Fluorescence studies of the replication initiator protein
RepA in complex with operator and iteron sequences
and free in solution
Rutger E. M. Diederix
1,2
, Cristina Da
´
vila
1,2
, Rafael Giraldo
2
and M. Pilar Lillo
1
1 Departamento de Biofı
´
sica, Instituto de Quı
´
mica Fı
´
sica ‘Rocasolano’, CSIC, Madrid, Spain
2 Departamento de Microbiologı
´
a Molecular, Centro de Investigaciones Biolo
´
gicas, CSIC, Madrid, Spain
RepA is the DNA replication initiator protein of the
Pseudomonas plasmid pPS10. It is representative of a
family of plasmid replication initiators active in many
Gram-negative bacteria, including the initiators from
plasmids such as pSC101, F and R6K [1]. The opera-
tor region preceding the repA gene contains a partially

doi:10.1111/j.1742-4658.2008.06669.x
RepA, the replication initiator protein from the Pseudomonas plasmid
pPS10, regulates plasmid replication and copy number. It is capable of
autorepression, in which case it binds as a dimer to the inverted repeat oper-
ator sequence preceding its own gene. RepA initiates plasmid replication by
binding as a monomer to a series of four adjacent iterons, which contain the
same half-repeat as found in the operator sequence. RepA contains two
domains, one of which binds specifically to the half-repeat. The other is the
dimerization domain, which is involved in protein–protein interactions in
the dimeric RepA–operon complex, but which actually binds DNA in the
monomeric RepA–iteron complex. Here, detailed fluorescence studies on
RepA and an N-(iodoacetyl)aminoethyl-8-naphthylamine-1-sulfonic acid-
labeled single-cysteine mutant of RepA (Cys160) are described. Using time-
resolved fluorescence depolarization measurements, the global rotational
correlation times of RepA free in solution and bound to the operator and to
two distinct iteron dsDNA oligonucleotides were determined. These provide
indications that, in addition to the monomeric RepA–iteron complex, a
stable dimeric RepA–iteron complex can also exist. Further, Fo
¨
rster reso-
nance energy transfer between Trp94, located in the dimerization domain,
and N-(iodoacetyl)aminoethyl-8-naphthylamine-1-sulfonic acid-Cys160,
located on the DNA-binding domain, is observed and used to estimate the
distance between the two fluorophores. This distance may serve as an indica-
tor of the orientation between both domains in the unbound protein and
RepA bound to the various cognate DNA sequences. No major change in
distance is observed and this is taken as evidence for little to no re-orienta-
tion of both domains upon complex formation.
Abbreviations
(I)AEDANS, N-(iodoacetyl)aminoethyl-8-naphthylamine-1-sulfonic acid; FRET, Fo

times were determined by fluorescence anisotropy
decay experiments using the extrinsic fluorophore
N-(iodoacetyl)aminoe thyl-8-naphthylamine-1-sul fonic
acid (AEDANS), specifically coupled to Cys160 in the
single-cysteine mutant C160–RepA. The AEDANS
probe was also used as a Fo
¨
rster resonance energy
transfer (FRET) acceptor to monitor putative interdo-
main movements in RepA upon binding the various
DNA sequences. We show that, despite the extensive
structural rearrangement that is known to occur upon
monomerization and DNA binding to the iteron
sequence [3–6], an appreciable change in the inter-
domain organization is not actually observed. Finally,
it appears that monomerization does not occur effi-
ciently in very short oligonucleotides that contain few
bases more than the iteron sequence, and RepA binds
as a dimer instead.
Results
Labeling and characterization of C160–RepA
C160–RepA is a double-mutant of His
6
-tagged wild-
type RepA [4] in which two wild-type Cys residues
(C29 and C106) have been changed to Ser. The single
remaining Cys160 is located on the C-terminal DNA-
binding domain of RepA, also called the WH2
domain, which specifically recognizes the operator and
iteron sequences [1,2]. Most C160–RepA is expressed

