(Received for publication, April 2, 1996, and in revised form, October 2, 1996)
From the Centro de Biología Molecular
"Severo Ochoa" (CSIC-UAM), Cantoblanco, 28049 Madrid, Spain,
¶ Instituto de Química Física Rocasolano (CSIC),
Serrano 119, 28006, Madrid, Spain, and
Centro de
Investigaciones Biológicas (CSIC), Velázquez 144, 28006 Madrid, Spain
The strand-displacement mechanism of
Bacillus subtilis phage 29 DNA replication occurs
through replicative intermediates with high amounts of single-stranded
DNA (ssDNA). These ssDNA must be covered by the viral ssDNA-binding
protein,
29 SSB, to be replicated in vivo. To understand
the characteristics of
29 SSB-ssDNA complex that could explain the
requirement of
29 SSB, we have (i) determined the hydrodynamic
behavior of
29 SSB in solution and (ii) monitored the effect of
complex formation on
29 SSB and ssDNA secondary structure. Based on
its translational frictional coefficient (3.5 ± 0.1) × 108 gs
1, and its rotational correlation time,
7.0 ± 0.5 ns,
29 SSB was modeled as a nearly spherical
ellipsoid of revolution. The axial ratio (p = a/b) could range from 0.8 to 1.0 (oblate model,
a < b) or 1.0 to 3.2 (prolate model,
a > b). Far-UV CD spectra, indicated that
29 SSB is highly organized within a wide range of temperatures (15 to 50 °C), being mainly constituted by
-sheet elements (~50%, at pH 7). Complex formation with ssDNA, although inducing minimal changes on the global conformation of
29 SSB, had a clear
stabilizing effect against pH and temperature increase of the solution
samples. On the other hand,
29 SSB binding leads to non-conservative
changes of the near-UV CD spectra of ssDNA, which are consistent with different nearest-neighbor interactions of the nucleotide bases upon
complex formation. The above results will be compared to those reported
for other SSBs and discussed in terms of the functional roles of
29
SSB.
In all organisms studied so far, key cellular processes related to
DNA metabolism, such as DNA replication, DNA repair, DNA recombination,
or DNA transcription, occur via transient ssDNA1
intermediates whose structure must be properly
organized. This organization is achieved and maintained
by highly efficient, localized (sequence-specific), or generalized
(sequence-independent) protein contacts. Noncatalytic proteins that
bind ssDNA with a relatively high affinity in a sequence independent
manner have been referred to as single-stranded DNA-binding proteins,
SSBs, (1, 2, 3, 4). One of the most striking features of the SSBs is the lack
of any clear structural common characteristics among them. Thus, the
oligomeric state of these proteins in solution is highly variable. They
have being found as monomers, e.g. N4 SSB (5); 29 SSB
(6); dimers, e.g. SSBs from filamentous Fd phages (3); T7
gene 2.5 (7); tetramers, e.g. Escherichia coli SSB (8);
multioligomers, T4 gp32 (1); or hetero-oligomers, e.g. human
SSB (9) and Drosophila SSB (10). Moreover, although several
attempts to find structural motifs in SSBs have been made (11, 12),
only among SSBs of very related filamentous phages a limited sequence
homology has been found within the so-called DNA binding wing (see Ref.
3, for review). NMR and crystallographic studies showed a similar
three-dimensional structure of the SSBs of E.
coli and Pseudomonas aeruginosa filamentous phages Ff
and Pf3 (13, 14, 15), rather different from that of T4 gp32, the other
SSB2 whose structure (x-ray difraction) has
been reported (18). In addition, detailed analysis of a few SSB-ssDNA
complexes indicate that each protein interacts with ssDNA using a
particular set of residues (3, 15, 19, 20, 21) and induces different conformational changes in the DNA structure (see Refs. 2, 4, and 8, for
reviews). The functional characteristics of the SSBs complicate even
more this variability. Most of the SSBs activate DNA replication,
e.g. E. coli SSB, T4 gp32, T7 gene 2.5 protein,
29 SSB,
and human SSB (reviewed in Ref. 4). However, other SSBs, as those of
the filamentous phages, block DNA replication (22). As a consequence of
this diversity, the molecular nature of the SSB-ssDNA interactions is
not yet well understood.
