doi:10.1016/j.jmb.2004.03.059
J. Mol. Biol. (2004) 339, 41–51
Architecture and folding mechanism of the Azoarcus
Group I Pre-tRNA
Prashanth Rangan1, Benoı̂t Masquida2, Eric Westhof2 and
Sarah A. Woodson1*
1
T. C. Jenkins Department of
Biophysics, Johns Hopkins
University, 3400 N. Charles St.
Baltimore, MD 21218, USA
2
Institut de biologie moléculaire
et cellulaire, UPR9002, CNRS
Université Louis Pasteur, 15
rue René Descartes, F-67084
Strasbourg Cedex, France
Self-splicing RNAs must evolve to function in their specific exon context.
The conformation of a group I pre-tRNAile from the bacterium Azoarcus
was probed by ribonuclease T1 and hydroxyl radical cleavage, and by
native gel electrophoresis. Biochemical data and three-dimensional
models of the pre-tRNA showed that the tRNA is folded, and that the
tRNA and intron sequences form separate tertiary domains. Models of
the active site before steps 1 and 2 of the splicing reaction predict that
exchange of the external G-cofactor and the 30 -terminal G is accomplished
by a slight conformational change in P9.0 of the Azoarcus group I intron.
Kinetic assays showed that the pre-tRNA folds in minutes, much more
slowly than the intron alone. The dependence of the folding kinetics on
Mg2þ and the concentration of urea, and RNase T1 experiments showed
that formation of native pre-tRNA is delayed by misfolding of P3 – P9,
including mispairing beween residues in P9 and the tRNA. Thus,
although the intron and tRNA sequences form separate domains in the
native pre-tRNA, their folding is coupled via metastable non-native basepairs. This could help prevent premature processing of the 50 and 30 ends
of unspliced pre-tRNA.
q 2004 Elsevier Ltd. All rights reserved.
*Corresponding author
Keywords: group I ribozyme; hydroxyl radical footprinting; RNA
modelling; tRNA splicing
Introduction
Most stable RNAs are processed from longer
transcription units, and both coding and noncoding sequences contribute to processing and
maturation. Among self-splicing RNAs, interactions between the intron core and the exon
sequences contribute to recognition of the 50 and 30
splice sites.1 On the other hand, long RNAs have a
lower probability of folding correctly, because the
opportunity for misfolding increases.2 Interactions
with flanking sequences attenuate self-cleavage of
the hepatitis delta virus ribozyme,3,4 and selfsplicing of the Tetrahymena large subunit rRNA
and phage T4 td group I introns.5,6 In natural
RNAs, where the tendency to misfold is expected
to be minimized over the course of evolution,
Abbreviations used: RNase, ribonuclease; exoG,
external G-cofactor.
E-mail address of the corresponding author:
[email protected]
such alternative structure might persist because
they may serve a regulatory function.
To further understand the role of flanking
sequences in RNA processing, we investigated the
conformation and folding kinetics of a self-splicing
group I intron from the pre-tRNAile of the cyanobacterium Azoarcus.7 Like other group I introns
found in tRNA genes, the Azoarcus intron has minimal P1 and P10 base-pairing interactions that
define the 50 and 30 splice sites. Because the introns
are inserted in the anticodon loop, interactions
between the group I intron core and the mature
tRNA domain may contribute to folding and
splice site recognition. In a closely related tRNA
intron from the cyanobacterium Anabaena, the
anticodon stem-loop of the tRNA was observed
to increase self-splicing as much as 60-fold.8
The authors proposed that the P1 duplex
containing the 50 splice site is either stabilized
directly by the anticodon stem or sequesters the
P1 helix from other unfavorable base-pairing
interactions.8
Three-dimensional models of the pre-tRNA were
0022-2836/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
42
constructed that suggest a conformational change
in P9.0 between the first and second steps of the
splicing reaction. Beyond the anticodon stem, the
tRNA makes little contact with the intron domain.
To understand how the tRNA exons might
influence the stability and folding kinetics of the
pre-tRNA, the formation of secondary and tertiary
structure in the Azoarcus pre-tRNA was monitored
as a function of magnesium concentration. We
have found that the Azoarcus L-9 ribozyme folds
in 50 – 100 ms in a concerted fashion.9 Here, we
find that correct folding of the pre-tRNA is delayed
by non-native base-pairs between the tRNA and
the intron. Thus, the exons contribute to splice site
recognition, but increase the probability of misfolding in vitro.
Results
Modular structure of the pre-tRNA
To determine how the tRNA exons interact with
the intron, the conformation of the Azoarcus pretRNA was probed by partial digestion with
ribonuclease (RNase) T1 and hydroxyl radical.
Azoarcus Group I Pre-tRNA Folding Mechanism
Reactions were carried out in the absence of GTP,
to prevent self-splicing. Above 1 mM MgCl2, the
RNase T1 protection pattern (Figure 1) was
consistent with the phylogenetically predicted
secondary structures of the intron and the
canonical “cloverleaf” structure of the tRNA,
respectively.10,11 In the presence of Mg2þ, G9 and
G11 are protected, while cleavage at G6 is
enhanced, consistent with formation of the P1
splice site helix. Digestion of core intron residues
was the same as in the L-9 ribozyme, except that
G40 and G41 in P3 were partially accessible to
RNase T1 in the pre-tRNA, but fully protected in
the L-9 ribozyme.12 This may reflect residual misfolding of the pre-tRNA (see below).
Fe-EDTA and RNase T1 protection assays
showed that the pre-tRNA adopts independent
folds for the tRNA and intron domains. Although
the extent of protection in the 50 end of the tRNA
(up to the anticodon loop) varied somewhat
between experiments, protections of the variable
region and T-loop are typical of a folded tRNA.13
In 15 mM MgCl2, the protected regions within the
intron were identical with those observed for the
L-9 ribozyme,12 except that nucleotides 57– 60 in
J4/5 and 8– 12 were protected in the pre-tRNA.
