From the
Department of Biological Sciences,
University of Alberta, Edmonton, Alberta T6G 2E9, Canada and
¶The Department of Biochemistry, University of
Alberta, Edmonton, Alberta T6G 2H7, Canada
Received for publication, March 27, 2003 , and in revised form, May 12, 2003.
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
In F-like plasmids, the regulatory activity of FinP depends upon the action
of the plasmid-encoded protein, FinO
(16), and in F, whose
finO gene is interrupted by an IS3 insertion element,
transfer is completely derepressed
(17). FinO is not
plasmid-specific, and when supplied in trans, FinO from the related
plasmid R65 (18) or
R100 (19) can repress F
transfer. FinO is a 186-amino acid, 21.2-kDa basic cytoplasmic protein with a
highly -helical structure that adopts a novel protein fold
(20). FinO binds the
relatively unstable FinP molecule
(2022),
sterically inhibiting RNase E cleavage of the single-stranded spacer between
SL-I and SL-II (23,
24) and allowing the
steady-state concentration of FinP to increase to sufficient levels to mediate
repression of traJ expression
(25,
26). Indeed, the requirement
of FinO for inhibition of transfer and traJ expression can be
alleviated by providing FinP at elevated copy number in the cell
(14,
15). FinO also catalyzes
FinP/traJ mRNA duplex formation in vitro
(10,
20,
23), which is believed to
allow rapid sequestration of the traJ RBS and efficient inhibition of
traJ expression in vivo
(14,
15).
"Kissing" between loops of RNA stem-loop structures is commonly the first interaction to occur during the process of RNA/RNA duplex formation (reviewed in Refs. 27 and 28). F-like conjugative plasmids encode eight different alleles of FinP, with the highest variability in the loops (8, 12, 21). The loop sequences of FinP and traJ mRNA are therefore thought to be responsible for mediating the plasmid specificity of the F-like FinOP systems (12, 14, 15) and are thought to be the initial site of interaction between the sense and antisense RNAs. Although the loop sequences of FinP in F-like plasmids vary, a common motif, 5'-YUNR-3' (where Y represents pyrimidine, N is any base, and R is purine), is found in several finP alleles (8), which is a key structural motif in the loops of many antisense RNA molecules (29).
The structural features of SL-I and SL-Ic of FinP antisense RNA and traJ mRNA, respectively, that influence FinO-mediated duplex formation in vitro were characterized. Duplex analyses employing EMSAs using in vitro synthesized RNAs and purified FinO protein were used to measure apparent second-order association rate constants (kapp) of duplex formation for a variety of interacting RNA partners. Our studies demonstrate that both in vitro and in vivo, FinO can overcome a variety of sequence and structural changes to FinP SL-I and traJ mRNA SL-Ic in order to promote duplex formation.
![]() |
EXPERIMENTAL PROCEDURES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
In Vitro TranscriptionsAnnealed templates prepared as
described above were added to a final concentration of 300 nM in
20-µl transcription reactions. For labeled reactions, GTP/ATP/CTP were
added to a final concentration for each of 2.5 mM, and UTP was
added to a final concentration of 0.1 mM, along with 1050
µCi of [-32P]UTP (PerkinElmer Life Sciences). Twenty-six
units of RNA Guard (Amersham Biosciences) were added to each reaction, along
with 1x transcription buffer (Roche Applied Science) supplemented with
0.01% (v/v) Triton X-100 (Sigma) and 0.1 mg/ml bovine serum albumin (RNase
Free; Roche Applied Science). Twenty units of T7 RNA polymerase (Roche Applied
Science) were added, and the reactions were incubated at 37 °C for 2 h.
Ten units of DNase I (RNase Free; Roche Applied Sciences) were added, and the
reactions were incubated for a further 15 min to digest the template DNA.
One-fifth volume of RNA load dye (96% (v/v) deionized formamide, 0.05% (w/v)
each xylene cyanol and bromphenol blue, 20 mM EDTA) was added, and
the RNA was heated to 95 °C for approximately 5 min and then cooled on
ice. The RNA was electrophoresed on an 8% (29:1), 8 M urea
polyacrylamide gel at 250 V for
2 h. The radioactive RNA band was
visualized by exposure to Kodak X-Omat R film for several minutes and then
excised and purified as described above, except the purified RNA was dissolved
in 10 µl of Milli-Q water after precipitation. To make unlabeled RNA, all
procedures were the same, except GTP/ATP/CTP/UTP were added to transcription
reactions at a final concentration of 2.5 mM. The unlabeled RNA was
visualized by staining in ethidium bromide and gel-purified as described
above.
