http://www.abbs.info e-mail:[email protected] ISSN 0582-9879 ACTA BIOCHIMICA et BIOPHYSICA SINICA 2002, 34(5): 533-543 CN 31-1300/Q |
Mini Review |
Methods
for Structural and Functional Analysis of an RNA Hexamer of Bacterial Virus
phi29 DNA Packaging Motor
(
Department of Pathobiology and Purdue Cancer Center, Purdue University, West Lafayette,
IN 47907, USA )
One
prominent feature in the assembly of all linear ds-DNA viruses is that their
lengthy genome is packed with a swift velocity into the pre-formed protein
coating and packaged to a near crystalline density (for review, see references[1-5]).
This DNA motion, which is an
energetically unfavorable process,
is accomplished by an ATP-hydrolyzing motor involving a connector and
two nonstructural components with certain characteristics typical of ATPases.
In phi29, one of the nonstructural
proteins for DNA packaging is an RNA (pRNA) molecule[6-9].
The connector is a 12-subunit hollow truncated cone cylinder having a central
channel with a diameter of about 4-nm through which DNA enters the procapsid
during packaging[10, 11]. The 120-base pRNA (Fig.1) encoded by the
virus, binds to the connector[12,
13] and is not present in the mature phi29 virion. The requirement for
pRNA in phi29 assembly appears to be very specific in that pRNAs from other
phages cannot replace the phi29 pRNA in in vitro packaging assays[14]
and that a single base mutation can render the pRNA inactive[15].
Fig.1 pRNA secondary structure
adapted
from[56] with
permission from J Biol Chem.
Phi29
is the most efficient and the best-characterized in vitro DNA packaging
model system. Most essential components required to reconstitute the phi29 in
vitro DNA packaging activity have been well defined. Direct force
measurements have shown that the phi29 packaging motor is the most powerful of
all reported molecular motors,
producing a force of 57 pico-Newtons[22, 23]. In
addition, the crystal structure of
the 12-subunit (gp10) connector has been solved[11, 24]. One
nonstructural component is a protein,
gp16, that has been shown
to contain consensus ATP-binding domains[25] and is able to
hydrolyze ATP. It has been found that one ATP is needed to package two base
pairs of DNA[25].
This
review will focus on the work produced in the author's laboratory and will
concentrate on methods and approaches that were used to investigate pRNA
structure and function.
In
1997, a comprehensive paper was published to describe the role of phi29 pRNA in
DNA packaging[16] which provided an elegant model for explaining the
mechanism of connector (portal vertex) rotation and the quantification of
energy (ATP) usage[16]. The model presented advocated that phi29 DNA
packaging is accomplished by a mechanism similar to driving a bolt with a hex
nut, which consists of six
DNA-packaging pRNAs (Fig.2).
Recent research[11,19,20,22,23]
in this area supports the conclusion about consecutive action of six pRNAs in
driving the DNA rotation machine.
Fig.2 A model to depict the sequential action
of pRNAs in phi29 DNA packaging motor
The hexagon represents the phi29
connector and the surrounding pentagon represents the capsid. Six protrusions
represent six pRNAs. The variable pRNA patterns portray the pRNA in serial
energetic states. For example,
pRNA 4 in panel A is in a contracted conformation, and pRNA 1 is in a relaxed
conformation. Arrows marks the
different transition states of pRNA 1. Steps A to G show the six steps of
rotation. Each step rotates 12°, since a five to six-fold symmetry
mismatch generates 30 equivalent positions, and 360°/30
= 12°. The portal vertex turns 72°
after six steps. For example, pRNA 1 moves from vertex a in A to
vertex b in G, and rotates 72°.
Each step consumes one ATP to induce one conformation change of pRNA, and six ATPs are used for the
transition from one vertex to another. 30 ATPs are used for each 360°
rotation (Reprinted from[16] with permission from J Virology).
The
pRNA contains two functional domains:
one for connector binding and the other for DNA translocation. Mutagenesis as well as chemical and
nuclease probing have revealed that the pRNA binds to the connector leaving the
essential 5′/3′domain
free for interaction with other components, such as gp16,
DNA-gp3 or other components on the procapsid. It has been reported that
the C18C19A20 bulge of the pRNA is
solvent-exposed when pRNA is bound to procapsid[27] and is critical
for DNA translocation[16, 28-31].