Figure 1. Structural probing of
the pre-tRNA. Small arrowheads
indicate Mg2þ-dependent change in
RNase T1 cleavage: blue, protected
in 20 mM MgCl2; red, requires more
than 20 mM MgCl2; black, enhanced
cleavage in MgCl2; white, protected
more strongly in intron than pretRNA. Shaded regions are protected
from hydroxyl radical cleavage in
Mg2þ; hatched regions are protected
only
in
pre-tRNA.
Blue,
Cm ¼ 0:4 – 0:7 mM, nH < 1; green,
Cm < 1:5 mM; nH < 3; yellow, Cm ¼
1:9 – 2:4: See Table 1 for details. The
degree of protection at the 50 end of
the pre-tRNA was variable; these
results are not reported. Intron
residues are numbered according
to Tanner & Cech;18 tRNA residues
are
numbered
according
to
Steinberg et al.44
43
Azoarcus Group I Pre-tRNA Folding Mechanism
These protections are consistent with known
tertiary contacts between the catalytic core and the
docked P1 helix.14 – 16 As the folded structures of
the intron and tRNA account for all the solventinaccessible regions detected by us, extensive
tertiary interactions between these two domains of
the pre-tRNA are unlikely.
It has been suggested that docking of P1 is
stabilized by the tRNA anticodon stem.8 To gain
insight into the structural interface between the
intron and anticodon stem, we modelled the
pre-tRNA in the states preceding the two steps of
self-splicing (Figure 2). In step 1, the external
G-cofactor (exoG) attacks the 50 splice site.17
Cleavage occurs 30 of U36 (third base of the
anticodon of the tRNA), which forms the conserved UoG pair at the 50 splice site in P1. After
step 1, exoG is ligated to the 50 end of the intron
and the 50 exon is held in place by the three basepairs between the intron guide sequence and
residues of the anticodon loop. The P1 loop is
unwound, allowing A37 from the 30 exon to replace
A1 in base-pairing with U8. Before step 2, the 30 terminal G of the intron (G205 or VG) moves into
the G-binding pocket. Finally, the 30 hydroxyl
group of U36 attacks the 30 splice site phospho-
diester bond (G205/A37) and ligates the two
exons.
We built a model in which G205 (VG) in step 2
replaces the G-cofactor in step 1, a process
promoted by the fold of the P9.0 –J7/9.0 region
(Figure 2). In step 1, the N2 group of G204 interacts
with the Hoogsteen edge of G179, and G205 forms
a sheared pair with A178. This conformation
allows for the docking of exoG onto the second
base-pair of P7 (C176vG129), above the bulged
residue A128. Then, upon hydrolysis and subsequent unwinding of the cleaved part of P1, the
covalently attached exoG is released from the
G-binding site, while VG205 binds to the conserved G-binding site (C176vG129) and A178 in
J7/9.0 is extruded from the helix, allowing the
204/179 pair to stack on the first base-pair of P7
(Figure 2). This conformation allows the WC edge
of VG205 to interact with base-pair C176vG129 in
the deep groove of P7. This model allows exchange
of exoG and VG without disrupting interactions
between the P1 splice site helix and the intron core.
Hierarchical assembly of the pre-tRNA
We previously concluded that the Azoarcus L-9
Figure 2. Three-dimensional
models of the pre-tRNA. Models of
the pre-tRNA conformation before
step 1 ((a) and (c)) and step 2 ((b)
and (d)) of self-splicing were
constructed
from
comparative
sequence analysis as described in
the text. (c) and (d) Details of the
active site, viewed from the same
orientation as (a) and (b). Bases
represented as balls-and-sticks in
(c) correspond to exoG (white) and
bases from P1 (green). Ut36, which
forms the UoG pair at the catalytic
site,
is
red. The
base-pair
G129vC176, which binds exoG
during the first step of splicing, is
depicted in purple. The vG is in
orange and lies at the end of P9.
(d) vG has moved to the G-binding
site during step 2, replacing exoG.
The color code is the same as in (c),
except that the scissile phosphate
group (A206) is represented in
orange.
44
ribozyme folds in a hierarchical manner, in so far
as helices in the catalytic core assemble at tenfold
lower concentrations of Mg2þ than are required to
stabilize the tertiary structure.9 At first glance, the
pre-tRNA appeared to follow a similar pathway,
judging from changes in the extent of RNase T1
and hydroxyl radical protection with increasing
concentration of MgCl2. Much of the pre-tRNA
secondary structure is formed in 10 mM Tris – HCl
(pH 7.5) or Tris – HCl plus 20 mM MgCl2 (blue
arrowheads in Figure 1). By contrast, residues in
P3, P6, P7 and P8 that form the core of the intron
were protected from RNase T1 cleavage only in
higher concentrations of Mg2þ (red arrowheads in
Figure 1). For these residues, the midpoints ðCm Þ
for individual transitions were between 0.1 mM
and 0.7 mM MgCl2 (Figure 3A and Table 1).
Typically, even higher concentrations of Mg2þ
were needed to stabilize tertiary interactions
(Cm ¼ 0:4 – 2:5 mM Mg2þ; Table 1).
Azoarcus Group I Pre-tRNA Folding Mechanism
Unlike the L-9 ribozyme, in which tertiary
folding is cooperative, the sensitivity of individual
tertiary contacts in the pre-tRNA to the concentration of Mg2þ varied (Table 1). Hydroxyl radical
footprinting experiments showed that tertiary
interactions within the tRNA and P3, P8 and P9
displayed the lowest midpoints (0.4 – 0.7 mM) and
least cooperativity (Hill coefficient nH < 1) with
respect to the concentration of Mg2þ (blue shading;
Figure 1). Contacts in P4 –P6, the triple helix and
J8/7 were more cooperative (nH < 3; green;
Figure 1). Docking of P1 and P2, which protects
residues in P1, P2 and J4/5 from hydroxyl radical,
required the highest concentrations of Mg2þ ðCm ¼
2:4 mMÞ and were moderately cooperative (yellow;
Figure 1). The variability in the Mg2þ-dependence
of individual tertiary contacts and the partial overlap in the range of concentrations of Mg2þ needed
for helix assembly and tertiary structure suggested
the presence of metastable intermediates in the
equilibrium (and kinetic) folding pathway of the
pre-tRNA.