EMSAs for Apparent Second Order Association Rate Constant DeterminationEMSA analyses for determination of kapp values were performed essentially as described (20, 23). Briefly, 60 fmol of 32P-labeled RNA was incubated with 600 fmol of its unlabeled complementary RNA in a 50-µl reaction containing 1x TMN buffer (20 mM Tris-HCl, pH 7.5, 10 mM magnesium acetate, 100 mM NaCl). Plasmid R65 FinO, purified as described (22), was added to a final concentration of 6 µM to the reactions where appropriate. Reactions were incubated at 37 °C, and 5-µl aliquots were withdrawn at various times, mixed with 10 µl of ice-cold TMN stop solution (1x TMN containing 30% (v/v) glycerol and 0.05% (w/v) bromphenol blue), and kept on ice. The samples were then electrophoresed on 8% (29:1) nondenaturing polyacrylamide gels containing Tris/glycine buffer (25 mM Tris-HCl, pH 8.08.3, 190 mM glycine) at a constant 160 V for 65 min at room temperature. Gels were dried and then exposed on Molecular Dynamics Storage Phosphor screens overnight. Free and duplexed RNA species were visualized and quantified using an Amersham Biosciences PhosphorImager 445 SI and ImageQuaNT software. k1 values were derived from log plots of the percentage of free labeled RNA versus time of incubation to determine the time required for 50% of the free labeled RNA to form a duplex. kapp values were then calculated from k1 and the concentration of the RNA species in excess, essentially as described (23, 30).
EMSA for Detection of FinO Binding to SL-I and
SL-I(tails)The association equilibrium constant
(Ka) for FinO binding to 32P-labeled
FinP SL-I or derivatives thereof was performed as described
(22), except 6 fmol of
32P-labeled RNA were used per reaction instead of 7.5 fmol.
Quantification of unbound and FinO-bound RNAs and calculation of the
association constants were performed exactly as described
(22,
23).
Construction of Recombinant PlasmidsThe plasmids used in this study and the relevant details and sources of each are listed in Table II. Isolation of all plasmid DNA was performed using a rapid alkaline extraction technique (31). All clones constructed during the course of this work were sequenced using the DYEnamic ET fluorescent sequencing system according to the manufacturer's instructions (Amersham Biosciences) to confirm that the correct DNA sequence was present in each clone. All restriction enzymes used for DNA cloning were purchased from Roche Applied Science. The plasmid pUC180GGA was constructed to express FinP(C16G/C17G/U18A) from its own promoter from the high copy number vector pUC18. The plasmid pUC180 contains a 180-base EcoRI/HindIII fragment derived from F, which encodes wild-type FinP, expressed from its own promoter in the absence of transcription from PtraJ (24). This plasmid served as a template for site-directed mutagenesis of FinP using the mutagenic primers MGU53 and MGU54 (Table I) and Pfu Turbo® (Stratagene) to create pUC180GGA. All procedures were performed according to the manufacturer's instructions (Stratagene), except the plasmid was transformed into rubidium chloride-competent E. coli to propagate the DNA (32). The presence of the mutation was confirmed by sequencing as described above. pLT180GGA was created by inserting the 180-base EcoRI/HindIII fragment containing FinP(C16G/C17G/U18A) from pUC180GGA into similarly digested pT7-3 (33), allowing this mutant FinP antisense RNA to be expressed from its own promoter in a moderate copy number plasmid.
|
Mating AssaysE. coli MC4100 containing the
finP F-derivative plasmid pSLF20
(Table II) was transformed with
the control plasmid pT7-3 or one of the test plasmids (pLT180, pLT180GGA) that
express FinP in trans from its own promoter
(Table II). To provide FinO in
trans, the plasmid pSnO104 (Table
II) was transformed into the test strains where appropriate. Donor
cultures were grown with appropriate antibiotic selection at 37 °C with
agitation to midlog phase (A600 0.61.0), and
0.5 A600 equivalents were pelleted and washed with fresh
LB broth (Difco) to remove antibiotics and then resuspended in 500 µl of
fresh LB broth. The recipient strain ED24
(Table II) was grown to midlog
phase without antibiotic selection. Aliquots (100 µl) of donor and
recipient cultures were mixed with 800 µl of fresh LB and then incubated at
37 °C for 45 min with no agitation. Cultures were vortexed vigorously for
15 s and placed on ice to disrupt mating pairs. 10-fold serial dilutions were
made using ice-cold 1x PBS (137 mM NaCl, 2.7 mM
KCl, 4.3 mM Na2HPO4·7H2O,
1.4 mM KH2PO4), and 10 µl of each dilution
were inoculated on agar plates to select for donor and transconjugant cells.
