ABSTRACT
M1 RNA, the catalytic RNA subunit of RNase P from
Escherichia coli,
has been covalently linked at its 3
'
terminus to oligonucleotides (guide sequences) that guide the enzyme to target
RNAs through hybridization with the target sequences. These constructs (M1GS
RNAs) have been used to determine some minimal features of model substrates. As
few as 3 bp on the 3
'
side of the site of cleavage in a substrate complex and 1 nt on the 5
'
side are required for cleavage to occur. The cytosines in the 3
'
terminal CCA sequence of the model substrates are important for cleavage
efficiency but not cleavage site selection. A purine (base-paired or not) at the 3'
side of the cleavage site is important both for cleavage site selection and
efficiency. M1GS RNAs provide both a simple system for characterization of the reaction governed by M1 RNA and a tool for gene therapy.
RNase P is a ribonucleoprotein complex responsible for the maturation of the 5' termini of tRNAs (
1
,
2
). It catalyzes a hydrolysis reaction that removes a 5' leader sequence from tRNA precursors (pretRNA) and several other small RNAs. In
Escherichia coli
, RNase P consists of a catalytic RNA subunit [M1 RNA: 377 nucleotides (nt)] and
a protein subunit (C5 protein: 119 amino acids) (
1
,
2
). In the presence of a high concentration of salt, such as 100 mM Mg
2+
, M1 RNA acts as a catalyst and cleaves pretRNAs
in
vitro
in the absence of C5 protein (
3
). The addition of C5 protein dramatically increases the rate of cleavage by M1 RNA
in vitro
and is required for RNase P activity and cell viability
in vivo
(
4
).
Studies of substrate recognition by M1 RNA and RNase P (
1
,
5
-
10
) have led to the development of a general strategy of gene targeting in which
M1 RNA and RNase P can be used as tools to cleave any specific mRNA sequence.
In a small model substrate for M1 RNA [Fig.
1
A (boldface regions) and Fig.
1
B], the 5' proximal sequence (the 5' leader and 5' proximal acceptor stem sequence) base pairs to the 3' proximal sequence (the 3' proximal acceptor stem sequence) (
5
). This 3' proximal sequence is called an external guide sequence (EGS) because it
can base pair with the targeted sequence and guide M1 RNA to cleave the
substrate (Fig.
1
B). Subsequent studies carried out in our laboratory have demonstrated that the
mRNAs encoding
E.coli
[beta]-galactosidase and
S.aureus
nuclease A can be cleaved by M1 RNA and
E.coli
RNase P
in vitro
with custom-designed EGSs (
6
). One limitation of this method is the relatively weak binding of the target
RNA to M1 RNA (
8
). In order to increase cleavage efficiency and strengthen substrate binding, we
constructed a derivative of M1 RNA, M1GS RNA, by linking a guide sequence (GS)
covalently to the 3' end of M1 RNA (Fig.
1
C) (
11
). Similar constructs have also been described recently in studies of M1 RNA-pretRNA conjugated molecules (
12
-
17
).
DNA templates for
in vitro
transcription of RNA substrates phe7 (identical to phe7-G), phe3, phe7-A, phe7-C and phe7-U were constructed (
18
) by annealing the T7 promoter-containing oligonucleotide, OliT7 (5'-TAATACGACTCACTATAG-3') with oligonucleotides:
Oli58 (5'-TCCGGGCGGTCCTATAGTGAGTCGTATTAA-3'),
Oli51 (5'-GGCGGTCCTATAGTGAGTCGTATTAA-3'),
Oli84 (5'-TCCGGGTGGTCCTATAGTGAGTCGTATTAA-3'),
Oli85 (5'-TCCGGGGGGTCCTATAGTGAGTCGTATTAA-3') and
Oli86 (5'-TCCGGGAGGTCCTATAGTGAGTCGTATTAA-3')
respectively. The RNA substrates were synthesized by T7 RNA polymerase (Promega
Inc., Madison, WI) from these DNA templates, according to the manufacturer's
instructions. RNA substrate phe7-1 (5'-CGCCCGGA-3') was synthesized chemically with a 380B DNA
synthesizer (Applied Biosystems). All RNA substrates synthesized either chemically or enzymatically were subjected to gel purification in 15% polyacrylamide gels that contain 7 M
urea.
To generate radioactive RNA substrates, substrates were either uniformly labeled
with [[alpha]-
32
P]GTP by T7 RNA polymerase, 5'-end-labeled with [[gamma]-
32
P]ATP in the presence of T4 polynucleotide kinase or 3'-end-labeled with [
32
P]pCp in the presence of T4 RNA ligase.