Pre-tRNA folding intermediates
Figure 3. Magnesium-dependent folding of pre-tRNA.
A, Native gel electrophoresis of uniformly 32P-labelled
pre-tRNA. The RNA was equilibrated in 25 mM
Na-Hepes (pH 7.0), 10% (v/v) glycerol, 0.1% (w/v)
xylene cyanol and increasing concentrations of MgCl2
for 15 minutes at 50 8C, then loaded onto a native 10%
polyacrylamide gel containing 3 mM MgCl2 (see
Materials and Methods). Band I, folding intermediates;
N, native RNA. B, Magnesium-dependence of folding
equilibrium. The relative saturation of RNase T1 and
Fe-EDTA footprinting and the fraction of native
pre-tRNA ðfN Þ from native gels were fit to the Hill
equation (Table 1). (O) RNase T1 (nucleotides 134– 138),
Cm ¼ 0:14 ^ 0:01, nH ¼ 1:2 ^ 0:1; (X) native gel,
Cm ¼ 0:12 ^ 0:01, nH ¼ 3:2 ^ 0:6; (A) Fe-EDTA (nucleotides 57 – 58), Cm ¼ 1:8 ^ 0:1, nH ¼ 2:6 ^ 0:5:
To determine whether the pre-tRNA forms metastable intermediates, native and non-native forms
of the pre-tRNA were resolved by native gel
electrophoresis.18 Because the Azoarcus L-9
ribozyme folds rapidly ðt ¼ 50 msÞ, 40% of the
ribozyme RNA reaches the native state (N) during
the brief time that the sample encounters Mg2þ in
the gel running buffer.9 The native RNA appears
as a sharp band in the gel, and further refolding is
largely arrested after the RNA enters the gel
matrix. By contrast, only 10% of the pre-tRNA
folded correctly in the deadtime of the native
gel assay (Figure 3B), when the RNA was
not pre-incubated in buffer containing Mg2þ.
The remainder of the RNA was trapped in
misfolded intermediates (I), which appeared as a
cluster of bands with slower gel mobility
(Figure 3B).
The fraction of native pre-tRNA in the gel
increased if samples were annealed in Mg2þ before
electrophoresis (Figure 3B). The midpoint of the
folding transition assayed by native PAGE was
0.12(^ 0.01) mM, similar to the Mg2þ concentration
range over which individual G bases became
protected from RNase T1 (Figure 3A). Therefore,
formation of native helices within the core of the
intron increased the fraction of pre-tRNA able to
rapidly form the native tertiary structure. However, the Hill coefficients for saturation of RNase
T1 protection at individual G bases varied from 1.2
to 3.4, compared to 3.2 from native PAGE. This
variability in the folding of different residues, plus
the presence of slowly migrating species on the
native gel in low concentrations of Mg2þ or at
short times, indicate that the pre-tRNA adopts
alternative conformations that compete with the
native state.
45
Azoarcus Group I Pre-tRNA Folding Mechanism
Table 1. Mg2þ-dependence of folding equilibrium
Secondary structureb
Nucleotidesa
Cm (mM)
P1
P2
J3/4
J4/5
J5/5a
P4
P6a
P6
P7
P3
J8/8a
P8a
P8
J8/7
P9.0
P9
tRNA
tRNA
8–12
26–27
48–49
57–58
62
88–89
99–100
121–124
127–129
134–138
147–148
150–151
162–163
169–168
179–181
191–192
t47 –t48
t58 –t60
Tertiary structurec
nH
0.38 ^ 0.05
3.4 ^ 1.8
0.22 ^ 0.04
0.21 ^ 0.03
0.14 ^ 0.04
1.9 ^ 0.4
1.2 ^ 0.2
1.12 ^ 0.3
0.62 ^ 0.06
0.66 ^ 0.05
0.56 ^ 0.16
0.35 ^ 0.03
2.1 ^ 0.5
2.6 ^ 0.5
1.6 ^ 0.7
1.3 ^ 0.18
Cm (mM)
nH
2.3 ^ 0.3
2.4 ^ 0.3
1.1 ^ 0.1
1.9 ^ 0.1
1.5 ^ 0.1
1.5 ^ 0.9
1.4 ^ 0.1
1.2 ^ 0.1
2.0 ^ 0.3
2.7 ^ 0.6
1.3 ^ 0.4
2.8 ^ 0.5
3.4 ^ 0.8
3.1 ^ 0.5
4.4 ^ 1
2.4 ^ 0.7
0.5 ^ 0.1
0.6 ^ 0.1
1.5 ^ 0.3
0.8 ^ 0.2
1.3 ^ 0.1
0.6 ^ 0.1
0.8 ^ 0.1
0.7 ^ 0.1
0.4 ^ 0.1
3 ^ 0.8
1.1 ^ 0.1
1.3 ^ 0.3
1.2 ^ 0.3
1.5 ^ 0.7
a
Range of residues on which quantification is based; this range may be narrower than the protected region shown in Figure 1.
Relative extent of RNase T1 cleavage (see Materials and Methods). Data were fit to the Hill equation; Cm is the midpoint of the
transition; nH is the Hill coefficient. Errors are from the precision of the fit. The variation between two trials was 10–15%.
c
Relative extent of hydroxyl radical-dependent cleavage.
b
Mg21-dependence of folding kinetics
The Mg2þ-dependence of the folding kinetics
was measured using native PAGE (Figure 4). In
20 mM MgCl2, 10% of the pre-tRNA folded in the
deadtime of the assay (, 30 s), and the remainder
reached the native state at a rate of 0.47 min21
(Figure 4A). Thus, the average folding time of the
pre-tRNA is much longer than that of the
ribozyme,9 consistent with the presence of metastable intermediates.
