A 5? cytosine binding pocket in Puf3p specifies
regulation of mitochondrial mRNAs
Deyu Zhua,1, Craig R. Stumpfb,1, Joseph M. Krahna, Marvin Wickensb,2, and Traci M. Tanaka Halla,2
aLaboratory of Structural Biology, National Institute of Environmental Health Sciences, Research Triangle Park, NC 27709; andbDepartment of Biochemistry,
University of Wisconsin-Madison, Madison, WI 53706
Edited by Keith R. Yamamoto, University of California, San Francisco, CA, and approved October 2, 2009 (received for review November 26, 2008)
A single regulatory protein can control the fate of many mRNAs
with related functions. The Puf3 protein of Saccharomyces cerevi-
siae is exemplary, as it binds and regulates more than 100 mRNAs
that encode proteins with mitochondrial function. Here we eluci-
determinant of Puf3p specificity is an unusual interaction between
a distinctive pocket of the protein with an RNA base outside the
‘‘core’’ PUF-binding site. That interaction dramatically affects bind-
ing affinity in vitro and is required for regulation in vivo. The Puf3p
structures, combined with those of Puf4p in the same organism,
illuminate the structural basis of natural PUF-RNA networks. Yeast
Puf3p binds its own RNAs because they possess a ?2C and is
excluded from those of Puf4p which contain an additional nucle-
otide in the core-binding site.
crystal structure ? RNA ? translational regulation
proteins to confer regulation. New networks of control have
begun to emerge from analyses of regulatory proteins, their
targets, and their biological functions; for example, webs of
interacting 3?-UTR regulatory proteins and mRNAs control
stem cells in Caenorhabditis elegans (1, 2), the cell cycle during
oocyte maturation and early development in Xenopus (2–4), and
RNA regulation in neurons (5).
The yeast Puf3 protein and mitochondrial mRNAs provide a
particularly dramatic example of a single protein that binds and
regulates a large number of functionally related mRNAs. In the
budding yeast, Saccharomyces cerevisiae, many nuclear-encoded
mRNAs with mitochondrial functions contain a conserved se-
quence in their 3?-UTRs (6) that binds Puf3p (7). Indeed, many
mRNAs of mitochondrial function coimmunopurify with Puf3p
from yeast extracts (8). Loss of this single PUF protein impairs
mitochondrial morphology and maintenance (9) and disrupts
localization of its target mRNAs to the vicinity of mitochondria
(10). Puf3p binds to and destabilizes COX17 mRNA, which
the mitochondrial respiratory complex (7, 11, 12). Thus, by
controlling a large battery of mRNAs, this single PUF protein is
vital for regulation of the organelle.
Puf3p is a member of the PUF family of 3?-UTR–binding
proteins, which regulate translation and mRNA degradation in
diverse eukaryotes (13, 14). Typically, PUF proteins use 8
structural repeats to bind to RNA sequences containing a
UGUR sequence (R, purine) at the 5? end and more variable 3?
sequences (8, 13, 15–24). Each of 8 ?-helical repeats contacts
primarily a single base (21, 25, 26). Although Puf3p is very
similar to other PUF proteins and has most of the same residues
that contact RNA in other PUF proteins, it selectively interacts
with and controls mitochondrial mRNAs.
We sought to understand the structural basis of the specificity
of Puf3p for mitochondrial mRNAs. To do so, we analyzed the
structure of Puf3p bound to 2 sites in one of its mRNA targets,
COX17, analyzed this interaction biochemically, and tested the
ontrol of mRNA stability, translation, and location are
pervasive. Elements in the 3?-UTR interact with regulatory
biological importance of the RNA base at the ?2 position for
regulation in vivo.
General Puf3p Architecture and the Core Region. Crystal structures
of the RNA-binding domain of Puf3p in complex with the
COX17 mRNA sequences for binding site A, CUUGUAUAUA,
and site B, CCUGUAAAUA (Fig. 1A), were determined at a
resolution of 3.2 and 2.5 Å, respectively. Puf3p interacts with the
8 bases of COX17 mRNA starting from the conserved UGUA
sequence much as human Pumilio1 (PUM1) binds to the equiv-
alent RNA [Fig. 1B and supporting information (SI) Fig. S1]
(21). The Puf3p and PUM1 structures are very similar [root
mean squared deviation (RMSD) of 1.3 Å over 307 CA atoms]
with equivalent curvatures; neither is flattened or twisted like
Puf4p and FBF (refs. 26 and 43; Fig. S2).