Plasmids pTK117, pM1[Delta](1-54), pM1[Delta](1-163), pM1[Delta](62-108), pM1[Delta](94-281), pM1[Delta](156-290) and
pM1[Delta](169-377) are derivatives of pUC19, in which the DNA sequences coding for M1 RNA and mutant
M1 RNAs with deletions from nt 1 to 54, 1 to 163, 61 to 108, 94 to 281, 156 to
290 or 169 to 377 are under the control of the T7 RNA polymerase promoter (
19
). The DNA templates for M1PHE (identical to M1PHE-c), M1PHE-C, M1PHE-CC, M1PHE-CCA (identical to M1PHE-CCA-c), M1PHE-a, M1PHE-g, M1PHE-u, M1PHE-CCA-a, M1PHE-CCA-g
and M1PHE-CCA-u were constructed by the polymerase chain reaction (PCR). In the
PCR, the DNA sequence for M1 RNA in plasmid pTK117 was used as the template and
OliT7 was used as the 5' primer oligonucleotide. The 3' primers which contain the appropriate guide sequences were:
Oli52 (5'-
Oli57 (5'-
Oli56 (5'-
Oli55 (5'-
Oli83 (5'-
Oli82 (5'-
Oli81 (5'-
Oli89 (5'-
Oli88 (5'-
Oli87 (5'-
The 3' proximal sequences of 10 nt serve as the primers for the PCR with the
pUC19 sequence. The underlined sequences and the bold sequences correspond to
the 3' ACCA sequence and the guide sequences respectively.
The DNA templates for ribozymes [Delta](1-54)M1PHE, [Delta](1-163)M1PHE, [Delta](62-108)M1PHE, [Delta](94-281)M1PHE, [Delta](156-290)M1PHE
and [Delta](169-377)M1PHE were constructed by PCR in which the DNA sequence coding
for M1 RNA in pM1[Delta](1-54), pM1[Delta](1-163), pM1[Delta](62-108), pM1[Delta](94-281), pM1[Delta](156-290) and
pM1[Delta](169-377) was used as the template respectively. The primers were OliT7
and OliTK13.
In all the M1GS ribozymes described here, the linker sequence connecting the 5' end of the guide sequence and the 3' end of M1 RNA is a 24 nt-long sequence of pUC19 (5'-TATGACCATGATTACGCCAAGCTT-3'). The enzymatic activity of M1GS
RNA is not affected significantly when the length of the linker sequence varies
from 24 to 50 nt (
11
,
17
).
The corresponding RNA enzymes were synthesized from the DNA templates by T7 RNA
polymerase. The synthesized RNA was subjected to gel purification in 4%
polyacrylamide gels which contain 7 M urea. Subsequently, RNA enzyme samples
were eluted from the gel slices and precipitated in the presence of ethanol.
Eventually the RNA enzymes were resuspended into buffer C (50 mM Tris, pH 7.5,
100 mM NH
4
Cl, 10 mM MgCl
2
). Prior to the assays for RNA enzymatic activity, these RNA enzymes were heated
to 75oC for 3 min, and then allowed to renature by cooling slowly to room
temperature.
RNA enzyme (20 nM) and substrate (50 nM), either uniformly labeled or end-labeled, were incubated for 30 min at 37 or 50oC in buffer A (50 mM Tris, pH 7.5, 100 mM NH
4
Cl, 100 mM MgCl
2
) or buffer B (50 mM Tris, pH 7.5, 100 mM NH
4
Cl) that contains MgCl
2
at various concentrations. Reactions were stopped by the addition of 8 M urea
and the cleavage products were then separated on 20% polyacrylamide gels that contain 7 M urea. C5 protein
was purified from
E.coli
as described previously (
20
). The RNase P holoenzyme from
E.coli
was assembled by mixing M1 RNA and C5 protein at a molar ratio of 1:20.
Studies with M1 RNA-pretRNA conjugates have shown that the interactions between M1 RNA and its conjugated pretRNA substrate are similar to those
observed when the ribozyme and the substrate are separated (
13
-
15
). However, it is also important to establish that the cleavage reactions
catalyzed by M1GS RNA share the same characteristics with those catalyzed by M1
RNA. Four sets of experiments were carried out to examine the features of the
M1PHE cleavage reactions.
