Virology 318 (2004) 463 – 473
www.elsevier.com/locate/yviro
Minireview
Viral RNA polymerase scanning and the gymnastics of Sendai virus
RNA synthesis
Daniel Kolakofsky, * Philippe Le Mercier, Frédéric Iseni, and Dominique Garcin
Deparment of Genetics and Microbiology, University of Geneva School of Medicine, Geneva CH1211, Switzerland
Received 19 August 2003; returned to author for revision 8 October 2003; accepted 23 October 2003
Abstract
mRNA synthesis from nonsegmented negative-strand RNA virus (NNV) genomes is unique in that the genome RNA is embedded in an N
protein assembly (the nucleocapsid) and the viral RNA polymerase does not dissociate from the template after release of each mRNA, but
rather scans the genome RNA for the next gene-start site. A revised model for NNV RNA synthesis is presented, in which RNA polymerase
scanning plays a prominent role. Polymerase scanning of the template is known to occur as the viral transcriptase negotiates gene junctions
without falling off the template.
D 2003 Elsevier Inc. All rights reserved.
Keywords: RNA polymerase; nonsegmented negative-strand RNA virus
This minireview summarizes current understanding of
Sendai virus (SeV) RNA synthesis, drawing on work from
other nonsegmented negative-strand RNA viruses (NNV),
and observations of bacterial RNA polymerase and RNA
pol II. We first describe NNV genome organization and our
current understanding of its expression, then what is known
about the structure of the viral proteins that carry out RNA
synthesis, and then what is known about cellular RNA
synthesis that is likely to apply to NNV RNA synthesis as
well. New experiments on VSV and SeV RNA synthesis
are discussed, and a general model for NNV RNA synthesis
that integrates the new results is presented, in which RNA
polymerase scanning of the template plays a prominent
role.
NNV genome organization and expression
Sendai virus, a Respirovirus of the Paramyxovirinae
subfamily of the Paramyxoviridae, is a model NNV. The
other subfamily, the Pneumovirinae, contains respiratory
* Corresponding author. Department of Genetics and Microbiology,
University of Geneva School of Medicine, CMU, 9 Ave de Champel,
Geneva CH1211, Switzerland. Fax: +41-22-379-5702.
E-mail address: Daniel.Kolakofsky@medecine.unige.ch
(D. Kolakofsky).
0042-6822/$ - see front matter D 2003 Elsevier Inc. All rights reserved.
doi:10.1016/j.virol.2003.10.031
syncytial virus (RSV) and human meta-pneumovirus. Other
families of the order Mononegavirales that carry out their
RNA synthesis in the cytoplasm include the Rhabdoviridae
(with vesicular stomatitis virus (VSV) and rabies virus), and
the Filoviridae (Ebola and Marburg viruses). NNV RNA
synthesis requires the action of three viral proteins, the
nucleocapsid protein N, which assembles the genome and
antigenome RNAs into helical nucleocapsids (N:RNAs),
and the P and L proteins that form the core viral RNAdependent RNA polymerase (vRdRp). The genomic RNA
of NNV functions firstly as a template for mRNA synthesis,
and then as a template for a full-length complementary copy
(the [+] antigenome), which like the genome is found only
as N:RNA nucleocapsids (Fig. 1). For the Paramyxovirinae,
each N subunit of the N:RNA is associated with precisely
six nucleotides and this stoichiometry is functionally important (Calain and Roux, 1993; Egelman et al., 1989). This
‘‘rule of six’’, however, does not apply to the other NNV.
During infection, ostensibly the same vRdRp carries out
both types of RNA synthesis on ostensibly the same N:RNA
templates. The first is termed transcription, in which vRdRp
responds to cis-acting signals and ‘‘stutters’’ (see below) to
polyadenylate and terminate each mRNA. As a consequence, vRdRp restarts at successive mRNA start sites to
synthesize each mRNA in turn. During RNA replication, in
contrast, RNA synthesis and assembly with N occur concomitantly, and vRdRp ignores all polyadenylation/termina-
464
D. Kolakofsky et al. / Virology 318 (2004) 463–473
Fig. 1. Schematic representation of SeV RNA synthesis. An electron micrograph of a full-length, resting SeV nucleocapsid, a helical assembly of 2564 N
subunits and 15,384 nucleotides with 196 turns (Egelman et al., 1989), is shown at the top. The donut shapes on the lower left of the photo show single turns of
an unraveled helix lying flat on the grid that offer a perpendicular view. The genome and antigenome nucleocapsids (N:RNAs) are represented as long boxes
with two widths; the central wider region indicates the entire coding region (from the start codon of N to the stop codon of L). Only the N/P junction is
indicated, and the drawing is not to scale. The narrow regions at the ends represent the noncoding regions where the two bipartite replication promoters are
found (red boxes below). The bent arrows show the sites and direction of RNA synthesis. le, tr, and mRNA synthesis are independent of Nj availability,
whereas genome/antigenome synthesis and assembly with N takes place concurrently. There are two views of how vRdRp gains access to the N mRNA start
site. In one view, vRdRp can initiate de novo at either the 3Vend or the N mRNA start site (but not at other mRNA start sites). In the other view, all RNA
synthesis starts at the genome 3Vend, and vRdRp reinitiates at the N start site only after releasing the le RNA.
tion signals to form a full-length antigenome nucleocapsid
(Lamb and Kolakofsky, 2001).