The C18C19A20 bulge might be directly involved
in interacting with ATP, gp16, DNA-gp3 or capsid components. It is predicted
that pRNA is part of an ATPase and possesses at least two conformations;
a relaxed and a contracted one. Alternating between contraction and relaxation
driven by ATP hydrolysis, each
member of the hexameric RNA complex helps rotate the translocating machine.
The
requirement of an intermolecular loop/loop interaction between individual pRNA
molecules during DNA packaging has led to the belief that pRNA forms a hexamer[19,
20] and supports a pRNA sequential action model. The pRNAs may need to communicate with
each other during DNA packaging to ensure that the motion is consecutive. Inter-pRNA interactions via loops might
serve as a link to pass a signal to adjacent pRNAs, regulating sequential conformational changes and/or
interactions. Thus base pairing between the right and left-hand loops might be
necessary to transfer a conformational change from one pRNA to an adjacent one.
Four
major approaches have been undertaken in this lab to study the structure and
function of pRNA.
The
first approach was the development of a highly sensitive in vitro phi29
virion assembly system for the assay of pRNA activity[32, 33]. With
this system, up to 5×109
infectious virions per ml can be obtained in the presence of pRNA, yet not a single infectious virion is
detected in the absence of pRNA.
Therefore, a system with a
dynamic range of more than 9 orders of magnitude and with a sensitivity level
of as few as 0 infectious virions can be used for the analysis of pRNA
structure and function.
The
second approach in analyzing pRNA function was the construction of circularly
permuted pRNAs (cp-pRNA) in which any internal base of the pRNA could be
reassigned to serve as new 5′-
or 3′-termini[28, 34] (Fig.3). The
circular permutation system greatly facilitated the construction of mutant pRNA
via PCR and enabled the labeling of any specific internal base by
radioisotopes, fluorescence[35]
or photoaffinity agents.
Fig.3 Tandem DNA for the synthesis of
circularly permuted pRNA (cp-pRNA)
Adapted
from[34] with permission from Virology.
The
fourth approach to analyze pRNA structure and function was the design of new methods
to determine the stoichiometry of pRNA in DNA packaging[18, 19, 38].
To determine the role of pRNA in DNA packaging, it is crucial to know how many copies of the pRNA are
involved in each DNA packaging event.
Three novel methods have been developed to determine the stoichiometry
of the pRNA, and have led to the
conclusion that six pRNAs are present in each DNA translocating motor. These
methods include: (a), binomial distribution (Yang Hui Triangle)[18, 38]
(Fig.4); (b), comparing slopes of concentration dependence[18, 33]; and (c), finding the common multiple of 2, 3, and 6 by
using a set of two interlocking pRNAs,
three interlocking pRNAs and six inter-locking pRNAs[19].
Fig.4 Stoichiometry determination by binomial
distribution
Z represents the total pRNA number, initially assigned a theoretical value
from 1 to 12, per procapsid
to be determined. The empirical curve from mutant pRNA P8/P4 falls between the theoretical curves for Z = 5 and Z = 6 (Reprinted from[18]
with permission from J Virology).
Although
nucleotide derivatives have been found in RNA, the primary sequence of the RNA molecule is nevertheless as
simple as DNA, since both are
composed of four nucleotides. All DNA molecules appear as double helices, while RNA has a diverse structure.
Intriguingly, small RNA molecules, containing only the four nucleotides
A, G, C, and U, exhibit versatile biological
functions. Such versatility is
ascribed to the flexibility and complexity in RNA structural folding. NMR and
X-ray crystallography have been used to obtain a physical tertiary structure of
RNAs. Currently, NMR can only be applied to an RNA
molecule with a size of less than 40 nucleotides. X-ray crystallography of structural RNAs has proven
difficult. The difficulty, uncertainty and time-span in obtaining
a diffractable RNA crystallographic structure, as well as the impossibility of using NMR for large
RNAs, compel the use of
alternative approaches to obtain information on RNA structures.
