ABSTRACT
In the initiation of reverse transcription in retroviruses, nucleocapsid (NC)
protein accelerates the rate of annealing of transfer RNA replication primer to
a complementary sequence on the genomic RNA. In this report, we have probed the
conformational changes induced by HIV-1 NC protein and domain deletion mutants in a structurally well-characterized transfer RNA, yeast tRNA
Phe
, as a model for the natural primer. One molar equivalent of recombinant 71 amino acid HIV-1 nucleocapsid protein (NC 1-71) is sufficient to completely inhibit the Pb
2+
-ribozyme activity of tRNA
Phe
at 25
o
C, pH 7.0 and 15 mM MgCl
2
. Zn
2
HIV-1 NC proteins which lack one or both flexible terminal domains also inhibit the ribozyme activity.
1
H NMR spectra acquired for Mg
2+
-tRNA
Phe
suggest that NC 1-71 and NC 12-55 (lacking residues 1-11 and 56-71) inhibit the lead-ribozyme activity by only modestly altering
the active site region rather than inducing large-scale unfolding of the molecule. In the absence of Mg
2+
, the extent of destabilization of tRNA
Phe
is greater but appears to be confined to internal regions of the acceptor and T
[psi]C helices, as evidenced by the selectively enhanced exchange rates for imino protons associated with these base
pairs. These findings show that NC destabilizes the folded form of tRNA
Phe
and by extension, other complex RNAs, in tertiary and secondary structural
regions most susceptible to thermally-induced denaturation.
The nucleocapsid (NC) of retrovirus particles is composed of a single dimeric
genomic RNA, ~3000 NC protein molecules, 10-50 molecules of reverse transcriptase and a variety of small
cellular RNAs, including transfer RNAs, 5S RNA and ribosomal RNAs (
1
). Associated with the 70S genomic RNA dimer is a specific cellular transfer
RNA, called replication primer tRNA. The earliest stages of proviral cDNA
synthesis requires that the 3'-end of the acceptor helix region of the replication primer tRNA be
stably base paired to a complementary primer binding site sequence (PBS)
located on each subunit of the genomic RNA dimer. This primer-template complex is used by reverse transcriptase to initiate reverse
transcription to synthesize minus strand strong stop cDNA. In
in vitro
reconstituted reactions, retroviral NC protein promotes the annealing of the
cognate tRNA to the PBS and this is at least partly responsible for enhancing
the overall yield of cDNA product (
1
-
9
). In addition, one proviral mutant of Rous sarcoma virus which contained a
missense mutation in NC domain was shown to package genomic RNA onto which the
replication primer was not stably associated (
4
).
Retroviral NC proteins bind to both DNA and RNA and appear to preferentially
interact with single-stranded nucleic acid sequences (
2
,
3
,
10
,
11
), although an interaction of NC proteins with RNA and DNA duplex molecules has
also been reported (
12
-
14
). NC proteins appear to be kinetically competent to destabilize or denature
double-stranded nucleic acids and can therefore be considered helix-destabilizing proteins (
3
,
15
). HIV-1 NC 1-71 and Mason-Pfizer monkey virus NC protein have been shown to strongly
accelerate the renaturation of two complementary nucleic acid strands (
15
-
18
). NC proteins have also been shown to facilitate strand-transfer DNA synthesis from both model (
19
) and viral RNA templates (
12
,
20
) in reactions which require base pairing between the nascent cDNA product and
complementary R (repeated) sequences to effect template jumping by the enzyme.
In another strand exchange activity, NC proteins have been shown to activate
multiple turnover by a number of ribozymes by accelerating the rate of
association of the ribozyme with its substrate as well as increasing the
product dissociation rate (
21
,
22
). Thus, NC proteins appear to facilitate (i.e., lower the kinetic barrier to)
the formation of the most stable intermolecular RNA duplex formation by
destabilizing intramolecular duplex regions of equal or lower thermodynamic
stability (
15
).