The observed folding rate of the Azoarcus pretRNA increased significantly as the concentration
of Mg2þ was raised from 0 to 1 mM, and continued
to increase more slowly over the range of 1 – 15 mM
Mg2þ (Figure 4B). These results were not consistent
with a two-state transition. Instead, the results
indicated a kinetic intermediate that is stabilized
by Mg2þ. Population of this intermediate at higher
concentrations of Mg2þ accelerates the formation
of native pre-tRNA. Candidates for such an
ensemble of intermediates were structures in
which the helices in the core of the intron are
assembled, because RNase T1 assays (Figure 1 and
Table 1) reported the formation of base-pairs in
the intron core over the same range of Mg2þ concentrations (0 –1 mM MgCl2,) in which the folding
rate increased.
In the Tetrahymena ribozyme, misfolding of the
core is stabilized by native and non-native tertiary
interactions in the P4 – P6 domain and peripheral
helices that must reorganize to reach the native
state of the core.19 – 21 Consequently, the folding
rate decreases with increasing concentration of
Mg2þ, because the folding intermediates are more
stable and less likely to unfold.22,23 For the pre-
tRNA, however, kobs continued to increase with
increasing concentration of Mg2þ, even beyond
saturation of the folding equilibrium at 3 mM
MgCl2 (Figure 4B). This suggests that reorganization of tertiary contacts is not rate determining.
Mispairing of intron and tRNA
To gain insight into whether folding of the pretRNA requires the unfolding of non-native
intermediates, folding reactions were done in the
presence of 0 –4 M urea (Figure 5). The folding
rates of some proteins and RNA are enhanced by
the addition of sub-denaturing concentrations of
urea, because misfolded intermediates are
destabilized relative to the native structure.24 – 26
Consistent with this expectation, the observed pretRNA folding rate increased fourfold in 4 M urea,
from 0.25 min21 to 1.1 min21. The amplitude of the
folding reaction changed very little over this
range, as 4 M urea is slightly below the midpoint
of the equilibrium denaturation. Because there
was little change in the apparent stability of the
native state under these conditions, the faster folding kinetics is most likely due to the destabilization
of misfolded intermediates. Together with the
Mg2þ-dependence of the kinetics, these results
suggest that formation of native pre-tRNA is
limited primarily by the rearrangement of
secondary structure elements, rather than
reorganization of tertiary interactions as in the
Tetrahymena ribozyme.
Since the results above suggested a slow
rearrangement in the secondary structure, RNase
T1 reactions were carried out at different times
during folding at 32 8C as described in Materials
46
Figure 4. Magnesium-dependence of folding kinetics.
A, Folding kinetics of the pre-tRNA at 50 8C was
measured by native PAGE: (B) 0.07 mM MgCl2, (W)
0.08 mM MgCl2, (X) 0.2 mM MgCl2, (O) 15 mM MgCl2.
The data were fit to a first-order rate equation
(kobs ¼ 0:035, 0.086, 0.28, and 0.47 min21, respectively).
B, Observed folding rates (X) as a function of Mg2þ concentration were fit to a three-state model, as described
by equation (1) (see Materials and Methods). Parameters
of the fit were k2 ¼ 0:38 min21 ; k22 ¼ 0:25 min21 ;
Cm;1 ¼ 0:22 mM; Cm;2 ¼ 39 mM; n ¼ 1:7: Error bars
represent the statistical uncertainty of individual
progress curves. The amplitude of folding reactions (W)
was fit to the Hill model as in Figure 3A, with Cm ¼
0:07 mM and n ¼ 5:3. The extent of folding after
60 minutes was 65 –70% due to residual misfolded RNA.
and Methods. In 15 mM MgCl2, most G bases were
protected from RNase T1 within the shortest time
of the assay (1.5 minutes; Figure 6B). However, G
bases in P3, P7, P9.0 and in the tRNA were protected more slowly (0.05 – 0.15 min21; Table 2), at a
rate roughly similar to the global folding rate
measured by native gel electrophoresis at this
temperature (kobs ¼ 0:17 min21 ; 32 8C). Residues in
P8 (G150-151) achieved 60% of maximal protection
in the first minute. This was followed by a slower
base-pairing transition (0.05 min21).
Delayed formation of helices P9.0 and P7 and the
resulting sensitivity of these residues to RNase T1
Azoarcus Group I Pre-tRNA Folding Mechanism
Figure 5. Acceleration of folding in urea. A, Folding of
pre-tRNA in 15 mM MgCl2 was monitored as a function
of urea concentration. The data were fit to a first-order
rate equation: (X), no urea; (W) 2 M urea. B, Observed
rate constants (X) as a function of urea concentration
were fit to a linear free energy model for the forward
rate constant in equation (3). Parameters were
kf ð0Þ ¼ 0:27 min21 ; mf ¼ 0:1 kcal mol21 M21 ; T ¼ 323 K.
Error bars as in Figure 4. The extent of folding after 15
minutes (W) was fit to the linear extrapolation model
(equation (2)). The fit yielded approximate values for
DGNU ð0Þ ¼ 9 kcal mol21 ; m ¼ 2 kcal mol21 M21 :
can be explained by interactions between nucleotides 175 –184 in the intron and a complementary
sequence between nucleotides 58– 67 in the tRNA
(Figure 6C). This non-native helix would disrupt
the native secondary structure of the tRNA as well
as P9.0 and P7, and could explain why the pretRNA is more prone to becoming trapped in
misfolded states than the L-9 ribozyme. By contrast, a shorter precursor that includes only the
anticodon stem of the tRNA folds and self-splices
very rapidly (P.R., unpublished results). We noted
that a GGUG sequence in P3 could exchange places
with an identical sequence in P8, destabilizing P3.