3.1
Genetic analysis by truncation,
insertion, deletion and
mutation
The
establishment of the highly sensitive in vitro phi29 assembly system
(Section 2) greatly facilitated the genetic analysis of pRNA structure[32,
33]. Taking advantage of the circularly permuted pRNA system (Fig.2)
(Section 2)[28, 34],
the technique of two-step PCR,
and the relatively small size of the pRNA (120 bases), mutant pRNAs can be easily constructed
with truncation, deletion, insertion and mutation targeting any
desired position[29]. A plasmid DNA with two tandem RNA coding
sequences linked with three bases AAA were used as templates to generate PCR
DNA fragments with primer pairs containing either the T7 or SP6 promoter and
mutations to pRNA. The DNA fragments from PCR were used to transcribe mutant or
circularly permuted pRNAs in vitro with either T7 or SP6 RNA polymerase.
In combination with the aforementioned in vitro assembly assay
system, dozens of mutant pRNAs can
be obtained and tested in one or two weeks.
By
the use of the truncation and deletion techniques, it was revealed that three
nucleotides, U72U73U74,
predicted to form a bulge located at a three-helix junction (Fig.2), function to provide flexibility in pRNA folding[28]. Three
other nucleotides, C18C19A20,
were shown to be present on the surface of the pRNA as a bulge that is not involved in procapsid binding but
is essential for DNA packaging[30].
3.2
Phylogenetic analysis
Phylogenetic
analysis of RNA is used to compare the sequences of RNA molecules with
identical or similar functions from different species. A common secondary
structure for RNA molecules with a similar function is deduced from such
analyses. The theory behind such logic is that RNA structure plays a critical
factor in RNA function. Nature
would select the most stable molecule with the best-fit structure or with
acceptable base co-variations. Later on,
such phylogenetic analysis of species from nature would be expanded into
molecules made artificially, such
as complementary modification or SELEX that will be described below.
Phylogenetic
analysis revealed that pRNAs from bacteriophages SF5, phi29,
PZA, M2, NF, GA1 and B103,
which have a very low sequence identity and few conserved bases, very
impressively show similar predicted secondary structures[14, 29].
The requirement for pRNA in phi29 assembly is very specific in that pRNAs from
other phages cannot replace the phi29 pRNA in in vitro packaging[14]
and that a single base mutation can render the pRNA completely inactive[15].
Thus, similar structures do not
translate into identical function. Interestingly, phylogenetic analysis revealed that the right (upper) loop
of each pRNA was complementary to the left (lower) loop within the same
molecule[29]. Complementary modification studies reveal that the
pairing is inter-molecular[19, 20] rather than intra-molecular
(pseudoknot[39]) and that two G/C pairs are sufficient to mediate
the interaction[19, 20].
3.3
Complementary modification
Ano
ther approach to confirm base-pairing in predicted RNA structure is
complementary modification. Before the conclusion that “G
pairs to C”
in an RNA structure is drawn, at
least three mutants should be constructed and analyzed. First, mutants with either the G changed to A (or U) or the C
changed to U (or A) should be inactive.
In addition, a mutant with
both the Gs changed to As (or Us) and the Cs changed to Us (or As) should
restore the activity.
Computer
predictions of the phi29 pRNA secondary structure[40] showed that
the 5′ and 3′
ends are paired. An extensive series of helix disruptions by base substitutions
almost always resulted in the loss of DNA packaging activity. Additional compensatory mutations that
restored the predicted base pairings rescued the activity of pRNA[15, 28,
39]. Such complementary modification has led to the conclusion that bases
1-3
are paired with bases 117-115;bases
7-9
are paired with bases 112-110;bases
14-16
are paired with bases 103-101;and
bases 76-78
are paired with bases 90-88
(Fig.1). This second site suppression confirmed the existence of a helical
structure that is essential for pRNA function.
Complementary
modification has also been used to study inter-pRNA loop/loop interactions in
dimers[19, 20, 29]. A series of mutant pRNAs carrying mutated right
and/or left-hand loop sequences were constructed such that loop sequences were
non-complementary. Each inactive mutant was mixed with another inactive mutant
such that the loop sequences were complementary in trans, allowing the
formation of intermolecular base pairing. All mutant pRNAs that had unpaired
right and left loops, such as pRNA A-b′,
were inactive in phi29 assembly when used alone. However, when two inactive pRNAs that were
trans-complementary in their right and left loops, for example pRNA A-b′and
B-a′, were mixed in an equimolar ratio, full
activity was restored. The observed activity of a mixture of two inactive
mutant pRNAs confirmed that the right loop interacted with the left loop
intermolecularly to form an RNA dimer.