The structures of Puf3p in complex with COX17 site A and site
B RNA sequences are nearly identical, with an RMSD of 0.2 Å
over 343 CA atoms. The backbone structure of the RNA is
different for the 2 sequences at positions 5 and 6 (Fig. 1B),
however. [As a convention, here we number the PUF protein
RNA recognition sequences beginning with the 5? uracil (U1) of
the conserved UGU sequence.] The ribose group at position 5
in the site B RNA is in a C2?-endo conformation, and A5 and A6
are presented in a syn orientation rather than in the typical anti
orientation (Fig. 2A). Thus for A5 and A6, Puf3p recognizes the
Hoogsteen edge of the base instead of the Watson–Crick edge,
These structural rearrangements allow the edges of the bases of
A5 and A6 in site B to align well with bases U5 and A6 in site
A (Fig. S3). A hydrogen bond between Y695 and the phosphate
group between A5 and A6 may facilitate this structural change.
A similar sequence (UGUAAAUA) is present in the human
Pumilio2 (PUM2) binding site in p38? (27). We predict that a
similar RNA backbone rearrangement may be necessary for
interaction with this and another p38 mRNA recognition se-
quence, UGUAGAUA; indeed, a tyrosine equivalent to Puf3p
Y695 is conserved in PUM2.
To evaluate the functional significance of the different RNA-
binding sites, we determined the binding affinity of Puf3p for
COX17 site A and site B RNAs using electrophoretic mobility
shift assays (Table 1). Based on RNA saturation, 94% of the
Puf3p molecules were active in RNA binding (see Methods).
Author contributions: D.Z., C.R.S., J.M.K., M.W., and T.M.T.H. designed research; D.Z. and
C.R.S. performed research; D.Z., C.R.S., J.M.K., M.W., and T.M.T.H. analyzed data; and D.Z.,
C.R.S., J.M.K., M.W., and T.M.T.H. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Data deposition: The atomic coordinates and structure factors have been deposited in the
Protein Data Bank, www.pdb.org (PDB ID codes 3K4E and 3K49).
1D.Z. and C.R.S. contributed equally to this manuscript.
2To whom correspondence may be addressed. E-mail: firstname.lastname@example.org or wickens@
This article contains supporting information online at www.pnas.org/cgi/content/full/
December 1, 2009 ?
vol. 106 ?
no. 48 www.pnas.org?cgi?doi?10.1073?pnas.0812079106
Puf3p binds tightly to both RNAs, but binds ?10-fold more
tightly to site B (Kdof 3.0 nM vs. 0.31 nM). Two bases (positions
?1 and 5) differ between sites A and B; thus, we tested binding
to an RNA with position 5 of site B changed to a uracil (A5U site
B). Comparing binding to this RNA (1.6 nM) with binding to site
B, we found a 5-fold decrease in affinity when uracil is at position
5. Cytosine at position ?1 results in a modest 2-fold tighter
binding versus uracil at that position. The tight binding for each
site is consistent with in vivo experiments demonstrating delayed
deadenylation and decay of the COX17 transcript when either of
the binding sites was destroyed (12).*
A Distinctive Binding Pocket in Puf3p.Thekey,andmostdistinctive,
feature of Puf3p binding is its interaction with the conserved
cytosine at the ?2 position of Puf3p targets, 2 bases 5? of the
conserved UGU sequence (8) (Fig. 2B). This cytosine is bound
by the protein in a pocket between the N-terminal ends of the
RNA-binding helix of repeat 8 and the terminal helix in repeat
8?. Hydrogen bonds that appear to be specific for a cytosine are
formed between the Watson–Crick edge of the base and main
chain and side chain atoms of S866 and K819 (Fig. 2B). L864
forms a stacking interaction with the base, and the main chain
nitrogen atom of A865 interacts with the 2? OH of C(?2).
Inspecting sequence alignments, the cytosine-binding pocket is
Aspergillus species. Analogous pockets may be present in other
proteins, but not be evident in primary sequence alignments.
the terminal helix and the RNA-binding helix of repeat 8 in PUM1
and Puf4p are positioned closer to each other than the equivalent
helices in Puf3p, and S866 is replaced by a tyrosine in PUM1 and
to fill what would have been the additional binding pocket, pre-
cluding an analogous interaction.
To explore the importance of this additional binding pocket
for RNA recognition, we first determined the binding affinity of
RNAs with different bases at the ?2 position (Table 1). RNAs
with a uracil, guanine, or adenine at the ?2 position bound
?70-fold more weakly than the equivalent RNA with a cytosine
at that position. An 8-base RNA starting from U1 (?5? site B)
bound ?400-fold more weakly than the corresponding 10-base
RNA. These data demonstrate that the binding pocket for the
?2 position is important for high-affinity binding.