Donors containing pSLF20 alone were selected on Maconkey-Lactose plates
(Difco) containing 200 µg/ml streptomycin. Donors containing pSLF20 and any
one of the plasmids pT7-3, pLT180, or pLT180GGA were selected on
Maconkey-Lactose plates containing 200 µg/ml streptomycin and 5.0 µg/ml
ampicillin. All donor constructs containing pSnO104 were selected on
Maconkey-Lactose plates to which chloramphenicol was added to a final
concentration of 50 µg/ml, in addition to the other antibiotics listed
above. All transconjugants were selected on L1-spectinomycin plates
(25) containing 100 µg/ml
spectinomycin. All plates were incubated at 37 °C for 1236 h until
visible colonies appeared. Donor and transconjugant cells were then counted,
and the ratio of transconjugants to donors was calculated, allowing mating
efficiency to be compared with the control of conjugative transfer of pSLF20
alone, which was set at 100% mating efficiency.
Immunoblot AnalysisCell pellets corresponding to 0.1 A600 equivalents were boiled in SDS sample buffer (34) for 5 min, and the supernatants were electrophoresed on 15% (29:1) SDS-polyacrylamide gels using the Bio-Rad Mini-Protean® system. Proteins were transferred to Immobilon-P membranes (Millipore Corp.) at 100 V for1hat4 °C using Towbin buffer (35). Membranes were blocked overnight at 4 °C with 10% (w/v) skim milk (Difco) dissolved in TBST (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 0.1% (v/v) Tween 20 (Caledon Laboratories)). Primary antibodies diluted in 10% skim milk in TBST were added to blots and incubated for 1 h at room temperature. The following dilutions of polyclonal antisera (raised in rabbits) were used: anti-FinO, 1:50,000; anti-TraJ, 1:15,000. Blots were washed at room temperature four times for 15 min each with TBST. The secondary antibody used was horseradish peroxidase-conjugated donkey anti-rabbit IgG (Amersham Biosciences) at a 1:10,000 dilution. Blots were incubated for 1 h at room temperature with the secondary antibody and then washed as described above. Blots were developed with Renaissance Western blot Chemiluminescence Reagent Plus (PerkinElmer Life Sciences) and exposed to Eastman Kodak Co. X-Omat R film for varying times to visualize the signals.
Northern Blot AnalysisTotal RNA was isolated via a modified
hot phenol method as described
(23,
24) from strains grown in
liquid cultures (LB broth) at 37 °C to an A600 of
0.81.0. RNA (30 µg) was denatured for 5 min at 95 °C in
formamide RNA load dye and then electrophoresed on an 8% (29:1), 8
M urea polyacrylamide gel and transferred to Zeta-Probe nylon
membranes (Bio-Rad) as described
(24). The blots were
prehybridized for 4 h using the same conditions as described
(24), except 200 µg/ml each
of boiled E. coli strain W tRNA type XX and sonicated calf thymus DNA
(Sigma) were added to the hybridization solution. The FinP-specific probe
primer A (Table I) was
5'-end-labeled with T4 polynucleotide kinase (Roche Applied Science) and
[-32P]ATP (PerkinElmer Life Sciences), and
10 pmol of
the probe was added to the blots in fresh hybridization solution. Incubation
proceeded overnight at 37 °C, and the blots were then washed as described
(23) and exposed on an
Amersham Biosciences storage phosphor screen. Bands corresponding to FinP were
visualized using an Amersham Biosciences PhosphorImager 445 SI and ImageQuaNT
software.