The requirement for a SSB is particularly important during
strand-displacement DNA replication, where great amounts of ssDNA are
produced. This ssDNA corresponds to the strand that is not used as
template and is displaced as the DNA polymerase progresses. One of the
best characterized strand-displacement DNA replication systems is that
of Bacillus subtilis phage 29 (see Refs. 23, 24, 25, for
reviews). In vivo,
29 DNA replication cannot procced in
the absence of the viral product of gene 5, the viral SSB (12, 26).
29 SSB is an extremely abundant protein in infected cells
(~2.5 × 106 copies/cell), representing ~3% of
the total protein content (26). In vitro,
29 SSB
stimulates DNA replication, at least at two levels: (i) by increasing
the amount of dNTP incorporated (12, 26) and (ii) by increasing the DNA
elongation velocity of
29 DNA polymerase under conditions in which
the DNA opening is impaired (27). This activity has been proposed to be
related to the helix-destabilizing properties of
29 SSB (27).
29
SSB becomes essential for in vitro viral DNA replication
when the amount of DNA is reduced (28), the closest condition to the
initial steps of the viral infection.
29 SSB binds to ssDNA in a cooperative way
(Keff = 105
M
1,
= 50-70) covering 3-4 nucleotides
per monomer, likely as a single array of protein units along the ssDNA
(6). When visualized by electron microscopy, the
29 SSB-DNA complex
shows a continuous and homogeneous structure, and nucleosome-like
structures were not apparent (6, 12). Protein binding induces a 50%
reduction in ssDNA length (29) and a 95% reduction of the intrinsic
tyrosine fluorescence emission of the protein (6). This observation has
been the starting point for the analysis of
29 SSB ssDNA binding
domain as described in the accompanying paper (30).
The functional requirement of 29 SSB during the DNA replication
stage of
29 infective cycle might be related to the structural properties of
29 SSB that would lead to a particular organization of
complex with the ssDNA of the replicative intermediates. In this paper,
we have analyzed the global conformation of
29 SSB in solution by
analytical ultracentrifugation, steady-state and time-resolved
fluorescence anisotropy measurements. Conformational changes in the
protein and ssDNA structure induced upon complex formation were studied
by means of circular dichroism techniques. The results of these studies
indicate that
29 SSB differs markedly from other well known
SSBs.
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 analytical ultracentrifugation,
time-resolved fluorescence, and circular dichroism (CD) analysis was 5 mM Tris-HCl, pH 7.5 (20 °C) (buffer A), while the buffer
used in steady state fluorescence studies was 50 mM
Tris-HCl, pH 7.5 (20 °C), 4% glycerol (buffer B). M13mp18 ssDNA was
from Boehringer Mannheim and oligonucleotides dT20,
dA16, and polynucleotides poly(dT300) and
poly(dA300) from Pharmacia Biotech Inc. Nucleotide
concentrations were quantified as indicated by the supplier. 29 SSB
was purified by differential ammonium sulfate fractionation of
29-infected cellular extracts and further phosphocelulose and
DEAE-cellulose chromatography, essentially as described (27). Protein
concentration was estimated by the Lowry method (31), using bovine
serum albumin as a standard.
Analytical
ultracentrifugation experiments were carried out in an Optima XL-A
analytical ultracentrifuge (Beckman Instruments Inc.) equipped with a
UV-VIS optical scanner, using a Ti60 rotor and double sector 12-mm
centerpieces of epon charcoal. In sedimentation equilibrium
measurements, samples of 29 SSB ranging from 10 to 30 µM were centrifuged in buffer A at 20 °C and at two
speeds, 25,000 and 30,000 rpm. When sedimentation equilibrium was
reached, absorbance scans (at 230 nm) were taken as a function of
radial distance. The base-line offsets were determined by scanning the samples at a higher speed (42,000 rpm). To obtain the relative molecular weight (Mr) of
29 SSB, the
experimental data were fitted to the equation which describes the
radial distribution of concentration of an ideal solute at
sedimentation equilibrium (see, for example, Ref. 32) using the
programs XLAEQ and EQASSOC (supplied by Beckman; see Ref. 33). The
value of the partial specific volume of
29 SSB was estimated from
its amino acid composition (34) as 0.739 ml g
1. In
velocity sedimentation experiments,
29 SSB (ranging in concentration from 10 to 30 µM) was centrifuged at 60,000 rpm and
20 °C and radial scans at 230 nm were taken at 10-min intervals
during 3 h. The analysis of the experimental data was performed in
two ways: (i) by taking the derivative of the absorbance data to
calculate the peak of the moving boundary (XLAVEL program, supplied by
Beckman), and (ii) by using the second moment analysis of Goldberg (35) to locate the boundary position (VELGAMMA program, Beckman). The rate
of the motion of the boundary was used for the estimation of the
sedimentation coefficient. This value was corrected to standard
conditions (water, 20 °C; 33) to obtain the
s20,w values.
