Báo cáo khoa học: Association of RNA with the uracil-DNA-degrading factor has major conformational effects and is potentially involved in protein folding pot

Association of RNA with the uracil-DNA-degrading factor
has major conformational effects and is potentially
involved in protein folding
Angela Bekesi
1
, Maria Pukancsik
1
, Peter Haasz
1
, Lilla Felfoldi
1
, Ibolya Leveles
1
, Villo Muha
1
,
Eva Hunyadi-Gulyas
2
, Anna Erdei
3
, Katalin F. Medzihradszky
2,4
and Beata G. Vertessy
1,5
1 Institute of Enzymology, Biological Research Centre, Hungarian Academy of Sciences, Budapest, Hungary
2 Proteomics Research Group, Biological Research Centre, Hungarian Academy of Sciences, Szeged, Hungary
3 Department of Immunology, Eo
¨
tvo
¨
s Lora

´
nd University of Sciences, Budapest, Hungary
4 Department of Pharmaceutical Chemistry, University of California, San Francisco, USA
5 Department of Applied Biotechnology, Budapest University of Technology and Economics, Hungary
Introduction
Genomic information stored in DNA is under constant
threat from spontaneous chemical modifications,
occurring under normal physiological conditions. One
of the most frequent spontaneous base transitions
is hydrolytic deamination of cytosine to form uracil
[1–3]. This alteration is mutagenic, as it will convert to
Keywords
conformational states; cotranslational
folding; RNA-assisted folding; RNA binding;
uracil-DNA-degrading factor
Correspondence
A. Bekesi and B. G. Vertessy, Karolina
Street 29, H-1113 Budapest, Hungary
Fax: +36 1 466 5465
Tel: +36 1 279 3116
E-mail: ;

Website:
(Received 7 June 2010, revised 29 October
2010, accepted 4 November 2010)
doi:10.1111/j.1742-4658.2010.07951.x
Recently, a novel uracil-DNA-degrading factor protein (UDE) was identi-
fied in Drosophila melanogaster, with homologues only in pupating insects.
Its unique uracil-DNA-degrading activity and a potential domain organiza-
tion pattern have been described. UDE seems to be the first representative

of a new protein family with unique enzyme activity that has a putative
role in insect development. In addition, UDE may also serve as potential
tool in molecular biological applications. Owing to lack of homology with
other proteins with known structure and ⁄ or function, de novo data are
required for a detailed characterization of UDE structure and function.
Here, experimental evidence is provided that recombinant protein is present
in two distinct conformers. One of these contains a significant amount of
RNA strongly bound to the protein, influencing its conformation. Detailed
biophysical characterization of the two distinct conformational states
(termed UDE and RNA–UDE) revealed essential differences. UDE cannot
be converted into RNA–UDE by addition of the same RNA, implying
putatively joint processes of RNA binding and protein folding in this con-
formational species. By real-time PCR and sequencing after random clon-
ing, the bound RNA pool was shown to consist of UDE mRNA and the
two ribosomal RNAs, also suggesting cotranslational RNA-assisted fold-
ing. This finding, on the one hand, might open a way to obtain a conform-
ationally homogeneous UDE preparation, promoting successful
crystallization; on the other hand, it might imply a further molecular func-
tion of the protein. In fact, RNA-dependent complexation of UDE was
also demonstrated in a fruit fly pupal extract, suggesting physiological rele-
vance of RNA binding of this DNA-processing enzyme.
Abbreviations
ANS, 8-anilinonaphthalene-1-sulfonate; EMSA, electrophoretic mobility shift assay; GAPDH, glyceraldehyde-3-phosphate dehydrogenase;
NLS, nuclear localization signal; UDE, uracil-DNA-degrading factor protein; UDG, uracil-DNA glycosylase.
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 295
a stable point mutation (G:C fi A:T). An alternative
pathway for uracil appearance in DNA is thymine-
replacing misincorporation, which is usually efficiently
prevented by the action of dUTPase, which sanitizes
the dUTP ⁄ dTTP pool [4]. Repair of uracil-substituted

DNA is initiated by uracil-DNA glycosylases (UDGs)
in an almost ubiquitous manner. The major UDG
enzyme, the product of the ung gene, excises both
deaminated cytosine and thymine-replacing uracil with
high efficiency. In Drosophila melanogaster, the well-
annotated genome provides unequivocal evidence for
lack of ung [5]. This situation, together with the lack
of dUTPase, which is responsible for preventing
thymine-replacing uracil incorporation into DNA, in
larval tissues may allow accumulation of uracil-
containing DNA at least at specific developmental
stages [6]. By the use of affinity chromatography for
searching additional uracil-DNA-recognizing factors in
late larval extracts, a specific uracil-DNA-degrading
protein (UDE), with homologues only in pupating
insect genomes, was identified and partially character-
ized [7,8].
The DNA-degrading activity of UDE necessarily
involves the formation of a functional DNA–protein
complex. In fact, the DNA-binding ability of UDE
was evident in the first experiments [7], but sequence
homology searches did not reveal any already
described nucleic acid-binding sequence motifs. No 3D
structure is yet available for any UDE-homologue pro-
tein. Trials for structural modelling by fold recognition
methods suggested a novel fold for the UDE protein.
Alignment of UDE homologue sequences identified
five conserved motifs, the first two segments of which
are extended and show significant homology with each
other (termed 1A and 1B) [7]. De novo 3D modelling

of the 1A ⁄ 1B fragments revealed the existence of two
helical bundles with an extended surface formed by
conserved positively charged residues [8]. These results,
together with DNA-binding-induced protection against
limited proteolysis, suggested that DNA binding may
occur in this region [8]. Detailed characterization of
the interaction between UDE and nucleic acids is
expected to provide important insights into specific
aspects of the novel activity.
In the present work, we aimed to characterize
UDE–nucleic acid complexes. Unexpectedly, we found
that a portion of recombinant UDE copurified with
strongly bound RNA. On the basis of this finding, we
decided to investigate the effects of RNA binding on
protein structure and stability in detail. Recombinant
UDE is eluted in two distinct peaks with different
chromatographic techniques, suggesting the existence
of two distinct conformational states of the enzyme:
RNA-bound (RNA–UDE) and RNA-free (UDE).
RNA binding has pronounced effects on thermal sta-
bility and the cooperativity of protein unfolding, as
shown by thermal transitions monitored by CD and
fluorescence spectroscopy. Possible effects of RNA
binding on UDE conformation were assayed by CD
spectroscopy and by titration of hydrophobic surface
patches. The DNA-binding and RNA-binding abilities
of the enzyme forms were measured by electrophoretic
mobility shift assay (EMSA). Characterization of the
bound RNA pool by real-time RT-PCR revealed an
abundance of UDE mRNA as well as of domain V of

23S rRNA. These results, together with the character-
istics of the two distinct conformational states, suggest
putative cofolding of UDE and RNA during transla-
tion. Physiological relevance of RNA–UDE binding
was also demonstrated.
Results
Two distinct molecular populations of UDE can
be separated by Ni
2+
-affinity and size exclusion
chromatography
In a refined purification method with recombinant His-
tagged UDE (see Experimental procedures), gradient
elution was applied on a HisTrap column with AKTA
purifier. Two distinct fractions of intact full-length
UDE eluted at about 300 and 600 mm imidazole,
respectively (Fig. 1A), reflecting their altered affinity
for the Ni
2+
-column. SDS ⁄ PAGE indicated the same
electrophoretic mobility for UDE in both fractions,
arguing against protein degradation or observable
modifications (Fig. 1A, insert panel). MS analysis of
the two protein bands further confirmed the lack of
any detectable post-translational modifications (data
not shown). In the first fraction, UDE eluted together
with a large amount of nucleic acid, whereas the sec-
ond fraction was practically nucleic acid-free. Both
fractions were subjected to size exclusion chromatogra-
phy, where the nucleic acid-containing fraction eluted

in the exclusion volume (corresponding to an apparent
molecular mass larger than 600 kDa), and the nucleic
acid-free fraction eluted at a position corresponding to
an apparent molecular mass of 52 kDa (Fig. 1B). Pre-
viously, we reported purification of UDE by stepwise
Ni
2+
-agarose chromatography, which did not allow
separation of these two fractions [8]. Size exclusion
chromatography of this UDE preparation resulted in
two distinct peaks at the same positions as above. The
observation that full-length UDE can be separated into
two fractions by applying two different chromato-
graphic methods, Ni
2+
-agarose affinity and size
RNA-dependent conformational states of UDE A. Bekesi et al.
296 FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS
exclusion, may suggest the potential existence of two
separate conformational states of the protein, one of
these containing a significant amount of associated
nucleic acid. To address the potential significance of
these conformational states, we first wished to charac-
terize the bound nucleic acid.
Nucleic acid-binding UDE fraction contains RNA,
and not DNA
On native agarose gel, a smeared band was visualized
in the nucleic acid-containing UDE fraction (Fig. 1C).
Test digestions of this UDE preparation by DNase as
well as by RNase A were performed. Surprisingly,

DNase treatment did not significantly perturb the
detected nucleic acid band, whereas RNase treatment
completely eliminated it (Fig. 1C). The specific activi-
ties of RNase A and DNase were also checked on
plasmid DNA; DNase did indeed have strong activity,
whereas RNase A was not contaminated by DNase
activity (data not shown). We therefore conclude that
the nucleic acid content found in the recombinant
UDE preparation that copurifies with UDE on both
Ni
2+
-agarose and size exclusion chromatography is, in
A
B
CD
Fig. 1. Purification of recombinant UDE and its RNA binding. (A) Elution profile of recombinant UDE in Ni
2+
-affinity chromatography resulted
in two distinct peaks. Fat and thin lines correspond to absorbance of the eluate at 280 and 260 nm, which are characteristic for proteins and
nucleic acids, respectively. The dashed line shows the imidazole gradient applied to promote the elution of His-tagged UDE. Left and right
vertical axes indicate absorbance units and imidazole concentration, respectively. Two peaks eluted at 300 and 600 m
M imidazole, respec-
tively; the first contained significant amount of nucleic acid, and the second was practically free of nucleic acids. The insert shows the pro-
tein contents of the two corresponding fractions (1 and 2) analysed by SDS ⁄ PAGE. (B) Elution profile of the two UDE fractions in size
exclusion chromatography. Fat and thin lines show absorbance of the eluate at 280 nm and 260 nm, respectively. Black and grey lines corre-
spond to the nucleic acid-containing and nucleic acid-free UDE fractions obtained by the previous Ni
2+
-affinity chromatography, respectively.
The observed values of the elution volume characteristic for the two fractions were 9 and 14.5 mL, corresponding to more than 600-kDa
and about 52-kDa apparent molecular masses, respectively. (C) The nucleic acid content of the first fraction of UDE was detected on aga-

rose gel as RNA. Lane 1: nucleic acid containing UDE fraction from Ni
2+
-affinity chromatography. Lane 2: the same treated with DNase.
Lane 3: the same treated with RNase. Lane 4: the same treated with proteinase K. Marker positions are indicated on the right. (D) DNase
and RNase treatment of the nucleic acid containing UDE fraction analysed by size exclusion chromatography also identified the nucleic acid
as RNA. Fat and thin lines show absorbance of the eluate at 280 and 260 nm, respectively. Nucleic acid-containing UDE without treatment
(black), upon DNase treatment (grey) and upon RNase A treatment (light grey) were analysed by size exclusion chromatography. Note that
RNase A may degrade the nucleic acid content of UDE, resulting in a similar elution profile to that of the nucleic acid-free UDE (B). The large
peak at about 20-mL elution volume (light grey chromatogram) might be caused by RNA fragments or nucleotides.
A. Bekesi et al. RNA-dependent conformational states of UDE
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 297
fact, RNA. Therefore, we termed the RNA-containing
UDE fraction RNA–UDE and, in parallel, the RNA-
free UDE fraction as UDE.
RNA–UDE was also analysed by size exclusion
chromatography after DNase or RNase treatment
(Fig. 1D). DNase treatment did not cause significant
changes in the chromatogram; one single peak in the
exclusion volume with significant nucleic acid content
(compare absorbance values at 260 and 280 nm) was
detected, similarly to what was seen in the chromato-
gram of the untreated sample. In contrast, RNase
treatment resulted in a drastic decrease in this peak,
with the concomitant emergence of a small addi-
tional peak at the same position as in the case of
untreated RNA-free UDE. The straightforward sepa-
ration of these two UDE species by chromatography
suggested that these species may represent two confor-
mational states.
RNA content of RNA–UDE