3.4
Chemical modification
Chemical
modification was employed to probe pRNA structure. The modifying agents used
include dimethyl sulfate (DMS), which methylates A at N1, G at N7 and C at N3[41,
42];kethoxal,
which modifies G at N1 and N2[43];and
1-cyclohexyl-3-(2-morpholinoehtyl)-carbodiimide metho-p-toluene sulfonate
(CMCT), which attacks U at N3 and G at N1[41-43].
In principle, only unpaired bases are susceptible to chemical attack. The
chemicals alter unpaired specific functional groups of RNA bases and thus
provide information regarding base pairing, base stacking, and the tertiary
interactions of specific bases within an RNA. Locations of modified bases can
be identified by primer extension with reverse transcriptase[41, 44].
Chemical modification of a base is a good indication that the base is unpaired
and that the specific functional group is solvent-exposed, and thus is a
possible candidate for intermolecular interactions. Lack of modification will
most likely be due to base pairing, especially in helical regions, but may also be the result of tertiary
interactions or non-canonical base-base,
base-sugar, or base-phosphate interactions[43] in loop or
bulge regions. Chemical modification data can provide information on base
accessibility, which is helpful in
assessing predicted secondary structures,
evaluating 3-D molecular models,
and analyzing RNA/protein interactions.
Phi29
pRNAs including various mutants have been modified with DMS, CMCT, and kethoxal[27, 41, 43]. Chemical modification
showed that the sequence C18C19A20, which is essential for DNA packaging
but dispensable for procapsid binding,
is accessible to chemicals in monomers and dimers as well as procapsid-bound
pRNA[27, 30]. These results indicate that CCA, though not involved in procapsid
binding[28, 31], is
present on the surface of the pRNA as a bulge which may interact with other DNA
packaging components[30] (Fig.5). This conclusion is supported by
mutation studies on the CCA bulge.
Fig.5 Direct observation of pRNA three-dimensional structure with
cryo-AFM (atomic force microscopy) (A and B) and shape comparision with computer
models (C and D). E and F are drawings to depict the structure of monomers and
dimers, respectively
Reprinted
from[27] with permission from RNA.
3.5
Chemical modification interference
Chemical
modification interference has been performed to determine which pRNA bases are
involved in dimer formation. The monomer pRNA B-a′was
treated with either DMS or CMCT and then mixed with unmodified monomer A-b′in
order to test its competency in dimer formation. If the base is involved in
dimer formation, chemical modification of this base could interfere with the
ability of pRNA B-a′to
form a dimer with pRNA A-b′,
and thus this pRNA will be present in a fast migrating band representing
monomers in native gels. Chemical modification was performed, and RNAs (both
fast and slow migrating corresponding to pRNA monomers and dimers,
respectively) were isolated from gels. After isolation, both monomers and dimers were subjected
to primer extension to identify the modified bases. The concentration of the
chemicals was titrated to ensure that on the average only one base per pRNA was
modified[45]. The general theory behind the experiment was that a
pRNA B-a′containing
an interfering modified base would appear in the fast migrating monomer
band, while pRNA B-a′containing
a non-interfering modified base would appear in the slower migrating dimer
band. Chemical modification interference analysis reveals that bases U54,
G55, U59, C65, A66, A68,
U69, A70, C71, C84, C85,
C88, A89, A90 and C92 interfered
with dimer formation, and thus are involved in dimerization, while bases 72-81
were not involved[45] as shown in the computer model of dimer
(Fig.6).
Fig.6 Computer models of pRNA monomer (A), dimer (B), hexamer (C),
and pRNA/connector complexes (E and F). D is the crytall structure of connector
11, 72
Reprinted
from[56] with permission from J Biol Chem.