protein (up to 10–20 lm). Under denaturing condi-
tions, RepA can easily be concentrated 10- to 100-fold,
thus favoring the bimolecular labeling reaction greatly
under the conditions of $ 15-fold excess IAEDANS.
The CD spectrum of unlabeled or AEDANS-labeled
C160–RepA is indistinguishable from that of wild-type
RepA at 5 °C (Fig. 1A), indicating that the secondary
structure is not affected by the mutation or by
AEDANS labeling. Thermal denaturation analysis of
the protein variants suggests a lower stability of the
mutant (Fig. 1B). The C160–RepA variants show a
lower melting temperature than wild-type RepA (60
versus 67 °C for wild-type RepA), and the thermal
transition of unlabeled C160–RepA has a substantially
lower slope (reduced co-operativity) than wild-type
RepA and the labeled variant. However, room temper-
ature is well below the melting transition, and as the
experiments described here have been performed at or
below this temperature, it can safely be assumed that
the mutant protein is fully folded. This is supported by
the observation that the fluorescence emission spec-
trum of the unique Trp residue (W94), a sensitive indi-
cator of the folding state of the dimerization domain
of RepA [4], is unchanged in the mutant with
respect to that of wild-type RepA (Fig. 1C). Finally,
AEDANS C160–RepA and wild-type RepA show
identical binding to the operator sequence (Fig. 1D),
confirming that mutation and labeling do not affect
the function, and by implication therefore also the
structure, of RepA.

pH 6.0. [RepA] was $ 2 l
M.
A
B
C
D
Fig. 1. (A) Near- and far-UV CD spectra of wild-type RepA (solid
line) and C160–RepA both unlabeled (dashed line) and AEDANS-
labeled (dash-dots). The spectra were recorded at 5 °C with
$ 3.5 l
M wild-type and unlabeled C160–RepA, and 7 lM AEDANS
C160–RepA. The buffer spectrum is subtracted and the spectra
have been transformed to mean residual ellipticity units. (B) Ther-
mal denaturation curves for wild-type RepA (solid lines) and C160–
RepA unlabeled (dashed line) and AEDANS-labeled (dash-dots). The
temperature dependence of the ellipticity at 220 nm is shown, nor-
malized to help compare the different proteins. (C) Fluorescence
emission spectra (k
ex
= 295 nm) of wild-type RepA (solid line),
C160–RepA both unlabeled (dashed line) and AEDANS labeled
(dash-dots), recorded at 23.5 °C with $ 2 l
M protein and with
intensities normalized with respect to their emission maximum at
327 nm. (D) Binding of wild-type RepA (
) and AEDANS C160–
RepA (s) to 10 nm Alexa568-labeled 1IR, monitored by Alexa568
fluorescence anisotropy (k
ex
= 535 nm, k

of the ‘pure’ AEDANS fluorescence, i.e. the emission
spectrum excited at 375 nm (not shown). There is,
however, a clear increase in the fluorescence anisotropy
for each of the sequences (Fig. 3D–F), indicating a
decrease in the rotational mobility of AEDANS C160–
RepA. The anisotropy increase is slightly different for
each of the three sequences, and relates to an increased
global rotational correlation time for the AEDANS
probe caused by C160–RepA binding to DNA (see
below). Addition of DNA also induces a change in the
shape of the excitation spectra. This is caused by a
decrease in the Trp contribution to AEDANS fluores-
cence, as illustrated by the difference spectra between
free and bound RepA, which are typical of Trp
(Fig. 3A–C).
The increase in directly excited AEDANS anisotropy
matches very well with the decrease in W94 contribu-
tion to AEDANS fluorescence for each of the three
Table 1. Sequence of the oligonucleotides used in this study. IR (operator, half sites in bold), 1DR (single iteron underlined, with the half
site also present in the operator in bold, purported DnaA box dashed underlined), 1DR-short (single iteron underlined, with the half site also
present in the operator in bold).
Name Length (bp) Sequence
1IR 39 GAACAAGGACAGGGCATTGACTTGTCCCTGTCCCTTAAT
1DR 45 ATACCC
GGGTTTAAAGGGGACAGATTCAGGCTGTTATCCACACCC
1DR-short 30 GCCC
GGGTTTAAAGGGGACAGATTCAGGCC
A D
B E
C F

AEDANS C160–RepA as a function of [1IR]
(s). Data were fit using the quadratic bind-
ing equation (see Eqns 3–4). (E) and (F) as
(D), except they refer to titrations with 1DR
and 1DR-short, respectively. Experiments
were performed at 23.5 °C, in 0.15
M
(NH
4
)
2
SO
4
,15mM NH
4
-acetate, 0.03 mM
EDTA, 3% glycerol, pH 6.0.
Fluorescence studies of RepA R. E. M. Diederix et al.
5396 FEBS Journal 275 (2008) 5393–5407 ª 2008 The Authors Journal compilation ª 2008 FEBS
tested oligonucleotides (Fig. 3D–F). The change in flu-
orescence and anisotropy were fit simultaneously for
each titration. In the fits, the protein concentration
was left free, to serve as an indicator of stoichiometry.
In the case of 1IR, the fit resulted in a binding stoichi-
ometry of 2 : 1, i.e. dimer binding. The reactant con-
centrations were too high to obtain relevant
information on the binding affinity. For binding to
1DR, the best fit yielded a binding stoichiometry of
$ 1 : 1, i.e. monomer binding, with a K
d