On native agarose gel, the RNA content of RNA–
UDE appears as a smear at above 10 000-nucleotide
apparent size (Fig. 1C). However, upon treatment of
RNA–UDE with proteinase K, a band appeared at a
much lower apparent size; this could correspond to
degradation products of the RNA. The observed sig-
nificant gel shift also confirms the binding between the
protein and the RNA in RNA–UDE.
To analyse the composition of RNA–UDE, we per-
formed deconvolution of its UV–visible absorbance
spectrum, using the separately determined A
260 nm
⁄
A
280 nm
ratios (2 for RNA, and 0.5 for nucleic acid-
free UDE, respectively). On the basis of this analysis,
the ratio is about 17–22 RNA nucleotides per UDE
monomer. This also means that several UDE proteins
may bind to the same RNA molecule, suggesting non-
sequence-specific characteristics of the binding.
UDE does not show RNA-cleaving activity
UDE was described as a uracil-DNA-degrading factor.
It was of interest to investigate whether UDE can also
degrade RNA. The results shown in Fig. 1C reveal
that the RNA content found to be associated with
UDE during extraction and purification is present as
100–500 nucleotides long molecular species, indicating
that UDE may not exhibit RNase-like degrading activ-
ity. To further investigate this, we tested the effect of

both RNA–UDE and UDE on a double-stranded
in vitro synthesized RNA of 471 nucleotides. Figure 2
shows that neither RNA–UDE nor UDE possesses
RNA-degrading activity.
The above results indicated that UDE may exist in
two distinct conformations that may not be in dynamic
equilibrium, as these cannot be converted into each
other under our experimental conditions. In one con-
formational state, UDE is present in a complex with
RNA (RNA–UDE), whereas in the other state, the
protein is practically free of bound nucleic acids
(UDE). To determine the specific characteristics of
these two conformational states, we investigated ther-
mal stability, secondary structure, the presence of
exposed hydrophobic surface patches, and the binding
ability of RNA and DNA oligonucleotides with regard
to both RNA–UDE and UDE.
RNA binding significantly affects thermal
unfolding of UDE
We followed the thermal unfolding of gel-filtrated
UDE and RNA–UDE fractions by tryptophan fluores-
cence (Fig. 3A) and by CD spectroscopy (Fig. 3B). In
the case of tryptophan fluorescence, we monitored the
specific local milieu of the four tryptophans within the
sequence (Trp10, Trp107, Trp259 and Trp299 within
the nonconserved N-terminal region, motifs 1A, 3, and
4, respectively [7,8]), and during the CD measure-
ments, we followed the global change in the ratio of
detected secondary structure elements. Despite this
basic dissimilarity between the two methods, the deter-

mined T
m
values were different by only about 2 °C,
indicating that conformational changes during unfold-
ing can be faithfully monitored by both techniques.
We found that RNA–UDE had a considerably lower
melting temperature than UDE. Interestingly, this
lower thermal stability was coupled to much higher
cooperativity during unfolding, as shown by the steep
slope of the RNA–UDE melting transition (Fig. 3 and
Table 1). These results indicated that, in the absence of
Fig. 2. UDE does not have RNase-like activity. Both RNA–UDE and
UDE were tested for RNase-like activity, using 471-bp dsRNA. Incu-
bation times are indicated at the top, and marker positions on the
right side. The observed shift may be caused by the protein–RNA
complex. Note the absence of significant time-dependent degrada-
tion.
RNA-dependent conformational states of UDE A. Bekesi et al.
298 FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS
the nucleic acid ligand, UDE may lose some of its sec-
ondary and tertiary interactions, whereas in the con-
formational state characteristic of RNA–UDE, the
protein exists in a more ordered and potentially meta-
stable state. The lower melting temperature of RNA–
UDE may reflect dissociation of the RNA followed
immediately by the consequent unfolding of the UDE
protein devoid of RNA. Hence, the lower melting
temperature of RNA–UDE might indicate basic
differences between the UDE and RNA–UDE confor-
mations. Importantly, unfolding of RNA–UDE occurs

according to a simple two-state model, indicating an
intimate interaction between RNA and the protein.
Elevated cooperativity in the RNA–UDE state can be
explained by assuming that maintenance of the 3D
protein structure strongly depends on RNA binding.
Such interactions may originate from translational-
coupled folding of UDE.
For further confirmation of distinct conformations
within the two states, RNA isolated from RNA–UDE
(using Trizol; see Experimental procedures) was added
to UDE. In this sample, the melting curve shows the
same phenomenon as in the case of UDE alone
(Fig. 3A). The fact that the conformation specific to
RNA–UDE could not be restored simply by RNA
readdition under the experimental conditions also
suggests an intrinsic relationship between the RNA
binding and protein folding in RNA–UDE.
Interestingly, treatment of RNA–UDE with RNa-
se A did not result in significant changes of the melting
curve, indicating that the characteristics of the confor-
mational state initially detected on the RNA–UDE
sample are preserved, at least for the duration of the
experiments (Fig. 3B). However, the melting curve of
further purified RNase-treated RNA–UDE was charac-
teristic for UDE (Fig. 3B), indicating that a transition
from the RNA–UDE to the UDE conformational state
is possible, even though it is not an immediate process.
On the basis of these results, we propose a scheme of
possible transitions and relations between the two
conformational states of UDE protein (Fig. 4).

RNA–UDE shows an increased amount of helical
content as compared with UDE
To characterize the conformational differences between
RNA–UDE and UDE suggested by thermal stability
data, several potentially relevant spectroscopic meth-
ods were applied. Tryptophan fluorescence showed
practically the same emission maximum wavelength
(data not shown); however, CD spectroscopy revealed
significant differences. After accurate determination of
protein concentrations from UV spectra as well as the
Bradford assay (which were in agreement within 7%
standard error) and densitometry of SDS ⁄ PAGE
bands, CD spectra were measured in the 190–270-nm
far-UV range (Fig. 5A). Although spectra of both
RNA–UDE and UDE showed a-helical characteristics,
a clear difference was also detected between the two
spectra, reflected by: (a) different signal intensities; and
(b) different shapes, e.g. a slight red shift in the
position of the 208-nm peak (Fig. S1). Quantitative
evaluation of the CD data is shown in the bar graph
A
B
Fig. 3. Effect of RNA binding on the thermostability of UDE.
(A) Thermal denaturation followed by tryptophan fluorescence.
Normalized and corrected curves are shown for RNA–UDE (full
black circles), UDE (full grey squares), UDE + RNA mixture (open
triangles) and RNase-treated RNA–UDE after repurification by size
exclusion chromatography (open squares). (B) Thermal denaturation
followed by CD spectroscopy. Curves are shown for RNA–UDE (full
black circles), UDE (full grey squares), in situ treatment of RNA–

UDE with RNase A (open circles), and RNase-treated RNA–UDE
after repurification by size exclusion chromatography (open
squares). The lines show the results of sigmoidal fitting. Insert: CD
spectra of RNA–UDE (black) and UDE (grey) in the native (20 °C,
full symbols) and denatured (70 °C, open symbols) states; lines
show fitted spectra calculated by
CDSSTR software [9].
A. Bekesi et al. RNA-dependent conformational states of UDE
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 299
of Fig. 5A, where the secondary structural elements
termed Helix1, Helix2, Strand1 and Strand2 corre-
spond to the two subcategories of helices and sheets
(regular and distorted fractions) as defined in [9].
RNA–UDE showed approximately 20% higher a-heli-
cal content than UDE. As accurate measurement of
protein concentration, especially when the protein is in
complex with nucleic acids, is not trivial, the suggested
conformational changes between RNA–UDE and
UDE are much strengthened in the case of significantly
different concentration-independent intensive parame-
ters. Such differences are shown in Fig. S1 (shape of
the far-UV CD spectra) and in Fig. 3 (T
m
and cooper-
ativity of thermal unfolding).
To test the model represented in Fig. 4, the respec-
tive samples were produced and characterized by CD.
A UDE+RNA mixture failed to restore the CD signal
associated with RNA–UDE (Fig. 5A, top left). RNA
alone did not result in significant CD spectra, as

shown in Fig. S1. In situ RNase treatment of RNA–
UDE resulted in an intermediate spectrum (Fig. 5A,
top right), whereas after gel filtration, this sample pro-
vided practically the same spectrum as observed for
UDE (Fig. 5A, bottom left). The same findings were
obtained by quantitative evaluation (Fig. 5A, bar
graph, bottom right). These results are in agreement
with the thermal unfolding studies (Fig. 3) and rein-
force the model of the two conformational states.
The presence of hydrophobic surface cavities is
significantly increased in RNA–UDE as compared
with UDE
To determine whether the above described differences
in the conformational states of UDE are also reflected
on the protein surface, we evaluated the interactions of
UDE and RNA–UDE with the environmentally sensi-
tive protein dyes 8-anilinonaphthalene-1-sulfonate
(ANS) and Sypro Orange. ANS is known to bind to
NH
3
+
moieties of the proteins, and exhibits elevated
fluorescence if this binding occurs in a hydrophobic
microenvironment [10]. Sypro Orange is known as a
protein gel stain (it does not bind to either nucleic
acids or lipids [11]), but it can be used as an alternative
to ANS in the analysis of hydrophobic protein surfaces
Table 1. Melting temperatures and values of cooperativity characteristics for thermal transitions of different conformational states of UDE
protein. The value for cooperativity come from the dT parameter of the sigmoidal fit of the melting curve, which negatively correlates with
the cooperativity of the thermal transition. Error values were derived from the average of several independent measurements from different