Mispairing of these helices is expected to impede
the formation of tertiary contacts. Accordingly,
residues in P3 – P9 and the tRNA that were slowly
protected from RNase T1 folded less cooperatively,
as judged by hydroxyl radical footprinting
(Table 1).
47
Azoarcus Group I Pre-tRNA Folding Mechanism
Figure 6. Ribonuclease probing of folding intermediates. Folding of (50 -32P)-labelled RNA at 32 8C was initiated with
addition of 15 mM Mg2þ. At various times, aliquots were treated with 0.1 unit of RNase T1 for one minute at 32 8C as
described in Materials and Methods. A, A 6% polyacrylamide sequencing gel.; lane G, RNase T1 under denaturing conditions; lane C, no T1 control; 0 – 15 minutes, folding reactions. B, Band intensity versus time was fit to: fN ðtÞ ¼
fN ð0Þ þ Að1 2 expð2kobs tÞÞ as described in Materials and Methods. (V) G129, kobs ¼ 0:06 min21 ; (K) G150 – 151,
fN ð0Þ ¼ 0:75, kobs ¼ 0:05 min21 ; (W) G174, kobs ¼ 0:11 min21 : C, Proposed alternative pairings that cause misfolding of
the pre-tRNA. Red, sequences at the 30 end of the intron that are complementary to the 30 end of the tRNA; the blue
sequence in P8 could displace the 30 strand of P3 (green).
Discussion
Modelling of the architecture
The nature of conformational changes between
the first and second steps of splicing by group I
introns is a long-standing problem. Because G
nucleotides are positioned within the active site
by a single binding site in P7, 27 the G cofactor
(exoG) must be exchanged for the 30 -terminal G
(VG) between steps 1 and 2.28 In tRNA introns,
pairing of the 50 and 30 exons in the folded
tRNA leads to additional constraints. In most
introns, the discrimination of the 30 splice site
involves the P9.0 and P10 pairings, which
bring, respectively, the VG at the G-binding site
and the 50 and 30 exons in register. In some
Table 2. Folding kinetics monitored by RNase T1
protection
Nucleotides
f Nð0Þ
kobs (min21)
G98
G118
G127
G129
G134
G137–G138
G150–G151
G162–G163
G168–G169
G174
tG65–tG77
–a
0.5
0.43
–
–
–
0.74
0.33
–
–
–
0.08
0.05
0.14
0.11
0.11
0.15
0.05
0.10
0.06
0.06
0.05
Parameters from fits as described in Materials and Methods;
f N(0) is the fraction protected in the first minute relative to no
MgCl2 control sample; kobs is the observed rate constant for the
slow kinetic phase. Errors are 10– 20%.
a
None observed.
introns (belonging to IB4, IA1, and IA2
subgroups), an additional residue between the 30
end of P7 and the 50 end of P9.0 forms the P9.0a
pair with the N residue of a conserved AANA
sequence upstream of the 30 end of P9.0.14,29,30 The
latter introns have a large P9 domain. In the
Azoarcus intron, that conserved sequence is
absent and it is suggested that the additional A
between P70 and P9.0 forms a GoA pair with the
terminal G in the first step. After cleavage, the
resulting accomodation promotes bulging out of
the A and insertion of the VG in the G-binding
site. The models presented here give a picture of
the catalytic site just before steps 1 and 2 of selfsplicing and suggest how G exchange might be
achieved.
Experiments on the Tetrahymena intron showed
that the 50 exon– intron complex is long-lived, and
that its stability depends on esterification of
VG.31,32 These data and direct measurements of
substrate association33 implied that P1 remains
docked with J4/5 during the catalytic steps of
splicing, even while undergoing molecular
motions like the unwinding of the 50 end of the
intron. Since the docking of P1 and J4/5 is maintained, we conclude that the main conformational
changes between steps 1 and 2 are more likely to
be centered around P9.0. The structure proposed
for J9/9.0 and J7/9.0 in the Azoarcus pre-tRNA is
the simplest one that accomodates all known
interations. A confomational change in J7/9.0
allows G205 to slide down and replace exoG in
the active site, without disrupting P1 docking
interactions.
Folding pathway of the pre-tRNA
As for the L-9 ribozyme RNA, nuclease and
48
hydroxyl radical footprinting show that the
equilibrium folding pathway of the pre-tRNA is
hierarchical, in that helices in the intron core form
before native tertiary contacts. Although many
elements of the secondary structure are stable in
low-salt buffer (U), assembly of helices in the core
of the intron (IC) requires moderate amounts of
Mg2þ (, 0.1 mM). The transition from U to IC is
accompanied by a large contraction of the RNA
chain as monitored by small-angle neutronscattering (Perez-Salas et al.34). At higher concentrations of Mg2þ (, 2 mM), tertiary interactions
stabilize a solvent-inaccessible core and docking of
the P1 splice site helix (N).
As summarized in Figure 7, only a small fraction
of the pre-tRNA (, 10%) forms the native structure
within the deadtime of the native gel assay. From
our previous work on the Azoarcus ribozyme,9 we
expect that direct transitions from U to IC to N
occur in 50 –100 ms. This is supported by X-raydependent
hydroxyl
radical
footprinting
experiments, which revealed partially saturated
native tertiary interactions throughout the pretRNA within this time-frame (P.R. & S.W.,
unpublished results).
In comparison with the ribozyme, the Azoarcus
pre-tRNA has a greater tendency to form nonnative base-pairs, resulting in metastable
intermediates that are observed by native gel
electrophoresis (Figure 7). The increase in the
folding rate of the pre-tRNA between 0.1 mM and
1 mM Mg2þ correlates with the assembly of helices
in the intron core, as detected by RNase T1
protection. Since the native gels monitor the
fraction of intermediates that can form N rapidly,
the net folding rate increases as the ensemble of I
states becomes more like the native state. Urea
accelerates folding presumably by destabilizing
intermediates that are incorrectly base-paired. It is
not known whether the intermediates must unfold
Azoarcus Group I Pre-tRNA Folding Mechanism
completely in order to reach N, or whether basepairing rearrangements occur in the context of a
compact state.