DNA complexes. Because the fraction of non-fluores-
cent donor molecules does not contribute to the
Trp fi AEDANS energy transfer process, a correc-
tion of the FRET efficiencies for the presence of non-
fluorescent W94 is required (see Eqn 1, Experimental
procedures). After doing so, it appears that the differ-
ence in FRET efficiency between free RepA and its
DNA complexes is actually relatively minor (see
Table 2). Accordingly, the resulting distance calculated
between W94 and AEDANS–C160 does not display
large variations between the different species.
However, there are a number of caveats that should
be taken into account. First, there are several tyrosine
residues in RepA. As the fluorescence was excited at
280 nm, there is the possibility that some of the five
tyrosines present in the W94-containing N-terminal
domain of RepA also contribute to the experimental
FRET efficiency, by Tyr fi Trp energy transfer. As
the distance between W94 and the nearest Tyr residue
is $ 15 A
˚
[5], this contribution is negligible, however.
This conclusion is well supported by the apparent lack
of contribution of Tyr to the excitation spectrum of
acceptor AEDANS indicated in the excitation differ-
ence spectra seen in Fig. 3A–C. Second, the Fo
¨
rster
and donor-acceptor distances determined here, relate
to the R

lute distance between 0.7 and 1.3 times the value of
R(
2
/
3
), presented in Table 2.
Nevertheless, the R(
2
/
3
) value in the complex with
1DR is in excellent agreement with the distance mea-
sured between the C
b
atoms of both residues in the
structural model of RepA [2] based on the monomer–
iteron complex structure of the homologous RepE pro-
tein [7]. W94 and C160 are each located on one of the
Table 2. Fluorescence and FRET parameters of the W94–
AEDANS–C160 pair and resulting average inter-probe distances, in
free RepA and RepA bound to various cognate DNA sequences.
FRET efficiency was determined using Eqn (1), and assuming
e
W 94
280 nm

e
AEDANS
340 nm
= 1 and e

3
)(A
˚
)
(± 7)
b
free RepA 0.29 1.00 0.7 22
+ 1IR 0.14 0.48 0.8 20
+ 1DR 0.21 0.72 0.6 23
+ 1DR-short 0.16
a
0.55
a
0.8
a
20
a
Values based on extrapolations from binding curves and as such
not experimentally confirmed.
b
Using R
0
(
2
/
3
) = 25 ± 1 A
˚
.
R. E. M. Diederix et al. Fluorescence studies of RepA

of global and local re-orientational motions, respec-
tively. This was carried out for both W94 in wild-type
RepA and AEDANS-labeled C160–RepA. AEDANS
has a much longer fluorescence lifetime than Trp,
allowing a much greater level of confidence in the
determination of correlation times pertaining to the
global rotational motion. Nevertheless, the global rota-
tional information obtained from Trp fluorescence
anisotropy decays (see Fig. S1 and Table S1) shows a
trend in agreement with the data from the AEDANS
experiments. Furthermore, despite the relatively poor
photon-counting statistics, the local dynamics of W94
have been characterized from the Trp decays. In
Fig. 4, anisotropy decays (k
em
= 480 nm) are shown
for the different AEDANS C160–RepA species,
together with best fits assuming a bi-exponential
function for r(t) (see Experimental procedures). The
A