The translational frictional coefficient, f, of 29 SSB
was obtained by the classical Svedberg equation (see Ref. 36 for review),
![]() |
(Eq. 1) |
The global conformation of 29 SSB in solution was modeled
with rigid ellipsoids of revolution (the axis being a,
b = c). Two models were considered, the
prolate ellipsoid, in which a > b, and the oblate ellipsoid, characterized by
a < b. In both cases, the rotational
parameters of the protein are related to the hydrated protein volume
V and asymmetry (given by the axial ratio p = a/b) in the following way,
![]() |
(Eq. 2) |
![]() |
(Eq. 3) |
![]() |
(Eq. 4) |
![]() |
(Eq. 5) |
![]() |
(Eq. 6) |
The experimentally determined value of global (7.0 ± 0.5 ns; see Ref. 30) was introduced for axial ratio ranging from 0.1 to 1 (oblate model) and 1 to 10 (prolate model), in Equation 2 to
obtain the V values of the set of ellipsoids compatible with the experimental rotational parameters. The theoretical
translational frictional coefficient,
ftheo corresponding to each of these ellipsoids was obtained as 6
RsphF. Where
Rsph stands for the radius of a sphere with the
same volume V as the ellipsoid, and F is the translational Perrin factor whose value also depend on p
(41). The F values for the prolate and oblate ellipsoids,
Fobl and Fpro, respectively, where calculated as,
![]() |
(Eq. 7) |
![]() |
(Eq. 8) |
The comparison of the experimentally determined
fobs and the estimated
ftheo provides the range of ellipsoids
compatible with the translational and rotational hydrodynamic behavior
of 29 SSB. The hydration extent of the protein,
(gram of
H2O/g of protein) was estimated as
=
(V
Vanh)/Vanh, where
Vanh (estimated from Ranh = 0.723 × 10
8 Mr(null)/1;3;
see Ref. 36 for further details) corresponds to the anhydrous volume of
a sphere whose Mr is that of
29 SSB.
CD spectra were recorded on a Jasco
Model J-700, at 0.2-nm intervals and, unless otherwise stated, at
25 °C and at a scan speed of 20 nm/min. The absorbance of the
samples was kept below 0.1 over the measuring range. 29 SSB (6.5 µM) CD spectra were obtained with 0.1-cm cells in the
absence and presence of an excess of ssDNA (350 µM
nucleotide). DNA binding was tested by parallel measurements of the
quenching of the intrinsic protein fluorescence upon DNA addition as
described previously (6). When required, small amounts of 2.5 M NaOH from a concentrated stock were added to the
indicated amount of free or bound
29 SSB (the volume change at the
end of the titration was <1%), the pH being simultaneously measured.
The effect of temperature on the protein conformational stability was
analyzed by recording the spectra after maintaining the protein or
protein-ssDNA solution for 5 min at the different temperatures.
Controls were made to ensure that the variations in the pH due to
temperature change did not affect the far-UV CD spectra of the protein.
Plotted CD data, as mean residue ellipticity, corresponded to an
average of three scans. Corrections for the contribution of the buffer
and DNA to the optical activity were made when necessary. The relative
amount of
-helix,
-sheet, and
-turns of
29 SSB where
estimated with the CONTIN (version 2) program (42) by fitting the
29
SSB spectra, between 190 and 240 nm, to those of 16 proteins of known
secondary structure (43) as outlined by Provencher and Glöckner
(44). Alternatively, convex constraint analysis, as described by the
Fasman group (45, 46), were also made, but the best fits were not so
accurate. ssDNA spectra (dT20, dA16, poly(dT),
or poly(dA)) were recorded in buffer A, with 1-cm path length cells.