To confirm the importance of the binding pocket on the
protein side, we mutated S866 to alanine and determined the
binding affinity of the mutant protein for COX17 RNA (Table
1). The mutant protein appears to be properly folded because it
elutes from a gel filtration column at the same position as
wild-type protein, and the circular dichroism spectra are essen-
tially identical (Fig. S4). The S866A mutant protein bound
?400-fold more weakly than native protein to the COX17 site B
RNA (Kd129 nM vs. 0.31 nM), indicating that this single amino
acid residue contributes critically to high-affinity binding.
Regulation by Puf3p in Vivo Requires Binding Sites with ?2C. To
assess whether the cytosine at the ?2 position is important for
regulation in vivo, we analyzed the decay of reporter mRNAs
containing either wild-type or mutant forms of the COX17
3?-UTR. We compared the stabilities of COX17 mRNA bearing
either the wild-type 3?-UTR or a mutant 3?-UTR in which the
base at ?2 was changed from cytidine to adenine in both site A
and site B (Table 1). (Mutation of either site disrupts binding in
vitro.) Both genes were placed under control of the GAL
promoter. To measure mRNA half-life, the genes were induced
in galactose-containing media, and then repressed by the addi-
tion of glucose. Decay of the mRNA after transcriptional
inhibition was quantified using Northern blot analysis (Fig. 3).
In wild-type cells, the wild-type reporter mRNA decayed
rapidly (half-life of ?4 min), similar to what was reported
previously (Fig. 3A) (7, 12, 28, 29). The C(?2)A double mutant
was stabilized approximately 4-fold in the same cells (half-life
?17 min). In the ?puf3 strain, wild- type and mutant reporters
were both stabilized equivalently (Fig. 3B; half-life ?12 min).
*An earlier report of a preference for site A (12) is undermined by the use of GST fusion
proteins with 5–10% apparent activity.
COX17 mRNA sequences. (A) Ribbon diagram of the RNA-binding domain of
S. cerevisiae Puf3p in complex with a COX17 site B RNA fragment (CCU-
colored according to atom type (carbon, white; nitrogen, blue; oxygen, red;
sulfur, yellow; phosphorus, orange). Protein side chains that contact the RNA
(B) Superposition of CA traces of the Puf3p:COX17 site B RNA complex (blue)
and PUM1:RNA complex (gray; PDB ID: 1M8Y) aligned over the entire protein
structures. The COX17 site A RNA from the structure of a complex with Puf3p
is shown in gold.
Crystal structures of the Puf3p RNA-binding domain in complex with
and RNA are colored as in Fig. 1A. (A) Puf3p recognition of bases A4–A6 in
Recognition of RNA by the Puf3p RNA-binding domain. The protein
Zhu et al.PNAS ?
December 1, 2009 ?
vol. 106 ?
no. 48 ?
These data permit 2 conclusions: (i) The presence of a C at the
?2 position is important for biological regulation in vivo (Fig.
3A),and (ii) the effect of the ?2 base is dependent on Puf3p
Bioinformatic analysis supports the biological importance of
the cytosine at ?2. Cytosine was enriched at the ?2 position
among the mRNAs associated with Puf3p in coimmunopurifi-
cation experiments; 82% of the mRNAs (144 of 174) associated
with Puf3p had a ?2C, compared with the 18% prevalence of
cytosine in the yeast genome (P ? 1.37 ? 10?77). Indeed, the
cytosine at ?2 was part of the binding motif initially identified
simply as a conserved sequence in the 3?-UTRs of mitochondrial
mRNAs (6) and later shown to bind Puf3p (7, 8). The cytosine
at ?2 is evolutionarily conserved as well. To examine conser-
vation, we analyzed homologs of the mRNAs that bind Puf3p in
S. cerevisiae (8) in related yeast species (S. paradoxus, mikatae,
bayanus, castellii, kudriavzevii, and kluyveri). A cytosine at ?2 is
more highly conserved among these species than are the rare A,
G, or U identities (P ? .001). A cytosine at ?2 also is enriched
in mRNAs that demonstrate a strict dependence on Puf3p for
localization to mitochondria (P ? .0001) (9). In summary, both
the effects on RNA stability and the conservation of the ?2C
support the view that the ?2C is important for Puf3p function
Puf3p Versus Puf4p: Two Determinants of Binding and Exclusion. The
Puf3p-RNA structures suggest how different PUF proteins
expressed in the same cell can specifically regulate only their own
RNA target sequences. Puf3p binds ?60-fold more weakly to the
Puf4p-binding site in HO RNA and ?200-fold more weakly to
the Mpt5p-binding site in HO RNA compared with COX17 site
A (Table 1). Both Puf4p and Mpt5p target sequences lack a ?2C
and contain inserted or mismatched bases downstream of the
UGU sequence (Table 1). The crystal structure of Puf4p with
RNA identified an additional base between bases 6 and 7
(referred to here as base 6?) as a key determinant for Puf4p
recognition (26). Our new understanding that the key determi-
nant for Puf3p binding is a ?2C allowed us to evaluate the
importance of these 2 features for selective PUF protein target
recognition by Puf3p and Puf4p.