First, a set of M1GS RNAs were constructed by linking the guide sequence
covalently to the 3' end of several M1 RNA deletion mutants (such as M1 RNA mutants with a
deletion from nt 1 to 54, 1 to 163, 62 to 108, 94 to 281, 156 to 290 and 169 to
377) (
19
,
22
). These mutants did not exhibit RNase P catalytic activity with pre-tRNA substrates when assayed
in vitro
(
19
,
22
). As shown in Figure
2
(lanes 3-7), the M1GS RNAs derived from these mutants were unable to cleave
substrate phe7, indicating that mutations in the catalytic domain of M1 RNA
also abolish the enzymatic activity of M1GS RNA.
In the second set of experiments, the effects of the concentrations of divalent ion (Mg
2+
) on the enzymatic activity of M1GS RNA were investigated. While ribozyme M1PHE
cleaved phe7 in low concentrations of Mg
2+
(10-20 mM MgCl
2
) (
11
,
17
), M1PHE exhibited much more significant enzymatic activity in high
concentrations of Mg
2+
(100 mM) (Fig.
3
, lanes 2-4). A similar dependence on Mg
2+
was observed in the cleavage reactions catalyzed by M1 RNA with pretRNA
substrates and small model substrates (
3
,
23
,
24
). It has been reported that the optimal concentration of Mg
+
for M1 RNA activity is reduced as the ionic strength (i.e., the concentration
of monovalent ions) in the assay buffer increases (
24
). Similar results were also obtained in reactions catalyzed by M1GS RNA (
11
,
16
,
17
).
The cleavage we observed can, in principle, be catalyzed either by the same M1GS
RNA molecule that base pairs to substrate phe7 (
cis
-cleavage) or by another M1GS RNA molecule that does not bind to the
substrate (
trans
-cleavage). However, efficient
trans
-cleavage can only be observed in high concentrations of Mg
2+
. It has been shown previously that at low concentration of Mg
2+
, cleavage of a substrate by an M1GS RNA is much more efficient than cleavage of
the same substrate by M1 RNA in the presence of a separated EGS (
11
,
16
,
17
). These results indicated that
trans
-cleavage is very much diminished and
cis
-cleavage is responsible for the results we report here. Similar
observations have also been reported in the studies of the cleavage of M1 RNA-pretRNA conjugates (
13
-
15
).
The third set of experiments indicated that the cleavage activity of M1PHE ribozyme was greatly stimulated by C5 protein (Fig.
3
, lane 5), as is that of M1 RNA (
3
).
Finally, the cleavage of phe7 by M1PHE occurred at the +1 position in the
substrate, yielding two cleavage products of 5 and 7 nt respectively (Fig.
4
, lane 6). The cleavage site is identical to that in the reaction with pretRNA
Phe
and M1 RNA (
7
). Ribozyme M1PHE-1, the guide sequence of which targets the cytosine at the fourth position
(-1 site) instead of the guanine at the fifth position (+1 site) (Fig.
1
C), cleaves phe7 at the -1 site, yielding two cleavage products of 4 and 8 nt (Fig.
4
, lane 10). Further analysis of the cleavage products indicated that cleavage
results in a 5' phosphoryl and 3' hydroxyl group, as does cleavage of pretRNA substrates by M1 RNA
(data not shown). These observations showed that the cleavage mechanism of M1GS
RNA is similar to that of M1 RNA.
The complex formed between M1PHE and substrate phe7 resembles a structure
equivalent to an acceptor stem, a 3' CCA sequence and a 5' leader sequence of a pre-tRNA molecule (Fig.
1
B and C). By systematically deleting parts of the 5' leader sequence, the 3' CCA sequence and the helix structure of the substrate-EGS complex, the minimal requirements for cleavage by M1GS
RNA of a model substrate were determined. This was achieved in three sets of
experiments.
The first set of experiments was designed to study the effect of the 3' CCA sequence on cleavage activity of M1GS RNA. Deletion of this sequence
in substrate ptRNA
Tyr
moderately reduces the cleavage rate by M1 RNA (
22
,
25
) (Table
1
). A ribozyme, M1PHE-CCA, was constructed in which the guide sequence did not contain the 3' CCA sequence. Cleavage of substrate phe7 by this ribozyme was at
least 250-fold slower than that by M1PHE (Fig.
4
, compare lanes 7 and 9 and Table
1
). These results were consistent with previous observations that the 3' CCA sequence is important for cleavage of a helix-like model substrate, pAT-1, by M1 RNA and RNase P (
5
,
7
,
8
). The less severe effects of the 3' CCA deletion observed with a pre-tRNA substrate could be explained as the loss of interactions
between M1 RNA and the substrate due to the deletion of segments of the tRNA
structure. Binding of this substrate by M1 RNA can still be achieved by
interactions between M1 RNA and the other, remaining, domains of the pretRNA
molecule.