UDE protein preparations. ND, non determined.
RNA–UDE UDE
UDE + isolated
RNA
In situ RNase-treated
RNA–UDE
RNase-treated
RNA–UDE, purified
Melting temperature (°C)
CD 50.2 ± 1.5 55.2 ± 0.6 ND 48.3 ± 0.5 55.3 ± 1.3
Fluorimetry 51.5 ± 0.6 56.6 ± 0.5 57.05 ± 0.13 ND 55.5 ± 0.3
Relative cooperativity (1 ⁄ dT )
CD 0.58 ± 0.03 0.21 ± 0.03 ND 0.45 ± 0.06 0.22 ± 0.04
Fluorimetry 0.60 ± 0.05 0.34 ± 0.02 0.27 ± 0.01 ND 0.36 ± 0.01
Fig. 4. Scheme of possible transitions between the two conforma-
tional states. The conformational state of RNA–UDE (black moon-like
shape, top left) is not destroyed immediately upon RNase treatment
(bottom left). After removal of RNA fragments (grey curves, bottom
left), the protein conformation changes into one characteristic for
UDE (grey circular segment, bottom right). UDE can bind to RNA
(grey curve, top right), but cannot be transformed into the specific
RNA UDE complex present in the RNA-UDE state.
RNA-dependent conformational states of UDE A. Bekesi et al.
300 FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS
[12]. Addition of either ANS or Sypro Orange to
RNA–UDE led to a drastic increase in the fluorescence
emission of the dyes, whereas a much smaller incre-
ment was induced upon mixing of the dyes with UDE
(Fig. 5B). A clear difference was also observed in the
positions of emission maxima characteristic for

A
B
Fig. 5. Conformational differences between RNA-UDE and UDE. (A) Secondary structure elements characteristic for RNA-UDE and UDE calcu-
lated from far-UV CD spectra. Top left: spectra for RNA-UDE (full black circles), UDE (full grey squares), and UDE + RNA mixture (open trian-
gles). Top right: spectra for RNA-UDE (as above), UDE (as above), and RNase-treated RNA-UDE (open black circles). Bottom left: spectra for
RNA-UDE (as above), UDE (as above), and RNase-treated and repurified RNA-UDE (open grey circles). Lines on the spectra indicate fitted
curves calculated by
CDSSTR software at the DICHROWEB server. The ratio of secondary structure elements, indicated on the bar graph (bottom
right), was calculated from several independent spectra in each case. The terms Helix1, Helix2, Strand1 and Strand2 correspond to the two
subcategories of helices and sheets (regular and distorted fractions) as defined in [9,71]. On the bar graph, RNA-UDE (black), UDE (grey),
UDE + RNA mixture (hatched grey), RNase-treated RNA-UDE (hatched black) and RNase-treated and repurified RNA-UDE (light grey) are
shown. (B) Altered hydrophobic surface patches in RNA-UDE and UDE. Left panel: spectra of 1.5 l
M RNA-UDE (black curves) and 1.5 lM UDE
(grey curves) mixed with 200 l
M ANS. Spectral maxima were at 471 ± 2 and 482 ± 1 nm for RNA-UDE and UDE, respectively. Right panel:
ANS titration. RNA-UDE (full black circles), UDE (full grey squares), RNase-treated (open circles), and UDE + RNA mixture (open triangles).
Maximal fluorescent signal intensities of the individual spectra are shown for each titration point. Lines show hyperbolic fitting of the data.
A. Bekesi et al. RNA-dependent conformational states of UDE
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 301
RNA–UDE and UDE complexed with both dyes [for
ANS, 471 and 482 nm, respectively (Fig. 5B); and
for Sypro Orange, 567 and 583 nm, respectively
(Fig. S2A)]. These results indicated that the number of
hydrophobic surface cavities may be much increased in
RNA–UDE as compared with UDE. We can exclude
the possibility that putative dimerization of UDE may
hide the hydrophobic surface exposed in the RNA–
UDE state, as full-length UDE was shown to be in a
monomeric state in solution [8].
For a detailed analysis, titration experiments were

performed with both dyes (Figs 5B and S1B). The
maximum emission values were plotted against dye
concentration, and the results were fitted with hyper-
bola (Figs 5B and S1B; Table 2).
The apparent dissociation constant for ANS and
RNA–UDE was three-fold smaller than that for UDE,
and the emission maximum was red-shifted from 468
to 478 nm, whereas, in the case of UDE, the emission
maximum was not shifted, also indicating altered ANS
binding fashion of the two conformational states of
UDE protein. Upon treatment of RNA–UDE with
RNase, only a slight decrease was detected in the fluo-
rescent signals (Fig. 5B), indicating that the conforma-
tional state characteristic for RNA–UDE was not
completely disrupted. When isolated RNA, in an
amount equivalent to that in RNA–UDE, was added
to UDE, we observed a significant increase in fluores-
cence; however, the extent of the signal was about half
that seen with the RNA–UDE sample, and the appar-
ent dissociation constant was more than two-fold
higher (Fig. 5B and Table 2).
Although titration with Sypro Orange resulted in
better-quality signals, evaluation of these curves can
only provide relative data, because the dye concentra-
tion data is not given. Sypro Orange complexed with
RNA–UDE showed a similar apparent dissociation
constant, but a much higher fluorescent signal inten-
sity, more than 20-fold higher than that with the Sypro
Orange–UDE complex (Fig. S2B; Table 2). The emis-
sion maximum was red-shifted from 567 to 587 nm,

whereas, in the case of UDE, the emission maximum
was not shifted. These results also support the altered
surface hydrophobicity in the two conformational
states.
Upon treatment of RNA–UDE with RNase, very
similar large fluorescent signals were detected, indicat-
ing that the conformational state characteristic for
RNA–UDE was not disrupted in the RNase-treated
mixture (Fig. S2B). This observation is in agreement
with the results of ANS titration measurements
(Fig. 5B), the melting experiments (Fig. 3), and CD
spectra (Fig. 5A).
When isolated RNA was added to UDE, we
observed a significant increase in fluorescence; how-
ever, the extent of the signal was much smaller than
that with the RNA–UDE sample, and the apparent
dissociation constant was seven-fold higher (Table 2;
Fig. S2B). This finding also suggests that the confor-
mational state of RNA–UDE, which showed a high
fluorescence intensity in this assay, could not be recon-
structed by simple addition of the isolated RNA com-
ponent to UDE, potentially indicating that the
conformation of RNA–UDE may originate in de novo
protein folding during translation.
In summary, the results obtained in complexation
experiments with ANS and Sypro Orange are in excel-
lent agreement with the hypothesis of the two confor-
mational states presented in Fig. 4, and are also in line
with the findings of the thermal unfolding and CD
studies.

RNA binding causes similar protection against
limited proteolysis as DNA binding
Previously, we performed limited proteolysis of UDE
and its complex with DNA, using AspN endoprotein-
ase and high-specificity chymotrypsin, and revealed the
contribution of N-terminal motifs 1A and 1B to DNA
binding as well as the relative compactness of the
C-terminal segment containing motifs 2, 3, and 4 [8].
Here, we aimed at characterization of proteolytic frag-
ments independently in the RNA–UDE and UDE
fractions, using AspN and ArgC endoproteinases
(Fig. 6). In UDE, several preferred proteolytic cleavage
sites were identified by MS (summarized in Fig. 6C),
whereas in RNA–UDE, protection against both pro-
teinases was evident on motifs 1A and 1B. The rela-
tive compactness of the C-terminal part containing
motifs 2, 3 and 4 was confirmed in both RNA–UDE
Table 2. Hydrophobic surface titration of different conformational
states of UDE protein, using ANS and Sypro Orange dyes. Parame-
ters were obtained by hyperbolic fitting of the data. F
max
values indi-
cate the maximal fold increase in fluorescence intensity. K
app
values
are apparent dissociation constants for dye–protein complexes.
Protein
Parameters
ANS Sypro Orange
F

max
(· 1000) K
app
F
max
(· 1000) K
app
UDE 18.7 ± 0.3 208 ± 5 3.8 ± 0.2 34 ± 7
RNase-treated
RNA–UDE
31.5 ± 1.0 62 ± 10 71.6 ± 3.3 44 ± 5
RNA–UDE 42.5 ± 0.5 77 ± 2 68.3 ± 4.2 47 ± 7
UDE + RNA 27.6 ± 0.9 161 ± 9 46.6 ± 7.2 300 ± 75
RNA-dependent conformational states of UDE A. Bekesi et al.
302 FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS
and UDE, as preferred cleavage sites could not be
identified in this region. To characterize the relation-
ship between the RNA-binding and DNA-binding
sites, DNA was added to both RNA–UDE and UDE.
The proteolysis patterns were rather similar in both
RNA-containing and DNA-containing complexes, with
one remarkable difference in AspN digestion. Here,
RNA binding produced significant protection at the
Asp333 site within the nonconserved C-terminus,
whereas DNA binding did not induce any protection
(Fig. 6B). These results suggest partially, but not fully,
overlapping sites for RNA and DNA binding in UDE
protein. Furthermore, addition of DNA in three-fold
molar excess to RNA–UDE results in exactly the same
patterns as those characteristic for the UDE–DNA

complex (Fig. 6A,B), suggesting that DNA is able to
replace RNA in RNA–UDE, in agreement with the
overlap of the respective binding sites. These findings
can be interpreted within the previously introduced
hypothesis that RNA–UDE and UDE represent two
distinct conformational states, although they do not
offer independent support in this respect.
Characterization of the nucleic acid-binding
ability of UDE
In order to provide a quantitative description of UDE
DNA-binding and RNA-binding properties, we
applied EMSA. In the first set of experiments, we
assessed the interaction of UDE and RNA–UDE with
DNA oligonucleotides (Fig. 7). When a double-
stranded 30mer oligonucleotide (see Experimental pro-
cedures) was titrated with UDE, we detected two
shifted bands (termed complex 1 and complex 2,
respectively) on 6% native TBE ⁄ PAGE (Fig. 7A).
Densitometry of the bands corresponding to the free
and complexed oligonucleotides resulted in a sigmoidal
decrease for the amount of free oligonucleotide, and a
A
B
C
Fig. 6. Comparison of RNA-binding and DNA-binding surfaces by limited proteolysis. (A, B) Limited proteolysis using ArgC and AspN endo-
proteinases. Molecular mass marker positions in kDa are indicated on the left. U-pl, uracil-substituted plasmid DNA mixed with the protein
samples prior to digestion. The first lane of each gel represents untreated samples; further lanes correspond to 10-min, 1-h, 3-h and 5-h time
points. The black cross and the black star indicate RNA protection against first cleavage at the C-terminus by ArgC and AspN, respectively.
Note that the DNA binding does not protect these sites. (C) Scheme of preferred cleavage sites. Cleavage sites were identified by MS. The
scheme was designed to be strictly proportional to the real sizes. His: His-tag. 1A, 1B, 2, 3, and 4: conserved motifs of UDE. NLS: nuclear