Paramyxo-, Rhabdo-, and Filoviridae genomes are 11-18
kb long and contain 5 – 10 genes in tandem, separated by
conserved junctional sequences that act as mRNA start and
polyA/stop sites (Fig. 1). The first mRNA (usually N) starts
around 50 nt from the genome 3Vend, and the 50 nt upstream
are called the leader region (le, Fig. 1). The last (L) mRNA
ends a similar but variable distance from the genome 5V end,
and is followed by the trailer region (tr, Fig. 1). Exactly how
NNV RdRp switches between its dual function as transcriptase and replicase remains unclear. However, a self-regulatory model of viral RNA synthesis was proposed after the
discovery that short le and tr RNAs were the first products
transcribed from viral genomes and antigenomes, respectively (Leppert et al., 1979). More importantly, unlike the
synthesis of genomes and antigenomes (and similar to the
synthesis of mRNAs), le and tr RNA synthesis not only
occurred, but was enhanced in the absence of on-going
protein synthesis (Blumberg et al., 1981). As on-going
protein synthesis is required for genome synthesis at any
time of infection, the component limiting for replication was
assumed to be unassembled N protein (Nj), as each mature
genome is associated with many N subunits (each 15 384-nt
long SeV genome has 2564 N subunits). This assumption
was later elegantly shown to be the case for VSV (Arnheiter
et al., 1985). Moreover, as the site for the initiation of
nucleocapsid assembly on VSV le RNA mapped to the first
14 nucleotides at its 5Vend (Blumberg et al., 1983), nucleocapsid assembly was proposed to initiate on nascent le and
tr RNAs before vRdRp had released these chains at or near
the le/tr-gene junctions. The sequences responsible for
measles virus nucleocapsid assembly are also contained
within the le region (Castaneda and Wong, 1990). Antigenome RNA synthesis and assembly would then become
coupled, and this vRdRp becomes a ‘‘replicase’’ that synthesizes complementary N:RNAs. In the absence of this
coupling, vRdRp would release the le chain near the N
mRNA start site, and be free to initiate N mRNA synthesis.
Upon N mRNA initiation (including mRNA 5V-end modification), this vRdRp becomes a ‘‘transcriptase’’ that responds
to transcriptional signals, that is, that stutters to form a
mature polyA tail and releases the mRNA, and thus synthesizes each mRNA in turn (Iverson and Rose, 1981;
McGeoch, 1979; Rose, 1980; Schubert et al., 1980). A
self-regulatory model was proposed, in which the relative
levels of transcription and replication are controlled by the
availability of Nj (Kolakofsky, 1982). In the absence of
sufficient levels of Nj to permit its assembly on the nascent
le chain, vRdRp only transcribes the genome, to generate
more mRNA that generates more Nj. When Nj levels rise,
some of the vRdRp are devoted to genome replication,
D. Kolakofsky et al. / Virology 318 (2004) 463–473
which depletes the pool of Nj and keeps its steady-state
levels low.
This simple self-regulatory model, proposed >20 years
ago, is generally accepted in broad outline. However, the
model has been difficult to pin down unambiguously and
important details are missing. There have been some
advances. For example, by reconstituting SeV genome
replication in a synchronous fashion in vitro, it has been
possible to show that the replicase traverses the N:RNA at a
constant speed from beginning to end (albeit at 2 nt/s) even
when mRNA synthesis is poorly processive. More importantly, genome RNA synthesis and assembly were found to
occur concurrently (Gubbay et al., 2001). Similarly, although we know little of how the coupling of genome
synthesis and assembly constrains vRdRp to ignore template
stutter signals and act as a replicase, the assembly process
itself has been better described. VSV, rabies virus, and SeV
Nj expressed by themselves tend to aggregate nonspecifically, often with RNA (Horikami et al., 1992; Masters and
Banerjee, 1988; Peluso and Moyer, 1988; Schoehn et al.,
2001); the Nj that assembles on nascent genome/antigenome RNA is chaperoned by an oligomeric P protein
[probably P4 – Nj for SeV (unpublished), and P2 – Nj or a
dimer thereof for rabies virus (Mavrakis et al., 2003)]. There
is biochemical evidence that P tetramers interact with each
other in vitro (Tarbouriech et al., 2000a), and the addition of
supplemental P4 to P4 – L/N:RNA transcription reactions
strongly enhances SeV mRNA synthesis (Curran, 1996).
This enhancement requires the binding of P4 to the template.
Thus, both P4 and P4 – L appear to be present in the
transcription elongation complex (TEC). The replication
elongation complex, by analogy, would contain P4 – Nj
along with P4 – L. The P4 of the P4 – Nj assembly complex
may similarly interact with that of the polymerizing P4 – L
during replication, and the ability of vRdRp to stutter in
response to polyA signals may be modulated in this way
(see below).