High-resolution solution structural studies of synthetic peptides and virally
produced NC from HIV-1 (55 amino acids)
reveal that while the two Zn(II) coordination domains (residues 13-51) adopt well-defined, structurally homologous folded structures, the other
regions of the protein, including the interfinger and the N-terminal domains, do not appear to adopt a defined secondary or tertiary
structure in solution (
23
-
25
). In contrast, another study suggests that the orientation of the finger
domains relative to one another is fixed with the zinc-fingers proximate to one another, primarily as a result of a well-defined linker conformation (
26
,
27
). With respect to HIV-1 NC-RNA interactions, Dannull
et al
. (
28
) showed that the N-terminal zinc-finger and flanking basic regions are required for specific binding
to an HIV-1 genomic RNA fragment containing the major packaging signals. The
sequence non-specific interaction of NC proteins with transfer RNA and many other
nucleic acid ligands appears to involve many of the same determinants, except
that the zinc-fingers are largely dispensible and readily substituted with a diglycyl
peptide linker (
29
). These studies suggest that the zinc-fingers play an accessory role in organizing the immediately adjacent
basic regions found in the N-terminal flanking and interfinger regions thereby providing a large
electrostatic stabilization to the complex (
1
-
3
,
9
,
10
; R.K. and D.P.G., unpublished results).
In this work we have structurally probed the interaction of HIV-1 NC 1-71 and domain-deleted forms NC 1-71, e.g., NC 12-55, with the extensively studied model tRNA,
yeast tRNA
Phe
using the Pb
2+
-ribozyme activity (
30
) and
1
H NMR spectroscopy (
31
-
35
). The use of yeast tRNA
Phe
as a model is justified for several reasons. Most importantly, previous studies
have shown that HIV-1 NC protein binds without sequence specificity and with nearly identical
affinities and stoichiometries to several folded transfer RNAs, including yeast
tRNA
Phe
, beef liver and synthetic forms of the tRNA packaged by HIV-1, tRNA
Lys,3
(
10
). [Recombinant HIV-1 NC 1-71 induces the same extent of conformational changes in a mixed
preparation of tRNA from
Escherichia coli
(cf.
3
), yeast tRNA
Phe
and synthetic human tRNA
His
as determined by circular dichroism spectroscopy (R. K. & D. P. G., unpublished results).] Thus, general features of the NC-tRNA interaction may be recognized from the study of yeast tRNA
Phe
. Secondly, the extensive crystallographic,
1
H NMR structural and dynamic, thermodynamic and chemical kinetic properties
established for this tRNA in the presence and absence of Mg
2+
make it possible to probe the NC-tRNA in molecular detail. Since tRNA
Phe
has well-defined regions of single-stranded, duplex and tertiary structure, we have discussed our
findings in the context of a proposed mechanism of activation of tRNA annealing
to the PBS and other generic nucleic acid chaperonin-like activities (
15
,
16
,
22
) characteristic of NC protein.
All buffers were prepared with doubly distilled and deionized Milli-Q water. NC proteins were prepared and characterized as described using
methods exactly analogous to those previously outlined for HIV-1 NC 1-71 (
2
) and MPMV NC 1-96 (
16
). NC 1-71 is a 71 residue recombinant protein corresponding to Met-377-Phe-447 of Pr55 gag polyprotein precursor encoded by the BH10-R3 HIV-1 proviral clone (formerly NC71) (
2
,
3
). The virally processed, mature form of nucleocapsid protein in many HIV-1 strains is a 55-residue N-terminal fragment of NC 1-71, corresponding to Met-377-Asn-431 and is denoted NCp7 (
2
). The two Cys-X
2
-Cys-X
4
-His-X
4
-Cys (X is any amino acid) zinc-finger regions of HIV-1 NC 1-71 encompass residues 15-28 and 36-49. NC 1-57 lacks the flexible C-terminal domain (residues 58-71) of NC 1-71 (
3
), NC 12-71 lacks the strongly basic N-terminal domain (residues 1-11), and NC 12-55 lacks both N-terminal and C-terminal domains. NCp7C corresponds to
residues 13-71 with the zinc-finger region residues 15-28 and 35-50 each replaced with a diglycyl linker and a C-terminal Leu added (
9
).The domain deletion NC proteins NC 12-71 and NC 12-55 were prepared from recombinant expression plasmids exactly
analogous to pT7nc71.hiv1 (
3
) and were constructed via PCR amplification of the appropriate coding regions
and subcloned in the
Nde
I-
Bam
HI sites of pET-3a (
3
,
36
). All NC proteins used in the current study were purified to homogeneity and
chararacterized by amino-acid analysis and as to their Zn(II) content and reduced thiol
concentration. HIV-1 NC 1-57 was prepared as previously described (
3
). NC 1-57 and NC 12-71 were found to contain 1.6 (+-0.2) and 1.8 (+-0.2) g
.