Five scans were averaged and the data were plotted as
L
R (
) per mol of nucleotide. Stacking
interactions between neighboring nucleotides were analyzed by
increasing the temperature of the solution. Spectra were taken after
keeping ssDNA samples 5 min at the desired temperature. To analyze the
effect of
29 SSB binding on the ssDNA structure, spectra were
recorded after adding small aliquots of a concentrated protein stock
(~7 mg/ml) to a fixed amount of the indicated ssDNA (~110
µM nucleotide). Buffer and protein optical activity were substracted to obtain the corrected ssDNA spectra. Saturation was
considered to be reached when further protein addition did not induce
any change in CD spectra. The nucleotide site size, that is, the number
of nucleotides covered by a protein monomer, was calculated from the
ratio DNA/SSB that induced the maximal change at wavelengths
corresponding to the maximal activity of free ssDNA.
Steady-state measurements of the
intrinsic fluorescence of 29 SSB were performed at 25 °C in a
SLM-8000D photon counting L-format fluorimeter fitted with
Glan-Thompson polarizers at
exc = 276 nm and
em = 308 nm (slits set to 2 and 4 nm,
respectively; Ref. 47). The anisotropy, <r>, of
29 SSB
fluoresence emission was obtained as described in the accompanying
paper (30). Complex formation with ssDNA was monitored by measuring the
increase in steady-state anisotropy together with the quenching of the
fluorescence emission at magic angle (I54) and during
reverse titrations of a fixed amount of
29 SSB (10 µM)
with increasing quantities of ssDNA. Corrections for dilution and inner
filter effect due to ssDNA absorption were made when necessary as
described previously (6).
Structural Features of Free and Bound 29 SSB
The solution conformation of
29 SSB was approached by analyzing hydrodynamic parameters directly
dependent on the protein aggregation state, size, and shape. The
initial step in the modeling process was the estimation, by analytical
ultracentrifugation techniques (sedimentation equilibrium and
sedimentation velocity), of the molecular weight and the sedimentation
and frictional coefficients of
29 SSB. Table I
summarizes the results obtained.
|
The experimental data corresponding to the radial distribution of 29
SSB concentration after sedimentation equilibrium were best fitted to
equations described for a monomeric behavior of a protein. An example
for 12 µM SSB centrifuged at 30,000 rpm is depicted in
Fig. 1. The fitting procedures (
2 = 6.4 × 10
3) gave an average of 13,600 ± 400, for the Mr of
29 SSB, very close to the
calculated from its amino acid composition (13,343; see Ref. 26). This
confirms the results previously obtained with indirect techniques (gel
filtration chromatography; Ref. 6), and therefore,
29 SSB can be
confidently considered as a monomer in solution, at least in the
concentration range used for the analysis of its functional activities
and ssDNA binding properties.
Sedimentation velocity experiments were carried out to determine the
sedimentation and diffusion coefficients,
s20,w and
D20,w, respectively. The sedimentation
process of a 60,000 rpm centrifugation of 29 SSB was analyzed by
following the movement of the boundary as described elsewhere (see Ref. 48, for a review). For the above indicated protein range, the estimated
s20,w and
D20,w (procedures not shown) were
1.7 ± 0.1 S and (1.18 ± 0.07) × 10
6
cm2 s
1, respectively. By including the
Mr and s20,w
values in the Svedberg equation (see "Materials and Methods,"
Equation 1) the translational frictional coefficient,
fobs, was computed as (3.5 ± 0.1) × 108 gs
1 (Table I).
It is well known that the value of the frictional coefficient does not
allow, by itself, the distinction between a symmetric but rigid
structure from a highly asymmetric but flexible conformations (see for
examples, Refs. 36, 48, and 49). Therefore, another hydrodynamic
parameter was considered to restrict the size and shape of 29 SSB,
the rotational depolarization of its intrinsic fluorescence anisotropy.
As described in the accompanying paper (30), the decay of
29 SSB
anisotropy is well described by two widely separated rotation
correlational times,
1 = 0.14 ± 0.06 ns
(
1 = 0.24 ± 0.03) and
2 = 7.0 ± 0.5 ns (
2 = 0.76 ± 0.03). Since
1
2, the second one can be
considered as corresponding to the global motion (
2
global) of the entire protein.