Table 1
To characterize the role of each nucleotide in the 3' CCA sequence, a set of M1GS ribozymes with a deletion of each nucleotide
in the 3' CCA sequence was constructed and tested for cleavage of phe7 (Fig.
4
and Table
1
). The salient features of the results are as follows. (i) Cleavage catalyzed by
all M1GS RNA constructs tested here was always at the correct position (i.e.,
between the -1 and +1 sites), indicating that the 3' CCA sequence is not essential either for cleavage or cleavage site selection. (ii) Deletion of the 3' terminal adenine led merely to a 2-fold reduction of the cleavage rate. In contrast,
deletion of either of the cytosines resulted in a 10-fold reduction of the cleavage rate.
In the second set of experiments, an RNA substrate, phe7-1, was synthesized chemically in which the 5' leader sequence was replaced with a single nucleotide (Fig.
5
). Subsequently, this substrate was labeled at the 3' end with [
32
P]pCp in the presence of T4 RNA ligase. Cleavage of this substrate by M1PHE
yielded two cleavage products of 7 and 1 nt respectively, as expected (Fig.
5
). Further kinetic analysis revealed that the cleavage rate with this substrate is
~50-fold slower than that of phe7 by M1PHE (Table
1
). This observation indicated that the 5' leader sequence, except for the nucleotide at the -1 position, is not required for cleavage or cleavage site
selection but is important for cleavage efficiency. These results are
consistent with observations of cleavage reactions with pretRNA and pAT-1-like model substrates catalyzed by M1 RNA (
26
,
27
).
The site of cleavage by RNase P of a pretRNA is nearly always at the junction
between a single- and a double-stranded region. Guanine at position +1 is optimal for recognition
by the enzyme (
1
,
29
). However, it is not clear whether the interaction between M1 RNA and the
nucleotide at position +1 is alone sufficient for positioning the scissile band
in the active center. Nor is it known if the identity of the nucleotide at
position +1 in a minimal, model substrate is important for M1 RNA recognition
of the cleavage site. A set of RNA substrates was derived from phe7 in which
the guanine at the +1 position was replaced with the other three bases (Fig.
1
C). Accordingly, a set of ribozymes was also derived from M1PHE in which the
cytosine in the guide sequence of the ribozyme that base pairs with the guanine
of phe7 was replaced with the other three bases. RNA substrates and ribozymes
were incubated together and cleavage products were separated in denaturing
gels. The salient features of the results (Fig.
6
and Table
2
) are as follows. (i) Optimal cleavage was observed when the nucleotide at the
+1 site was base-paired. For example, significant cleavage was found in the reactions with
phe7-G and M1PHE-C (lane 5), phe7-A and M1PHE-U (lane 7), phe7-C and M1PHE-G (lane 13) and phe7-U and M1PHE-A (lane 19). (ii) Selection of the cleavage
site was influenced by the identity of the nucleotide at the +1 site. When
there was a purine at this position, cleavage at the correct position (i.e.,
between -1 and +1 position) occurred regardless of whether the +1 position was not
base-paired [e.g., cleavage of phe7-G by M1PHE-A and M1PHE-G (lanes 3 and 4) and phe7-A by M1PHE-G (lane 8)]. In contrast, cleavage occurred
at the +2 position when a pyrimidine was at the +1 position and was not base-paired [e.g., cleavage of phe7-C by M1PHE-U, M1PHE-A and M1PHE-C (lanes 12, 14 and 15) and phe7-U by M1PHE-U and M1PHE-C (lanes 17 and 20)]. These
observations suggest that a purine nucleotide is a determinant for cleavage
site selection by M1 RNA. (iii) The identity of the nucleotide at the +1 site was also important for cleavage efficiency. For example, the cleavage rate of the reaction with phe7-G and M1PHE-C was ~10-fold higher than that with phe7-C and M1PHE-G. This observation also explains why very
little cleavage at the +2 position (cytosine) of the substrates tested here was found when the
base pair at the +1 site was disrupted.
An important question regarding the catalytic mechanism of M1 RNA is how this
ribozyme recognizes numerous substrates of different sequences and structures. Previously, systematic deletion analyses of a pretRNA molecule were carried out to determine minimal substrate
requirements for M1 RNA (
5
-
8
). Tethering substrates to M1 RNA has provided an additional unique approach to
the study of substrate recognition (
11
,
13
,
14
). In this report, we demonstrated that an RNA substrate, 8-12 nt long, can be cleaved by M1 RNA, provided that a guide sequence
which can form a 3-7 bp duplex with the substrate is linked to M1 RNA. Thus, part of the
minimum substrate requirement for cleavage by M1 RNA is an RNA duplex as short
as 3 bp.