at Zn(II) by atomic absorption and 5.7 (+-0.2) and 6.8 (+-0.2) free thiols by DTNB reactivity, respectively. NC 12-55 was found to contain 0.8 (+-0.2) g
.
at Zn(II) and 3-4 free thiols. Apo-
S
-methylated NC 12-71 was prepared from Zn(II)
2
NC12-71 exactly as described previously for apo-
S
-methylated gene 32 protein (
37
). Recombinant T4 gene 32 protein was prepared as described (
38
). Native yeast tRNA
Phe
was obtained from Sigma. The synthetic peptide, NCp7C was the generous gift of
Dr W. Sundquist, University of Utah.
Buffer 1 contained 11 mM NaP
i
, 111 mM NaCl and 16.5 mM MgCl
2
, pH 7.0 and buffer 2 contained 11 mM NaP
i
and 20 mM NaCl, pH 7.0. To obtain Mg
2+
containing samples of the tRNA
Phe
, 5-10 mg of the lyopholized tRNA
Phe
was resuspended in 0.5 ml buffer 1 and then dialyzed extensively against the
same buffer. To prepare the Mg
2+
free form, lyopholized tRNA
Phe
was resuspended in 0.5 ml buffer 2 supplemented with 9 mM Na
3
EDTA, incubated at room temperature for 30 min, followed by exhaustive dialysis
against buffer 2 to ensure the removal of EDTA. The samples were checked for purity and possible degradation during the study by denaturing polyacrylamide
gel electrophoresis.
tRNA
Phe
concentrations were determined using [epsilon]
258
= 5.64 * 10
5
M
-1
.cm
-1
.
All other reagents were from Bio-Rad and Sigma and were used without further purification.
Native yeast tRNA
Phe
was 5'-end labelled by dephosphorylation with alkaline phosphatase
followed by phosphorylation with [[gamma]-
32
P]ATP and polynucleotide kinase (
39
). Lead acetate stocks (10 mM) were prepared in DEPC H
2
O and stored at -70oC to avoid formation of an insoluble lead carbonate species (
30
). All tRNA preparations were renatured by heating to 70oC for 1 min in 10 mM Tris-HCl, pH 8.0, and slow cooled to room temperature prior to use. A
typical 40 [mu]l cleavage reaction contained 2 [mu]M cold tRNA and ~100 nCi end-labelled tRNA, 15 mM morpholinopropanesulfonic acid (MOPS)
(pH 7.0), 1.5 mM spermine, 15 mM MgCl
2
, 200 [mu]M Pb(OAc)
2
and NC protein at the indicated concentrations (
30
). The reactions were initiated by the addition of 200 [mu]M Pb(OAc)
2
after all other components had equilibrated at the reaction temperature
(typically 5 min, 25oC). At appropriate time intervals (from 5-60 min), 5 [mu]l aliquots were removed and added to 5 [mu]l loading buffer (7 M urea, 50 mM EDTA, 0.04% bromophenol
blue and 0.04% xylene cyanol FF). The cleavage products were then analyzed on a
15% polyacrylamide/7 M urea denaturing gel, the gel dried and subjected to
quantitation using a Betascope 603 (Betagen).
1
H NMR spectra at 500 MHz were recorded on tRNA samples dissolved in the
indicated buffer and 90% H
2
O/10% D
2
O on Varian UnityPlus 500 MHz spectrometer using a shaped pulse sequence (
40
) to minimize excitation of the solvent water. Up to 14 000 transients were
recorded at each temperature under solution conditions given in the figure
legends. Chemical shifts are reported relative to external trisilyl propionic
acid (TSP) under identical solution conditions.