B

C D
Fig. 4. Anisotropy decays R(t)
(k
ex
= 375 nm, k
em
= 480 nm) of AEDANS

analogous decays with k
em
= 530 nm, with corre-
sponding best fits and tabulated parameters, are sup-
plied in Fig. S2 and Table S1.
The AEDANS data confirm the presence of discrete
complexes under the conditions of the experiment,
and that binding is complete, in agreement with the
binding curves (Fig. 3), except for the case of the
complex with 1DR-short, which under these condi-
tions should contain $ 20% free RepA. As expected,
the global rotational correlation time, /
2
, increases
upon binding of RepA to its cognate DNA. Apart
from the RepA–1DR-short complex, the observed val-
ues easily fall within the range reasonably expected
from molecules of this size and shape (Table 3). The
expected values for free RepA and the dimeric RepA–
1IR complex were calculated on basis of hydro-
dynamic shapes and volumes corresponding to prior
[3] sedimentation velocity measurements as shown in
Fig. 5. Both can be characterized as rigid elongated
shapes. For the monomer RepA–1DR complex, the
structure modeled on the homologous mRepE–DNA
crystal structure [7] was used directly to calculate the
expected global rotational correlation time. The calcu-
lated values for the RepA–1DR-short complex pertain
to a monomer, i.e. the modeled structure as above,
but with a truncated oligonucleotide having 30 bp

i
± 0.002
hsi
c
(ns)
± 0.4
b
1
± 0.05
/
1
(ns)
±3
b
2
±0.05
/
2
(ns)
±10
/
2
calc (ns)
(max–min)
Free RepA 0.209 13.1 0.234 4.5 0.766 56 (42–89)
d
+ 1IR 0.239
b
12.2 0.156 2.7 0.844 109 (61–131)
d

(upper) and RepA–1IR complex (lower), with axial ratio and volumes
corresponding to frictional ratios determined from prior sedimenta-
tion velocity analysis (3) and molecular mass (23) respectively. The
modeled structure of monomeric RepA–1DR is shown in two orien-
tations (center). For clarity, the purported structure of RepA mono-
mer with 1DR-short is not shown. The length of 1DR-short only
allows for five nucleotides (half a helical turn) to protrude from
either end of the protein–DNA interface.
R. E. M. Diederix et al. Fluorescence studies of RepA
FEBS Journal 275 (2008) 5393–5407 ª 2008 The Authors Journal compilation ª 2008 FEBS 5399
correlation, in line with the fact that it is the only
complex involving monomeric RepA.
Finally, we note that the range of global rotational
correlation times calculated for the dimer–operator
complex of the F plasmid RepE protein, which is
highly homologous to RepA and of which the crystal
structure is known [9], is shorter (57–83 nucleotides)
than observed here for the RepA–1IR complex. This
could mean that there are significant differences
between the RepE– and RepA–operator complexes,
which are possibly related to the different spacing
between the half repeats in both operator DNA
sequences [9].
Oligomerization state of free and complexed
RepA determined by homoFRET
In order to understand the oligomerization state of
RepA in the different DNA complexes better, homo-
FRET experiments were carried out. Herein, use is
made of C160–RepA specifically labeled with Atto532.
In a double-labeled sample, FRET is expected to occur

lengths (red-edge excitation). The enhanced fluores-
cence depolarization in the double-labeled dimers with
respect to the single-labeled samples is a clear indica-
tion of homoFRET in the double-labeled samples [16].
The increase in steady-state anisotropy observed upon
decreasing the degree of Atto532-labeling from 60% to
10% is also observed when excess 1DR is added to
60% labeled Atto532 C160–RepA, but not upon the
addition of excess 1IR and 1DR-short (Fig. 6B). This
means that addition of 1DR abolishes the homoFRET,
by inducing RepA monomerization. In fact, the addi-
tion of 1IR and 1DR-short causes a small decrease in
anisotropy which may be related to enhanced homo-
FRET caused by slight rearrangement of the mono-
mers in the RepA dimers or by minor aggregation.
Thus, RepA is dimeric free in solution and when
bound to its operator sequence, but also when bound
to 1DR-short. In the presence of excess 1DR, mono-
merization of RepA takes place.
Discussion
One of the striking properties of RepA is that it is able
to recognize two types of DNA sequence, either the
operator – with inverted repeats – or the iteron, in
which the same 6 bp sequence half-site found in the
operator is specifically recognized. Upon binding to
the operator, RepA remains dimeric; it thus retains its
symmetry matching the inverted repeats of the oligo-
nucleotide. When this symmetry is absent, i.e. for the
A
B