RNA Secondary Structure PredictionsSecondary structure
predictions and G values of SL-I, SL-Ic, and their derivatives
were performed using the Mfold version 3.1 algorithm
(36,
37). The RNA sequences were
analyzed at the Rensselaer Polytechnic Institute Mfold server (available on
the World Wide Web at
bioinfo.math.rpi.edu/
mfold) using standard settings. The secondary structure of FinP and
traJ184 mRNA were experimentally determined previously
(10) and were used as a
reference with which to compare the predicted structures of the individual
stem-loop constructs employed in this study.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
A single base pair mismatch in the stem of SL-I and two single base pair
mismatches in the stem of SL-Ic results in lower stability of these stems
compared with the more extensively base-paired stems in SL-II and SL-IIc
(Fig. 1). The lower predicted
free energy of unfolding of SL-I (G = 10.1 kcal/mol)
and SL-Ic (
G = 8.6 kcal/mol) compared with SL-II
(
G = 28.2 kcal/mol) and SL-IIc (
G =
23.3 kcal/mol) suggests that intermolecular base pairing between the
stems of SL-I and SL-Ic during the formation of a stable FinP/traJ184
mRNA duplex are more likely to occur than between the stems of SL-II and
SL-IIc (10).
Contribution of the Loop Residues of SL-I to RNA/RNA Duplex
FormationThree regions of the loop of SL-I were chosen to test for
their contribution to SL-I/SL-Ic duplex formation. All mutations were
transversions that disrupted the expected Watson-Crick base pair interactions
between the loops. The predicted secondary structures of all of these
constructs are shown in Fig. 2.
One- and two-base transversion mutations in these regions resulted in no
noticeably obvious alterations to duplex formation ability (data not shown);
therefore, 3- and 4-base transversion mutations were examined. The first
mutation examined lies within the 5' side of the loop of SL-I,
5'-C16G/C17G/U18A-3', which is referred to as SL-I
(1618)
throughout this work. This mutation alters 3 of the 6 bases that comprise the
predicted anti-RBS of FinP (Fig.
1). The second mutation is located on the 3' side of the
loop of SL-I, 5'-C21G/A22U/A23U-3', which is referred to as
SL-I(2123). The last extends across the top of the loop of SL-I,
5'-U18A/C19G/A20U/C21G-3', which is referred to as
SL-I(1821). When compared with the kapp for
SL-I/SL-Ic duplex formation under identical conditions, the
kapp for SL-I(1618)/SL-Ic duplex formation was
reduced by 52% in the absence of FinO and by 55% in the presence of FinO
(Fig. 3, A and
B; Table
III). SL-I(1821)/SL-Ic duplex formation exhibited a
kapp reduced by 35% in the absence of FinO and by 60% in
the presence of FinO, whereas SL-I(2123)/SL-Ic revealed a
kapp reduced by 55% in the absence of FinO and by 51% in
the presence of FinO, when compared with the kapp for
SL-I/SL-Ic duplex formation under the same conditions
(Fig. 3, A and
B; Table
III). These results suggest that the level of complementarity
between loop residues of SL-I and SL-Ic affects FinO-mediated duplex formation
in vitro. The observation that the kapp values
for duplex formation of all of the interactions tested between the SL-I loop
mutants and SL-Ic were 1019-fold higher in the presence of FinO than in
the absence of FinO (Table III)
reveals that FinO can overcome as many as four mismatches in the loop-loop
base pairing interaction to promote duplex formation in vitro.
|
|
|
The Effect of Stem Mutations on SL-I and SL-Ic Duplex
FormationThe bulged A12:A27 base pair mismatch in SL-I and the
corresponding U85:U100 mismatch in SL-Ic
(Fig. 2) were examined for
their contribution to duplex formation. SL-I(A27U) (G =
14.3 kcal/mol) and SL-Ic(U85A) (
G = 12.1
kcal/mol) were made to increase the stability of the stems while maintaining
full intermolecular complementarity between the stems of the two RNAs. The
kapp for SL-I(A27U)/SL-Ic(U85A) duplex formation in the
absence of FinO was reduced by 32%, whereas in the presence of FinO, the
kapp showed a 74% reduction, compared with the
kapp for SL-I/SL-Ic duplex formation
(Fig. 4A,
Table III). These results
suggest that the overall stability of the stem regions of SL-I and SL-Ic
influences their transition to a stable duplex. To create more drastic
mutations affecting stem complementarity and to provide insight into the
direction of progression of duplex formation, SL-Ic(TSR) and SL-Ic(BSR) were
constructed. SL-Ic(TSR) has 5 base pairs in the stem immediately below the
loop reversed in orientation, resulting in noncomplementarity with the
corresponding region in SL-I (Fig.