Pb
2+
-catalyzed cleavage of yeast tRNA
Phe
is an exquisitely sensitive assay of tRNA tertiary structure (
30
). At neutral pH, the active species Pb(OH)
+
is precisely coordinated in a pocket formed by eight residues at the junction
between the D and T[psi]C loops in tRNA in such a way that it can increase the nucleophilicity of
the oxygen atom of the 2'-hydroxyl of ribose 17. Brown
et al
. (
41
) propose that the resulting 2' oxygen nucleophile then attacks phosphate 17 to give a pentacovalent
phosphorus intermediate which decays to form the 2',3'-cyclic phosphate and, after protonation, the 5'-hydroxyl product. The high specificity of the
cleavage for tRNA
Phe
is dependent on the conformation of the elbow region of the `L' shaped molecule
so that the metal ion can be precisely oriented with respect to the cleavage
site (
30
,
42
).
Figure
1
compares the kinetics of Pb
2+
-catalyzed cleavage of 5'-end labelled tRNA
Phe
to give a 5'-end labelled 17mer which results from cleavage 3' to U17. In the absence of NC protein, the lead-catalyzed cleavage rate is observed as first-order and occurs at rates comparable with
previous findings (
30
). In marked contrast, in the presence of stoichiometric NC 1-71, the lead cleavage reaction becomes completely inhibited. This result
suggests that NC causes a conformational change in some regions of the tRNA
Phe
which prevents the binding and/or hydrolysis activity of the Pb
2+
.
In an effort to define the structural domains of NC protein which are required
for binding to tRNA
Phe
as measured by the inhibition of the lead-catalyzed hydrolysis reaction, a variety of N- and C-terminal domain deletion molecules were tested at a
concentration equimolar to tRNA
Phe
(Table
1
). Zn(II)-complexed forms of NC 1-57, NC 12-71 and NC 12-55 were found to give approximately the same extent
of protection as full-length NC 1-71, inhibiting the rate lead cleavage by at least 20-fold under these conditions.
The zinc-finger structures of NC protein were next destroyed using two derivatives
of NC 12-71 to investigate the importance of the intact zinc-fingers for this inhibitory activity. One derivative is a synthetic
peptide, NCp7C, in which each of the zinc-finger regions (residues 15-29 and 35-51) is replaced by a Gly-Gly linker (
9
). Apo
S
-methylated NC 12-71 was prepared by incubating Zn
2
NC 12-71 with methylmethanethiosulfonate to form the mixed disulfide -Cys-S-S-CH
3
derivative which forms with expulsion of the bound zinc (
37
). This creates an unfolded or extended conformation of zinc domain in which no
intramolecular Cys-Cys disulfide bonds are formed. The capacity of these two molecules to
protect tRNA
Phe
from lead cleavage is lost (Table
1
). This suggests that in the context of the N-terminal domain deletion molecule NC 12-71, the zinc-finger domain is required to efficiently alter the tRNA
Phe
conformation.
Table 1
The data presented above show that full-length NC 1-71 and domain deletion NC 12-55 inhibits the Pb
2+
-ribozyme activity of tRNA
Phe
. Inhibition of ribozyme activity of tRNA
Phe
could be due to a subtle change in disposition of active site residues, a
change in the Pb
2+
coordination sphere thereby lowering the Pb
2+
-binding affinity for the site, direct competition of NC protein and Pb
2+
for the Pb
2+
-binding site, or result from a much larger scale destabilization or
unfolding of the molecule. To distinguish among these possibilities, the
conformation tRNA
Phe
was studied in the presence and absence of NC 1-71 and NC 12-55 with and without Mg
2+
by
1
H NMR spectroscopy.
Initial experiments were carried out at 22oC in the presence of 111 mM NaCl, 16.5 mM total Mg
2+
, pH 7.0. These are solution conditions which are largely comparable with the
lead-cleavage reactions described above as well as previously published
1
H NMR spectra of tRNA
Phe
which investigated the thermal melting of yeast tRNA
Phe
in the absence and presence of Mg
2+
(
35
,
43
,
44
). These studies showed that in the presence of 16.5 mM Mg
2+
, yeast tRNA
Phe
melts in a single cooperative transition between 65 and 75oC, consistent with earlier optical and fluorescence spectroscopic studies.