To do so, we analyzed Puf3p and Puf4p binding to a series of
sites (Fig. 4; Table S1). We focused on 2 determinants: the identity
of the ?2 base and the presence or absence of the 6? nucleotide (8,
26). Thus, we prepared 4 RNAs with either U or C at the ?2
position and either with or without an ‘‘extra’’ base at 6?.
Puf3p bound tightly (Kd3.0 nM) to an RNA that matched its
optimal site, with a ?2C and without a 6? nucleotide. The
presence of a cytosine at ?2 enhanced binding ?70-fold relative
to an identical RNA with uracil (Fig. 4, Top; Table S1). Similarly,
Puf3p had a ?30-fold preference for an RNA lacking the 6?
nucleotide relative to one containing it. Puf3p bound the worst
to an RNA that was suboptimal in both determinants; the
presence of a ?2U and a 6? nucleotide reduced binding by
Puf4p exhibited a reciprocal pattern. It bound tightly to its
optimal site (Kd 7.2 nM) but poorly to sites lacking the 6?
nucleotide (Fig. 4, Bottom; Table S1). Puf4p did not discriminate
and ?2C regardless of whether the 6? nucleotide was present.
Together, these data demonstrate that the presence of a C at ?2
and the presence or absence of a 6? nucleotide account substan-
and noncognate sites.
Table 1. RNA-binding analyses of wild-type and mutant Puf3p
COX17 site A
COX17 site B
A5U site B
C(?2)U site A
C(?2)G site A
C(?2)A site A
C(?2)A site B
?5? site B
COX17 site B
HO RNA Puf4p BS
HO RNA Mpt5p BS
3.0 ? 0.3
0.31 ? 0.07
1.6 ? 0.09
205 ? 14
224 ? 11
234 ? 9
58.3 ? 7
119 ? 6
129 ? 19
181 ? 12
607 ? 44
0.5‡(5‡vs. site B)
188‡(vs. site B)
384‡(vs. site B)
416‡(vs. site B)
*Numbers indicate positions in the core recognition sequence. Bases that differ from the sequence of COX17 site A RNA are in bold.
†Krelreports the affinity of wild-type or mutant Puf3p for the RNA sequence, relative to the affinity of wild-type Puf3p for COX17 site A RNA, unless indicated
otherwise in parentheses.
‡Asterisks denote statistically significant differences (P ? .05).
decay in vivo. Representative Northern blots illustrating the decay from the
steady state of COX17 mRNAs bearing either the wild-type or mutant 3?-UTR
in wild-type (A) and ?puf3 (B) strains. The mutant 3?-UTR carries C(?2)A
mutations in both of its Puf3p-binding sites. The control RNA is ScR1.
A cytosine at the ?2 position is critical for Puf3p-mediated mRNA
www.pnas.org?cgi?doi?10.1073?pnas.0812079106Zhu et al.
Our results begin to provide a structural explanation for net-
works of PUF protein control in vivo. Each PUF protein must
bind its own set of target RNAs and be excluded from those of
other PUF proteins. A comparison of Puf3p and Puf4p is
particularly instructive, because the structures of both PUF-
RNA complexes are now known. These 2 PUF proteins bind
almost completely nonoverlapping sets of mRNAs. Of the 213
RNAs associated with Puf3p and the 176 RNAs associated with
Puf4p, only 4 associate with both proteins, and those 4 mRNAs
have 2 separate elements for each protein (8). The data reported
here permit a structural model of binding and exclusion of these
2 PUF proteins. Puf3p requires a cytosine in the ?2 position for
high-affinity binding, and that cytosine is accommodated by a
specific binding pocket in repeats 8 and 8? of the protein (Fig. 5).