.
Fluorescence studies of RepA R. E. M. Diederix et al.
5400 FEBS Journal 275 (2008) 5393–5407 ª 2008 The Authors Journal compilation ª 2008 FEBS
iteron sequence, RepA binds as a monomer instead of
a dimer.
When operator or iteron DNA is added to AE-
DANS C160–RepA, discrete complexes are formed
(Fig. 3), characterized by higher AEDANS fluores-
cence anisotropy values and decreased apparent Trp-
AEDANS FRET (see below). RepA binds operator
DNA (1IR) with a clear stoichiometry of 2 : 1, i.e. the
protein binds as a dimer. With the iteron sequence
1DR, which includes an additional stretch of bases
(see Table 1), a stoichiometry of 1 : 1 is found, i.e.
monomer binding. When the number of bases flanking
the iteron sequence is considerably shorter, as with
1DR-short, the binding affinity is significantly
decreased (2.9 lm), and nears that of non-specific
DNA [6]. Still, a discrete complex is formed in this
case, as corroborated by fluorescence anisotropy decay
measurements.
Fluorescence anisotropy decay analysis is a potent
method to obtain information on the local and global
dynamics of species in solution. Here, it is used to
characterize the discrete species discussed above. For
free RepA and RepA in complex with 1IR or 1DR,
experiments were performed with AEDANS. The anal-
ysis, summarized in Table 3, yields global rotational
correlation times for free RepA and the complex with
1IR corresponding to species involving dimeric RepA,

despite the fact that it contains the full 22 bp iteron.
This last conclusion is corroborated by the observa-
tion that inter-monomeric homoFRET is observable
with 1DR-short, but not 1DR (Fig. 6). That dimer-
binding to iterons is, in principle, possible has previ-
ously been shown by us. According to EMSAs carried
out under crowded conditions, a fraction of RepA
dimers was observed to bind to the 1DR sequence [6].
This fraction is obviously much larger in the case of
1DR-short, and the extra bases on the longer, mono-
mer-binding, oligonucleotide 1DR seem to play a role
in aiding monomerization. The presence of bases
downstream of the iteron sequence has also previously
been shown to promote binding of Rep to pSC101
[18].
The related replication initiator protein p from R6K
plasmid is a well-documented case where not only
monomers, but also dimers, are known to bind to the
iteron [19]. Interestingly, dimers of p protein occupy a
much shorter stretch of the iteron sequence than do
monomers; whereas almost the entire 22 bp iteron
sequence is occupied by the p monomers, only half of
this – notably including the specific 6 bp recognition
sequence (repeat) – is occupied when dimeric p protein
is bound [19]. This may occur here as well. As there is
only one half of the inverted repeat of the operator
sequence present in 1DR-short, it is likely that only
one of two WH2 DNA-binding domains in RepA
dimers is involved in binding. This also makes sense
energetically, the RepA dimer binds operator DNA

FEBS Journal 275 (2008) 5393–5407 ª 2008 The Authors Journal compilation ª 2008 FEBS 5401
initiation. It should be mentioned that the four iterons
in pPS10 are contiguous, thus limiting the degree to
which the flanking sequences may contribute to bind-
ing. In other replicons, however, there are spacer
sequences between the iterons, which in addition may
have some sequence variability [22]. It would be inter-
esting to see whether our findings for RepA can be
extrapolated to other Rep proteins.
An attractive feature of using AEDANS as an
extrinsic label is that, besides its use to analyze the
global rotational correlation times of macromole-
cules, it is useful as a FRET acceptor for intrinsic
Trp residues. RepA fortunately has only one Trp,
making this use of the AEDANS probe more mean-
ingful and helping interpretation of the FRET in
terms of distances between the two fluorophores.
Moreover, W94 and C160 are located on the dimer-
ization and DNA-binding domains of RepA, respec-
tively, allowing us to interpret any observed changes
in FRET in terms of relative movements between
the two domains.
It emerges that the average estimated distance
between the C160–AEDANS and W94 is $ 22 A
˚
in
the free RepA dimer, and this distance decreases by a
few angstroms upon binding either the 1IR, or 1DR-
short oligonucleotides and increases slightly upon
monomerization and binding to 1DR (see Table 2).

RepA and C160–RepA
In all cases, the concentration of protein is expressed in
monomer units. What is referred to as wild-type RepA is the
His
6
-tagged variant of RepA, which was expressed and
purified as described previously [4]. This protein is indistin-
guishable from that without His-tag, except that it has a
higher solubility [3,4]. It was therefore used without sub-
sequent removal of the tag. C160–RepA also has the
His
6
-tag and is a single-cysteine variant of wild-type RepA
in which two of the three wild-type Cys residues (C29, C106)
have been successively replaced by Ser using the PCR-based
QuickChange Kit (Stratagene, Cedar Creek, CA, USA).
Mutations were verified by sequencing. C160–RepA was
expressed as wild-type RepA [4] but almost all C160–RepA
was present in the form of insoluble aggregates. The protein
was isolated by solubilization of the inclusion bodies and
purification by Ni(II)-affinity chromatography under dena-
turing conditions, similarly as described previously [4]. This
results in pure protein, exhibiting a single band on
SDS ⁄ PAGE. After purification, the protein was reduced by
addition of 2 mm 2-mercaptoethanol and exchanged to
unfolding buffer (5.6 m guanidinium hydrochloride, 0.56 m
(NH
4
)
2