However, in 30 mM Na
+
and the absence of Mg
2+
, optical and fluorescence techniques results show five transitions (cf.
45
) which merged to give basically three distinguishable transitions (numbered I,
II and III with increasing
t
m
) detectable by NMR (
43
). Transition I reports on the loss of tertiary interactions and denaturation of
the acceptor and T[psi]C helices, while transitions II and III report on the melting of the
dihydrouridine (D) and anticodon helices, respectively.
Inspection of the high-field and low-field regions of Mg
2+
-tRNA
Phe
(22oC, 16.5 mM Mg
2+
)
in the absence or presence of 0.15 and 0.4 molar equivalents of NC 1-71 revealed a significant decrease in intensity of the T54 methyl proton
resonance, in addition to other smaller changes in the peak positions and
intensities, e.g., m
5
C49 resonance intensity (data not shown). Although these spectral changes are
small, they may well be selective since the Y37 methyl resonance in the
anticodon loop is not affected at these NC 1-71 to tRNA
Phe
ratios. Inspection of the low-field region of the same spectra revealed only minor generalized
broadening of the imino proton resonances with the overall shape of the
spectrum largely unchanged (data not shown). We attribute this line broadening
to formation of higher molecular weight aggregate(s) since the titration of
tRNA
Phe
with NC 1-71 was terminated at 0.4 molar equivalents due to a visible turbidity in
the sample.
NC 12-55 forms nucleic acid complexes which resist precipitation to a larger
degree than the parent NC 1-71 due at least partly to its lower binding affinity and smaller
electrostatic contribution to the binding energy to nucleic acids (
36
). Figure
3
shows the high-field region of the spectrum of tRNA
Phe
acquired at high Mg
2+
, 22oC alone or in the presence of increasing molar ratios of NC 12-55. Essentially the same spectral changes are observed as previously
described for NC 1-71 with the T54 and m
5
C49 methyl resonances both appearing to decrease in intensity. These again are
by far the most dramatic changes. The imino proton region of the same spectra
showed only a generalized broadening of the resonances and therefore did not
provide any new or additional information than was provided by the NC 1-71 titration (data not shown).
Figure
In order to obtain more information on the perturbations of the imino proton
region which occur upon NC 12-55 binding, yeast tRNA
Phe
was prepared in the absence of added Mg
2+
. Heershap
et al
. (
33
) suggest that in the absence of Mg
2+
and moderate Na
+
concentrations, the tRNA
Phe
molecule maintains all of the secondary and many of the tertiary interactions
associated with the Mg
2+
-conformer; in particular, the T[psi]C-acceptor stems and D-anticodon stems remain coaxially stacked upon one another
just as they are in the Mg
2+
form. The low-field region of tRNA
Phe
in 11 mM NaP
i
, 20 mM NaCl, pH 7.0 at 22oC, is shown in Figure
4
in the absence and presence of 2.0 and 4.0 molar equivalents of NC 12-55. As expected (
32
-
34
,
43
,
44
), the
1
H NMR spectrum of yeast tRNA
Phe
in the absence of Mg
2+
and 20 mM Na
+
is readily distinguished from that obtained in the presence of Mg
2+
(data not shown).
Figure
Figure
4
shows that the addition of 2 or 4 molar equivalents of NC 12-55 again appears to induce generalized broadening of the entire imino
proton envelope, much like that noted above, although to a larger extent. In
order to determine if there may be more selective broadening of a subset of
resonances, the free tRNA
Phe
spectrum was subjected to an artificial line broadening function of 30 Hz, and
superimposed on the actual spectrum acquired in the presence of 4.0 molar
equivalents of NC 12-55 (Fig.
4
,
LB30 scan). Selective resonance broadening is clearly observed with the major
changes summarized as follows. The resonances associated with peaks A, C, G and
S are either unambiguously broadened or markedly shifted in the NC 12-55 complex. These correspond to imino protons associated with the U6-A67, A5-U68, U7-A66 base pairs in the acceptor helix and tertiary imino
proton interaction in the pseudouridine loop, Y55-P58, respectively. In addition, one or more resonances associated with the
degenerate peak D/E as well as peak H are also broadened or lost. Here, it
cannot be determined which imino proton resonance(s) are selectively broadened
due to degeneracy of resonances from various helical stems in the molecule.