The values of global, within the experimental
uncertainties, were used to simulate (using Equations 2, 3, 4, 5, 6, 7, 8 under "Materials and Methods") a set of rotationally
compatible ellipsoids of revolution with axial ratios p
(=a/b) ranging from 0.1 to 1 (oblate mode) or 1 to 10 (prolate mode). The theoretical f,
ftheo, of each of these ellipsoids are
represented in Fig. 2, A and C. Those values of p giving ftheo in the
range of fobs could account for the shape of
29 SSB. Thus, a nearly spheric oblate ellipsoid, p = 0.8 to 1 (Fig. 2A), and a set of prolate ellipsoids up to p ~ 4.5 (Fig. 2C) would satisfy
simultaneously the experimental
global and
fobs. Thus, both, a disc-shape and a fibrillar
shape can be excluded for
29 SSB solution conformation. Since the
dry volume of the protein can be easily estimated from the
Mr (Table I), it is possible to compute the
corresponding hydration factors
(see "Materials and Methods")
from the hydrodynamic volume of each of the ellipsoids. All of the
compatible oblate ellipsoids (Fig. 2B) and the prolate
ellipsoids with p < ~3.2 (Fig. 2D) would have a reasonable hydration (
values below 0.15 g of water/g of
protein were considered as higly unlikely). The best approximation to
the dimensions (a × b) of
29 SSB,
corresponding to the mean values of the compatible p, would
be a = 34 ± 2 Å, b = c = 38 ± 4 Å (V = 25 × 103
Å3) for the oblate ellipsoid (p = 0.9 ± 0.1), and a = 56 ± 16 Å, b = c = 28 ± 8 Å (V = 23 × 103 Å3) for the prolate ellipsoid
(p = 2 ± 1).
Circular
dichroism (CD) spectroscopy is a powerful tool for the analysis of the
global secondary structure of proteins in solution due to its
dependence on the type and amount of -helical,
-sheets, and
unordered conformations (50, 51). Moreover, aromatic amino acids have
also a characteristic CD activity toward the near UV region of the
spectra, that is more evident when these residues are not free to
rotate (52, 53).
Fig. 3A shows the far UV CD spectrum at
neutral pH and 25 °C of free and bound 29 SSB. A maximum at 196 nm and two minima at 210 and 223 nm can be clearly identified in both
cases. These spectra were analyzed by the comparative method of
Provencher (42, 44) using as reference proteins whose structure is
known (43). The goodness of the fitting procedure is shown in Fig. 3,
B and C, as the random distribution of the
weighted residuals. Both forms of
29 SSB were found to be highly
structured. The free protein contains a relatively high amount of
-structures (47%
-sheet and 23%
-turn) and a low amount of
-helix (10%).
Complex formation with ssDNA had minimal effects on the CD spectra of
29 SSB: the difference on the content of the secondary structure
elements of
29 SSB between the free and poly(dT)-bound forms of the
protein being within the experimental error (Fig. 3A). As
expected, since
29 SSB has no sequence specificity (6), similar
results were observed with ssDNA of heterogeneous sequence, M13 ssDNA
(not shown). Nonconformational changes upon complex formation have been
also reported for SSBs of filamentous phages fd (54), Ike (55), or Pf3
(56). The ellipticity in the near UV region of the spectrum
(wavelengths above 250 nm) which has been described (52, 53) to
correspond to the emission of the aromatic residues of the protein (in
the case of
29 SSB, 3 Tyr and 7 Phe), was found almost negligible
for both, free and complexed
29 SSB (not shown). Therefore, and in
agreement with the results obtained by analyzing the intrinsic features
of
29 SSB fluorescence (30),
29 SSB tyrosines could not be
strongly restricted by interactions with surrounding residues.