allowed to proceed for 2 h at room temperature, and then
12.3 mg of glutathione was added to quench the reaction.
The reaction mixture was exchanged for fresh unfolding
buffer by extensive ultrafiltration. The labeling efficiency
was close to 50%, as judged from UV ⁄ Vis spectroscopy.
AEDANS C160–RepA was refolded in the same way as
unlabeled RepA. Similarly, C160–RepA was labeled with
maleimide Atto532 (Atto-Tec, Siegen, Germany), with 60%
labeling efficiency. Here, the degree of labeling in the folded
Fluorescence studies of RepA R. E. M. Diederix et al.
5402 FEBS Journal 275 (2008) 5393–5407 ª 2008 The Authors Journal compilation ª 2008 FEBS
protein was important and varied from 60% to 10% by
mixing the appropriate amounts of Atto532-labeled and
unlabeled C160RepA before refolding. Correct refolding
of Atto532 C160RepA was conrmed by determining its
binding efciency and stoichiometry to Alexa647-labeled
1IR, using FRET (not shown).
DNA purication and labeling
1DR (Table 1) was prepared as described previously [36].
The 1IR and 1DR-short duplexes were prepared by anneal-
ing their constituent complementary strands (Sigma-Geno-
sys, Cambridge, UK) in equimolar amounts. The duplexes
were puried using a MA7 column (BioRad, Hercules, CA,
USA) followed by desalting using C
18
Sep-PaK columns
(Waters, Milford, MA, USA). Also, 5Â-amine modied vari-
ants of the 1IR and 1DR-short oligonucleotides were rst
reacted with NHS-Alexa568 and Alexa647 (Molecular
Probes) respectively, according to the manufacturers

Trp residue, (W94, the donor) and AEDANS (the acceptor)
was quantied by the (ratio)
A
approach [23,24] using excita-
tion spectra measured at an emission wavelength where
W94 uorescence does not contribute (480 nm). This per-
mits several simplications, among them disregard of uncer-
tainties in the degree of labeling. Because of signicant
static quenching of the W94 donor upon complex forma-
tion, a correction was necessary to account for the fraction
(d
+
) of uorescent donor remaining. The FRET efciency
was calculated as follows:
E ẳ
1
d
ỵ
F
280 nm
F
340 nm


e
AEDANS
280 nm
e
AEDANS
340 nm

structure [5], and thus any contribution of Tyr was not
taken into account in the efciency calculations. The W94
uorescence decay could be analyzed assuming three dis-
crete lifetime components with $ 80% of the uorescence
from one of these species (s $ 4 ns) (see below and
Table S1). The decay can thus be approximated as mono-
exponential, allowing use of simple FRET theory.
The Fo
ă
rster radius, R
0
,inA

, for the W94AEDANS
pair was calculated as follows:
R
0
ẳ 0:211n
4
U
D
j
2
Jị
1=6
2ị
Here, n is the refractive index, with a value of 1.4. The
orientation factor j
2
depends on the relative orientation of

(k
ex
= 295 nm) of free and bound unlabeled C160–RepA
and wild-type RepA, with the protein concentration con-
stant. The ratio of the apparent F
W94
of bound and free
RepA was then used to estimate the fraction of fluorescent
donor remaining upon DNA complexation (d
+
in Eqn 1).
For the R
0
calculation we used the quantum yield deter-
mined for free C160–RepA (see above).
Titrations were performed with increasing amounts of
DNA added to AEDANS C160–RepA, or with increasing
amounts of wild-type RepA or unlabeled C160–RepA
added to Alexa568–1IR. Fresh solutions were prepared for
each data point, and equilibrated 10 min before measuring.
Binding curves were fit to the quadratic expression given in
Eqn (3) for the amount of RepA–DNA complex, with
[RepA] divided by the expected stoichiometry n. The
amount of bound protein or DNA is related to the instru-
ment signal (AEDANS C160–RepA anisotropy and the flu-
orescence intensity ratio for excitation at 280 and 340 nm
F
280 nm
⁄ F
340 nm