2). SL-Ic(BSR) has 6 base pairs at the bottom of the stem reversed
in sequence in the same fashion (Fig.
2). The single-stranded tail regions were not included in these
constructs, ensuring that only the effects on intermolecular stem/stem
interactions were examined. SL-I/SL-Ic(TSR) duplex formation in both the
presence and absence of FinO was minimal, and a kapp could
not be calculated in either case because less than 20% of the
32P-labeled free RNA in the reactions was converted to a duplex
(Fig. 4A).
SL-I/SL-Ic(BSR) duplex formation in the absence of FinO revealed a
kapp that was reduced by 84% relative to the
kapp for SL-I/SL-Ic duplex formation
(Fig. 4A;
Table III). In the presence of
FinO, the kapp for SL-I/SL-Ic(BSR) duplex formation was
reduced by 66% compared with the kapp for SL-I/SL-Ic
duplex formation (Fig.
4A; Table
III). These results suggest that stable duplex formation between
SL-I and SL-Ic can proceed only if intermolecular complementarity extends from
the loop through the top of the stem. The virtually identical
kapp values for SL-I/SL-Ic(
tails) and
SL-I/SL-Ic(BSR) duplex formation also suggests that a region of
noncomplementarity at the bottom of the stem has no significant effect on the
ability of FinO to promote duplex formation between these constructs in
vitro.
|
Detection of SL-I/SL-Ic Kissing ComplexesSince kissing between loop regions is normally the first interaction in most antisense/sense RNA-pairing reactions, it was decided to determine whether a SL-I/SL-Ic kissing dimer could form and be detected by EMSA analysis. SL-IcR was created such that the loop region was completely complementary to SL-I, but the stems and tails were not complementary (Fig. 2). In the presence and absence of FinO, no stable kissing intermediate was detectable (Fig. 4B), suggesting that that any initial kissing complex that forms between SL-I and SL-Ic is transient and unstable and is not detectable by EMSA analysis. These results also confirm the observations resulting from the SL-I/SL-Ic-(TSR) duplexing experiments described above. The formation of a stable SL-I/SL-Ic duplex requires complementarity in both the loops and as much as half of the stem in the region immediately below the loops of both RNA molecules.
Contribution of the Single-stranded Tail Regions of SL-I to
RNA/RNA Duplex FormationSince the single-stranded tails
of FinP SL-I and SL-II have been shown to influence the ability of FinO to
bind FinP with high affinity
(21), the contribution of
these regions to duplex formation in vitro was tested.
SL-I/SL-Ic(tails) duplex formation showed a kapp
reduced by 68% in the absence of FinO, and a kapp reduced
by 72% in the presence of FinO, relative to the kapp for
SL-I/SL-Ic duplex formation under identical conditions
(Fig. 5A;
Table III). Analysis of
SL-I(
tails) duplex formation with SL-Ic(
tails) revealed a 55%
decrease in kapp in the absence of FinO and an 81%
reduction in kapp in the presence of FinO, compared with
the kapp for SL-I/SL-Ic duplex formation under identical
conditions (Fig. 5A;
Table III). These values are
comparable with the values obtained for SL-I/SL-Ic(
tails) duplex
formation, suggesting that complementarity of the single-stranded tail
regions, rather than possible structural alterations to the molecules upon
removal of these regions, affects duplex formation in vitro. These
results suggest that the presence of the single-stranded regions flanking SL-I
and SL-Ic makes important contributions to the FinO-mediated formation of the
RNA/RNA duplex in vitro. In order to ensure that any decrease in
kapp was the result of alterations in complementarity of
the interacting RNAs and not due to an inability of FinO to bind them, EMSA
analysis was performed to determine whether FinO could bind to SL-I and
SL-I(
tails). As shown in Fig.
5B, FinO was able to bind to both RNA molecules, with a
Ka of
8.6 x 106
M1 and 3.5x106
M1, for binding SL-I and SL-I(
tails),
respectively. These Ka values are higher than
those reported in a previous study, which may be attributable to the fact that
our study employed native FinO, whereas the previous study employed a
glutathione S-transferase-FinO fusion protein
(21). Regardless, our results
confirm that FinO could bind the SL-I constructs employed in the duplex
assays.
|
Contribution of the Anti-RBS of FinP to Its in Vivo
FunctionAs described above, alteration of a portion of the
anti-RBS of FinP within the loop of SL-I moderately reduced the efficiency of
duplex formation with its complementary RNA, SL-Ic, in vitro.