Although ssDNA binding did not apparently change the
global conformation of 29 SSB, it had, however, a clear effect on
its stability. Changes in the far UV spectra at different pH monitored for free and bound
29 SSB in buffer Tris-HCl are shown in Fig. 4. It should be mentioned that reliable spectra could
not be recorded at pH < 6.0 because
29 SSB became insoluble. A
pH increase up to 7.0 had no effect on the protein structure, as
indicated by the invariability of the CD spectra. In the absence of
ssDNA (Fig. 4A), from pH 7.0 to 9.3 the same molar
ellipticity (
) was observed for wavelengths ranging from 190 to 210 nm, although great changes were observed in the region spanning between
210 and 240 nm. Thus a 20% reduction respect to the values at pH 7.5 was recorded for pH 8.0, while a 30% increase was recorded for pH 9.3. These oscillations likely correspond to less stable conformations of
29 SSB. Higher pH values greatly affected the protein structure. The
apparition of a minimum around 200 nm and the progressive anulation of
from this wavelength are consistent with a mostly random coiled organization, which reaches the maximum value at pH 11.5. Similar results were obtained with sodium cacodylate-HCl or glycine-NaOH buffers (not shown). The increase in the salt concentration (due to the
NaOH added to modify de pH) was discarded as responsible of those
structural changes, since additions of even a 3-fold excess of NaCl did
not induce any effect on the
29 SSB CD spectra at neutral pH
(results not shown).
CD spectra of the bound 29 SSB (at a binding density of ~1) was,
on the contrary, relatively insensitive to the solution pH (Fig.
4B). The same maxima and the same
values were recorded between pH 6.0 and pH 10.5. A further increase in the solution pH
started to induce changes in the global protein structure, as indicated
by the reduction of the optical activity values around 196 nm, which,
nevertheless, are still less affected respect to those in the absence
of ssDNA. pH 11.5 also denatured the protein structure.
Temperature had a less dramatic effect on 29 SSB structure. The far
UV spectra of
29 SSB was almost constant from 15 °C up to
55 °C (Fig. 5A). Even at temperatures as
high as 70 °C the protein retained a large fraction of its tertiary
structure. Again, ssDNA had a stabilizing effect, alleviating the
effects of temperature increase (Fig. 5B).
The Steady-state Intrinsic Anisotropy of
Although ssDNA binding did not significantly
change the global 29 SSB structure, it could restrict the rotational
properties of the protein. To test this possibility, binding was
followed by monitoring the differential changes in the vertical and
horizontal components of the protein emission during reverse titrations
of a constant amount of protein with poly(dT). Reduction of
fluorescence intensity at magic angle orientation of the polarizers,
I54 was recorded in parallel to follow ssDNA binding. An
example (20 µM
29 SSB) is shown in Fig.
6. The anisotropy, increased from 0.174 to 0.214, a
value very close to the limiting anisotropy of
29 SSB, that is, the
remaining anisotropy that would be recorded by time-resolved techniques
if the protein was immobile (r
= 0.210; 30).
In addition, this maximal value is still far from the expected value
for immobilized tyrosines (ro = 0.32; 57-58).
Changes in the anisotropy of a single molecule of
29 SSB could not
be analyzed due to the special binding features of this protein: it
cannot bind to short oligonucleotides (see above and Ref. 6).
Effects of Complex Formation on ssDNA Structure
Complex formation between 29 SSB and ssDNA was also studied by
analyzing the changes in the CD spectra of DNA induced by protein
binding and comparing them with those induced by temperature increase.
The spectra were recorded above 250 nm, where, as already mentioned,
29 SSB has a negligible optical activity. As it can be deduced form
Fig. 7A,
29 SSB induced a great alteration
of the M13 ssDNA secondary structure. Both a progressive red shift of
the spectra and a reduction of the circular dichroism (
) values
took place when increasing amounts of protein were added. At the
saturation point, the two exciton maxima moved from 249.0 to 251.0 nm
and 276.0 to 281.0 nm, the activities being reduced by ~2.2-fold
(negative band) and by ~0.1-fold, 10% (positive band). The crossover
moved from 261.5 to 266.5 nm. Those effects could be reflecting
reorganizations of the contiguous nucleotides. To distinguish
unstacking from other orientation effects, spectra of free
M13 ssDNA were recorded at higher temperatures, since, as described for
different ssDNAs, temperature decreases the stacking degree of neighbor
nucleotides (59, 60, 61, 62). The conservative changes (no displacement of the
maxima and the crossover) induced by temperature increase (Fig.