The Sendai virus RNA polymerase
P, the polymerase cofactor
NNV L proteins (approximately 2200 aa) are thought to
contain all the catalytic sites of vRdRp (NMP polymerization, 5V end guanylylation and methylation), but L does
not bind to the N:RNA by itself. L in virions and infected
cells is stably complexed with the more abundant P
protein, which in SeV (and probably all Paramyxovirinae)
is a parallel coiled-coil tetramer (Fig. 2) (Curran et al.,
1995a; Tarbouriech et al., 2000a; Tarbouriech et al.,
2000b). L is unstable in the absence of P, and coexpression
of P and L is required for active P4 – L complex formation
(Horikami et al., 1992). Paramyxovirus P genes are complex, and are divided in half by editing sites; specialized
vRdRp stutter sites that induce vRdRp to add pseudo-
465
templated guanylylates to the mRNA (Lamb and Kolakofsky, 2001). The net result of these nucleotide insertions in
the mRNA is that different carboxyl segments (from
overlapping ORFs) are fused to a common P aminoterminal segment (called PNT). X-ray and NMR studies
have determined the structure of most of the longest of
these carboxyl modules (called PCT), that of the SeV P
protein (Blanchard et al., 2003; Marion et al., 2001;
Tarbouriech et al., 2000b). Although never expressed
independently in nature, PCT by itself is sufficient for
all aspects of transcription in vitro (Curran, 1996), and has
been dubbed the ‘‘polymerase cofactor’’. The overall size
and shape of PCT and its subdomains have been studied
by small angle X-ray and neutron scattering (Blanchard et
al., 2003; Tarbouriech et al., 2000a). PCT appears to be a
rod ca. 160 Å long with a diameter of 36 Å (Fig. 2). The
polymerase cofactor module contains a long a-helix that
forms a stable coiled-coil tetramer of approximately 100
Aj long. The coiled coil is buttressed at its NH2-terminus
by short a-helical segments, and is followed by a flexible
linker of ca 30 aa, and then the X domain that contains the
major site(s) for stable N:RNA binding via N-tail (see
below). A 95-aa X protein is also expressed during some
infections (Curran and Kolakofsky, 1988), and it is the
structure of this monomer that was determined by NMR.
The structure of the measles virus X domain has also been
determined, by crystallography (Johansson et al., 2003).
Both X domains are structurally well conserved, and their
most prominent feature is a triple a-helical bundle at the
C-terminus that is predicted to be present in the X domains
of all the Paramyxovirinae (Curran et al., 1995a). The
three helices show an asymmetric charge distribution and
this bundle contains grooves between the helices to which
N-tail may bind either through hydrophobic interactions or
via its conserved, highly negatively charged peptide, forming a four-helix bundle (see below). As expected from the
different methods used, the X-ray study stresses the atomic
details of the crystallized protein, whereas NMR spectroscopy stresses the highly dynamic nature of this protein
domain in solution. This latter property is likely to be
important for P protein function in vRdRp (Blanchard et
al., 2003).
The region where L is bound to P, mapped broadly by
deletion analysis (Curran et al., 1994; Smallwood et al.,
1994) (residues 412– 445, Fig. 2) and more finely by alanine
scanning mutagenesis (Bowman et al., 1999) (relevant side
chains shown in Fig. 2), lies in the carboxyl half of the
coiled coil. Unlike classic coiled coils that have hydrophobic
interfaces of uniform diameter, the internal channel here is
of variable size and widest at the center where water
molecules are also found. Because RNA synthesis occurs
one nucleotide at a time whereas displacement of P on the
template occurs in steps of N subunits (6 nt at a time, Calain
and Roux, 1993), P4 is proposed to cartwheel across the
N:RNA as L transcribes the template RNA (Curran, 1998).
This would require some mobility at the P– L interface,
466
D. Kolakofsky et al. / Virology 318 (2004) 463–473
Fig. 2. The SeV P protein and viral RNA synthesis. The entire SeV P protein is shown schematically above, divided into N-terminal (PNT) and C-terminal
(PCT) segments at the GGG glycine codon (#317) into which G insertions occur during mRNA editing (dark blue circle). The Nj chaperone site in PNT, and
the tetramerization and L binding region, and the independently folding X domain (the strong N:RNA binding site) of PCT are highlighted. The regions of PNT
that are dispensable for transcription and replication are indicated, as well as two long regions predicted to be unstructured (Karlin et al., 2002). A structural
model of PCT or the polymerase cofactor tetramer based on the work of Tarbouriech et al., is shown below. The side chains of the charged residues critical for L
binding (K408, R409, E412, K415, and E416) are shown as sticks and their location on the face of the bottom helical ribbon is indicated by coloring. Note that
all these side chains point to the outside. P4 is proposed to interact with N:RNA in two ways; indirectly via L which interacts with N-core and the template
RNA, and directly via the X domains which binds N-tail. It is this latter interaction that tethers P4 to (mostly inactive) N:RNAs isolated from virions and
infected cells (Ryan et al., 1990). Two Nass subunits from an imaginary N:RNA are positioned below the L binding site. N-core is simply represented as a
sphere whose diameter is taken from the spacing between two turns of the N:RNA super-helix (53 Aj) (Egelman et al., 1989). Our representation of L
(approximately 2200 aa) is definitely not to scale. During both RNA synthesis and vRdRp scanning, the coiled coil of P is proposed to act as a molecular axle
that allows L to remain fixed to and slide along the template RNA while the X domains cartwheel along successive N-tails.
which may be provided by the plasticity of the non-canonical coiled coil due to its increased radius, and the exceptional mobility of the interacting side chains at the P –L
interface, as indicated by their temperature factors (Tarbouriech et al., 2000b). The coiled coil of P is proposed to act as
a molecular axle that allows P4 rotation while L slides along
the genomic RNA.