If pre-tRNA misfolding involves improper basepairing between the P9 region and downstream
nucleotides in the 30 half of the tRNA, then sequential folding should yield more native pre-tRNA,
because the correct P9 base-pairs will have a
chance to form before the 30 exon is transcribed. In
preliminary experiments, co-transcriptional folding reactions at 37 8C yielded a burst of 40% native
pre-tRNA, compared with 5– 10% in refolding
reactions (data not shown). Proteins may stimulate
folding of the pre-tRNA. Thus, a greater fraction
of pre-tRNAile is expected to fold correctly in vivo
than in vitro.
A consequence of non-native interactions
between the tRNA exons and the intron core is
that neither the tRNA nor the intron can fold
correctly until this mispairing is resolved. Because
the Azoarcus intron and many tRNAs fold rapidly
on their own, we imagine that native tertiary
structure will follow quickly once the correct
base-pairs are in place. Thus, folding of the
intron and tRNA is effectively coupled via
reorganization of the secondary structure.
Although this interaction between P9 and the
T-stem is not conserved among group I introns in
bacterial tRNAs, it is interesting to speculate
whether such interactions serve a biological
function by coordinating folding of the intron and
exons. For example, this could help prevent 50 and
30 processing of pre-tRNAs that are not competent
to self-splice.
Materials and Methods
RNA preparation
In vitro transcription of the 284 nt pre-tRNAile
from pAZ-PREt was carried out as described.9,35
The pre-tRNA contains the sequences of the mature
tRNA domain ending in CCA (79 nt) plus the 205 nt
intron.
RNase T1 cleavage
Figure 7. Summary of folding mechanism. About
80 – 90% of the pre-tRNA collapses to misfolded intermediates (I), which may involve mispairing between
tRNA and intron residues. Rearrangement of secondary
structure elements increases the time needed to form
native tertiary structure in the tRNA and the intron
domains. Direct transitions through IC are estimated to
require 100 ms (see the text).
Equilibrium footprinting reactions were carried out
essentially as described,9 except that the RNA was
annealed for 15 minutes before the addition of RNase
T1. The counts in each band as a function of Mg2þ concentration were fit to the Hill equation. For time-dependent studies, the 32P-labelled RNA was incubated at
32 8C and folding was initiated with addition of MgCl2
(15 mM
final
concentration).
Aliquots
(10 ml;
150,000 cpm) were removed at various times and treated
with 0.1 unit of RNase T1 for one minute at 32 8C.
Digestion was stopped with 1 mM aurin tricarboxylic
acid (Sigma) and an equal volume of buffered formamide. The reactions were loaded onto a sequencing 6%
(w/v) polyacrylamide gel. Dried gels were exposed to
Phosphorimager screens and quantified using ImageQuant software (Molecular Dynamics). The counts in
each band were normalized to their intensity in unfolded
49
Azoarcus Group I Pre-tRNA Folding Mechanism
(no Mg2þ) and folded RNA (15 mM MgCl2). The relative
extent of protection was plotted as a function of time
and fit to:
fN ðtÞ ¼ fN ð0Þ þ Að1 2 expð2kobs tÞÞ
where f ð0Þ is the fraction of native RNA after 30 seconds
and A ¼ fmax 2 f ð0Þ.
DfN ¼
with respect to Mg2þ. Because the ions are responding to
changes in the electrostatic field of the RNA rather than
mass action,38 the Hill constant reflects the relative
stabilization of U, I and N by Mg2þ, rather than the precise
number of excess magnesium ions associated with the
folded RNA.
The urea dependence of the kinetics was fit to the
linear extrapolation model:39
ðfN ð0Þ þ mN ½DÞ þ ðfU ð0Þ þ mU ½DÞexp½ð2DGUN þ m½DÞ=RT
1 þ exp½ð2DGUN þ m½DÞ=RT
Hydroxyl radical footprinting
Fe(II)-EDTA cleavage reactions were carried out
essentially as described,9 except that the 32P-labelled
RNA was incubated for 15 minutes at 50 8C before
addition of Fe(II)-EDTA.
Native gel folding assay
To measure the magnesium-dependence of folding,
uniformly labelled [32P]RNA was incubated in 25 mM
Na-Hepes (pH 7.0), 10% glycerol, 0.1% xylene cyanol,
and 0 –15 mM MgCl2 for 30 minutes at 50 8C, then loaded
onto a 10% polyacrylamide gel containing 3 mM
MgCl2.18 For urea denaturation, RNA was incubated in
buffer plus 15 mM MgCl2 and 0 – 8 M urea for 15 minutes
at 50 8C. Samples were electrophoresed at 15 W for three
to four hours at 2 – 8 8C. The fraction of native RNA, f N,
was determined from the ratio of counts in band N
(Figure 1) to total counts in the lane. The fraction of
native RNA as a function of Mg2þ concentration was
normalized to the extent of folding at saturation and fit to
the Hill equation. The maximum extent of folding at 50 8C
was typically 85%, presumably reflecting a small population that remained trapped in misfolded conformations.
ð2Þ
where ½D is the concentration of urea, and the first two
terms represent the urea-dependence of the native (N)
and unfolded (U) states, respectively. The Mg2þ-dependence of the folding kinetics below 4 M urea were fit to:
kobs < kf ¼ kf ð0Þexpðmf ½D=RTÞ
ð3Þ
where kf is the observed rate of the forward reaction, kf(0)
is the rate in 0 M urea, and mf reflects the dependence of
the transition state free energy on urea.
Molecular modelling
The crystal structure of tRNA Asp40 was interactively
added onto the 3D model of Azoarcus group I intron9
using the package of programs Manip.41 Models were
subjected to restrained geometrical least-squares refinement using the program Nuclin/Nuclsq40,42 in order to
ensure geometry and stereochemistry with correct distances between interacting atoms and to avoid steric
clashes. Interactive modelling followed by refinement
steps were performed iteratively until solution data
reported in this work could be explained satisfactorily.