7A, inset) were clearly different from the effects of
29
SSB. In addition, binding was tested on poly(dT) and poly(dA), ssDNAs
described as having low and high, respectively, degree of stacking
(63, 64). In the case of poly(dT), again, a red shift and a reduction
(~35% in this case) of the
intensities were detected upon
complex formation (Fig. 7B). Therefore, it can be concluded
that the ssDNA in the complex with
29 SSB still remains organized,
although the nearest-neighbor interactions differ from the ones
established in the free DNA, likely reflecting, apart from a different
degree of stacking, a different geometry of the contiguous bases. These
changes should be corresponding to the factors leading to a global 50%
reduction of the ssDNA length measured by electron microscopy
techniques (12).
29 SSB did not induce any change in the poly(dA)
spectra (not shown), suggesting an impairment of intercalating
interactions between the aromatic ring of
29 SSB tyrosines and the
bases of the ssDNA, reinforcing the hypothesis presented in the
accompanying paper (30) of the involvement of these residues in complex
formation.
We also analyzed whether the changes in the CD intensity () were
compatible with the results obtained from the analysis of the changes
in the intrinsic
29 SSB fluorescent emission upon complex formation.
Thus, the value of the nucleotide site size, n, the number
of nucleotides covered by a protein monomer, was computed from the
relationship between the changes in
at the wavelengths
corresponding to the exciton bands of the free DNA and the
29
SSB/poly(dT) ratio (Fig. 8). Saturation, given by a
maximum change of
, started at a
29 SSB/poly(dT) ratio of 0.28-0.29 which gave a value of n of about 3.4-3.6, in
agreement with the results from fluorescent techniques (6). No effects on the conformation of short oligonucleotides dT20,
dA16 were detected even at very high doses of
29
SSB (not shown), confirming the impairment of
29 SSB to bind these
DNAs, that was also observed by fluorescent techniques (6).
SSBs are critical for the proper metabolism of ssDNA. Although
they have been found in all the organisms studied so far (4), and
different efforts have been made to understand their mechanism(s) of
action, only very few3 of them have been
well characterized in terms of the molecular interactions that
contribute to complex formation with ssDNA: filamentous phages Ff gene
V proteins (3, 13, 15, 19), E. coli SSB (see Ref. 21, for
review), and T4 gp32 (18, 19, 20). A certain conservation in the global
folding of the ssDNA binding domain (-sheet strands connected by
loops) has been proposed (16, 17). However, only among the SSBs of very
related organisms (E. coli Ff phages and P. aeruginosa phage Pf3) is this conservation clear (3, 15).
Therefore, the particular residues involved in SSB-ssDNA contacts can
be considered as being rather specific, likely as a consequence of
their differences in the global structure of the SSBs and in their DNA
binding parameters.
B. subtilis phage 29 DNA replication occurs
via a strand displacement mechanism which takes place with the
transient formation of high amounts of ssDNA (see Refs. 23 and 24, for
reviews). The elongation step of
29 DNA replication requires a
viral-encoded SSB (
29 SSB), very abundant in infected cells (4% of
the total protein content; 26). The participation of
29 SSB in viral
DNA replication seems to be rather specific in vivo, since
it cannot be substituted by the host SSB. The thermodynamic parameters
defining complex formation between
29 SSB and ssDNA, i.e.
maximal quenching, nucleotide site size, affinity, cooperativity, have
been described in a wide variety of experimental conditions (6, 30).
However, very little information was available with respect to the
structural features of the free and bound
29 SSB and to the effects
of binding on the ssDNA structure.
In this paper, we propose that the hydrodynamic behavior of 29 SSB
in solution (see Table I), analyzed by translational (sedimentation
equilibrium and sedimentation velocity centrifugations) and rotational
measurements (time-decay of its intrinsic fluorescence anisotropy), is
compatible with a monomeric protein of ~13,600 kDa that could be
modeled as a nearly spherical ellipsoid of revolution. The axial ratio
p (=a/b) of
29 SSB could range from
0.8 to 1.0 (oblate model) or 1.0 to 3.2 (prolate model). CD analysis of
29 SSB spectra between 190 and 240 nm allowed us to estimate that
29 SSB is highly structured (47%
-sheets, 10%
-helices, 23%
-turns, at neutral pH). This high amount of
-structures is in the
order of that derived from the CD of T4 gp32 (18) or Ff SSBs (see Ref.