Genome replication, in contrast to transcription, requires
the participation of at least part of the 317-aa aminoterminal segment of SeV P encoded upstream of the editing
site (PNT). This segment of P contains a site near its aminoterminus (residues 33– 41, defined by deletion analysis) that
forms a complex with Nj. This site is thought to chaperone
Nj during the nascent chain assembly step of genome
replication, and to ensure specificity of assembly (Curran
et al., 1995b). Both the Nj chaperone domain and the
surrounding sequence that has properties of an acidic
activation domain are required for genome replication, but
the remainder of the amino-terminal segment (residues 78–
320) can be deleted without destroying its activity (at least
for mini-genome synthesis) (Curran et al., 1994) (Fig. 2).
The unexpected non-lethality of this massive deletion may
be related to the finding that this segment in measles virus
(and probably all the Paramyxovirinae) is intrinsically
unstructured (Karlin et al., 2002). This segment of SeV P
is similarly predicted to be unstructured, and the four Nj
chaperone domains of P4 appear to be connected to the rigid
coiled coil of the polymerase cofactor module by a long
D. Kolakofsky et al. / Virology 318 (2004) 463–473
flexible linker (which is nonessential for activity under some
conditions). This arrangement might prove useful in providing the flexibility required to position Nj subunits onto
the growing end of the nascent nucleocapsid, as the product
RNA emerges from the surface of vRdRp for assembly
(bottom cartoon, Fig. 3).
467
The L protein polymerase
There is no detailed structure for any NNV L protein (ca.
2200 aa) as yet. However, NNV L proteins contain conserved sequence motifs characteristic of other RNA polymerases (Poch et al., 1989), and mutational analysis has
Fig. 3. Four states of the SeV RdRp elongation complex. In these cartoons of SeV RdRp elongation complexes, the Nass subunits of the template are shown as
spheres representing N-core, and N-tail is shown either as a bolt of lightening (disordered) or a filled rectangle (induced fold) when associated with the X
domain of P (open rectangle). The genome RNA is shown as a horizontal line, from which the N subunits are transiently displaced during passage of vRdRp.
This displacement is thought to occur so that L can enclose the hybrid (and this simplifies the cartoon), but the model applies as well if L accesses the
nucleotide bases without N displacement. The nascent RNA is shown as a bent arrow, below the genome RNA. The L protein polymerase is shown as a large
circular object in the center, whose precise shape indicates the degree to which the putative active site cleft of L encloses the nascent RNA – genome RNA
hybrid, and thereby increases vRdRp processivity, i.e., its ability to synthesize long chains efficiently. Only the polymerase-cofactor module (PCT) of the
tetrameric P protein is shown, and only two of the four P polypeptides (those in the act of binding N-tail) are drawn for clarity. The coiled-coil of P is shown as
a pair of vertical lines, to which the carboxyl X domains that bind N-tail are shown as open boxes connected to the rigid coil. The bottom cartoon of the
replication elongation complex also shows PNT of the trailing P4 – Nj complex that is thought to participate in replication. The zig-zag lines represent flexible,
disordered regions (see Fig. 2). For further explanation, see text.
468
D. Kolakofsky et al. / Virology 318 (2004) 463–473
confirmed their importance for polymerase function. Given
the conservation in overall structure of all RNA polymerases
determined to date (Bressanelli et al., 1999; Cramer et al.,
2001; Gnatt et al., 2001; Hansen et al., 1997; Korzheva et
al., 2000; Tao et al., 2002; Zhang et al., 1999), it is
reasonable that the basic architecture of NNV L proteins
will be similar. The catalytic domain of nucleic acid
polymerases is organized around a central cleft in an
arrangement reminiscent of a right hand (Ollis et al.,
1985). The remainder of the cleft accommodates the nascent
chain –template chain hybrid (the hybrid) that maintains
polymerase register during RNA synthesis (8– 9 bp long
for cellular DdRp (Nudler et al., 1997), and postulated to be
7 bp long for SeV RdRp (Hausmann et al., 1999a). One of
the most striking features of recent bacterial and yeast DdRp
structures in different phases of RNA synthesis is the
network of contacts that interconnect side chains in the
active site cleft with the hybrid and the nascent RNA exit
channel. The central feature of this interaction network is a
moveable ‘‘clamp’’ that contacts all the above. Comparison
of these structures reveals several changes related to transcription, including the movement of this massive clamp by
>30 Å between the proposed (open) initiation state and the
transcribing complex (closed state), in which it largely
surrounds the hybrid (Cramer et al., 2000; Cramer et al.,
2001; Gnatt et al., 2001; Landick, 2001). Five polypeptide
segments, termed switches, connect the clamp to the stationary parts of DdRp. The switches are disordered when
the clamp is open, but they fold cooperatively upon interacting with the hybrid (and the bridge helix) when the clamp
closes in the TEC (Cramer et al., 2001). The bridge helix
spans the active site cleft and positions the nascent RNA 3V
end in the active site. This induced fit interaction is thought
important in maintaining catalysis and TEC processivity,
especially in negotiating pause sites. Polymerase pausing,
the prelude to arrest and termination for cellular DdRp
(Landick, 1999), is triggered by a conformational change
in the interaction network that partially opens the active site
cleft and misaligns the hybrid with the key active site
residues (Landick, 2001). This inhibits nucleotide addition,
either directly, or by allowing backtracking of the TEC.