The color views were generated with the program
Drawna.43 The accessibilities of the C40 atoms to the
hydroxyl radicals were computed using the program
Naccess.†
Folding kinetics
Uniformly labelled pre-tRNA was pre-incubated for
two minutes at 50 8C in 25 mM Na-Hepes (pH 7.0), 10%
glycerol, 0.1% xylene cyanol. Folding was initiated by
addition of 0.1 – 20 mM MgCl2. Aliquots (2 ml) were
removed at various times and immediately loaded onto
a 10% polyacrylamide gel as described above. Where
indicated, RNA was incubated in 0 – 4 M urea, then folding reactions were begun with the addition of 15 mM
MgCl2. Progress curves were fit to a first-order rate
equation as above.
The simplest mechanism that described the folding
kinetics (after the initial burst of 0 – 10% N) was a threestate model UIN, in which transitions between the
unfolded (U) and intermediate (I) ensembles were
assumed to be faster than the transitions from I to N. The
Mg2þ dependence of folding was fit to equation (1), where
C is Mg2þ concentration, Cm,1 and Cm,2 are the midpoints
of the U to I and I to N transitions, respectively, n is the
Hill constant, and k2 and k22 are the forward and reverse
rate constants for the I to N transition:
kobs ¼
Cn
k2 Cn
k22 C
þ
C þ Cm;2
þ Cnm;1
ð1Þ
A binding model was used to approximate coupling
between Mg2þ binding and RNA folding transitions.36,37
Only the U to I transition was assumed to be cooperative
Acknowledgements
This work was supported by grants from the
National Institutes of Health (to S.A.W.) and the
Institut Universitaire de France (to E.W.).
References
1. Cech, T. R. (1993). Structure and mechanism of the
large catalytic RNAs: group I and group II introns
and ribonuclease P. In The RNA World (Gesteland,
R. F. & Atkins, J. F., eds), pp. 239 –269, Cold Spring
Harbor Library Press, Plainview, NY.
2. Thirumalai, D. & Woodson, S. A. (1996). Kinetics of
folding of protein and RNA. Acc. Chem. Res. 29,
433 –439.
3. Perrotta, A. T. & Been, M. D. (1998). A toggle duplex
in hepatitis delta virus self-cleaving RNA that
stabilizes an inactive and a salt-dependent pro-active
ribozyme conformation. J. Mol. Biol. 279, 361– 373.
4. Chadalavada, D. M., Knudsen, S. M., Nakano, S. &
Bevilacqua, P. C. (2000). A role for upstream RNA
† http://wolf.bms.umist.ac.uk/naccess/
50
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Azoarcus Group I Pre-tRNA Folding Mechanism
structure in facilitating the catalytic fold of the
genomic hepatitis delta virus ribozyme. J. Mol. Biol.
301, 349– 367.
Woodson, S. A. & Cech, T. R. (1991). Alternative
secondary structures in the 50 exon affect both
forward and reverse self-splicing of the Tetrahymena
intervening sequence RNA. Biochemistry, 30,
2042– 2050.
Semrad, K. & Schroeder, R. (1998). A ribosomal
function is necessary for efficient splicing of the T4
phage thymidylate synthase intron in vivo. Genes
Dev. 12, 1327– 1337.
Reinhold-Hurek, B. & Shub, D. A. (1992). Selfsplicing introns in tRNA genes of widely divergent
bacteria. Nature, 357, 173– 176.
Zaug, A. J., McEvoy, M. M. & Cech, T. R. (1993). Selfsplicing of the group I intron from Anabaena pretRNA: requirement for base-pairing of the exons in
the anticodon stem. Biochemistry, 32, 7946– 7953.
Rangan, P., Masquida, B., Westhof, E. & Woodson,
S. A. (2003). Assembly of core helices and rapid
tertiary folding of a small bacterial group I ribozyme.
Proc. Natl Acad. Sci. USA, 100, 1574– 1579.
Michel, F., Jacquier, A. & Dujon, B. (1982). Comparison of fungal mitochondrial introns reveals extensive
homologies in RNA secondary structure. Biochimie,
64, 867– 881.
Sprinzl, M., Horn, C., Brown, M., Ioudovitch, A. &
Steinberg, S. (1998). Compilation of tRNA sequences
and sequences of tRNA genes. Nucl. Acids Res. 26,
148–153.
Rangan, P. & Woodson, S. A. (2003). Structural
requirement for Mg2þ binding in the group I intron
core. J. Mol. Biol. 329, 229– 238.
Latham, J. A. & Cech, T. R. (1989). Defining the inside
and outside of a catalytic RNA molecule. Science, 245,
276– 282.
Michel, F. & Westhof, E. (1990). Modelling of the
three-dimensional architecture of group I catalytic
introns based on comparative sequence analysis.
J. Mol. Biol. 216, 585– 610.
Strobel, S. A., Ortoleva-Donnelly, L., Ryder, S. P.,
Cate, J. H. & Moncoeur, E. (1998). Complementary
sets of noncanonical base pairs mediate RNA helix
packing in the group I intron active site. Nature
Struct. Biol. 5, 60 – 66.
Soukup, J. K., Minakawa, N., Matsuda, A. & Strobel,
S. A. (2002). Identification of A-minor tertiary interactions within a bacterial group I intron active site
by
3-deazaadenosine
interference
mapping.
Biochemistry, 41, 10426– 10438.
Cech, T. R. (1990). Self-splicing of group I introns.
Annu. Rev. Biochem. 59, 543– 568.
Emerick, V. L. & Woodson, S. A. (1994). Fingerprinting the folding of a group I precursor RNA.
Proc. Natl Acad. Sci. USA, 91, 9675– 9679.