However, knowing that imino resonances associated with D-stem base pairs G10-C25 (peak I) and C13-G22 (peak K) are unaffected by NC 12-55 binding makes it unlikely that the other immediately
adjacent D-stem base pairs C11-G24 and U12-A23 associated with peak D/E would be broadened. Therefore,
the lost resonance(s) associated from peak D/E may correspond to A52-U62 base pair in the T[psi]C helical stem. This is consistent with the finding that some or all
of the imino proton intensity associated with peak M, which contains the
acceptor G2-C71 and T-stem m
5
C49-G65 base pairs, and peak N, which contains imino resonances from two T-stem base pairs, G51-C63 and G53-C61, is lost or broadened significantly upon interaction
with NC 12-55. The G2-C71 imino proton resonance is probably not lost from under peak M,
since the immediately adjacent G3-C70 (peak O) and G4-U69 (peak R) resonances appear to be relatively unaffected. Finally,
one or more imino resonances associated with peak H, which derive from the U50-A64 T-stem base pair, as well as two anticodon base pairs are also lost or
broadened. Given the fact the no other anticodon stem imino resonances are
broadened (cf. resonance peaks L and P), it seems plausible that the lost
intensity associated with peak H derives from the U50-A64 base pair, consistent with the broadening of other T-stem imino proton resonances.
These spectral changes are superimposed upon a tertiary or `L-shape' arrangement of the tRNA
Phe
in Figure
5
. The interaction of NC 12-55 with the Mg
2+
-free conformer of tRNA
Phe
appears to induce increased solvent exchange rates associated with the
pseudocontinuous acceptor-T[psi]C coaxially stacked helical stem at the top of the molecule with the
terminus of the acceptor helix largely intact. This parallels exactly what
occurs when Mg
2+
-free tRNA is subjected to increasing temperature: the middle A-U base pairs of the acceptor helix are lost first, with the terminal
G-C base pairs of the acceptor helix undergoing rapid exchange only at
higher temperature (
34
,
35
). In fact, the spectra that we observe for Mg
2+
-free tRNA
Phe
in the presence of 2 and 4 molar equivalents of NC 12-55 bear a qualitative resemblance to those acquired at 36 and 43oC (
44
), respectively under similar solution conditions. These results are consistent
with NC 12-55 acting essentially as a non-specific denaturant, selectively destablizing the least stable
parts of the molecule, rather than recognizing some special structural feature
of the tRNA fold. For example, this effect of NC 12-55 contrasts sharply with those structural changes induced in tRNA
Phe
in the presence of elongation factor Tu-GTP, which induces selective broadening of resonances associated with the
terminus of the acceptor helix (
47
,
48
).
Figure
In the initiation of retrovirus strong stop minus strand reverse transcription,
17-19 nucleotides of the 3'-end of the replication primer transfer RNA must base pair to
a complementary primer binding site sequence located on the genomic RNA. This
sequence of nucleotides extends through the acceptor and T[psi]C helical stems. In order for tRNA to stably base pair to the PBS, two
limiting models of how NC protein might lower the kinetic barrier to
intermolecular duplex formation can be considered. One requires formation of
formally single-stranded tRNA and genomic RNA intermediates stabilized by binding of helix-destabilizing or melting proteins. Another requires rather small
changes in average conformation but one in which transient base pair opening or
local melting of one or both molecules is sufficient to nucleate duplex
formation. As a first step in understanding this process as well as other RNA
chaperonin-like activities of NC proteins at the molecular level (
15
,
16
,
22
), we have investigated the extent of the conformational changes induced by NC
proteins with the model tRNA, tRNA
Phe
, using lead ribozyme activity and imino proton NMR spectroscopy as structural
probes.
In order to effect self-hydrolysis by Pb
2+
, the Pb
2+
ion must be precisely coordinated in a pocket formed by eight residues in the
elbow region so that the cleavage between the D17 and G18 can occur with high
efficiency and specificity (
49
). When a stoichiometric amount of HIV-1 NC protein is added to these reactions, the rate of lead cleavage is reduced >= 20-fold (Fig.