The 3' CCA sequence and 5' leader sequence (upstream of nt -1) of substrates are important for cleavage efficiency but
not required for cleavage. The importance of each nucleotide in the 3' CCA sequence was further examined. These studies indicated that the
entire 3' CCA sequence, but especially the CC sequence, is important for efficient
cleavage of the model substrates we used. These observations are consistent
with the notions that interactions between M1 RNA and this sequence bring the
cleavage site in proximity to the active site (
9
,
10
).
It is noteworthy that the cleavage rate of the minimal substrate phe7 by M1 RNA
in the presence of an EGS is at least 10-fold slower than that of the same substrate by M1PHE, and is at least 100-fold slower than that of pretRNA by M1 RNA (unpublished
experiments). Therefore, it is to be expected that in
E.coli
possible substrates that contain very short RNA helices are processed at
extremely low rates, if at all.
Disruption of the RNA duplex at the 3' side of the site of cleavage in M1GS RNA significantly diminishes
cleavage efficiency. The identity of the base at the +1 position is also a
determinant for cleavage site selection. When a purine is at the +1 position,
cleavage occurs correctly (i.e., between the -1 and +1 site), regardless of whether the +1 position is base-paired. In contrast, aberrant cleavage occurs when a pyrimidine is
at the +1 position and is not base-paired. Furthermore, cleavage of a substrate with a purine at the +1
position is more efficient than that with a pyrimidine of +1. Therefore, a
purine nucleotide seems to be the dominant determinant for cleavage site
selection by M1GS RNA. This observation is consistent with the notion that
guanine at the +1 position, which has been found in most pretRNAs, serves as
the guide nucleotide for M1 RNA and RNase P cleavage (
29
).
It has been shown that M1 RNA binds to both the 3' CCA and 5' leader sequence, and brings the cleavage site in proximity to the
catalytic center of M1 RNA (
9
,
10
,
22
). This conclusion is consistent with our results that deletion of the 5' leader sequence (except for the nucleotide at -1) segment of substrates and the 3' CCA sequence lead to a reduction of 50- and 250-fold in cleavage rate respectively. Our results also
demonstrated that the interactions between the helix and M1 RNA are sufficient to position the cleavage site into
the catalytic center.
The helix binding domain of M1 RNA has not been defined. In the case of the well-studied group I intron ribozyme, the ribozyme registers the RNA helix that
contains the substrate through interactions between 2' hydroxyl groups of the substrate and the ribozyme (
32
) and determines the cleavage site by recognizing a wobble G-U pair (
33
). However, it is likely that the helix binding domain of M1 RNA interacts in a
critical way with the functional groups of bases (
34
) in substrates for RNase P, as well as with the 2' OH groups at several positions (
8
,
35
), since the identity of the nucleotide at position +1 is important,
Additionally, results of chemical footprinting experiments show that
denaturation of the acceptor stem (to make bases more accessible) occurs during
the cleavage reaction (
36
). Such a denaturation step can only occur after binding to M1 RNA, which
recognizes initially the intact helix structure of a model substrate or the
acceptor stem portion of a tRNA.
We thank all members of our laboratory, especially Cecilia Guerrier-Takada for providing M1 RNA deletion mutants, Venkat Gopalan for C5 protein,
Ying Li for sharing unpublished results, Yan Yuan, Venkat Gopalan and Paul Eder
for many helpful discussions and for reviewing the manuscript. F.L. is a Parke-Davis postdoctoral fellow of the Life Sciences Research Foundation. This work
has been supported by USHPS GM19422 to S.A.
+
Present address: Program in Infectious Diseases, School of Public Health,
University of California, 140 Earl Warren Hall, Berkeley, CA 94720, USA
Substrate
Enzyme
K
m
([mu]M)
k
cat
(min
-1
)
k
cat
/
K
m
phe7
M1PHE
0.07 +- 0.01
0.17 +- 0.02
2.6 +- 0.2
phe7
M1PHE-A
0.1 +- 0.02
0.14 +- 0.02
1.4 +- 0.1
phe7
M1PHE-CA
0.09 +- 0.02
0.01 +- 0.002
0.1 +- 0.04
phe7
M1PHE-CCA
0.41 +- 0.04
0.004 +- 0.001
0.01 +- 0.004
phe7-1
M1PHE
-
-
0.05 +- 0.009
phe3
M1PHE
-
-
0.01 +- 0.002
REFERENCES
Return