Polymerase backtracking is also the proposed prelude to
NNV RdRp pseudo-templated transcription (Hausmann et
al., 1999b). Cellular RNAP are thought to bind to their
templates in a more open-handed conformation, and the
initiating complex converts to a more closed, stable TEC
that ‘‘grips’’ the template after formation of the hybrid and
the entry of the nascent RNA 5Vend into the RNA exit
channel (promoter clearance).
Cellular and NNV RNA synthesis, and polymerase
processivity
Synthesis of mature mRNAs is not guaranteed when
TEC leaves a promoter in either bacteria or eucaryotes, as
TECs (lacking elongation factors) are prone to extensive
pausing and arrest during elongation. Eucaryotic elongation
factors such as TFIIF, elongin, and ELL all suppress
polymerase pausing via interaction with RNAPII (TFIIF is
thought to act by preventing the displacement of the nascent
RNA 3Vend from the active site) (Price, 2000; Shilatifard et
al., 2003). There is much evidence that cellular factors
stimulate NNV RNA synthesis (reviewed in Gubbay et
al., 2001). In a similar vein, bacteriophage E Q-protein is
stably incorporated into E. coli DdRp shortly after initiation,
and fundamentally alters its elongation properties (Roberts
et al., 1998). Q-modified DdRp display diminished pausing
and are resistant to all downstream termination sites, much
like NNV RNA synthesis that is coupled to concurrent
assembly of the nascent chain. Mutations that uncouple this
phage property suggest that Q-induced anti-termination is
due to the stabilization of DdRp interactions with the
DNA –RNA hybrid that optimize the alignment of nucleic
acids in the catalytic site (Santangelo et al., 2003).
NNV transcriptases respond to cis-acting ‘‘stutter’’ signals that cause TEC to copy the same template base(s)
repetitively (pseudo-templated transcription). Other transcriptase properties, like termination and initiation of the
next mRNA, depend on the extensive vRdRp stuttering that
forms the mRNA polyA tail (Barr and Wertz, 2001). The
repetitive copying of template nucleotides requires cycles of
realignment of the two strands of the hybrid (Hausmann et
al., 1999a; Vidal et al., 1990), and this realignment implies
conformational changes in the active site cleft – hybrid
interactions that would permit this realignment (loosening
of induced fit beyond that necessary to permit simple
polymerase translocation after each nucleotide addition).
Replicases, in contrast, do not (or very rarely, Bilsel and
Nichol, 1990; Castaneda and Wong, 1990; Hausmann et al.,
1996) respond to stutter signals. One way in which the
coupling of nascent chain synthesis and assembly could
commit vRdRp to replication is by simply locking this
induced fit in place, preventing pseudo-templated transcription of any kind and maintaining near-absolute vRdRp
processivity. The coupling of NNV nascent chain synthesis
and assembly could act in a manner similar to the interaction
of eucaryotic elongation factors with RNAP II, or bacteriophage E Q-protein and E. coli DdRp.
SeV RdRp that has initiated at the genome 3Vend (and
read through the le/N junction in the absence of nascent
chain assembly) will respond to a gene-end signal placed
80 nt downstream of the le/N junction as efficiently as the
transcriptase at natural gene junctions (Le Mercier et al.,
2002). Although SeV RdRp that initiates le chains retains
this property of a transcriptase, it is relatively non-processive (le/N chains >300 nt long are not detected; Vidal and
Kolakofsky, 1989). A similar situation is found for VSV
polR mutants, which read through the le/N junction at high
frequency (Perrault et al., 1983). In contrast, a bona fide
SeV or VSV transcriptase that has initiated the N mRNA
produces matures mRNAs at a frequency of >90%. SeV
D. Kolakofsky et al. / Virology 318 (2004) 463–473
and VSV RdRp become processive independent of nascent chain assembly upon initiation (and capping) of the
N mRNA, and they do so while retaining their ability to
respond to stutter signals. How this processivity is acquired is unknown, but work with VSV has indicated that
processivity is acquired only after the modification of the
nascent mRNA 5Vend when it is approximately 50 nt long
(Stillman and Whitt, 1999). Similarly, VSV RdRp does
not respond to a polyA/stop site within the first 50
nucleotides of the gene-start site (Whelan et al., 2000).
mRNA 5V-end modification has been proposed to be a
quality-control checkpoint for VSV mRNA synthesis,
similar to that described for nuclear mRNA synthesis
(Lis, 1998). This checkpoint could act as well to ensure
that other interactions have been formed, which may be
required for the transcriptase to negotiate gene junctions
(see below).