The
chemical psoralen can intercalate into RNA or DNA helices and, upon irradiation
with 320-400
nm light, freeze (in helix or
pseudoknot) uridines of RNA or the thymidines of DNA by covalent attachment46
if they are in close proximity (in helix or pseudoknot)[47, 48]. The
sites of crosslinks can be determined by primer extension[49] and/or
mung bean nuclease treatment[50]. The psoralen derivative, AMT (4′-aminomethyl-4,
5′, 8-trimethyl psoralen), was used to
crosslink pRNA due to its solubility[49]. Psoralen crosslinks only
RNA or DNA but not protein, which
is different from the azido group (see below) which crosslinks non-specifically
to both protein and nucleic acids. Psoralen, however, can
induce intra-molecular crosslinks within the pRNA even in the presence of other
proteins, such as procapsid or gp16. Thus, pRNA conformations in different
environments can be detected.
Psoralen crosslinks can also be reversed by 254 nm irradiation. With the
use of a 2-dimensional gel electrophoresis[46, 48, 51] and 5′-end
radiolabeled cp-pRNAs, pRNA
conformational change in the presence of different packaging components can be
investigated. Psoralen crosslinking
experiments revealed that pRNA had at least two conformations--one
that was able to bind procapsid and the other that was not able to bind. In the absence of Mg2+, the region comprising bases C67
to A70 was in close proximity to bases U31 to U36, since these two areas were crosslinked
together by psoralen[49].
3.7
Photoaffinity crosslinking with GMPS/Aryl azide
Aryl
azides contain functional groups that are chemically inert in the absence of
light, but can be converted to a reactive nitrene after long wavelength UV
irradiation[52, 53]. Thus, aryl azides can be incorporated into RNA
to obtain structural data[54]. Aryl azide has been specifically
attached to the 5′-end
of pRNAs or cp-pRNAs. For this 5′-end
labeling, 5′-thiophosphate
pRNA or cp-pRNA is synthesized by in vitro transcription in the presence
of excess GMPS (guanine-monophosphorothioate) over GTP[53]. GMPS is
an efficient primer in RNA synthesis with T7 RNA polymerase but cannot be used
by this enzyme for chain elongation. The 5′-thio-pRNA
and the 5′-thio-cpRNAs
are then treated with azidophenacyl bromide to produce the 5′-azido-pRNA
and 5′-azido-cp-pRNAs, respectively, by the nucleophilic displacement of bromine[53].
The azido group is converted to a reactive nitrene by long wavelength UV
irradiation, which is then
inserted into nearby bonds resulting in covalent crosslinks[52].
Since it is possible to generate active cp-pRNAs by assigning certain internal
sites of the pRNA as new 5′-
and 3′-termini (Section 2)[28, 34],
specific internal bases of the pRNA have been uniquely labeled with
photoaffinity crosslinking agents to analyze inter- and intra-molecular
interactions. When necessary, the 5′-end
of the RNA can also be labeled with [32P]. Crosslinked RNAs were
separated from uncrosslinked RNAs by denaturing gel electrophoresis, and crosslink sites were determined by
primer extension[45, 55]. Bases identified as crosslink sites by
primer extension indicate that these bases are in close proximity to the
photoagent labeled base. The use of cp-pRNAs allows the identification of
intra-molecular contacts throughout the pRNA molecule, and such data have been used as
distance constraints in molecular modeling studies[28, 34, 45, 55]
(Section 3.13).
Intra-molecular
crosslinking of monomers[45] revealed that G108 neighbors
C10 and G11;G75
is near bases 26-30, while G78 is near U31.
The azidophenacyl group is only 0.9 nm in length, but experimental data has demonstrated that the
cross-linking group can reach distances of 1.2 nm (Norman Pace,
personal communications). These distances have been used as constraints
in the computer modeling of the pRNA monomer structure (Fig.7).
Fig.7 Comparison of chemical modification patterns
of monomer (A) and dimer (B)
The black arrow, gray square, and
double-lined arrow indicate a strong, moderate, and weak modification of
bases, respectively. C is a model
to portray the formation of dimer. The four base-pairs (45-48/85-82
in gray boxes) were modified in monomers, but were protected from chemical
modification in dimers (Adapted from[27, 45, 56] with permission
from RNA and J Biol Chem).