localization signal at the end of the sequence [17]. Black and grey arrows indicate the preferred cleavage sites of ArgC and AspN, respec-
tively. Solid arrows: preferred sites even in the presence of nucleic acids. Dotted arrows: sites protected by both RNA and DNA binding.
Dashed arrows: sites protected by RNA binding, but not by DNA binding. Numbering of residues follows the sequence of the physiological
form derived from the NCBI database. R* indicates the arginine present only in the His-tagged protein. In the absence of straightforward
identification of close cleavage sites, arrows list multiple potential sites.
A. Bekesi et al. RNA-dependent conformational states of UDE
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 303
sigmoidal increase for complex 2, whereas complex 1
was only transiently present, in a manner characteristic
for intermediate states (Fig. 7B). Data were fitted with
the Hill equation (Fig. 7C; see Experimental proce-
dures), providing an apparent K
d
value of
50±13lm, and n = 4.9 ± 0.2. The value of the
apparent Hill coefficient may indicate either complexes
with higher stoichiometry or strong cooperativity [13]
(see Experimental procedures). Figure 7A shows only
two bands corresponding to complex forms; however,
the upper band, located in the well without significant
migration, may contain additional complex forms pro-
posed by the apparent Hill coefficient. To further
address the stoichiometry issue, a similar EMSA was
A
B
C
D
E
Fig. 7. DNA-binding ability of UDE and RNA–UDE characterized by EMSA. (A) UDE causes a significant shift in the electrophoretic mobility
of a 30mer oligonucleotide. One-micromolar 30mer double-stranded oligonucleotide with a single uracil at the mid-position was titrated with

increasing amounts of UDE, up to 5 l
M, in native 6% TBE ⁄ PAGE. Arrows on the left show the positions of free oligonucleotide and two dis-
tinct complex forms. (B) Densitometry of the bands corresponding to the three detected species. Band densities (free oligonucleotide, full
black squares; complex 1, grey circles; complex 2, open squares) were normalized and plotted against UDE concentration. (C) Evaluation of
binding. The total relative amount of bound oligonucleotide (black squares) was calculated from the Eqn (1) – [free], where [free] is the rela-
tive amount of free oligonucleotide, and plotted against UDE concentration. The line shows the curve fitted by the Hill equation. (D) EMSA
on agarose gel revealed the presence of several complex forms with higher stoichiometry. Two-micromolar 30mer (left panel) and 60mer
(right panel) oligonucleotides with a single uracil at the mid-position were compared on 1.5% agarose gel, with UDE concentrations up to
32 l
M. Symbols (rhomboids) at both sides indicate positions of complexes; five and seven distinct positions were observed for 30mer and
60mer oligonucleotides, respectively. (E) DNA-binding ability of UDE versus RNA–UDE. One-micromolar Cy3-labelled single-stranded uracil-
containing oligonucleotide was titrated with UDE (top panel) and RNA–UDE (bottom panel) with protein concentrations up to 5.5 l
M. Note
that, in the case of RNA–UDE, saturation was not observed.
RNA-dependent conformational states of UDE A. Bekesi et al.
304 FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS
performed in 1.5% agarose gel, allowing detection of
larger complexes separately (Fig. 7D). Indeed, we
observed five distinct bands in shifted positions. To
address whether the stoichiometry of UDE binding is
dependent on the length of the oligonucleotide, a
60mer oligonucleotide was also titrated in the same
manner. In this case, several additional bands were
detected, migrating even in the opposite direction,
owing to the putative excess positive charge provided
by the higher proportion of bound UDE (the calcu-
lated pI of the protein is 9.7). On the basis of these
results, we propose that UDE binds to DNA in a non-
sequence-specific manner in our experimental condi-
tions.

To compare nucleic acid binding of UDE and
RNA–UDE, a 30mer oligonucleotide, 5¢-labelled with
Cy3 fluorescent dye and containing a single uracil in
the mid-position (see Experimental procedures), was
used to allow specific detection of the DNA oligonu-
cleotide separately from the RNA content of RNA–
UDE (Fig. 7E). An apparently lower affinity for the
DNA oligonucleotide was observed for RNA–UDE
than for UDE. This result may indicate competition
between the UDE-bound RNA and the added DNA,
in agreement with the overlapping binding sites of the
two types of nucleic acid (see Fig. 6, limited proteoly-
sis). In conclusion, the data of Figs 6, 7E and 8 (see
below) directly suggest that uracil-DNA can displace
RNA from RNA–UDE.
As we observed that UDE is capable of binding
both DNA and RNA, it was of interest to compare
the relative affinities for these two different nucleic
acids. To this end, we applied single-stranded DNA
and RNA oligonucleotides with the same sequences
(see Experimental procedures). Figure 8 shows that
UDE has apparently comparable affinity for both oli-
gonucleotides. Although the fitted parameters show
some alterations (RNA binding is characterized by a
smaller n and, correspondingly, a smaller apparent K
d
),
the UDE concentration values at which 50% of the
oligonucleotides are bound are practically the same.
The possible difference between the stoichiometries of

the two complexes was investigated on 1.5% agarose
gel, and no significant alteration was observed (data
not shown).
Identification of RNA species copurifying with
the protein
Using several independent approaches, we have shown
major differences between the RNA–UDE and UDE
conformational states (Figs 3, 5, and S1). Under our
experimental conditions, the distinct conformations
characteristic for RNA–UDE and UDE may not be in
dynamic equilibrium, as UDE could not be converted
into RNA–UDE by simple addition of the same RNA
(Fig. 4). The results indicate that the specific confor-
mation of RNA–UDE is intimately linked to RNA
binding, suggesting putative RNA-assisted de novo
protein folding during translation.
To characterize the RNA species copurified with the
protein, the presence of three specific RNA molecules
was checked by reverse transcription followed by real-
time PCR. We targeted the mRNA of UDE, which is
abundant in this recombinant overexpression system;
domain V of the 23S rRNA, which is known to be
involved in assisting de novo protein folding during
translation [14–16]; and the mRNA of the abundant
housekeeping glyceraldehyde-3-phosphate dehydroge-
nase (GAPDH) protein (this latter was used as a con-
trol). The results shown in Fig. 9 confirmed the
presence of UDE mRNA as well as 23S rRNA in the
RNA–UDE species, whereas mRNA of GAPDH was
not present in a significant amount.

Fig. 8. Binding affinity of UDE for RNA and DNA oligonucleotides.
Two-micromolar 30mer RNA (left panel) and ssDNA (right panel) oli-
gonucleotides with the same sequence were titrated with UDE up
to 10 l
M in native 6% TBE ⁄ PAGE. The density of the bands corre-
sponding to the free oligonucleotides (lower position on the gel;
see arrow) was used to calculate the amounts of bound forms
(RNA, open squares; ssDNA, full squares). The line shows curves
fitted by the Hill equation. Apparent K
d
values were 11 ± 2 and
48 ± 6 l
M; stoichiometries were 2.8 ± 0.2 and 4.2 ± 0.2 for RNA
and ssDNA (ssT) oligonucleotides, respectively.
A. Bekesi et al. RNA-dependent conformational states of UDE
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 305
For analysis of the composition of the bound RNA
pool with an independent method, the isolated RNA
was converted into double-stranded cDNA, using both
random hexamer and oligo-dT primers in independent
experiments, randomly cloned, and 22 and 11 positive
colonies, respectively, were sequenced. Both UDE
mRNA and 23S rRNA were identified, confirming the
results of the real-time PCR experiments. Furthermore,
16S rRNA was also detected, whereas other RNA spe-
cies were not present at all (Table 3). The results con-
firmed that UDE can bind to RNA apparently
without sequence specificity, and select those RNA
species that are in close proximity to the site of trans-
lation and the coupled folding processes, further

strengthening our hypothesis that RNA–UDE may
undergo RNA-mediated de novo protein folding.
Real-time PCR experiments were performed on flag-
tagged recombinant UDE expressed in Sf9 insect cells.
Here, lack of genomic data for Spodoptera frugiperda
prevented straightforward PCR reactions for S. fru-
giperda mRNAs, and another experimental setup was
therefore employed. Namely, we performed real-time
PCR with UDE mRNA-specific primers in immuno-
precipitated samples, as compared with samples precip-
itated with preimmune serum, and found significant
enrichment of UDE mRNA in the immunoprecipitated
samples (Fig. S3).
Physiological relevance of UDE RNA binding
On the basis of several independent measurements, we
can conclude that recombinant UDE expressed in Esc-
herichia coli is associated with RNA that probably
contributes to protein folding. To determine whether
RNA binding of UDE occurs under physiological con-
ditions in the fruit fly, we investigated the presence of
UDE in nucleic acid-binding complexes of D. melanog-
aster pupal extracts. Total extract was fractionated on
a size exclusion column, and the presence of UDE was
monitored by western blotting in each fraction.
Fig. 10 also shows elution profiles of pupal extracts
after DNase or RNase A treatment. It is apparent that
both in intact and in DNase-treated pupal extracts,
most of the endogenous UDE appears within the
exclusion volume, at a very similar elution position to
that of recombinant UDE complexed with RNA

(Fig. 1). Treatment of pupal extract with RNase A,
however, dramatically shifted the elution profile of
UDE towards the RNA-free UDE elution position
(approximately 52 kDa; Fig. 1). This observation indi-
cates that a considerable fraction of endogenous UDE
within pupal extracts is associated with RNA (or
RNA-containing larger macromolecular complexes)
that can be partially dissociated by RNase A treatment
and gel filtration.
Discussion
The target of the present studies, UDE, is the first rep-
resentative of a novel enzyme family that requires
detailed structural and functional characterization.
With regard to structural aspects, domain analysis,
characterization of the secondary structure by CD
spectroscopy and structural modelling have already
been performed [8]. The present work was motivated
by efforts to understand the structural behaviour of
UDE alone and in complex with RNA, and to prepare
Fig. 9. Determination of the RNA content of RNA–UDE by real-
time RT-PCR. The RNA content of RNA–UDE was isolated and
reverse-transcribed, using a random hexamer primer. Real-time
PCR amplification curves are shown, obtained with primers specific
for 200-nucleotide segments of either domain V of 23S rRNA (grey
line), mRNA of UDE (black line), or mRNA of GAPDH (light grey
line). C
t
values detected at a threshold of 2000 are 20.4, 21.5 and
32.2 for the three targets, respectively, indicating that rRNA may
be present in more than two-fold excess over mRNA of UDE, and

that the presence of mRNA of GAPDH is negligible (< 0.5% as
compared with rRNA).
Table 3. Characterization of the bound RNA pool by cDNA
sequencing: the results of two independent experiments using two
alternative cloning strategies, as indicated. Data show the number
of colonies from which the indicated RNA species were identified.
Oligo-dT PCR
amplification
(Smarter)
Random hexamer without
PCR amplification
(SuperScript)
Number of
sequenced colonies
22 11
UDE mRNA 6 (27%) 0
23S rRNA 7 (32%) 8 (73%)
16S rRNA 9 (41%) 3 (27%)
RNA-dependent conformational states of UDE A. Bekesi et al.
306 FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS
conformationally homogeneous protein samples for
further crystallization experiments.
Detailed investigations of the nucleic acid copurified
with recombinant UDE protein revealed, unexpectedly,
that the major component contains RNA, and not
DNA. RNA binding to UDE may have different phys-
iological roles: (a) RNA may stabilize the conforma-
tion of UDE protein; (b) RNA may inhibit UDE
activity; (c) RNA may influence the cellular localiza-
tion of the protein; and (d) UDE may have an as yet