Treiber, D. K., Rook, M. S., Zarrinkar, P. P. &
Williamson, J. R. (1998). Kinetic intermediates
trapped by native interactions in RNA folding.
Science, 279, 1943– 1946.
Pan, J. & Woodson, S. A. (1999). The effect of longrange loop– loop interactions on folding of the Tetrahymena self-splicing RNA. J. Mol. Biol. 294, 955– 965.
Russell, R., Zhuang, X., Babcock, H. P., Millett, I. S.,
Doniach, S., Chu, S. & Herschlag, D. (2002).
Exploring the folding landscape of a structured
RNA. Proc. Natl Acad. Sci. USA, 99, 155– 160.
Pan, J., Thirumalai, D. & Woodson, S. A. (1999). Magnesium-dependent folding of self-splicing RNA:
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
exploring the link between cooperativity, thermodynamics, and kinetics. Proc. Natl Acad. Sci. USA,
96, 6149– 6154.
Rook, M. S., Treiber, D. K. & Williamson, J. R. (1999).
An optimal Mg(2 þ ) concentration for kinetic
folding of the Tetrahymena ribozyme. Proc. Natl
Acad. Sci. USA, 96, 12471– 12476.
Kiefhaber, T., Grunert, H. P., Hahn, U. & Schmid, F. X.
(1992). Folding of RNase T1 is decelerated by a
specific tertiary contact in a folding intermediate.
Proteins: Struct. Funct. Genet. 12, 171– 179.
Pan, J., Thirumalai, D. & Woodson, S. A. (1997).
Folding of RNA involves parallel pathways. J. Mol.
Biol. 273, 7 – 13.
Pan, T. & Sosnick, T. R. (1997). Intermediates and
kinetic traps in the folding of a large ribozyme
revealed by circular dichroism and UV absorbance
spectroscopies and catalytic activity. Nature Struct.
Biol. 4, 931– 938.
Michel, F., Hanna, M., Green, R., Bartel, D. P. &
Szostak, J. W. (1989). The guanosine binding site of
the Tetrahymena ribozyme. Nature, 342, 391– 395.
Been, M. D. & Perrotta, A. T. (1991). Group I intron
self-splicing with adenosine: evidence for a single
nucleoside-binding site. Science, 252, 434– 437.
Michel, F., Jaeger, L., Westhof, E., Kuras, R., Tihy, F.,
Xu, M. Q. & Shub, D. A. (1992). Activation of the
catalytic core of a group I intron by a remote 30 splice
junction. Genes Dev. 6, 1373– 1385.
Jaeger, L., Westhof, E. & Michel, F. (1993). Monitoring
of the cooperative unfolding of the sunY group I
intron of bacteriophage T4. The active form of the
sunY ribozyme is stabilized by multiple interactions
with 30 terminal intron components. J. Mol. Biol. 234,
331– 346.
Mei, R. & Herschlag, D. (1996). Mechanistic investigations of a ribozyme derived from the Tetrahymena
group I intron: insights into catalysis and the second
step of self- splicing. Biochemistry, 35, 5796– 5809.
Emerick, V. L., Pan, J. & Woodson, S. A. (1996).
Analysis of rate-determining conformational changes
during self- splicing of the Tetrahymena intron.
Biochemistry, 35, 13469– 13477.
Bevilacqua, P. C., Sugimoto, N. & Turner, D. H.
(1996). A mechanistic framework for the second step
of splicing catalyzed by the Tetrahymena ribozyme.
Biochemistry, 35, 648– 658.
Perez-Salas, U. A., Raugan, P., Krueger, S., Briber,
R. M., Thirumalai, D. & Woodson, S. A. (2004). Compaction of a bacterial group I ribozyme coincides
with the assembly of core helices. Biochemistry, 43,
1746– 1753.
Tanner, M. & Cech, T. (1996). Activity and thermostability of the small self-splicing group I intron in
the pre-tRNA(lle) of the purple bacterium Azoarcus.
RNA, 2, 74 – 83.
Gluick, T. C., Gerstner, R. B. & Draper, D. E. (1997).
Effects of Mg2þ, Kþ, and Hþ on an equilibrium
between alternative conformations of an RNA
pseudoknot. J. Mol. Biol. 270, 451– 463.
Fang, X. W., Thiyagarajan, P., Sosnick, T. R. & Pan, T.
(2002). The rate-limiting step in the folding of a large
ribozyme without kinetic traps. Proc. Natl Acad. Sci.
USA, 99, 8518– 8523.
Laing, L. G., Gluick, T. C. & Draper, D. E. (1994).
Stabilization of RNA structure by Mg ions. Specific
and non-specific effects. J. Mol. Biol. 237, 577– 587.
Bolen, D. W. & Santoro, M. M. (1988). Unfolding
free energy changes determined by the linear
Azoarcus Group I Pre-tRNA Folding Mechanism
extrapolation method. 2. Incorporation of delta G
degrees N –U values in a thermodynamic cycle.
Biochemistry, 27, 8069– 8074.
40. Westhof, E., Dumas, P. & Moras, D. (1985).
Crystallographic refinement of yeast aspartic acid
transfer RNA. J. Mol. Biol. 184, 119 – 145.
41. Massire, C. & Westhof, E. (1998). MANIP: an interactive tool for modelling RNA. J. Mol. Graph. Model.
16, 197– 205.
51
42. Westhof, E. (1993). Modeling the 3-dimensional
structure of ribonucleic-acids. Theochem. -J. Mol.
Struct. 105, 203– 210.
43. Massire, C., Gaspin, C. & Westhof, E. (1994).
Drawna—a program for drawing schematic views
of nucleic-acids. J. Mol. Graph. 12, 201– 206.
44. Steinberg, S., Misch, A. & Sprinzl, M. (1993).
Compilation of tRNA sequences and sequences of
tRNA genes. Nucl. Acids Res. 21, 3011 – 3015.
Edited by J. Doudna
(Received 18 December 2003; received in revised form 2 March 2004; accepted 3 March 2004)