3.8
Photo-crosslinking by phenphi
Unlike
psoralen, phenphi[(cis-Rh(phen)(phi)Cl2+ (phen =
1, 10-phenanthroline;and
phi = 9, 10-phenanthrenequinone diimine)] induces covalent bonds between
guanosine bases upon UV activation. Phenphi has also been shown to crosslink
pRNA and has revealed the close proximity of bases G75, G28
and G30 of pRNA[57].
3.9
Ribonuclease probing
Some
ribonucleases are sensitive to RNA secondary structure. For example, RNases T1
(specific for GpN linkages), U2
(specific for ApN linkages), and S1 prefer to cleave single-stranded RNA.
Nuclease V1 is specific for double stranded RNA. End-labeled pRNA and cp-pRNA
in various solutions containing Mg2+ or procapsid individually or in
combination, have been probed by
T1, U2 or V1 nucleases[14,
49].
T1
and V1 were used to distinguish the loops and helices of four RNAs with similar
function[14]. RNase footprinting has also been performed to detect
the sequences that bind procapsid[49, 58]. In addition, T1 nuclease
has been used to study changes in pRNA conformation[49]. Since the
activity of RNase T1 is Mg2+ independent, this enzyme was used to
investigate the conformational change of pRNA in the presence or absence of Mg2+[49].
A
Mg2+-induced pRNA conformational change was verified by T1
ribonuclease probing[49]. The pattern of partial digestion of pRNA
by T1 provided strong evidence for the presence of two conformations, dependent
on either the presence or absence of Mg2+. Without Mg2+,
strong cleavages by T1 were seen at bases G28, G30, and G34.
While in the presence of Mg2+, these three bases became more
resistant to T1 attack, indicating a conformational change or refolding of pRNA
stimulated by Mg2+[49].
3.10 Footprinting
Foot
printing is a technique derived from nuclease probing or chemical modification
and is particularly useful in probing the interaction of RNA with proteins. The
procapsid/pRNA complex was probed with nucleases A, T1 and V1[58].
The optimal concentration of enzymes was determined empirically to ensure, on
the average, one cleavage site per RNA molecule. Results of footprinting
studies revealed that bases 22-84
were protected from enzyme digestion[58] indicating that the region
from bases 22-84
contacts the procapsid.
3.11 SELEX (systematic evolution of ligands by exponential
enrichment)
In
vitro evolution is a powerful tool to study
consensus elements of RNA structure and function. Starting with a library
containing pRNA sequences with random mutations within a defined region, in
vitro evolution techniques allow the selection of pRNA variants that can
bind a specific ligand. Such selection for interacting species is based on
different primary structures that can adopt the same structural feature as wild
type RNA. SELEX allows screening for co-variation of several nucleotides and
can be used to reveal noncanonical interactions that are difficult to prove by
classic genetic and biochemical approaches[59, 60]. SELEX has been
used for the selection of pRNA sequences that bind procapsids and are involved
in intermolecular loop/loop interactions[61]. It was concluded that
the wild type pRNA sequence is the most suitable sequence for procapsid binding.
3.12 Images revealed by cryo-AFM (atomic force microscopy)
Atomic force microscopy has been used by several
investigators to detect images of RNA (Fig.8) in a denatured conformation. As
the first attempt to test whether this technique can be used to detect the 3D
structure of RNA in native conformation, cryo-AFM has been performed on the
phi29 pRNA monomer, dimer and trimer[27, 37, 45].
Fig.8 Computer model of pRNA dimer is in
accord with the results of chemical modification interference
Bases that are demonstrated to interfere
with dimer formation are shown as gray spacefill bases in the pRNA subunits.
The dimer model is in agreement with the empirical data by showing that these
bases are located at the interface of the two pRNAs of the dimer (adapted from[56]
with permission from J Biol Chem).
Fig.9 Computer model of pRNA dimer are in
accord with the results of intermolecular azidophenacyl photoaffinity
crosslinking
G82 (in black spacefill) in
one pRNA unit is in close proximity to G39, G40, A41,
C49, G62, C63, and C64 (in gray
wireframe) of the other pRNA unit, as determined by base-specific photo
affinity crosslinking (adapted from[56] with permission from J
Biol Chem).