Previous studies have shown that the loop nucleotides of SL-I and SL-II of
plasmid R1 FinP directly affect the ability of FinP to inhibit plasmid
transfer (14,
15). We wanted to determine
the effect of mutations in the anti-RBS of FinP on inhibition of TraJ
expression and F plasmid transfer. The plasmid pLT180GGA
(Table II) expresses FinP
(1618)
from a moderate copy number plasmid (pT7-3; 1030 copies/cell). This
plasmid was introduced into E. coli MC4100, with or without plasmid
pSnO104, which expresses plasmid R65 FinO in trans
(Table II). Plasmid pLT180,
which expresses wild-type FinP at a moderate copy number, was tested in the
same way, as was the negative control parental plasmid, pT7-3
(Table II). The
finP F-derivative pSLF20
(Table II)
(25) was present in all
strains, and pSLF20 conjugative transfer and expression of TraJ were tested in
the presence of the FinP derivatives as described under "Experimental
Procedures." Mating inhibition assays revealed that FinP
(1618),
when supplied in trans at medium copy number, was unable to fully
repress mating (Table IV). Its
efficiency was reduced by 100-fold and 175-fold, in the absence or
presence of FinO, respectively, compared with wild-type FinP under the same
conditions. The presence of FinO significantly enhanced mating repression
mediated by both wild-type FinP and FinP
(1618)
expressed in trans at medium copy number
(Table IV). Wild-type FinP
(pLT180) fully repressed TraJ accumulation in both the presence and absence of
FinO, whereas FinP
(1618)
(pLT180GGA) was able to fully repress TraJ accumulation only in the presence
of FinO, as determined by an immunoblot analysis
(Fig. 6). These results confirm
previous observations that the ability of FinP to inhibit conjugative transfer
of F-like plasmids is dependent upon gene dosage
(15). They also confirm the
results from the in vitro duplex formation assays described above,
which showed that FinO can overcome multiple base mutations in FinP SL-I and
promote SL-I/SL-Ic duplex formation in vitro when complete loop-loop
complementarity is absent. In vivo, it also appears that FinO can
compensate for suboptimal loop-loop base complementarity and promote fertility
inhibition when the tested FinP loop mutant is supplied at an elevated copy
number. Experiments with the very high copy number construct pUC180 gave
similar results as with pLT180 (data not shown).
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The plasmid R1 encodes a FinOP system very similar to the F plasmid.
Single-base mutations in the top portions of the loops of R1 FinP SL-I or
SL-II that altered potential loop-loop base interactions with traJ
mRNA significantly inhibited FinO-mediated repression of conjugative transfer
of R1, when these mutant FinP molecules were supplied in trans at
elevated copy number. However, FinO was able to mediate repression of
traJ expression under the same conditions, as measured by
-galactosidase assays of a traJ-lacZ translational fusion
reporter construct (15).
Whereas mutations in FinP can severely inhibit R1 fertility inhibition, FinO
appears to be able to promote direct inhibition of translation of
traJ by mutant FinP RNA. When single base transversions were made in
the loops of SL-I and SL-II simultaneously, FinO-mediated repression of both
traJ expression and conjugative plasmid transfer were significantly
reduced, suggesting that both loops play a role in FinP/traJ mRNA
duplex formation (15).
Interestingly, a single-base transversion mutation made in the 3'
portion of the loop of FinP SL-I had no negative effect on FinO-mediated
inhibition of traJ expression or plasmid transfer
(15). These results suggest
that the interaction of FinP and traJ mRNA in vivo relies
more on the bases located at the top of the loops than those situated on the
3' side (15). In the
present study, the inhibitory function of FinP in vivo was shown to
rely on interactions between the anti-RBS of FinP and the RBS of traJ
mRNA. When supplied in trans at medium copy number, FinP
(1618)
exhibited full negative regulatory function only in the presence of FinO.
These observations support the finding that FinO can compensate for loop
mutations in its RNA targets and promote duplex formation in vitro
and confirm that loop-loop base pairing between the anti-RBS of FinP and the
RBS of traJ mRNA is critical for the regulatory function of the FinOP
system in vivo. However, it must be stressed that under normal
physiological conditions (i.e. wild-type levels of FinP and
traJ mRNA), it is unlikely that pairing of such mutant FinP molecules
with traJ mRNA would occur because of the relatively low levels of
these molecules in vivo. In the present study, the inhibitory effect
of mutant FinP on plasmid transfer and traJ expression in
vivo probably depends completely on the presence of elevated FinP levels
in the cell. Indeed, sequence differences in the loops of FinP are responsible
for conferring plasmid specificity to the FinOP system of F-like plasmids
(12,
15,
41).