unidentified role in either stabilization, localization or
regulation of RNAs. Nuclear localization of UDE was
suggested by both nuclear localization signal (NLS)
analysis of green fluorescent protein fusion constructs
[17], and immunostaining of S2 cells (unpublished).
However, detailed characterization of UDE localiza-
tion in tissues awaits further studies. Here, we
addressed the possible impact of RNA binding on pro-
tein structure and substrate-binding ability.
Proteins that bind and process nucleic acids are
often capable of binding to both RNA and single-
stranded or double-stranded DNA, producing altered
biological functions, depending on the type of ligand.
Several proteins are known to have a function that is
primarily related to DNA; however, their RNA-
binding complexes may also have physiological rele-
vance. Activation-induced deaminase can deaminate
cytosines only in ssDNA, and not in RNA [18]; how-
ever, activation-induced deaminase expressed in Sf9
cells is bound to RNA, and elimination of this RNA is
necessary for its ssDNA-editing activity [19]. Several
transcription factor families with specific recognition
sequence elements (e.g. WT1 [20], TFIIIA [21], H-NS
[22], and Y-box binding proteins [23,24]) are also
involved in post-transcriptional regulation, owing to
their alternative RNA-binding ability. Human AP
endonuclease I, also referred to as Ref1 transcription
factor, also binds to and cleaves RNA [25]. Impor-
tantly, the physiological functions of several proteins
directly involve both DNA-binding and RNA-binding

ability (telomerase, RNA-dependent DNA polymerases,
and DNA-dependent RNA polymerases).
Furthermore, several metabolic enzymes and ribo-
somal proteins bind to their own mRNA and regulate
mRNA turnover. These include thymidylate syn-
thase [26–29], dihydrofolate reductase [30,31], the
A
B
C
Fig. 10. The physiological form of UDE shows significant RNA-
dependent complexation in pupal extract. Chromatograms were
obtained by size exclusion chromatography of pupal extracts
[untreated (A), DNase-treated (B), and RNase-treated (C)]. Black
and grey lines show absorbance values at 280 and 260 nm, respec-
tively, plotted against the elution volume. Inserts: western blot for
UDE present in the corresponding fractions. Arrows indicate the
expected elution positions of recombinant UDE and RNA–UDE cor-
responding to 52-kDa and larger than 660-kDa apparent molecular
masses, respectively. Note the significant shift in elution of UDE
towards the complex size for untreated and DNase-treated samples
versus RNA-treated samples, indicating RNA-dependent complexa-
tion of the protein in physiological conditions.
A. Bekesi et al. RNA-dependent conformational states of UDE
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 307
carboxyltransferase component of E. coli acetyl-CoA
carboxylase [32], amino acyl-tRNA synthetases (the
yeast aspartyl-tRNA synthase [33], threonyl-tRNA
synthetase [34], and lysyl-tRNA synthetase [35]), ribo-
somal proteins (rpL20 [36], rpS19 [37], rpS20 [38], and
rps15 [39]), RNA-modifying proteins (SGM methylase

[40] and ermC 23S rRNA methyltransferase [41]), and
histone H3 [42].
In the present study, we showed that UDE binding to
RNA occurs under physiological conditions, and can be
demonstrated in both the heterologous overexpression
system in E. coli and in fruit fly cellular extracts
(Fig. 10). In the E. coli recombinant system, the bound
RNA pool was shown to consist of 16S rRNA, 23S
rRNA, and UDE mRNA (Fig. 9; Table 3), potentially
implicating either an additional RNA-related function
or RNA-binding-coupled folding of UDE protein. In-
depth analysis revealed significant conformational dif-
ferences between RNA–UDE and UDE, the two distinct
fractions separable by independent chromatography
methods. Furthermore, we demonstrated that the spe-
cific conformation characteristic for RNA–UDE cannot
be generated simply by addition of the RNA to UDE
(Figs 3, 5, and S1), which also argued for the possibility
of RNA-assisted cotranslational folding.
Ligand binding often stabilizes the protein structures
by enhancing the folding of flexible or disordered
regions, especially in the case of nucleic acid-binding
proteins (e.g. bateriophage lambda integrase [43],
GCN4-bZIP transcription factor [44,45], and RNase P
[46]). Here, we found that RNA binding alters the
protein conformation, significantly enhancing its
a-helical content (Fig. 5), strengthens the cooperativity
of protein thermal unfolding, and lowers the melting
temperature of RNA–UDE as compared with UDE
(Fig. 3; Table 1), indicating essential differences

between the two conformational states. UDE is still
able to bind RNA, but in an altered manner as com-
pared with the RNA–UDE species. Similar conforma-
tionally heterogeneous RNA binding occurs within
ribonuclease III dsRNA complexes, where both two
major states have physiological roles [47].
As the conformation associated with RNA–UDE as
purified from cellular extracts cannot be formed by
reconstituting the UDE and RNA mixture under our
experimental conditions, but, upon RNase treatment,
RNA–UDE is able to slowly transform into the con-
formation characteristic for UDE, we suggest that the
origin of RNA–UDE complex formation is within the
translation-coupled folding mechanism, which may
involve the presence of both rRNAs and UDE
mRNA, and potentially the assistance of domain V of
23S rRNA [14,16].
De novo protein folding occurs mostly upon transla-
tion, and is intimately coupled to the translation
machinery [48,49]. Secondary structure formation is
already initialized in the tunnel of the large ribosomal
subunit, which may accommodate about 30 residues
of the nascent polypeptide chain, whereas tertiary
structure formation might occur on the surface of
the ribosome as the polypeptide chain reaches the exit
site [48–50]. Within the ribosomal tunnel, 23S rRNA
interacts with the nascent polypeptide chain, helping
secondary structure formation. Moreover, direct inter-
action of rRNA with proteins, especially with nucleic
acid-binding proteins, also has an effect on protein

refolding in vitro [51,52]. On the ribosomal surface,
large-scale folding mechanisms potentially involve the
assistance of chaperone proteins; for example, elonga-
tion factor-a is known to contribute to such chaperone
function [53,54].
The mRNA sequence also has an impact on de novo
protein folding, possibly because of the kinetic conse-
quences of rare codon usage [55], as well as of second-
ary structure elements present in the mRNA [56],
which might explain why silent mutations may influ-
ence protein folding [57–59] manifested as either func-
tional alterations [60–63] or subcellular localization
[64]. Other observations indicate that secondary struc-
ture elements are differently coded within the mRNA
sequences [65–67], and rare codon patterns of mRNAs
show correlations with the borders of secondary or ter-
tiary structural elements within the protein [50,68,69],
supporting the potential significance of mRNA in pro-
tein folding.
On the basis of the above detailed results and data
from the literature, we conclude that the RNA–UDE
complex may be formed via a cofolding mechanism.
The present results shed light on novel aspects of UDE
structure, providing important insights into its RNA-
dependent native fold; the protein is also identified as
a model for future studies of RNA-assisted protein
folding. Furthermore, these results indicate that UDE
may possess functional roles related to its RNA bind-
ing in addition to its uracil-DNA-degrading activity.
Experimental procedures

Materials
Oligonucleotides purchased from Eurofins MWG Operon
(Germany) are as follows: Cy3-labelled 30mer (Cy3, 5¢-AC
GAATTCGTAAUATGCCTACACT GGAGCA-3¢), Cy3-labe-
lled 60mer (Cy3, 5¢-CTCGCAAATGAACTGGGCGA
TGCGGTCGCACUACTTCACCTCGAAATCAACATCT
GAGTG-3¢), primers specific for domain V of 23S rRNA
RNA-dependent conformational states of UDE A. Bekesi et al.
308 FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS
(forward, 5¢-GCGGTGTTTGGCACCTCGAT-3¢; reverse,
5¢-AGTGATGCGTCCACTCCGGT-3¢), primers specific
for mRNA of UDE (forward, 5¢-GGAGAAGGCGTTAG
AGACGTTG-3¢; reverse, 5¢-CTCTGTCTTGGCATTCTT
GCTG-3¢), and primers specific for E. coli GAPDH (for-
ward, 5¢-CAACGCTTCCTGCACCACCA-3¢; reverse, 5 ¢-A
GCAGCACCGGTAGAGGACG-3¢). Unlabelled DNA
and RNA oligonucleotides for EMSA were obtained from
Biological Research Centre, Szeged, Hungary: ssDNA (5¢-A
CTGCCATGCCAXCGCCCTAACGATCCGCA-3¢;X=
T or U); reverse complementer (5¢-TGCGGATCGTTA
GGGCGATGGCATGGCAGT-3¢); and RNA with the
same sequence (5¢-ACUGCCAUGCCAUCGCCCUA
ACGAUCCGCA-3¢).
Protein expression and purification
The recombinant His-tagged UDE protein was produced
using the BL21(DE3)ung-151pLysS E. coli expression sys-
tem, as previously described [7], and was purified in the
first step on a HisTrap column (GE Healthcare, Budapest,
Hungary), and in the second step on a Superdex 200HR
size exclusion column (GE Healthcare) mounted in an

AKTA Purifier (GE Healthcare) System, with unicorn
software (GE Healtcare). Absorbance of eluates was moni-
tored at both 260 and 280 nm. Total cell lysate was pre-
pared from 500 mL of culture in 25 mL of lysis buffer
[50 mm Tris ⁄ HCl, 150 mm KCl, and 1 mm EDTA, pH 8.0,
complemented with 1 mm dithiothreitol, 0.1 mm phen-
ylmethanesulfonyl fluoride and proteinase inhibitor cock-
tail (Sigma, Budapest, Hungary) at a 200-fold dilution].
After loading, the HisTrap column was washed with buf-
fer A (50 mm Tris ⁄ HCl, 30 mm KCl, 1 mm EDTA, and
5mm imidazole, pH 7.5, complemented with 0.1 mm phen-
ylmethanesulfonyl fluoride and 1 mm dithiothreitol) until
baseline was reached, and then with 50% buffer B (buf-
fer A also containing 1 m KCl) to elute the nonspecifically
attached contaminating proteins; finally, a linear gradient
of buffer C (buffer A also containing 1 m imidazole) was
applied. UDE fractions were dialyzed against 25 mm
Tris ⁄ HCl, 150 mm KCl, and 1 mm EDTA (pH 7.5), con-
centrated on a Vivaspin centrifugal concentrator (Sarto-
rius-Membrane Ltd., Budakeszi, Hungary), and further
purified or stored on ice until use. Size exclusion chroma-
tography was performed as described previously [8], using
the dialysis buffer, except for samples used in CD mea-
surements, where 100 m m sodium phosphate (pH 7.5) was
applied.
Agarose gel electrophoresis
The nucleic acid content of UDE preparations was deter-
mined on 1% agarose gels, in standard Tris ⁄ acetate ⁄
EDTA (TAE) buffer, with application of a constant
voltage of 100 V, and final staining with ethidium

bromide. Samples of 10 lL containing 3–10 mgÆmL
)1
UDE were treated with either 0.45 mgÆmL
)1
proteinase K
(Sigma) or the same amount of DNase (Sigma) or RNa-
se A (Epicentre Biotechnologies, Madison, Wisconsin,
USA) for 15 min at room temperature, and 2 lLof
6 · loading dye (Fermentas, Biocenter Ltd., Szeged, Hun-
gary) was then added prior to loading. For EMSA, 1.5%
agarose gel was used with the same conditions.
Test for RNA-cleaving activity
RNA-cleaving activity was tested in the same manner as
previously described for uracil-DNA-cleaving activity [7],
with 471-bp dsRNA (produced by in vitro transcription
with a MEGAscript RNA interference kit on a PCR-
amplified template encoding Renilla luciferase) as potential
substrate. Samples were analysed on 1% agarose gel and
stained with ethidium bromide.
Heat-induced unfolding monitored by either
tryptophan fluorescence or CD spectroscopy
Fluorescence of protein tryptophan residues was measured
with a JOBIN Fluoromax-3 spectrofluorimeter, in 0.8 mL
cuvette (Hellma 108.004F-QS) thermostatted at 25 °C, with
excitation and emission wavelengths of 295 and 340 nm,
respectively, and excitation and emission slits set to 1 and
2 nm, respectively. Protein (3 lm) in Tris buffer (the same
as for dialysis) was heated from 25 to 70 °C at a rate of
1 °CÆmin
)1