In Puf4p, the would-be pocket is disrupted by a histidine residue
and a change in orientation of 2 ?-helices; indeed, Puf4p shows
no preference for a C at ?2 (Figs. 4 and 5; Table S1). Consistent
with this, the ?2 position is a cytosine in 82% of RNAs (144 of
174) immunopurified with Puf3p that have an identifiable bind-
ing site, but in only 12% (16 of 135) of those that bind Puf4p (8).
Conversely, Puf4p targets have an additional nucleotide between
positions 6 and 7. The presence of this nucleotide enhances the
binding of Puf4p and diminishes binding of Puf3p (Figs. 4 and
5; Table S1). This permits Puf4p to bind to its targets, while
excluding Puf3p. We suggest that Puf3p’s specificity ensures that
a battery of mitochondrial mRNAs are coordinately controlled,
and that other mRNAs are excluded from that regulation.
have a nucleotide other than C at the ?2 position (8). The
stabilities of 2 such mRNAs, COX15 and PET117, have been
examined and found to be unaffected by the deletion of puf3 (8,
30). Nonetheless, mRNAs of this type may still use Puf3p for
processes other than decay, such as translational repression or
The studies reported here and in the accompanying paper
illustrate the versatility of PUF protein specificity and affinity.
In the cases of FBF and yeast Puf4p, the protein imposes a
requirement for an extra base relative to PUM1, and that base
flips out away from the protein. In Puf3p, a specific binding
pocket imposes a requirement for a cytosine at the ?2 position.
Indeed, biological regulation by Puf3p requires the high-affinity
interaction due to the additional binding pocket. These permu-
tations of PUF scaffolds enable PUF proteins to bind mutually
exclusive sets of RNAs yet use largely overlapping sets of atomic
contacts to do so.
Protein Expression and Purification. A cDNA encoding the RNA-binding do-
main of Puf3p (amino acid 511–879) was amplified by PCR from S. cerevisiae
genomic DNA and inserted into the pTYB3 expression plasmid (New England
Biolabs) using NcoI and SapI restriction sites. The protein, in which Puf3p was
fused to the N terminus of an intein and a chitin-binding domain, was
expressed in Escherichia coli strain BL21(DE3) with 1 mM IPTG for 6 h at 33 °C.
Bacterial pellets were resuspended in sonication buffer (20 mM sodium phos-
phate [pH 8.0], 500 mM NaCl) and lysed by sonication in an ice bath after the
addition of 1 complete EDTA-free protease inhibitor tablet (Roche). The
agarose resin (New England Biolabs) for 40 min at 4 °C. The resin was washed
extensively with sonication buffer, then with high-salt buffer (20 mM sodium
phosphate [pH 8.0], 1 M NaCl), and finally with a sonication buffer containing
150 mM NaCl. The resin was then transferred to a 50-mL conical tube and
to initiate cleavage by the intein domain. The tube was sealed, and the resin
was incubated for 2–3 days at room temperature. The beads were transferred
into a 10-mL disposable column, and the cleaved Puf3p protein was collected.
The beads were washed 2–3 times with a 1/4-column volume of sonication
column buffer [10 mM Tris-HCl (pH 7.4), 50 mM NaCl, 6 mM ?-mercaptoetha-
nol). For Puf3p alone, the protein was loaded onto a Superdex 200 HR 10/30
column (GE Healthcare) and run at 0.5 mL/min in column buffer. Fractions
containing Puf3p were pooled and concentrated to 3.5 mg/mL. For the com-
plexes of Puf3p with RNA oligonucleotides, each RNA oligonucleotide (CCU-
GUAAAUA or CUUGUAUAUA; Dharmacon Research) was added at a 1:1
stoichiometry and incubated at 4 °C for 6 h before being loaded onto the
Superdex 200 HR 10/30 column. Fractions containing Puf3p with RNA oligo-
?2 mg/mL for crystallization.
S866A mutPuf3p cDNA was made using the QuikChange site-directed
were confirmed by sequencing the full cDNA insert. Mutant protein was
prepared as described above for the wild-type protein.
Crystallization and Data Collection. Crystallization experiments were con-
ducted using the hanging-drop vapor-diffusion method at 4 °C, with 2-?L
drops composed of 1 ?L of Puf3p:RNA complex solution and 1 ?L of well
site B (CCUGUAAAUA) complex, the well solution was 22% PEG4000, 0.1 M
sodium citrate (pH 5.2), and 1% dextran sulfate. Diffraction-quality crystals
graph illustrating relative binding of Puf3p and Puf4p to RNAs with either C
given in Table S1.