A common structural motif in prokaryotic antisense RNA systems is the 5'-YUNR-3' loop motif, which is thought to provide optimal alignment of bases on the 3' side of the loop with those in a complementary RNA (29). Mutations in the YUNR motif of hok RNA of the plasmid R1 hok/sok postsegregational killing system greatly reduced Sok antisense RNA/hok mRNA duplex formation in vitro, although complementarity between the interacting RNAs was maintained (29). In the present study, two of the three multiple loop mutations in FinP SL-I performed in this work altered the YUNR motif and significantly disrupted complementary Watson-Crick base-pairing interactions but led to only moderate decreases in duplex formation rates (Table III). Whether or not the loop mutations disrupted the YUNR motif, the decrease in in vitro duplex formation rates was approximately equivalent. In all cases, the presence of FinO resulted in higher kapp values for duplex formation, demonstrating its ability to promote duplex formation in vitro between RNAs with suboptimal complementarity in loop regions. These results suggest that whereas loop-loop pairing between FinP and traJ mRNA is important, the sequence, and possibly the structure, of the YUNR motif in the loops may play a smaller role than in other systems.
The presence of short single-stranded tails flanking both the 5' and 3' sides of SL-I influenced the ability of SL-I to duplex with SL-Ic in vitro. The removal of both single-stranded tails from SL-I and SL-Ic led to a decrease in FinO-catalyzed duplex formation, which was more significant than any of the loop mutations that were tested, suggesting that single-stranded regions in FinP and traJ mRNA are critical for efficient duplex formation. However, a reduced affinity between FinO and these RNA constructs cannot be ruled out as having an effect on duplex formation at this time. RNA I/RNA II interaction in ColE1 replication control (4) as well as the CopA/CopT interaction of plasmid R1 (38, 42, 43) rely on interactions between single-stranded regions for full activity, demonstrating the importance of such regions to antisense-sense RNA pairing. Considering the short length of complementary single-stranded regions in FinP and traJ mRNA and the importance of such regions to stable duplex formation in other systems (38, 42, 43), the requirement for complementarity in both the loop and single-stranded tail regions of these RNAs is not unexpected.
The presence of bulged nucleotides and mismatched bases in the stems of interacting RNAs is critical for antisense/sense RNA interactions both in vitro and in vivo for several plasmids. These regions are thought to allow breathing of the stems immediately below the loops in order to allow for efficient progression of stable duplex formation (4446). When the purine:purine mismatches in SL-I and SL-Ic were altered to A:U base pairs, maintaining their complementarity but increasing the predicted free energy of unfolding of the stems, the kapp for FinO-mediated duplex formation decreased relative to duplex formation between wild-type SL-I and SL-Ic. These results indicate that the bulged mismatched base pairs in the stems of SL-I and SL-Ic influence the progression of duplex formation. Alteration of complementarity between the stems of SL-I and SL-Ic revealed that, provided at least 5 base pairs immediately below the loops are complementary, stable duplex formation could occur, albeit at a reduced rate. Interaction of the antisense regulatory RNA DsrA with one of its targets, rpoS mRNA, exhibits a similar requirement. Complementarity between bases in the top of the stem of SL-I in DsrA and a specific region of the upstream leader of rpoS mRNA is required for efficient intermolecular pairing in order to promote translation of rpoS (47). Kissing intermediates formed by interacting RNA stem-loop constructs can often be detected readily by EMSA analysis (48, 49). Our inability to detect a SL-I/SL-IcR kissing intermediate by EMSA analysis under the conditions tested suggests that such a complex may be unstable and short-lived, unless initial loop-loop pairing can progress through the stems to form a more stable duplexing intermediate. Alternatively, a stable kissing intermediate may form but might only be detectable using more sensitive means, such as NMR analysis (50). One cannot exclude the possibility that a stable kissing intermediate, mediated by SL-I/SL-Ic and SL-II/SL-IIc interactions between whole FinP and traJ RNAs, may occur, although this possibility remains to be tested.