, using a Peltier thermostat. The heat-induced
change in the far-UV CD signal of 3 lm protein in 20 mm
sodium phosphate buffer (pH 7.5) was measured with a
JASCO 720 spectropolarimeter at 208 nm in a 1-mm-path-
length 0.2-mL cuvette (165-QS). Thermal unfolding was
carried out in the range 20–70 °C at a rate of 1 °CÆmin
)1
with a Neslab RTE-100 computer-controlled thermostat.
Spectra were recorded at 20 and 70 °C.
Raw data of thermal denaturation monitored by CD and
fluorescence emission were converted to fraction of native
protein F
N
, according to Eqn (1) [70]:
F
N
¼
ðS
U
þ m
U
TÞÀS
ðS
U
þ m
U
TÞÀðS
N
þ m
N

TÞ
ð1Þ
where S is the observed spectroscopic signal at temperature
T, S
N
and S
U
are the intercepts and m
N
and m
U
the slopes
of the pre-transitional and post-transitional baselines of the
raw data, respectively. F
N
– T plots were fitted by a sig-
moid function (Eqn 2) [70]:
F
N
¼
1
1 þ exp½ðT
M
À TÞ=dT
ð2Þ
where T
m
is determined as the temperature at which 50%
of the protein is unfolded, and the reciprocal value of dT is
a relative measure of cooperativity of the transition.

A. Bekesi et al. RNA-dependent conformational states of UDE
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 309
CD spectroscopy
Far-UV CD spectra (190–270 nm) were recorded on a
JASCO 720 spectropolarimeter with a 1-nm bandwidth,
50 nmÆmin
)1
scanning speed, and 1-mm-pathlength cuvettes
(Hellma 165-QS) thermostatted at 25 °C by a Neslab RTE-
100 computer-controlled thermostat. Both UDE and RNA–
UDE at 0.06 mgÆmL
)1
were measured in 20 mm potassium
phosphate buffer (pH 7.5). Five scans of each spectrum
were averaged, baseline was substracted, and the resulting
spectra were zeroed in the 263–270-nm range and converted
into De units (mean residual weight = 112.02) by the
cdtool software package [71]. Far-UV CD spectra were
evaluated at the dichroweb server [72], using cdsstr soft-
ware [9], which provided the best fit among all four avail-
able methods appropriate for our measurements (k2d,
selcon, contin, and cdsstr) for both UDE and RNA–
UDE on all the three datasets (set 4, set 7, and SP175 [73])
available and appropriate for our assay arrangement. Fig-
ures were created with origin software. As evaluation of
CD spectra is extremely sensitive for possible errors in deter-
mination of protein concentration, extreme accuracy was
required, so consensus values of UV spectra (using the
A
0.1%

= 0.782 coefficient calculated for the His-tagged
UDE sequence by protparam at the expasy server) and the
Bradford assay (Bio-Rad, Budapest, Hungary), as well as
comparative densitometry of SDS ⁄ PAGE bands, were used.
In the case of RNA–UDE, UV spectra have to be corrected
with the contribution of RNA according to Eqn (3), assum-
ing that the 260-nm ⁄ 280-nm absorbance ratios are 0.5 and
2.0 for pure protein and pure RNA, respectively.
X ¼
A
280 nm
À 0:5A
260 nm
0:75
ð3Þ
where X is a theoretical absorbance value corresponding to
the protein content of the RNA–UDE sample, and A
280 nm
and A
260 nm
are the measured absorbances at the corre-
sponding wavelengths.
Hydrophobic surface titration
The environmentally sensitive protein dyes ANS (Fluka,
Sigma, Budapest, Hungary) and Sypro Orange (Invitrogen,
Life Technologies, Budapest, Hungary) were used to char-
acterize hydrophobic surface patches of the proteins. Fluo-
rescence emission was measured with a JOBIN
Fluoromax-3 spectrofluorimeter, in a 0.8-mL cuvette (Hell-
ma 108.004F-QS), thermostatted at 25 °C. Excitation

wavelengths were 388 and 470 nm, emission spectra were
recorded in 400–570- and 500–670-nm ranges for ANS
and Sypro Orange, respectively. Excitation ⁄ emission slits
were set to 1 nm ⁄ 2 nm for both dyes. Titration experi-
ments were performed in Tris buffer (the same as for dial-
ysis), with 1.5 lm protein. Whole spectra were recorded in
each titration point, baseline was substracted, maximum
emission values were plotted against dye concentration,
and results were fitted with hyperbola (Table 2). RNase-
treated samples were obtained by overnight incubation of
0.7 mgÆmL
)1
RNA–UDE with 0.05 mgÆmL
)1
RNase A on
ice. Isolated RNA was added to UDE at 0.01 mg ÆmL
)1
,
the same concentration as present in the corresponding
dilution of RNA–UDE. A potential inner filter effect of
isolated RNA was excluded, as RNA–UDE showed the
same high signal intensity in the presence and in the
absence of the isolated RNA. Signals attributable to RNa-
se A or isolated RNA were measured in independent
experiments, and were subtracted from the raw data of
the corresponding samples.
Limited proteolysis
Protein samples (0.25 mgÆmL
)1
) were incubated with

2.5 lgÆmL
)1
AspN or ArgC endoproteases (Sigma) in the
absence or in the presence of 0.25 mgÆmL
)1
uracil-substi-
tuted DNA plasmid (prepared in the dut– ⁄ ung– K12 CJ236
E. coli strain [7]) in 100 mm Tris ⁄ HCl buffer (pH 8.0) also
containing 5 mm MgCl
2
. Reactions were run at room tem-
perature and terminated at different time points by the
addition of 1 mm phenylmethanesulfonyl fluoride or 5 mm
EDTA for ArgC or AspN digestion, respectively. The sam-
ples were stored at )20 °C until analysis by 12%
SDS ⁄ PAGE.
MS analysis of proteolytic fragments
SDS ⁄ PAGE-separated AspN and ArgC cleavage products
were in-gel digested by trypsin (4 h, 37 °C) and chymotryp-
sin, respectively. The digests were analysed by LC-MS ⁄ MS,
with an LCQ Fleet 3D ion trap mass spectrometer (Thermo
Fischer Scientific, Bio-Science Ltd., Budapest, Hungary) in
a data-dependent fashion, in triple-play mode [8,74,75].
Data analysis was performed with a mascot in-house data-
base search engine, first against the NCBInr 20080718
(6 833 826 sequences) database and then against the recom-
binant protein sequence (CG18410). Cleavage sites were
identified either from CID data, or from the apparent
molecular masses of the cleavage products and their
sequence coverage from the LC-MS ⁄ MS analysis.

EMSA
This was performed with either 1.5% agarose gel or 6%
native TBE ⁄ PAGE (TBE – 89 mm Tris ⁄ HCl, 89 mm boric
acid, and 2 mm EDTA; in PAGE, 1 : 29 bis-acrylamide ⁄
acrylamide was used). 30mer and 60mer oligonucleotides
were applied, containing a single uracil in the mid-position.
In some cases (see figure legends), oligonucleotide was
5¢-end-labelled with Cy3 fluorescent dye, allowing selective
detection independently from other nucleic acids. UDE and
RNA–UDE in increasing concentrations from 0 to 5 lm
RNA-dependent conformational states of UDE A. Bekesi et al.
310 FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS
were mixed with 1 or 2 lm oligonucleotides in 10-lL total
volumes in 25 mm Tris ⁄ HCl, 150 mm KCl, and 1 mm
EDTA (pH 7.5) buffer. Two microlitres of 6 · loading dye
(Fermentas) was added prior to loading. Electrophoresis
was performed for about 50 min at room temperature with
constant voltages of 120 and 100 V for TBE ⁄ PAGE and
agarose gels, respectively. In the case of nonlabelled oligo-
nucleotides, gels were stained with ethidium bromide for
10 min at room temperature, and then detected with a Uvi-
Tec gel-documentation system (Cleaver Scientific Ltd.,
Rugby, UK). Cy3-labelled oligonucleotides can be detected
directly, without any staining, using the same UV lamp
constituent of the UviTec system. In each case, densitome-
try was performed with uvidoc software provided by Uvi-
Tec. The concentrations of bound nucleic acid, calculated
by substracting the concentrations of free oligonucleotides
from the total concentration, were plotted against UDE
concentration and fitted according to the Hill equation

(Eqn 4; see Fig. 7C), providing apparent K
d
values as well
as an estimation of the stoichiometry or cooperativity [13]
of the binding [76,77].
½complex¼
c
oligo
 c
UDE
n
K
app
þ c
UDE
n
ð4Þ
Reverse transcription and real-time PCR
RNA was isolated from purified RNA–UDE with Trizol
reagent, according to the supplier’s protocol (Invitrogen).
In 50 lL reaction volumes, 1 lg of RNA template was
reverse transcribed by Moloney murine leukaemia virus
reverse transcriptase enzyme (New England Biolabs, Kvali-
tex Ltd., Budapest, Hungary), with random hexamer primer
(Fermentas) in the presence of RiboLock RNase inhibitor
(Fermentas), at 37 °C overnight. After 10 min of inactiva-
tion at 65 °C, series dilutions were made from reverse tran-
scribed samples (· 1, · 10, · 100, · 1000, and · 10 000),
1 lL aliquots of which were added to the PCR mixtures.
Immomix (Bioline GmbH, Luckenwalde, Germany) master

mix and EvaGreen (Fermentas) were used for PCR amplifi-
cation of three target sequences (rRNA, and mRNAs of
UDE and GAPDH), with a Stratagene Mx3000P real-time
PCR Thermal Cycler operated by mxpro software (Agilent
Technologies, Budapest, Hungary). Dissociation curves of
the positive controls and the products showed specific
amplification of the targeted regions in each case. C
t
values
were determined at a threshold fluorescence defined within
the exponential amplification stage.
Random cloning and sequencing
RNA was isolated from the freshly prepared RNA–UDE
fraction, immediately after Ni
2+
-affinity chromatography
with an RNeasy kit (Qiagen, Go
¨
do
¨
ll
}
o, Hungary); RNA was
converted into double-stranded cDNA with either a
SMARTER kit (Clontech; including oligodT primer and
PCR amplification) or a SuperScrip kit (Invitrogen; includ-
ing random hexamer primer without PCR amplification);
blunt ends were made by T4 DNA polymerase (NEB),
blunt end cloning was performed with a Zero blunt end
Topo cloning kit (Invitrogen), and the products of the clon-