2
; Table
1
). Since the hydrolysis rate at 25oC (~1 * 10
-5
s
-1
) is presumably far slower than the rate of dissociation of NC 1-71 from tRNA
Phe
given a binding affinity of ~10
6
M
-1
(
10
), then NCs may inhibit the cleavage reaction by directly competing with Pb
2+
ions for binding tRNA. Consistent with this, under conditions at which the rate
of lead cleavage is faster (increased temperature and/or increased lead
concentration), inhibition by NC is somewhat reduced (data not shown). The
sensitivity of this assay to conformational changes in tRNA is extremely high
since the lead cleavage reaction becomes inhibited when the last three base
pairs of the acceptor stem are broken (
42
). Thus, NC 1-71 may inhibit the rate of lead cleavage by binding to a region(s) of the
tRNA molecule which causes conformational changes which extend into the elbow
region of the molecule directly. Another possibility is that the conformational
changes induced by NC are located distant to the elbow region but result in
structural changes in the elbow since the folding of tRNA
Phe
is highly cooperative (
42
). A third possibility is that the conformational changes induced by NC are far
more extensive, resulting in denaturation of helical regions of the molecule.
To discriminate among these possibilities, we carried out
1
H NMR studies of the yeast tRNA
Phe
-NC protein complexes formed under a variety of solution conditions.
Experiments carried out in the presence of high Mg
2+
, conditions comparable with the lead cleavage reactions, reveal that the binding of NC 1-71 and NC 12-55 induces only small changes in the structure of the molecule, as reported on select methyl and methylene protons associated
with bases very close to the catalytic Pb
2+
binding pocket. These perturbations presumably subtly alter the orientation of
the binding or catalytic groups within the enzyme, thereby greatly reducing the
rate of the hydrolysis reaction. They are inconsistent with significant
unfolding of the molecule, at least at the protein:tRNA ratios used in this
study. These findings are fully consistent with our previous studies, which
showed that in the presence of Mg
2+
, tRNA was largely resistant to NC 1-71-induced conformational changes as monitored by near-UV CD spectroscopy at stoichiometric or lower concentrations
of NC 1-71 (
3
).
When a subset of tertiary interactions are minimal or absent (i.e., in low Na
+
and in the absence of Mg
2+
), the NMR studies suggest relatively modest changes in the secondary structure
content of the molecule induced by the binding of NC 1-71 or NC 12-55, with the major perturbations in the exchange rates of select
imino protons in the T[psi]C and acceptor stems of the molecule. These regions of the molecule are
precisely those which are most susceptible to temperature-induced destabilization (
43
). Our findings are fully consistent with a non-specific binding model where, at stoichiometric or slightly greater
concentrations, NC protein acts as a non-specific denaturant, i.e., it interacts with tRNA in a way in which the
weakest stabilizing interactions are weakened or broken. As an enhanced
exchange rate is generally believed to reflect enhanced rate of base pair
opening and if base pairs in the acceptor and T[psi]C helix are open for a longer fraction of time, they would become available
for intermolecular base pairing with another complementary sequence, such as the viral RNA primer binding site. Given the high
cooperativity of the coil to helix transition and the rate-limiting step of helix initiation, transient opening of base pairs in this
region of the molecule would enhance the nucleation rate or intermolecular
helix initiation rate and thereby enhance the rate of duplex formation as has
been observed experimentally (
15
,
16
). This does not in any way require a unique conformation of the bound NC-tRNA complex, as would be required by a formally site-specific binding model. For example, we note that we find no evidence that NC 1-71 or NC 12-55 binds at or near the anticodon domain of the
molecule, where a previously postulated high affinity site was thought to
localized on tRNA
Lys,3
(
10
,
50
). This is presumably due to the fact that these interactions in the anticodon
domain are significantly more stabilizing than those of acceptor and T[psi]C helices, at least in tRNA
Phe
.