Key determinants of Puf3p and Puf4p binding and exclusion. Bar
PUF-RNA complex, with 8 helices (tan circles, R1–R8) aligned with 8 bases
(purple hexagons, ?1 to ? 8), depicted in the center, represents the ‘‘base’’
PUF-RNA structure, as seen in human PUM1 (21). The 2 key determinants for
Puf3p and Puf4p are expanded above and to the right, respectively. See the
text for details.
Structural basis of the RNA selectivity of Puf3p and Puf4p. A core
Zhu et al.PNAS ?
December 1, 2009 ?
vol. 106 ?
no. 48 ?
with dimensions of 0.25 ? 0.15 ? 0.03 mm3were obtained by successive
microseeding (32). For the Puf3p:COX17 site A (CUUGUAUAUA) complex, the
were obtained by a combined procedure of cross-seeding (original crystal
seeds from a Puf3p:COX17 site B complex crystal), microseeding, and
Crystals were flash-cooled in liquid nitrogen after incubation in cryopro-
tectant solution containing the well solution supplemented with 10% ethyl-
1.0 Å) and Rigaku RU-007HF (wavelength 1.5418 Å), respectively. The data
were processed using HKL2000 (33). Data statistics are summarized in Table 2.
Structure Determination and Refinement. The structure of the Puf3p:COX17
site B complex was determined by molecular replacement using the coordi-
nates of the PUM1 structure (1M8Y) with the Phaser program (34). Swiss PDB
Viewer (35) was used to thread the aligned Puf3p sequence onto the model,
followed by manual model building using ‘‘O’’ (36). The model was refined
using PHENIX (37) with noncrystallographic restraints for both protein and
RNA. Parameters for the bases at positions 5 and 8, which have C2?-endo
geometry, were modified based on C3?-endo restraints from CNS (38). The
structure of the Puf3p:COX17 site A complex was determined using the
Puf3p:COX17 site B complex structure model. Electron density maps revealed
interpretable density for the RNA, with differences primarily in locations
where the RNA bases differ. Both structures contain 3 Puf3p:RNA complexes
in the asymmetric unit; we focus on the details of the Puf3p:COX17 siteB
complex (chains A and B), which was determined at higher resolution (2.5 Å).
Most details are observed in all complexes with either RNA. The structures
were analyzed using MolProbity (39). Model and refinement statistics are
summarized in Table 2. All ?–? torsion angles are within allowable regions of
the Ramachandran plot, and 98% and 99% are in the favored regions for
structures of Puf3p:COX17 site B and Puf3p:COX17 site A, respectively. There
are no incorrect sugar puckers or RNA backbone conformations.
Electrophoretic Mobility Shift Assays. Equilibrium dissociation constants for
Puf3p and S866A mutPuf3p proteins were determined by electrophoretic
macon and radiolabeled at the 5?-end using32P-?-ATP (PerkinElmer Life
Sciences) and T4 polynucleotide kinase (New England BioLabs) following the
manufacturer’s instructions. Puf4p was expressed and purified as described
Binding reactions included radiolabeled target RNA and serially diluted
concentrations of protein and were incubated overnight at 4 °C in buffer
containing 10 mM Tris-HCl (pH 7.4), 100 mM NaCl, 2 mM MgCl2, 3% glycerol,
5 ?g/mL of heparin, 0.01% IGEPAL, and 6 mM ?-mercaptoethanol. For the
Puf3p-binding assays, we used 30 pM for the COX17 site A, COX17 site B, and
RNAs and 300 pM for the other RNA oligonucleotides. Reactions were loaded
within 6 min after addition of Ficoll to 2.5% (vol/vol) and run on 20% 19:1
acrylamide:bisacrylamide gels with 0.5? TBE at 4 °C under 150 V constant
screens (GE Healthcare) and scanned on a Molecular Dynamics Typhoon
phosphorimager (GE Healthcare). The intensities of bands corresponding to
bound and free radiolabeled RNA were measured using GelEval (Frogsoft),
and the data were plotted and analyzed using IgorPro (WaveMetrics). The
equilibrium dissociation constants were calculated by fitting the plot of the
fraction bound versus protein concentration to the Hill equation, assuming a
Hill coefficient of 1. The mean and standard error of at least 3 independent
Prism software. A representative experiment is shown in Fig. S5. The condi-
tions for the assay were established by testing the effect on Kdof varying
incubation time (2–48 h), RNA concentration (10–300 pM), and time after the
addition of loading buffer (5–15 min). No significant difference in Kdwas
detected under these conditions. Determination of the percentage of active
protein was performed as described previously (40). Puf3p was 94% active,
and Puf4p was 91% active; no corrections were made to the measured Kd.
on CEN plasmids and transformed into wild-type and puf3? yeast strains.
mRNA half-life determinations were made by decay from steady-state exper-
iments as described previously with minor modifications (41). In brief, yeast
were grown to mid-log phase at 30 °C in galactose-containing media. Yeast
were concentrated in media containing galactose and incubated for 30 min.