ing reactions were transformed into XL1 Blue competent
cells by applying carbenicillin selection. Twenty-two and 11
colonies were selected for sequencing from samples pre-
pared using oligo-dT and random hexamer primer, respec-
tively. Plasmid DNA was isolated with a Plasmid mini prep
kit (Promega GmbH, Mannheim, Germany). Sequencing
was performed by MWG Operon. Sequences were analysed
by blast at the NCBI page.
Pupal extract preparation
D. melanogaster Oregon R strain fruit flies were kept on
standard cornmeal–yeast food, containing Nipagin as a fun-
gicide, at room temperature. About 1.5 mL pupae was col-
lected and stored at )80 °C until use. The pupae were
homogenized in 8 mL of lysis buffer [50 mm sodium phos-
phate, pH 7.5, 150 mm NaCl, 1 mm EDTA, 1 mm benzami-
dine, 1 mm dithiothreitol, 0.5 mm phenylmethanesulfonyl
fluoride, proteinase inhibitor cocktail at a 500-fold dilution
(Sigma), phosphatase inhibitor mix at a 1000-fold dilution
(Sigma)] with a pestle, a Potter-Elhvejm homogenizer, and
a Bandelin Sonopuls sonicator (2 · 1 min, 60% power,
50% intervals). After centrifugation at 4 °C and 20 000 g
for 15 min, the clear supernatant was removed by accurate
pipetting from under the fat layer and then pushed across a
sterile Whatmann filter (Millipore, Budapest, Hungary).
This clarified extract was aliquoted into three parts
(0.5 mL each); two of them were treated with either
5 lgÆmL
)1
RNase A (Epicentre) or 5 lgÆmL
)1

DNase
(Sigma; the mixture was completed with 5 mm MgCl
2
) for
5 min at room temperature, and then on ice until fraction-
ation on a Superdex 200 HR column, in 200 mm ammo-
nium acetate (pH 7.4) buffer, at 4 °C. One-millitre fractions
were collected, freeze-dried, redissolved in 25 lL of SDS
sample buffer, and separated by SDS ⁄ PAGE.
Western blotting
To produce a better-reacting new antiserum against UDE,
rabbits were immunized with recombinant RNA-free UDE
in the same way as described in [6,7]. Ten-microlitre sam-
ples of lyophilized gel-filtered fractions were subjected to
SDS ⁄ PAGE and transferred to a poly(vinylidene difluoride)
membrane (Millipore) in 10 mm Caps (pH 12.5) buffer with
10% methanol at 4 °C, at a constant current of 350 mA for
5 h. Blots were first stained with Ponceau dye, and then
developed with the antiserum at a dilution of 1 : 100 000;
this was followed by staining with the secondary antibody,
horseradish peroxidase-conjugated anti-rabbit IgG (GE
A. Bekesi et al. RNA-dependent conformational states of UDE
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 311
Healthcare), at a dilution of 1 : 2500, as described in [6].
For visualization, the enhanced chemiluminescence reagents
of Millipore were used.
Acknowledgements
We thank S. Bottka for the synthesis of unlabelled
DNA and RNA oligonucleotides, and M. Kova
´

cs, and
B. Taka
´
cs for equipment and advice in flag-UDE
expression and purification. A. Be
´
ke
´
si was supported
by NKTH-OTKA H07-BEL74200. This work was fur-
ther supported by the following grants: the Hungarian
Scientific Research Fund (OTKA K68229), Howard
Hughes Medical Institutes (#55005628 and #55000342),
Alexander von Humboldt Foundation, GVOP-3.2.1.
-2004-05-0412 ⁄ 3.0 and JA
´
P_TSZ_071128_TB_INTER
from the National Office for Research and Technology
of Hungary, and FP6 STREP 012127, FP6 SPINE2c
LSHG-CT-2006-031220 and TEACH-SG LSSG-CT-
2007-037198 from the EU to B. G. Ve
´
rtessy.
References
1 Lindahl T (1993) Instability and decay of the primary
structure of DNA. Nature 362, 709–715.
2 Pearl LH & Savva R (1996) The problem with pyrimi-
dines. Nat Struct Biol 3, 485–487.
3 Vertessy BG & Toth J (2009) Keeping uracil out
of DNA: physiological role, structure and catalytic

mechanism of dUTPases. Acc Chem Res 42,
97–106.
4 Nyman PO (2001) Introduction. dUTPases. Curr Pro-
tein Pept Sci 2, 277–285.
5 Aravind L & Koonin EV (2000) The alpha ⁄ beta fold
uracil DNA glycosylases: a common origin with diverse
fates. Genome Biol 1, RESEARCH0007, 1–8.
6 Bekesi A, Zagyva I, Hunyadi-Gulyas E, Pongracz V,
Kovari J, Nagy AO, Erdei A, Medzihradszky KF &
Vertessy BG (2004) Developmental regulation of dUT-
Pase in Drosophila melanogaster. J Biol Chem 279,
22362–22370.
7 Bekesi A, Pukancsik M, Muha V, Zagyva I, Leveles I,
Hunyadi-Gulyas E, Klement E, Medzihradszky KF,
Kele Z, Erdei A et al. (2007) A novel fruitfly protein
under developmental control degrades uracil-DNA. Bio-
chem Biophys Res Commun 355, 643–648.
8 Pukancsik M, Bekesi A, Klement E, Hunyadi-Gulyas E,
Medzihradszky KF, Kosinski J, Bujnicki JM, Alfonso
C, Rivas G & Vertessy BG (2010) Physiological trunca-
tion and domain organization of a novel uracil-DNA-
degrading factor. FEBS J 277, 1245–1259.
9 Compton LA & Johnson WC Jr (1986) Analysis of pro-
tein circular dichroism spectra for secondary structure
using a simple matrix multiplication. Anal Biochem 155,
155–167.
10 Matulis D & Lovrien R (1998) 1-Anilino-8-naphthalene
sulfonate anion-protein binding depends primarily on
ion pair formation. Biophys J 74, 422–429.
11 Steinberg TH, Haugland RP & Singer VL (1996) Appli-

cations of SYPRO orange and SYPRO red protein gel
stains. Anal Biochem 239, 238–245.
12 Krintel C, Morgelin M, Logan DT & Holm C (2009)
Phosphorylation of hormone-sensitive lipase by protein
kinase A in vitro promotes an increase in its hydropho-
bic surface area. FEBS J 276, 4752–4762.
13 Weiss JN (1997) The Hill equation revisited: uses and
misuses. FASEB J 11, 835–841.
14 Samanta D, Mukhopadhyay D, Chowdhury S, Ghosh
J, Pal S, Basu A, Bhattacharya A, Das A, Das D &
DasGupta C (2008) Protein folding by domain V of
Escherichia coli 23S rRNA: specificity of RNA–protein
interactions. J Bacteriol 190, 3344–3352.
15 Chattopadhyay S, Das B & Dasgupta C (1996) Reacti-
vation of denatured proteins by 23S ribosomal RNA:
role of domain V. Proc Natl Acad Sci USA 93, 8284–
8287.
16 Chowdhury S, Pal S, Ghosh J & DasGupta C (2002)
Mutations in domain V of the 23S ribosomal RNA of
Bacillus subtilis that inactivate its protein folding prop-
erty in vitro. Nucleic Acids Res 30, 1278–1285.
17 Merenyi G, Konya E & Vertessy BG (2010) Drosophila
proteins involved in metabolism of uracil-DNA possess
different types of nuclear localization signals. FEBS J
277
, 2142–2156.
18 Dickerson SK, Market E, Besmer E & Papavasiliou FN
(2003) AID mediates hypermutation by deaminating
single stranded DNA. J Exp Med 197, 1291–1296.
19 Bransteitter R, Pham P, Scharff MD & Goodman MF

(2003) Activation-induced cytidine deaminase deami-
nates deoxycytidine on single-stranded DNA but
requires the action of RNase. Proc Natl Acad Sci USA
100, 4102–4107.
20 Caricasole A, Duarte A, Larsson SH, Hastie ND, Little
M, Holmes G, Todorov I & Ward A (1996) RNA bind-
ing by the Wilms tumor suppressor zinc finger proteins.
Proc Natl Acad Sci USA 93, 7562–7566.
21 Ryan RF & Darby MK (1998) The role of zinc finger
linkers in p43 and TFIIIA binding to 5S rRNA and
DNA. Nucleic Acids Res 26, 703–709.
22 Brescia CC, Kaw MK & Sledjeski DD (2004) The
DNA binding protein H-NS binds to and alters the sta-
bility of RNA in vitro and in vivo. J Mol Biol 339,
505–514.
23 Skabkin MA, Evdokimova V, Thomas AA & Ovchinni-
kov LP (2001) The major messenger ribonucleoprotein
particle protein p50 (YB-1) promotes nucleic acid
strand annealing. J Biol Chem 276, 44841–44847.
RNA-dependent conformational states of UDE A. Bekesi et al.
312 FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS
24 Deschamps S, Viel A, Garrigos M, Denis H & le Maire
M (1992) mRNP4, a major mRNA-binding protein
from Xenopus oocytes is identical to transcription fac-
tor FRG Y2. J Biol Chem 267, 13799–13802.
25 Barnes T, Kim WC, Mantha AK, Kim SE, Izumi T,
Mitra S & Lee CH (2009) Identification of
apurinic ⁄ apyrimidinic endonuclease 1 (APE1) as the
endoribonuclease that cleaves c-myc mRNA. Nucleic
Acids Res 37, 3946–3958.