þ DNA½
T

ð3bÞ
signal
AEDANS
¼
RepA À DNA½
RepA½
T

n
!
S
bound
þ 1 À
RepA À DNA½
RepA½
T

n
!
S
free
ð4Þ
signal
Alexa568
¼
RepA À DNA½
DNA½

Time-resolved fluorescence intensity and depolarization
measurements were made using the time-correlated single-
photon counting technique, using the set-up described pre-
viously [29]. For Trp fluorescence depolarization measure-
ments (k
ex
= 297 nm, k
em
= 345 nm), the excitation light
source was a Ti : sapphire picosecond laser (Tsunami, Spec-
tra Physics, Mountain View, CA, USA), pumped with a
5 W Nd : YVO
4
diode laser (Millennia, Spectra Physics),
and associated with a third harmonic generator. The pulses
had 1–2 ps width and a repetition rate of 0.8–4 MHz, with
an average power of 20 lW reaching the cuvette. The tem-
perature of the sample was thermostated at 5 °C for
increased sample stability, required because of long acquisi-
tion times. The timing calibration was 6.1 ps per channel,
with 4096 data channels. Typical polarized decay curves
had $ 4000–10 000 counts in the peak (1–2 · 10
6
total pho-
tons). The quality of photon counting statistics in the Trp
experiments was limited by the sample stability. For
AEDANS measurements (T = 23.5 °C; k
ex
= 375 nm,
k

decay of the polarized components by:
RðtÞ¼
I
VV
ðtÞÀGI
VH
ðtÞ½
I
VV
ðtÞþ2GI
VH
ðtÞ½
ð6Þ
In this instrumental setup, G is a scaling factor which is
independent of the emission wavelength and, in general,
has values near 1. It takes in account small instabilities of
the laser and ⁄ or differences in accumulation times for the
two polarized intensities. It was determined by correlating
the steady-state anisotropy value measured separately, to
the anisotropy value resulting from integration of the I
VV
(t)
and I
VH
(t) traces.
Decays I
m
(t) were fit to a sum of n exponential functions
(n = 1, 2) by iterative convolution, using nonlinear global
least-squares methods from the program globals

gave very similar lifetime values. Therefore, anisotropy data
from two emission wavelengths (480 and 530 nm) were fit-
ted simultaneously, linking the corresponding component
lifetimes to get a better defined set of fitting parameters.
The adequacy of the analyses was determined from the
reduced weighted sum of squares of residuals, and visual
inspection of the distribution of weighted residuals. The
general expression used for the emission anisotropy param-
eters in the fits is given by Eqn (7), in which /
i
are correla-
tion times, and the pre-exponential factors b
i
are
normalized. In all the AEDANS anisotropy experiments
the time zero anisotropy, r
0
, from the fit had an average
value of 0.31 ± 0.015. In the case of Trp anisotropy analy-
sis, r
0
was kept fixed at a reasonable value, 0.28 [31], to
suppress the contribution of otherwise unaccounted for
scatter to useful parameters.
rðtÞ¼r
0
X
b
i
exp Àt=/

correlation times for each geometry are included in Table 3
for comparative purposes. The global rotational correlation
times for the F plasmid RepE–operator complex were cal-
culated in the same way, using its recently published crystal
structure [9]. Note that the correlation time actually mea-
sured depends on the orientation of the AEDANS absorp-
tion and emission transition dipoles with respect to the
protein axes. For a case where any of these are parallel to
the long axis, the value takes the maximum limit. For the
other extreme, i.e. transition dipoles perpendicular to the
protein long axis, the correlation time takes a value slightly
higher than the minimum value [32].
For free RepA and the complex RepA–1IR, no (homolo-
gous) structures are available. A prolate ellipsoid shape was
assigned to both structures, with dimensions based on fric-
tion coefficients determined previously by sedimentation
velocity experiments (f ⁄ f
0
= 1.2 for free and 1IR-bound
RepA), in which special care was taken regarding complex
dissociation during the analytical ultracentrifugation [3].
This value implies a ratio of long versus short axes of 4 in
the prolate ellipsoid, giving an observed average global
rotational correlation time that is 1.6–3.4 times higher than
expected if the molecule were spherical, depending on probe
orientation [33,34]. Just like for the cases where correlation
times were calculated using hydropro, the limiting mini-
mum and maximum rotational correlation times have been
included in Table 3. Partial specific volumes used for pro-
tein, DNA and hydration were 0.703, 0.55 and 0.28 mLÆg