The
goal of modeling pRNA structure is to organize collections of structural data
from crosslinking, chemical or
ribonuclease probing, chemical
modification interference, cryo-AFM
and other genetic data into a three-dimensional form. Since a large number of structural constraints are
available, computer programs can
successfully construct three-dimensional structures[56, 62, 63].
pRNA
monomer, dimer and hexamer (Fig.8) were produced on Silicon Graphics Octane and
IndigoⅡ
computers running IRIX 6.2 or 6.5,
using the programs NAHELIX,
MANIP, PRENUC, NUCLIN, and
NUCMULT[64, 65]. The modeling was performed based on the following
assumptions:(1)
All helices were modeled as regular A-form double helices. (2) Internal loops and mismatched bases
were constructed by maintaining the integrity of the double helix while
optimizing base pairing and stacking inside the loop, as suggested by most
structural data from X-ray and NMR analysis. (3) A general rule for the
modeling of the RNA hairpin loop has been proposed[66], which
involves maximal stacking on the 3′side
of the stem and enough nucleotides stacked on the 5′side
to allow loop closure, as found in the anticodon loop of tRNA. (4) Bulges less
than four bases in size were modeled either radiating out from stems to avoid
helical distortion, while larger loops were constructed protruding from the
stems or within the helical domain, causing the helical axis to bend.
Parameters for stacking energy are considered in order to decide whether bulges
should be protruding from or within the helical stems[67]. (5) Helix
untwisting or twisting, helix-helix interactions, triple base interactions[68], pseudoknots, or other higher order
structures have been built into the model at constant geometrical distances
while allowing certain torsion angle variation. The program regarding RNA
flexibility has been applied to the construction of the pRNA UUU bulge at the
three-helix junction. This three-base bulge has been found to provide
flexibility for the appropriate folding of pRNA. Conventional computer
algorithms involving the minimization of empirical energy functions have been
considered. Twelve angstroms has been considered as a maximum distance
constraint when bases are crosslinked by GMPS/aryl azide. Modified distance
geometry and molecular mechanics algorithms using simplified “pseudo
atom” representations have been considered to
generate structures consistent with data from crosslinking, chemical
modification and chemical modification interference. A constraint satisfaction
algorithm is combined with discrete representations of nucleotide conformations
to refine the disturbed area in order to ensure the normal representation of
all atoms.
X-ray
crystallography studies have revealed that the phi29 connector contains three
sections, a narrow end, a central section, and a wider end, with diameters of
6.6 nm, 9.4 nm, 13.8 nm, res-pectively (Fig.8)[11, 69]. The
hexameric pRNA model by Hoeprich and Guo[56] contains a central
channel with a diameter of 7.6 nm, that perhaps can sheath onto the narrow end
of the connector to perhaps be anchored by the central section of the
connector, which is wider than the central channel of the pRNA hexamer (Fig.8).
As
noted earlier, pRNA contains two functional domains (Fig.1);one
for connector binding and one for DNA translocation. The connector binding
domain is located in the middle of the pRNA primary sequence, i.e. bases 23-97, and the DNA translocation domain is
located at the 5′/3′
paired ends. It has been predicted that the connector protein (gp10) contains a
conserved RNA recognition motif (RRM), located between residues 148-214
of each gp10 monomer. This region
of gp10 is located at the narrow end of the dodecameric connector that
protrudes from the procapsid[70, 71]. The hexamer model by Hoeprich
and Guo[56] complies with the aforementioned data by showing that
pRNA bases 23-97
(colored green in Fig.6C, E &
F), within the connector binding
domain, interact with the
predicted RRM motifs of the connector (Fig.2E and F in blue), while the 5′/3′
paired region (Fig.2E in red and cyan),
comprising the DNA translocation domain, extends away from the connector.
Acknowledgements I would like to thank Jane
Kovach, Dan Shu and Stephen Hoeprich for the manuscript preparation, Drs. Mark Trottier and Chaoping Chen
for critical review, Dr. Zhifeng Shao for providing his AFM images, and Dr.
Michael Rossmann for his permission to use his published connector structure.