NNV RdRp scanning
Intrinsic or factor-independent termination sites in E. coli
act by forming 2j structures in the DdRp RNA exit channel
that pry open the active site cleft, such that the TEC
becomes unstable (Landick, 1999b). Most RNA polymerases are thought to dissociate from their templates upon
nascent chain release and loss of the hybrid that tethers it to
its template. However, a hallmark of NNV mRNA synthesis
is that vRdRp does not dissociate from its template upon
release of the mature mRNA. Rather, this transcriptase
scans the template (in either direction) for a nearby (re)start
site. RdRp scanning has long been suspected for the
Pneumovirus RSV, whose M-2 and L genes overlap by
68 nt (Collins et al., 1987) and then shown to occur
experimentally (Fearns and Collins, 1999). Scanning was
also unexpectedly found for the rhabdovirus VS(NJ)V
(Stillman and Whitt, 1998). Scanning at gene junctions
appears to be a general property of NNV, as the overlapping
arrangement of gene ends is found in filoviruses (Sanchez
et al., 1993) and some rhabdoviruses (Teninges et al.,
1993). RdRp scanning neatly explains why the intergenic
regions of many NNV are not conserved (and can vary from
1 to 52 nt in length); this would have little consequence
when scanning is efficient. Scanning also explains how
vRdRp is released from the template after completing the
last mRNA; vRdRp would simply run off the template 5V
end. Finally, scanning may also explain the unusual situation for the rubulavirus SV41, where M gene transcripts are
produced exclusively as M –F dicistronic read-throughs, yet
abundant monocistronic F mRNA is produced (Tsurudome
et al., 1991). In this case, however, scanning would have to
occur over much larger distances (i.e., 1400 nt from the start
of the M gene).
NNV RdRP scanning is presumably related to the
unusual nature of NNV templates that are composed mostly
of protein; the SeV template is 97% N protein by weight.
469
Moreover, N is thought to be as much a part of the TEC as
P4 – L itself (Perrault et al., 1983; Vidal and Kolakofsky,
1989). SeV and VSV P are found tightly bound to the
assembled N subunits of the template (Nass) via its hypervariable, protease-sensitive carboxyl tail (N-tail) (Iseni et al.,
1998; Ryan et al., 1990). The N proteins of the Paramyxovirinae are divided into a well-conserved, proteaseresistant core (N-core, the amino-terminal 80% of the
polypeptide) and the carboxyl tail. N-core by itself can
self-assemble and bind RNA, and N-core is sufficient to
drive mini-genome replication (a measure of nascent chain
assembly). However, templates composed of N-core are
apparently inactive for RNA synthesis (Curran et al.,
1995a). N-tail is thus required for template function, at least
during replication. ‘‘Tail’’ is an apt description of the
carboxyl domain of Nass, as this domain was recently shown
to be intrinsically disordered (Longhi et al., 2003). Many
polypeptides have now been found to contain little or no
ordered structure under physiological conditions, and a
majority also undergo some degree of folding in the
presence of their physiological partner(s). Bacterially
expressed measles virus N-tail was found to be intrinsically
disordered by biochemical and biophysical criteria, and to
undergo a structural transition (a significant increase in ahelical content of the complex) upon binding to PCT
(Longhi et al., 2003).
These results explain the protease hypersensitivity of Ntail, and they also help explain how vRdRp might scan the
N:RNA for a new start site so efficiently after releasing their
transcripts. The functional importance of disorder is thought
to reside in the advantages of flexible as compared to rigid
structures. In particular, induced folding can result in
complexes that are highly specific but of low affinity. For
acidic activation domains of transcription factors, affinity is
attained via multiple activation domains, often through
oligomerization (Melcher, 2000). For SeV, as long as two
X domains of P4 are in continual contact with N-tails, this
interaction will be stable. We note that the net change in
Gibbs free energy of P4 cartwheeling on successive N-tails
itself would be zero, and if X domain– N-tail and P4 –L
interactions are broken and remade simultaneously, the
activation energy barriers to P4 cartwheeling on the template might be minimal. This is exactly what is needed if the
X domains of P4 – L are to continually make and break
contact with successive Nass subunits within the TEC as the
L active site cleft slides along the genome RNA in search of
a new start site. Rotation of the coiled coil with the P
binding site of L is presumably coordinated with X domain/
N-tail cartwheeling, and the flexible linker domain of PCT
may be involved in this. During RNA synthesis, vRdRp
translocation is coupled to nucleotide addition (pseudotemplated transcription is of course an exception here)
and is unidirectional. During vRdRp scanning, the absence
of the hybrid permits vRdRp to traverse the template RNA
in both directions. Whether vRdRp scanning requires the
input of much chemical energy is unclear. It should be
470
D. Kolakofsky et al. / Virology 318 (2004) 463–473
noted in this context that, like P, N-tail is the site of
extensive phosphorylation and de-phosphorylation (Roux
and Kolakofsky, 1974;Vidal et al., 1988). The free energy
of phosphate hydrolysis can be harnessed to drive vRdRp
scanning.
tion of when vRdRp becomes scan-competent during NNV
RNA synthesis.
Recent experiments with VSV
When gene-start sites are introduced into the SeV
promoter for [ ] genome synthesis, this diminishes the
use of this 3Vend promoter for replication. Similarly, when
gene-start sites are removed from the promoter for antigenomes synthesis in SeV mini-genomes, these minigenomes compete more effectively with their helper
genomes for the limiting supply of replication substrates
provided by the latter. These experiments suggest that
mRNA start sites and 3Vend replication promoters compete
for a common pool of vRdRp (Le Mercier et al., 2003).