pez A,
Alfonso C, Rivas G, Dı
´
az-Orejas R & Giraldo R (2003)
Structural changes in RepA, a plasmid replication
R. E. M. Diederix et al. Fluorescence studies of RepA
FEBS Journal 275 (2008) 5393–5407 ª 2008 The Authors Journal compilation ª 2008 FEBS 5405
inititator, upon binding to origin DNA. J Biol Chem
278, 18606–18616.
4 Giraldo R, Andreu JM & Dı
´
az-Orejas R (1998) Protein
domains and conformational changes in the activation
of RepA, a DNA replication initiator. EMBO J 17,
4511–4526.
5 Giraldo R, Ferna
´
ndez-Tornero C, Evans PR, Dı
´
az-
Orejas R & Romero A (2003) A conformational switch
between transciptional repression and replication initia-
tion in the RepA dimerization domain. Nat Struct Biol
10, 565–571.
6Dı
´
az-Lo
´
pez T, Da
´

9–15.
13 Callis PR & Vivian JT (2003) Understanding the vari-
able fluorescence quantum yield of tryptophan in pro-
teins using QM–MM simulations. Quenching by charge
transfer to the peptide backbone. Chem Phys Lett 369,
409–414.
14 Dale RE, Eisinger J & Blumberg WE (1979) The orien-
tational freedom of molecular probes. The orientational
factor in intramolecular energy transfer. Biophys J 26,
161–194.
15 Lillo MP, Beechem JM, Szpikowska BK, Sherman MA
& Mas MT (1997) Design and characterization of a
multisite fluorescence energy-transfer system for protein
folding studies: a steady-state and time-resolved study
of yeast phosphoglycerate kinase. Biochemistry 36,
11261–11272.
16 Weber G & Shinitzky M (1970) Failure of energy trans-
fer between identical aromatic molecules on excitation
at the long wave edge of the absorption spectrum. Proc
Natl Acad Sci USA 65, 823–830.
17 Garcia de la Torre J, Huertas ML & Carrasco B (2000)
Calculation of hydrodynamic properties of globular
proteins from their atomic-level structure. Biophys J
78,
719–730.
18 Fueki T, Sugiura S & Yamaguchi K (1996) Open
strands to iterons promote the binding of the replication
initiator protein (Rep) of pSC101 to the unit sequence
of the iterons in vitro. Biochim Biophys Acta 1305, 181–
188.

istry 12, 4154–4161.
27 van der Meer BW (2002) Kappa-squared: from nuisance
to new sense. J Biotechnol 82, 181–196.
28 Szabo AG & Rayner DM (1980) Fluorescence decay of
tryptophan conformers in aqueous solution. J Am Chem
Soc 102, 554–563.
29 Lillo MP, Can
˜
adas O, Dale RE & Acun
˜
a AU (2002)
Location and properties of the taxol binding center in
microtubules: a picosecond laser study with fluorescent
taxoids. Biochemistry 41, 12436–12449.
30 Beechem JM, Gratton E, Ameloot M, Knutson JR &
Brand L (1991) The global analysis of fluorescence
intensity and anisotropy decay data: second generation
theory and programs. In Topics in Fluorescence
Spectroscopy, Vol. 2 (Lakowicz JR, ed.), pp. 241–305.
Plenum Press, New York, NY.
Fluorescence studies of RepA R. E. M. Diederix et al.
5406 FEBS Journal 275 (2008) 5393–5407 ª 2008 The Authors Journal compilation ª 2008 FEBS
31 Ruggiero AJ, Todd DC & Fleming GR (1990) Subpico-
second fluorescence anisotropy studies of tryptophan in
water. J Am Chem Soc 112, 1003–1014.
32 Lakowicz JR (1999) Principles of Fluorescence Spectro-
scopy, 2nd edn. Kluwer Academic, New York, NY.
33 Zorrilla S, Rivas G & Lillo MP (2004) Fluorescence
anisotropy as a probe to study tracer proteins in
crowded solutions. J Mol Recog 17, 408–416.

materials supplied by the authors. Any queries (other
than missing material) should be directed to the corre-
sponding author for the article.
R. E. M. Diederix et al. Fluorescence studies of RepA
FEBS Journal 275 (2008) 5393–5407 ª 2008 The Authors Journal compilation ª 2008 FEBS 5407