Several biological systems employ an accessory protein to promote RNA duplex formation, each using a different mechanism. The Rom protein of ColE1 binds to and stabilizes an initial RNA complex between RNAI and RNAII, driving the reaction toward stable duplex formation (5, 51). The E. coli Hfq protein is thought to form a nucleoprotein complex with Spot42 antisense RNA and its target, galK mRNA, cooperatively facilitating RNA/RNA pairing (52). The NCp7 nucleocapsid protein of HIV-1 has been shown to facilitate dimerization between the stem-loops of the dimerization initiation site of the HIV-1 genomic RNA by converting an initial unstable RNA loop-loop complex to a stable dimer (48, 49, 53). More recently, NCp7 was also shown to transiently melt the secondary structure of portions of the stems of human immunodeficiency virus TAR RNA and its DNA complement, cTAR (54). Clearly, accessory proteins that mediate RNA/RNA interactions use a variety of mechanisms to promote RNA pairing. Based upon its similarities to such systems, previous work done on the FinOP system, and the results presented in this work, we present a preliminary model of the mechanism of FinO-mediated duplex formation. FinO is able to destabilize double-stranded RNA, which, along with its RNA-RNA duplex catalysis activity, has been localized to a lysine-rich region within the N-terminal 44 amino acids of the protein.2 The highest affinity binding sites of FinO are SL-II of FinP and SL-IIc of traJ mRNA, although SL-I is also a target for binding (Fig. 5B) (21). Initial binding of FinP and traJ mRNA by FinO allows their loops to come into close proximity and begin base pairing, whereas its RNA destabilization activity begins to open the stems of both SL-I and SL-II. This destabilization of the stems should alleviate the topological restraints inherent in such RNA/RNA interactions, which impose a kinetic barrier to extended duplex formation (reviewed in Ref. 28). Thus, more extensive intermolecular interactions between FinP and traJ mRNA should occur in the presence of FinO. It is likely that destabilization of SL-II and SL-IIc by FinO is more critical than for SL-I and SL-Ic, considering the lower thermal stability of SL-I and SL-Ic imposed by the presence of mismatches in both of their stems. Once duplex formation initiates at the loops and tops of the stems, the single-stranded tail regions of both RNAs may begin to base pair, leading to extended duplex formation. Alternatively, once it has bound to each of its RNA targets, FinO might induce extended regions of single-stranded RNA via its destabilization activity. This function might be similar to that of the E. coli Hfq RNA chaperone protein, which binds to and partially destabilizes the secondary structure of stem-loops in the small, untranslated regulatory RNA OxyS. This destabilization has been hypothesized to facilitate interaction of OxyS with one of its targets, fhlA mRNA (55). Once a critical level of single-stranded RNA is achieved, rapid formation of a stable duplex between the FinP and traJ mRNA can occur.
An interesting observation that emerges from this work is that in vivo, minor changes to the RNA components of the FinOP system can cause significant changes to its function, although in vitro, structural changes to the portions of the RNAs are tolerated during duplex formation. The interaction between FinP, FinO, and traJ mRNA occurs concurrently with transcription of the tra operon, which is in turn influenced by a variety of other factors, including the concentration of these molecules (8). Likewise, the ability of FinOP to inhibit plasmid transfer does not perfectly correlate with the ability of the system to prevent traJ expression (15). It is likely that a delicate balance of factors influences the ability of FinO to promote the formation of a FinP/traJ mRNA duplex in vivo that ultimately results in inhibition of transcription of the tra operon. The exact mechanism underlying FinO-mediated FinP/traJ mRNA duplex formation remains to be elucidated.
![]() |
FOOTNOTES |
---|
Supported by studentships from the Canadian Institutes of Health Research
and the Alberta Heritage Foundation for Medical Research.
|| To whom correspondence should be addressed: Dept. of Biological Sciences, CW-405 Biological Sciences Bldg., University of Alberta, Edmonton, Alberta T6G 2E9, Canada. Tel.: 780-492-0672; Fax: 780-492-9234; E-mail: laura.frost{at}ualberta.ca.
1 The abbreviations used are: SL, stem-loops; UTR, untranslated region; EMSA,
electrophoretic mobility shift assay; RBS, ribosome binding site.
2 A. F. Ghetu, D. C. Arthur, M. J. Gubbins, R. A. Edwards, L. S. Frost, and
J. N. M. Glover, submitted for publication.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|