These experiments, as such, do not rule out that committed
transcriptases and replicases are present before they engage
the template, and that initiation at a gene-start site impedes
access of a potential replicase to the genome 3Vend, or
otherwise impedes replication. However, in mini-genomes
in which the gene-start site is displaced further downstream from the genome 3Vend (from positions 56 to 68),
the gene-start site is equally effective in reducing replication (Le Mercier et al., 2003). The precise position of the
gene-start site within the genomic promoter is thus not
critical for its ability to reduce 3Vend promoter efficiency,
and these results disfavor interference models. However, it
will be necessary to examine the properties of minigenomes in which the gene-start site is displaced even
further downstream from the genome 3Vend, to be more
certain that mRNA start sites and 3Vend replication promoters compete for a common pool of vRdRp, or generic
vRdRp.
Evidence has also been presented that supports the notion
that committed transcriptases and replicases are present
before they engage the template, but this evidence remains
equally open to alternative interpretations. For example, (1)
mutant VSV P proteins exist that are defective in transcription and not in replication (Pattnaik et al., 1997). However,
as (i) we do not know in what way these P proteins cannot
carry out transcription and (ii) at least some of these mutant P
proteins are not transcription defective in insect cell extracts
(Gupta et al., 2003), further conclusions are difficult to draw.
(2) Insect cell extracts containing coexpressed N, P, and L
(when combined with purified N:RNA templates) are more
efficient in replication than mixed extracts in which N + L
and P + L are coexpressed. Moreover, a tripartite complex
containing N, P, and L can be immunoprecipitated from the
former cell extracts (Gupta et al., 2003). This has been taken
as evidence that there are distinct replicases (containing N as
well as P and L) and transcriptases (that do not contain N) in
the absence of the template. However, at this stage, the active
complex in the triply expressed extract remains uncharacterized, and this extract may contain other complexes with
To date, there is no way of discriminating between
transcriptases and replicases until these vRdRp have been
committed to their specialized tasks. The question of when
this commitment occurs thus remains open. Le RNA is the
first product of VSV RNA synthesis in vitro (Colonno and
Banerjee, 1976), and whether this short promoter-proximal
transcript is synthesized by a transcriptase, a replicase, or a
generic vRdRp is at the center of a debate of how vRdRp
switches between transcription and replication. There has
long been the feeling that these vRdRp may be committed to
their specialized tasks before they engage the template (Barr
et al., 2002), and evidence has recently been presented that
in infected cells, VSV RdRp may initiate N mRNA synthesis directly, without first having initiated le RNA synthesis
(Whelan and Wertz, 2002). These authors propose that VSV
RdRp can recognize an overlapping set of cis-acting promoter sequences differently, such that its active site is
positioned to initiate at the genome 3Vend for replication,
or at the nearby N mRNA start site (ca. 50 nt downstream)
for transcription. How this alternate vRdRP-positioning
mechanism would respond to Nj availability (Arnheiter et
al., 1985) is not specified.
The notion that vRdRp can initiate N mRNA synthesis
independently of prior le RNA synthesis is based on experiments in which UV cross-links are introduced in the le
region in a variable manner within viable VSV, as barriers to
vRdRP movement during elongation. The effect of these
barriers on downstream mRNA synthesis was then examined. When this experiment was carried out in vitro, mRNA
synthesis was found to strictly depend on prior le RNA
synthesis. These results confirmed earlier work that found
that vRdRp initiates exclusively at the genome 3Vend in
vitro, using reconstituted VSV reactions (Emerson, 1982).
More importantly, these results show that VSV RdRp
cannot scan the template and reinitiate at the N start site
upon release of the le chain, at least in vitro. This vRdRp
must fall off the N:RNA upon release of the le chain.
However, when the same experiment is carried out in
infected cells in which Nj accumulation is prevented by
cycloheximide, exactly the opposite result is found; mRNA
synthesis is now independent of the number of UV crosslinks in the le region, hence, the conclusion that direct
initiation at N must have occurred in vivo (but not in vitro).
This conclusion, however, begs the question of why vRdRp
cannot initiate at N directly in vitro if it can do so in vivo.
Moreover, these experiments do not rule out other possibilities, which are in fact suggested by considering the ques-
The problem of determining when vRdRp is committed
to its specialized tasks
D. Kolakofsky et al. / Virology 318 (2004) 463–473
various combinations of N, P, and L that interact to initiate
replication.
(3) VSV le RNA and mRNA synthesis in vitro have
different optimal requirements for ATP (Helfman and Perrault, 1988). However, until we know how ATP hydrolysis
participates in the initiation of le and mRNA synthesis,
again, further interpretation is difficult.
(4) A spontaneous mutation in the VSV N protein (polR)
was found that (i) strongly increases vRdRp read-through of
the le/N junction, producing incomplete le/N transcripts of
various lengths, and (ii) these mutants synthesizes more N
mRNA than le RNA in vitro, in contrast to the wild-type
reaction (Chuang and Perrault, 1997). This is the best
evidence so far that vRdRp can directly initiate the N mRNA
in vitro, at least when the Nass of the template is mutated.