Transcription of the reporters was terminated by the addition of glucose.
Samples were collected and frozen. RNA was isolated using the MasterPure
was performed as described previously (42). Decay data were analyzed using
GraphPad Prism software. Site-directed mutants of Puf3p that affect RNA
binding and mRNA decay (29) are analyzed in Fig. S6.
ACKNOWLEDGMENTS. We are grateful to our colleagues for critical com-
ments on the manuscript. We thank A. Sigova of the University of Massachu-
setts Medical School for the yeast genomic DNA, A. Clark for advice on
purifying yeast genomic DNA, L. Pedersen for crystallography support, G.
Kissling for statistical support, and A. Steinberg and L. Vanderploeg for
sion studies for Puf3p. This work was supported in part by the Intramural
Research Program of the National Institutes of Health, National Institute of
Environmental Health Sciences (T.H.) and by extramural grants from the
this study was supported by the U.S. Department of Energy, Office of Science,
Office of Basic Energy Sciences, under Contract W-31–109-Eng-38.
sperm/oocyte decision in Caenorhabditis elegans. Annu Rev Cell Dev Biol 23:405–433.
2. Thompson B, Wickens M, Kimble J (2007) in Translational Control in Biology and
Cold Spring Harbor, NY), pp 507–544.
Mol Cell Biol 2:521–529.
4. Pique M, Lopez JM, Foissac S, Guigo R, Mendez R (2008) A combinatorial code for
CPE-mediated translational control. Cell 132:434–448.
RNA processing. Nature 456:464–469.
6. Jacobs Anderson JS, Parker R (2000) Computational identification of cis-acting ele-
ments affecting post-transcriptional control of gene expression in Saccharomyces
cerevisiae. Nucleic Acids Res 28:1604–1617.
7. Olivas W, Parker R (2000) The Puf3 protein is a transcript-specific regulator of mRNA
degradation in yeast. EMBO J 19:6602–6611.
8. Gerber AP, Herschlag D, Brown PO (2004) Extensive association of functionally and
cytotopically related mRNAs with Puf family RNA-binding proteins in yeast. PLoS Biol
9. Garcia-Rodriguez LJ, Gay AC, Pon LA (2007) Puf3p, a Pumilio family RNA-binding
in budding yeast. J Cell Biol 176:197–207.
Table 2. Data collection and refinement statistics
Puf3p:COX17 SiteB Puf3p:COX17 SiteA
a, b, c, Å
?, ?, ?, degrees
Number of atoms
Bond length, Å
Bond angle, degrees
151.13, 87.07, 125.18
90.0, 116.5, 90.0
150.18, 87.13, 124.91
90.0, 116.2, 90.0
*Values in parentheses are for the highest-resolution shell.
www.pnas.org?cgi?doi?10.1073?pnas.0812079106 Zhu et al.
10. Saint-Georges Y, et al. (2008) Yeast mitochondrial biogenesis: A role for the PUF Download full-text
RNA-binding protein Puf3p in mRNA localization. PLoS ONE 3:e2293.
11. Glerum DM, Shtanko A, Tzagoloff A (1996) Characterization of COX17, a yeast gene
involved in copper metabolism and assembly of cytochrome oxidase. J Biol Chem
12. Jackson JS, Jr., Houshmandi SS, Lopez Leban F, Olivas WM (2004) Recruitment of the
Puf3 protein to its mRNA target for regulation of mRNA decay in yeast. RNA 10:1625–
13. Wickens M, Bernstein DS, Kimble J, Parker R (2002) A PUF family portrait: 3?-UTR
regulation as a way of life. Trends Genet 18:150–157.
14. Wharton RP, Aggarwal AK (2006) mRNA regulation by Puf domain proteins. Sci STKE
15. Bernstein D, Hook B, Hajarnavis A, Opperman L, Wickens M (2005) Binding specificity
and mRNA targets of a C. elegans PUF protein, FBF-1. RNA 11:447–458.
16. Edwards TA, Pyle SE, Wharton RP, Aggarwal AK (2001) Structure of Pumilio reveals
similarity between RNA and peptide-binding motifs. Cell 105:281–289.