1 Earnshaw WC, Casjens SR. DNA packaging by the double-stranded DNA
bacteriophages. Cell,
1980, 21:
319-331
2 Black LW. DNA packaging in dsDNA
bacteriophages. Ann Rev Microbio,
1989, 43:
267-292
3 Guo P. Introduction:
Principles,
perspectives, and potential
applications in viral assembly. Seminars in Virology (Editor's Introduction), 1994, 5(1): 1-3
4 Guo P. The mechanism of DNA packaging
in double stranded DNA bacteriophages. Virologica Sinica, 1988, 2: 229-246
6 Guo P, Erickson S,
Anderson D. A small viral RNA is required for in vitro packaging
of bacteriophage f29
DNA. Science 1987; 236: 690-694
7 Guo P, Trottier M. Biological and biochemical properties of the
small viral RNA (pRNA) essential for the packaging of the double-stranded DNA
of phage f29.
Seminars in Virology,
1994, 5:
27-37
8 Guo P. Structure and function of
hexameric RNA to drive viral DNA packaging motor. Prog in Nucl Acid Res
& Mole Biol, 2002;
72: 415-418
9 Guo P. Inhibition of novobiocin, coumermycin A, oxolinic acid, nalidixic, ethidium bromide and ATP analogue on the packaging of f29 DNA in vitro. Virologica
Sinica, 1988, 2: 198-205
10 Jimenez J, Santisteban A,
Carazo JM, Carrascosa JL.
Computer graphic display method for visualizing three-dimensional biological
structures. Science,
1986, 232:
1113-1115
11 Simpson AA, Tao Y, Leiman
PG, Badasso MO, He Y, Jardine PJ, Olson
NH et al. Structure of the bacteriophage phi29 DNA packaging motor. Nature, 2000, 408(6813): 745-750
12 Guo P, Bailey S,
Bodley JW, Anderson D.
Characterization of the small RNA of the bacteriophage f29 DNA packaging machine. Nucleic
Acids Res, 1987, 15: 7081-7090
13 Garver K, Guo P. Boundary of pRNA functional domains and minimum pRNA
sequence requirement for specific connector binding and DNA packaging of phage
phi29. RNA, 1997, 3: 1068-1079
14 Bailey S, Wichitwechkarn J,
Johnson D, Reilly B, Anderson D, Bodley JW. Phylogenetic analysis and secondary structure of
the Bacillus subtilis bacteriophage RNA required for DNA packaging. J
Biol Chem, 1990, 265:
22365-22370.
15 Zhang CL, Tellinghuisen T,
Guo P. Conformation of the helical structure of the 5′/3′
termini of the essential DNA packaging pRNA of phage f29. RNA, 1995, 1: 1041-1050
16 Chen C, Guo P. Sequential action of six virus-encoded DNA-packaging
RNAs during phage phi29 genomic DNA translocation. J Virol, 1997, 71(5): 3864-3871
17 Trottier M, Zhang CL, Guo
P. Complete inhibition of virion assembly in vivo with mutant pRNA
essential for phage f29 DNA packaging. J Virol, 1996, 70: 55-61
18 Trottier M, Guo P. Approaches to determine stoichiometry of viral
assembly components. J Virol,
1997, 71:
487-494
19 Guo P, Zhang C, Chen
C, Trottier M, Garver K. Inter-RNA interaction of
phage phi29 pRNA to form a hexameric complex for viral DNA transportation. Mol
Cell, 1998, 2: 149-155
20 Zhang F, Lemieux S, Wu
X, StArnaud D, McMurray CT, Major F,
Anderson D. Function of hexameric RNA in packaging of bacteriophage
phi29 DNA in vitro. Mol Cell,
1998, 2:
141-147
21 Hendrix RW. Bacteriophage DNA packaging:
RNA gears in a DNA transport machine (Minireview). Cell, 1998, 94: 147-150
22 Smith DE, Tans SJ, Smith
SB, Grimes S, Anderson DL, Bustamante C. The bacteriophage phi29 portal motor can
package DNA against a large internal force. Nature, 2001, 413: 748-752
23 Davenport R.J. Crossover research yield
scents and sensitivity-watching a virus get stuffed. Science, 2001, 291: 2071-2072
24 Valpuesta JM, Fernandez JJ, Carazo JM, Carrascosa JL. The three-dimensional structure of a DNA translocating machine at 10 A resolution. Structure Fold Des, 1999, 7