3, for review) and much higher than the ~20% reported for E. coli SSB (67). In the case of M13 and Pf3 SSB or T4 gp32, the high
amount of
-elements has been confirmed by three-dimensional analysis
(13, 14, 15, 18). No positive band was detected at 230 nm, a difference
with the filamentous phages fd and Ike SSBs, which was proposed to
derive from interactions between the phenolic group of certain
tyrosines and local amides (55, 56, 57).
Complex formation induces minimal changes in 29 SSB spectra, the
global conformation of the bound protein remaining essentially unchanged. However, ssDNA had a significant stabilizing effect against
denaturation induced by pH and temperature increase. The major
structural effects induced upon complex formation seem to occur on the
ssDNA structure. Thus, great non-conservative changes (59, 61), defined
by a red shift and a reduction of the
values, occur in the CD
spectra of the ssDNA. These changes likely correspond, besides a
decrease of the stacking, to a different spatial geometry of the
neighbor nucleotides in the bound ssDNA. This different geometry might
be related to the 50% reduction of the ssDNA length determined by
electron microscopic visualization of the
29 SSB-ssDNA complexes
(12).
In the absence of three-dimensional data, the results presented here
constitute the first approximation to the global conformation of 29
SSB and its complex with ssDNA. A hypothetical model to illustrate the
ratio between the size of the protein and the ssDNA in the complex is
presented in Fig. 9. The protein is depicted as a
revolution ellipsoid assuming an axial ratio p = 2. Taking a standard nucleotide interspace of 4 Å, and the calculated
value for the nucleotide site size (n) of
29 SSB, 3-4 nt
(6), it is possible to discard
29 SSB-ssDNA interactions as ssDNA
wrapping around the protein core upon complex formation, as it has been proposed for E. coli SSB (8). Rather, it is likely that, in agreement to the relatively low ssDNA binding constant
(Keff = 1-3 × 105
M
1), ssDNA contacts with a limited region of
29 SSB. The location of this DNA binding domain is not presently
known. However, consistent with the steady-state and time-decay
fluorescence analysis presented in the accompanying paper (30), it
could contain the three tyrosines.
Certain features of the SSBs from the above indicated filamentous phage
SSBs, their small size (~9.7 kDa), small n (4 nt/monomer), a not so high Keff
(105-107 M1), and a
high contain of
-structure elements (3) that remain essentially
unchanged upon ssDNA binding (54, 55, 56), are close to those of
29 SSB.
However, a number of significant differences can be found that strongly
suggest that
29 SSB may have a particular structure and a particular
DNA binding domain. First, those SSBs form dimers in solution and the
ssDNA binding domain is shared by both monomers (13, 15, 19) and
therefore each dimer can contact two strands of ssDNA. Second, although
all contain high amounts of
-elements, the CD spectra of those SSBs
is quite different from that of SSB
29 SSB. Third, ssDNA binding of
these SSBs is moderately salt-dependent (68), indicating
that ionic interactions are mediating complex formation. An opposite
situation was found for
29 SSB-ssDNA complex, where the ionic
composition of the reaction buffers had small effects on its
Keff (6). Fourth, significant functional
differences exist, since binding of Ff SSB to ssDNA leads to an
inhibition of the phage DNA replication process (22) while complex
formation between the
29 SSB and the displaced ssDNA has an
stimulatory effect on
29 DNA replication (23, 24).
The structural and functional differences discussed above strongly
suggest that 29 SSB may have a unique organization. The absolute
requirement of
29 SSB during in vivo viral DNA
replication could be reflecting these particularities.
We are indebted to J. M. Lázaro and L. Villar (Centro de Biología Molecular "Severo-Ochoa," Spain)
for the purification of 29 SSB and Drs. M. P. Lillo and R. Dale
(Instituto de Química Física Rocasolano, Spain) for
their invaluable contribution in fluorescence and analytical
ultracentrifugation data analysis and interpretation. We also thank
Drs. C. W. Hilbers and B. J. M. Harmsen (University of Nijmegen, The
Netherlands) for their training in the circular dichroism techniques,
Dr. A. J. W. G. Visser (Wageningen Agricultural University, The
Netherlands) for the analysis of the secondary structure of free and
bound
29 SSB, and Dr. J. M. Andreu and J. Evangelio for their help
with the J700 CD apparatus.