17. Galgano A, et al. (2008) Comparative analysis of mRNA targets for human PUF-family
proteins suggests extensive interaction with the miRNA regulatory system. PLoS ONE
18. Macdonald PM (1992) The Drosophila pumilio gene: An unusually long transcription
unit and an unusual protein. Development 114:221–232.
19. Opperman L, Hook B, DeFino M, Bernstein DS, Wickens M (2005) A single spacer
nucleotide determines the specificities of two mRNA regulatory proteins. Nat Struct
Mol Biol 12:945–951.
20. Stumpf CR, Kimble J, Wickens M (2008) A Caenorhabditis elegans PUF protein family
with distinct RNA binding specificity. RNA 14:1550–1557.
21. Wang X, McLachlan J, Zamore PD, Hall TM (2002) Modular recognition of RNA by a
human pumilio-homology domain. Cell 110:501–512.
22. Wang X, Zamore PD, Hall TM (2001) Crystal structure of a Pumilio homology domain.
Mol Cell 7:855–865.
23. Zamore PD, Williamson JR, Lehmann R (1997) The Pumilio protein binds RNA through
a conserved domain that defines a new class of RNA-binding proteins. RNA 3:1421–
24. Zhang B, et al. (1997) A conserved RNA-binding protein that regulates sexual fates in
the C. elegans hermaphrodite germ line. Nature 390:477–484.
noncognate RNAs reveal molecular mechanisms for binding promiscuity. Structure
26. Miller MT, Higgin JJ, Tanaka Hall TM (2008) Basis of altered RNA-binding specificity by
PUF proteins revealed by crystal structures of yeast Puf4p. Nat Struct Mol Biol 15:397–
27. Lee MH, et al. (2007) Conserved regulation of MAP kinase expression by PUF RNA-
binding proteins. PLoS Genet 3:e233.
29. Houshmandi SS, Olivas WM (2005) Yeast Puf3 mutants reveal the complexity of
Puf-RNA binding and identify a loop required for regulation of mRNA decay. RNA
30. Ulbricht RJ, Olivas WM (2008) Puf1p acts in combination with other yeast Puf proteins
to control mRNA stability. RNA 14:246–262.
31. DeLano WL (2002) The PyMOL Molecular Graphics System (DeLano Scientific, San
32. Zhu DY, Zhu YQ, Xiang Y, Wang DC (2005) Optimizing protein crystal growth through
dynamic seeding. Acta Crystallogr D 61:772–775.
33. Otwinowski Z, Minor W (1997) Processing of X-ray diffraction data collected in oscil-
lation mode. Methods Enzymol 276:307–326.
34. McCoy AJ, Grosse-Kunstleve RW, Storoni LC, Read RJ (2005) Likelihood-enhanced fast
translation functions. Acta Crystallogr D 61:458–464.
35. Guex N, Peitsch MC (1997) SWISS-MODEL and the Swiss-PdbViewer: An environment
for comparative protein modeling. Electrophoresis 18:2714–2723.
36. Jones TA, Zou J-Y, Cowan SW, Kjeldgaard M (1991) Improved methods for building
protein models in electron-density maps and the location of errors in these models.
Acta Crystallogr A 47:110–119.
37. Adams PD, et al. (2004) Recent developments in the PHENIX software for automated
crystallographic structure determination. J Synchrotron Radiat 11:53–55.
38. Bru ¨nger AT, et al. (1998) Crystallography and NMR system: A new software suite for
macromolecular structure determination. Acta Crystallogr D 54:905–921.
39. Davis IW, et al. (2007) MolProbity: All-atom contacts and structure validation for
proteins and nucleic acids. Nucleic Acids Res 35:W375–W383.
Proc Natl Acad Sci USA 103:13635–13639.
41. Coller JM, Tucker M, Sheth U, Valencia-Sanchez MA, Parker R (2001) The DEAD box
helicase, Dhh1p, functions in mRNA decapping and interacts with both the decapping
and deadenylase complexes. RNA 7:1717–1727.
42. Seay D, Hook B, Evans K, Wickens M (2006) A three-hybrid screen identifies mRNAs
controlled by a regulatory protein. RNA 12:1594–1600.
43. Wang Y, Opperman L, Wickens M, Hall TMT (2009) Structural basis for specific recog-
nition of multiple mRNA targets by a PUF regulatory protein. Proc Natl Acad Sci USA,
Zhu et al. PNAS ?
December 1, 2009 ?
vol. 106 ?
no. 48 ?