In an attempt to determine the functional domains of NC 1-71 that are necessary and sufficient to change the conformation of tRNA
Phe
as measured by the inhibition of lead cleavage reaction, domain deletion
molecules (NC 1-57, NC 12-71 and NC 12-55) were tested and compared with intact wild-type Zn
2
NC 1-71. We find that deletion of the C-terminal and/or N-terminal domains of the NC 1-71, while significantly weakening the binding affinity
of the parent molecule to poly(A) (
2
,
36
) results in an undetectable change in ribozyme inhibitory activity under the
concentrations used in this study (Table
1
). To evaluate the importance of the zinc-fingers in HIV-1 NC protein in inhibiting the self-cleavage of tRNA, the activities of apo
S
-methylated NC 12-71 and NCp7C were compared with Zn
2
NC 12-71. Removal of Zn(II) from these molecules appeared to significantly
attenuate their ability to protect tRNA
Phe
from lead cleavage (Table
1
). Thus, in context of the N-terminally deleted NC 12-71 molecule, an intact zinc-finger domain(s) appear to confer additional binding energy
on the complex, consisent with our binding studies with non-specific homopolymeric nucleic acid lattices (
36
; R.K. and D.P.G., unpublished results). Interestingly, when either the N-terminal or C-terminal zinc coordination site is independently destroyed by tandem Cys -> Ser substitutions in the context of NC 1-71, both molecules effectively inhibit the lead
ribozyme activity of tRNA
Phe
(data not shown).
The basic domains which surround the zinc-fingers are highly conserved in amino acid composition but not sequence
among the retroviral NC proteins (
1
). The function of these regions in HIV-1 NC and Moloney Murine Leukemia Virus (MoMuLV) NC have been investigated
by Darlix and coworkers (
9
,
29
). They find that when either the N-terminal flanking region (residues 1-14) or the entire C-terminal flanking region (residues 52-72) is removed, the tight interaction of NC protein
with native primer tRNA
Lys,3
and the ability of NC to anneal primer tRNA to PBS and to dimerize the genomic
RNA are lost. However, if the short sequences containing the basic residues
flanking the zinc-finger regions are retained, then the activity of this protein is only
partially or not affected. For example, deletion of the N-terminal 12 and C-terminal 8 amino acids of their synthetic NCp7 (which has 72 amino
acids) to create NC 13-64 appears to have little or no effect on NC protein activity
in vitro
. In addition, these workers found that the two minimal basic domains
surrounding the first finger motif must be covalently linked to retain NC
protein activity since a mixture of peptides corresponding to the N- and C-terminal domains are still inactive (
9
). They suggest that the basic residues and the N-terminal zinc-finger cooperate to select and package the genomic RNA
in vivo
, with the zinc-fingers directing the spatial recognition of the target RNAs by the bi-partitate basic domain on either side of the zinc-finger structure. Our cleavage results with NCp7C and apo
S
-methylated NC 12-71 are consistent with this picture and suggest that a cooperative
interaction between the N-terminal domain and the zinc-finger region is required for efficient destabilization of tRNA
Phe
as well as tight binding to both viral (
51
) and model (
36
) RNA molecules.
This work was supported by a grant from the National Institutes of Health
(GM42569) to D.P.G. The NMR instrumentation was supported by a grant from the
National Science Foundation (BIR-9217413) to D.P.G. This work was performed in partial fulfillment of the
requirements for the PhD degree at Texas A&M University (to R.K.).
NC protein
Concentration ([mu]M)
k
obs
(* 10
5
s
-1
)
None
0
12.1 (+-1.4)
NC 1-71
2.0
<= 0.5
b
NC 1-57
2.0
<= 0.5
b
NC 12-71
2.0
<= 0.5
b
NC 12-55
2.0
<= 0.5
b
Apo
S
-CH
3
NC 12-71
0.5
12.4 (+-0.2)
Apo
S
-CH
3
NC 12-71
1.0
13.4 (+-0.2)
Apo
S
-CH
3
NC 12-71
2.0
15.7 (+-0.8)
Apo
S
-CH
3
NC 12-71
5.0
23.4 (+-1.0)
NCp7C
0.5
8.8 (+-0.4)
NCp7C
1.0
10.2 (+-0.7)
NCp7C
2.0
11.2 (+-0.8)
NCp7C
5.0
12.8 (+-0.9)
NCp7C
10.0
10.7 (+-0.5)



REFERENCES
Return
