosine
side chains. This value is very close to that of free tyrosines in
solution (0.13 ns; 60) and, therefore, can be considered as indicating
absence of restrictive interactions among tyrosine side chains and
other residues. The long correlation time (
long = 7.0 ± 0.5 ns) describes a slower motion of the whole protein.
According to the parameters of the anisotropy fitting (Table I), the
limiting anisotropy, r
, that would be
attained if the tumbling of the whole protein was arrested, was found
to be 0.21. From this value, the average angular oscillations of the
emitting residues of
29 SSB were estimated (50) as spanning about
50 °C. In summary,
29 SSB tyrosines are able to undergo fast
subnanosecond oscillations as if they were not blocked, although some
steric hindrance, likely due to the presence of the rest of the protein
bulk, partially limits the amplitude of these motions.
The
large quenching of the intrinsic fluorescence of 29 SSB upon complex
formation (Qmax = 95%; 21) precluded the
analysis of the effect of DNA interaction on the rotational correlation times of the anisotropy, the lifetimes of the total fluorescence or the
Stern-Volmer constants. However, the changes in the intrinsic fluorescence intensity as a result of DNA binding, monitored at different experimental temperatures, can be used to determine the
individual contributions of the intrinsic affinity and the cooperativity parameters to the global binding process, as well as to
obtain a rough estimation of the enthalpy and entropy values, as it has
been described for other SSBs (52, 61, 62, 63). The increase of the
reaction temperature in the range in which the global protein
conformation is stable (see "Results" and Ref. 23) has opposite
effects on the intrinsic binding constant, K, and on the
cooperative parameter,
, the effective binding constant
(Keff = K
) being largely
independent of temperature. Thus, while a ~10-fold decrease of
K was estimated from 10 to 42 °C, an interesting
simultaneous 4-fold increase is observed for
and, as a result,
Keff is reduced only by ~2-fold. Therefore, thermally-induced brownian motions seem to favor a more accessible conformation of the protein domain involved in monomer-monomer interactions. The van't Hoff analysis of temperature effect on K
allowed a gross estimation of the energies driving the
global ssDNA binding process. The linearity of the van't Hoff plots, characteristic of non-site specific DNA interactions (53, 64) indicated
that, at low salt concentration,
Hobso (~
5 kcal
mol
1) was temperature independent
(
Cpobso
= 0),
allowing the estimation of
Sobso
as ~8 cal
K
1 mol
1.
Altogether, our results are consistent with fluorescent tyrosines being
located either on the surface of the protein or in a pocket whose size
and shape allow the quencher to gain access to the tyrosine residues.
The fact that the nucleotide site size has a relatively small value
(3-4 nt are covered by protein monomer; 21) supports the possibility
of a discrete domain involved in complex formation with ssDNA.
Furthermore, tyrosine mobility does not seem to be highly restricted by
interactions with neighboring residues. This localization might
contribute to complex formation, either indirectly, by establishing
protein contacts between contiguous monomers, or through direct
interactions with the DNA bases. Preliminary results support the last
hypothesis, since 29 SSB point mutants in each of the three
tyrosines have an impaired ability to bind ssDNA.2 In this sense, the presence of a
net negative charge surrounding the tyrosines could be the reason for
the relatively low affinity of this protein for ssDNA, a property which
may be directly related to its functional roles during
29 DNA
replication (20, 21). In addition, the relative independence of the
effective DNA binding constant of
29 SSB of the solution temperature
could provide an adaptative advantage for phage
29 development in
the very different habitats, e.g. soil, field crops,
different foodstuffs, seewater (see Ref. 65, for a review), in which
the cellular host, B. subtilis, is able to
grow.
We are indebted to J. M. Lázaro and L. Villar (Centro Biología Molecular "Severo Ochoa", Madrid)
for the purification of 29 SSB and Drs. M. P. Lillo and R. Dale
(Instituto de Química Física Rocasolano, Madrid) for
their invaluable contribution in data analysis and interpretation. We
also thank Drs. J. C. Brochon and P. Tauc (Laboratoire of Biochimie
Moleculaire et Celulaire. CNRS, Orsay, Paris) for making available to
us the picosecond spectrometer and for their kind help with the
measurements.