26 Liu J, Schmitz JC, Lin X, Tai N, Yan W, Farrell M,
Bailly M, Chen T & Chu E (2002) Thymidylate
synthase as a translational regulator of cellular gene
expression. Biochim Biophys Acta 1587, 174–182.
27 Lin X, Mizunuma N, Chen T, Copur SM, Maley GF,
Liu J, Maley F & Chu E (2000) In vitro selection of an
RNA sequence that interacts with high affinity
with thymidylate synthase. Nucleic Acids Res 28, 4266–
4274.
28 Chu E & Allegra CJ (1996) The role of thymidylate
synthase as an RNA binding protein. Bioessays 18,
191–198.
29 Chu E, Koeller DM, Casey JL, Drake JC, Chabner
BA, Elwood PC, Zinn S & Allegra CJ (1991) Autoregu-
lation of human thymidylate synthase messenger RNA
translation by thymidylate synthase. Proc Natl Acad Sci
USA 88, 8977–8981.
30 Tai N, Schmitz JC, Chen TM & Chu E (2004) Charac-
terization of a cis-acting regulatory element in the pro-
tein-coding region of human dihydrofolate reductase
mRNA. Biochem J 378, 999–1006.
31 Tai N, Schmitz JC, Liu J, Lin X, Bailly M, Chen TM
& Chu E (2004) Translational autoregulation of thymi-
dylate synthase and dihydrofolate reductase. Front Bio-
sci 9, 2521–2526.
32 Meades G Jr, Benson BK, Grove A & Waldrop GL
(2010) A tale of two functions: enzymatic activity and
translational repression by carboxyltransferase. Nucleic
Acids Res 38, 1217–1227.
33 Frugier M & Giege R (2003) Yeast aspartyl-tRNA syn-

thetase binds specifically its own mRNA. J Mol Biol
331, 375–383.
34 Moine H, Romby P, Springer M, Grunberg-Manago
M, Ebel JP, Ehresmann B & Ehresmann C (1990) Esc-
herichia coli threonyl-tRNA synthetase and tRNA(Thr)
modulate the binding of the ribosome to the transla-
tional initiation site of the thrS mRNA. J Mol Biol 216,
299–310.
35 Lanker S, Bushman JL, Hinnebusch AG, Trachsel H &
Mueller PP (1992) Autoregulation of the yeast lysyl-
tRNA synthetase gene GCD5 ⁄ KRS1 by translational
and transcriptional control mechanisms. Cell 70, 647–
657.
36 Allemand F, Haentjens J, Chiaruttini C, Royer C &
Springer M (2007) Escherichia coli ribosomal pro-
tein L20 binds as a single monomer to its own mRNA
bearing two potential binding sites. Nucleic Acids Res
35, 3016–3031.
37 Schuster J, Frojmark AS, Nilsson P, Badhai J, Virtanen
A & Dahl N (2010) Ribosomal protein S19 binds to its
own mRNA with reduced affinity in DiamondBlackfan
anemia. Blood Cells Mol Dis 45, 23–28.
38 Parsons GD, Donly BC & Mackie GA (1988) Muta-
tions in the leader sequence and initiation codon of the
gene for ribosomal protein S20 (rpsT) affect both trans-
lational efficiency and autoregulation. J Bacteriol 170
,
2485–2492.
39 Springer M & Portier C (2003) More than one way to
skin a cat: translational autoregulation by ribosomal

protein S15. Nat Struct Biol 10, 420–422.
40 Kojic M, Topisirovic L & Vasiljevic B (1996) Transla-
tional autoregulation of the sgm gene from Micromo-
nospora zionensis. J Bacteriol 178, 5493–5498.
41 Denoya CD, Bechhofer DH & Dubnau D (1986) Trans-
lational autoregulation of ermC 23S rRNA methyltrans-
ferase expression in Bacillus subtilis. J Bacteriol 168,
1133–1141.
42 Lee KH, Lee NJ, Hyun S, Park YK, Yang EG, Lee
JK, Jeong S & Yu J (2009) Histone H3 N-terminal pep-
tide binds directly to its own mRNA: a possible mode
of feedback inhibition to control translation. Chembio-
chem 10, 1313–1316.
43 Kamadurai HB, Subramaniam S, Jones RB,
Green-Church KB & Foster MP (2003) Protein folding
coupled to DNA binding in the catalytic domain of
bacteriophage lambda integrase detected by mass
spectrometry. Protein Sci 12, 620–626.
44 Berger C, Jelesarov I & Bosshard HR (1996) Coupled
folding and site-specific binding of the GCN4-bZIP
transcription factor to the AP-1 and ATF ⁄ CREB DNA
sites studied by microcalorimetry. Biochemistry 35,
14984–14991.
45 Wang X, Cao W, Cao A & Lai L (2003) Thermody-
namic characterization of the folding coupled DNA
binding by the monomeric transcription activator
GCN4 peptide. Biophys J 84, 1867–1875.
46 Guo X, Campbell FE, Sun L, Christian EL, Ander-
son VE & Harris ME (2006) RNA-dependent folding
and stabilization of C5 protein during assembly of

the E. coli RNase P holoenzyme. J Mol Biol 360,
190–203.
47 Gan J, Tropea JE, Austin BP, Court DL, Waugh DS &
Ji X (2005) Intermediate states of ribonuclease III in
complex with double-stranded RNA. Structure 13,
1435–1442.
48 Das D, Das A, Samanta D, Ghosh J, Dasgupta S,
Bhattacharya A, Basu A, Sanyal S & Das Gupta C
(2008) Role of the ribosome in protein folding. Biotech-
nol J 3, 999–1009.
49 Kramer G, Boehringer D, Ban N & Bukau B (2009)
The ribosome as a platform for co-translational pro-
A. Bekesi et al. RNA-dependent conformational states of UDE
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 313
cessing, folding and targeting of newly synthesized pro-
teins. Nat Struct Mol Biol 16, 589–597.
50 Komar AA (2009) A pause for thought along the co-
translational folding pathway. Trends Biochem Sci 34 ,
16–24.
51 Choi SI, Han KS, Kim CW, Ryu KS, Kim BH, Kim
KH, Kim SI, Kang TH, Shin HC, Lim KH et al.
(2008) Protein solubility and folding enhancement by
interaction with RNA. PLoS ONE 3, e2677.
52 Kim HK, Choi SI & Seong BL (2010) 5S rRNA-
assisted DnaK refolding. Biochem Biophys Res Commun
391, 1177–1181.
53 Caldas TD, El Yaagoubi A & Richarme G (1998)
Chaperone properties of bacterial elongation factor
EF-Tu. J Biol Chem 273, 11478–11482.
54 Kudlicki W, Coffman A, Kramer G & Hardesty B

(1997) Renaturation of rhodanese by translational elon-
gation factor (EF) Tu. Protein refolding by EF-Tu flex-
ing. J Biol Chem 272, 32206–32210.
55 Zhang G, Hubalewska M & Ignatova Z (2009) Tran-
sient ribosomal attenuation coordinates protein synthe-
sis and co-translational folding. Nat Struct Mol Biol 16,
274–280.
56 Tuller T, Waldman YY, Kupiec M & Ruppin E (2010)
Translation efficiency is determined by both codon bias
and folding energy. Proc Natl Acad Sci USA 107, 3645–
3650.
57 Komar AA, Lesnik T & Reiss C (1999) Synonymous
codon substitutions affect ribosome traffic and protein
folding during in vitro translation. FEBS Lett 462, 387–
391.
58 Crombie T, Boyle JP, Coggins JR & Brown AJ (1994)
The folding of the bifunctional TRP3 protein in yeast is
influenced by a translational pause which lies in a
region of structural divergence with Escherichia coli in-
doleglycerol-phosphate synthase. Eur J Biochem 226,
657–664.
59 Cortazzo P, Cervenansky C, Marin M, Reiss C, Ehrlich
R & Deana A (2002) Silent mutations affect in vivo
protein folding in Escherichia coli. Biochem Biophys Res
Commun 293, 537–541.
60 Kimchi-Sarfaty C, Oh JM, Kim IW, Sauna ZE, Calcag-
no AM, Ambudkar SV & Gottesman MM (2007) A
‘silent’ polymorphism in the MDR1 gene changes sub-
strate specificity. Science 315, 525–528.
61 Tsai CJ, Sauna ZE, Kimchi-Sarfaty C, Ambudkar SV,

Gottesman MM & Nussinov R (2008) Synonymous
mutations and ribosome stalling can lead to altered
folding pathways and distinct minima. J Mol Biol 383 ,
281–291.
62 Knobe KE, Sjorin E & Ljung RC (2008) Why does the
mutation G17736A ⁄ Val107Val (silent) in the F9 gene
cause mild haemophilia B in five Swedish families? Hae-
mophilia 14, 723–728.
63 Duan J, Wainwright MS, Comeron JM, Saitou N,
Sanders AR, Gelernter J & Gejman PV (2003) Synony-
mous mutations in the human dopamine receptor D2
(DRD2) affect mRNA stability and synthesis of the
receptor. Hum Mol Genet 12, 205–216.
64 Zalucki YM & Jennings MP (2007) Experimental
confirmation of a key role for non-optimal codons in
protein export. Biochem Biophys Res Commun
355, 143–
148.
65 Xie T & Ding D (1998) The relationship between syn-
onymous codon usage and protein structure. FEBS Lett
434, 93–96.
66 Oresic M & Shalloway D (1998) Specific correlations
between relative synonymous codon usage and protein
secondary structure. J Mol Biol 281, 31–48.
67 Mukhopadhyay P, Basak S & Ghosh TC (2007) Synon-
ymous codon usage in different protein secondary struc-
tural classes of human genes: implication for increased
non-randomness of GC3 rich genes towards protein sta-
bility. J Biosci 32, 947–963.
68 Thanaraj TA & Argos P (1996) Ribosome-mediated

translational pause and protein domain organization.
Protein Sci 5, 1594–1612.
69 Marin M (2008) Folding at the rhythm of the rare
codon beat. Biotechnol J 3, 1047–1057.
70 Takacs E, Grolmusz VK & Vertessy BG (2004) A
tradeoff between protein stability and conformational
mobility in homotrimeric dUTPases. FEBS Lett 566,
48–54.
71 Lees JG, Smith BR, Wien F, Miles AJ & Wallace BA
(2004) CDtool – an integrated software package for cir-
cular dichroism spectroscopic data processing, analysis,
and archiving. Anal Biochem 332, 285–289.
72 Whitmore L & Wallace BA (2008) Protein secondary
structure analyses from circular dichroism spectroscopy:
methods and reference databases. Biopolymers 89, 392–
400.
73 Lees JG, Miles AJ, Wien F & Wallace BA (2006) A ref-
erence database for circular dichroism spectroscopy cov-
ering fold and secondary structure space. Bioinformatics
22, 1955–1962.
74 Nemeth-Pongracz V, Barabas O, Fuxreiter M, Simon I,
Pichova I, Rumlova M, Zabranska H, Svergun D, Pe-
toukhov M, Harmat V et al. (2007) Flexible segments
modulate co-folding of dUTPase and nucleocapsid pro-
teins. Nucleic Acids Res 35, 495–505.
75 Varga B, Barabas O, Kovari J, Toth J, Hunyadi-
Gulyas E, Klement E, Medzihradszky KF, Tolgyesi
F, Fidy J & Vertessy BG (2007) Active site closure
facilitates juxtaposition of reactant atoms for initia-
tion of catalysis by human dUTPase. FEBS Lett 581,

4783–4788.
76 Steinmetzer K & Brantl S (1997) Plasmid pIP501
encoded transcriptional repressor CopR binds asymmet-
RNA-dependent conformational states of UDE A. Bekesi et al.
314 FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS
rically at two consecutive major grooves of the DNA.
J Mol Biol 269, 684–693.
77 Kanaya E, Nakajima N & Okada K (2002)
Non-sequence-specific DNA binding by the
FILAMENTOUS FLOWER protein from Arabidopsis
thaliana is reduced by EDTA. J Biol Chem 277, 11957–
11964.
Supporting information
The following supplementary material is available:
Fig. S1. Validation of the difference between the CD
spectra of RNA–UDE and UDE.
Fig. S2. Comparative characterization of hydrophobic
surface cavities present in RNA–UDE and UDE using
Sypro Orange dye.
Fig. S3. Real-time RT-PCR analysis of the RNA content
of recombinant UDE expressed in insect cells.
This supplementary material can be found in the
online version of this article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.

A. Bekesi et al. RNA-dependent conformational states of UDE
FEBS Journal 278 (2011) 295–315 ª 2010 The Authors Journal compilation ª 2010 FEBS 315