However, other evidence (cited above) indicates that (wild
type) VSV RdRp cannot directly initiate at N in vitro
without prior synthesis of le RNA. The unexpected finding
that polR mutants have wild-type growth phenotype (Perrault et al., 1983) is cited as evidence that VSV RdRp can
directly initiate at the N start site (at least in polR mutants),
as this can explain the absence of an obvious growth defect
in VSV polR. However, efficient vRdRp scanning for the N
start site can equally explain the absence of an obvious
growth defect.
A general model for the control of NNV RNA synthesis
In the end, there is as yet no compelling evidence for or
against the existence of committed transcriptases and replicases before they engage the template. It is nevertheless
useful to consider a general model of NNV RNA synthesis
based on the premise of a generic vRdRp that initiates at the
genome 3Vend. This model can accommodate all the available information, including the results of Whelan and
Wertz. Four forms of vRdRp are proposed, two forms that
are poorly processive, one incompetent and the other
competent to scan the template for a new start site, a
transcriptase, and a replicase, and these are shown in
cartoon form in Fig. 3.
1. The top cartoon shows the TEC shortly after initiating at
the genome 3Vend. This vRdRp is poorly processive
because it readily responds to pause sites that open the
active site cleft of L, and this promotes release of the
nascent RNA, and is scanning incompetent as P– N-tail
interactions have not yet been formed. This vRdRp falls
off the template upon nascent RNA release (and joins the
pool of P4 –L off the template), if RNA release occurs
before (or after) vRdRp has reached the N start site. If
this situation predominates for VSV in vitro, N mRNA
synthesis will depend on precise le RNA synthesis.
2. If P – N-tail interactions are formed before the nascent le
chain is released (second cartoon), this probably does
not affect vRdRp processivity, but it prevents the loss of
471
P4 –L from the N:RNA upon RNA release, and thus
permits vRdRp scanning for a nearby start site. If this
situation predominates for VSV in vivo, N mRNA
synthesis will be independent of precise le RNA synthesis.
RdRp scanning should not be impeded by UV cross-links;
RNA synthesis stops at cross-links because the polymerase cannot incorporate an NTP opposite the modified base.
3. Only when vRdRp has initiated the first mRNA,
modified its 5Vend, and formed P – N-tail interactions if
they have not already been formed, does this vRdRp
becomes more processive while retaining its responsiveness to stutter sites (transcriptase). This vRdRp synthesizes mature polyadenylated mRNAs and negotiates
gene junctions efficiently, and thus synthesizes each
mRNA in turn.
4. The bottom cartoon shows the replication elongation
complex, which can be formed from either of the first
two TECs by the coupling of RNA synthesis and
assembly. The coupling is proposed to prevent loosening
of the L active site cleft – hybrid interaction that would be
required to permit realignment of the two chains of the
hybrid during pseudo-templated transcription. This
vRdRp would be extremely processive, as it does not
respond to stutter signals (or does so very rarely). There
would be no need for P – N-tail interactions to tether
vRdRp to its template during replication. Nevertheless,
as N-tail is required for template function (see text), it
presumably is involved in other tasks as well, for
example, coordinating the proposed transient displacement of Nass from the template RNA.
Two P tetramers are proposed to cooperate for efficient
SeV RNA synthesis; the P4 –L polymerase itself, along
with an upstream P4 for the transcriptase (not shown), and
a P4 –Nj complex for the replicase. Only the carboxyl half of
the leading P4 is shown in the first three cartoons (and only
two of the four chains are shown for clarity), and the aminoterminal half of the trailing P4 (in complex with Nj in the act
of assembling the nascent chain) is shown as well in the
replication complex. This already-crowded representation is
likely to be oversimplified, as there is new evidence that L
proteins can directly interact with each other, and that this
interaction helps RNA synthesis (Smallwood et al., 2002).
Although the cartoon in Fig. 3 implies otherwise, it is not
impossible that the trailing P4 for replication is complexed
with both Nj and L, as these binding sites are at opposite ends
of P4. Two L proteins could interact during RNA synthesis,
the trailing (presumably non-polymerizing) L could help
stabilize the leading L as both P tetramers rotate on/in their
respective L proteins and the entire replication elongation
complex (P4 –Nj –L plus P4 –L) rolls along the template.
The above model can form the basis for new experiments
to examine NNV RNA synthesis. We can now look forward
to studies that examine vRdRp scanning that is unique to
NNV, and which may shed light on whether this process
also occurs at the genome 3Vend.
472
D. Kolakofsky et al. / Virology 318 (2004) 463–473
Acknowledgments
We thank Nicolas Tarbouriech and Rob Ruigrok
(Grenoble) for providing the tetramer structure in Fig. 2
and the electron micrograph in Fig. 1. The work in the
authors’ lab is supported by The Swiss National Fund for
Research.
References
Arnheiter, H., Davis, N.L., Wertz, G., Schubert, M., Lazzarini, R.A., 1985.
Role of the nucleocapsid protein in regulating vesicular stomatitis virus
RNA synthesis. Cell 41, 259 – 267.
Barr, J.N., Wertz, G.W., 2001. Polymerase slippage at vesicular stomatitis
virus gene junctions to generate poly(a) is regulated by the upstream 3Vauac-5Vtetranucleotide: implications for the mechanism of transcription
termination. J. Virol. 75, 6901 – 6913.
Barr, J.N., Whelan, S.P., Wertz, G.W., 2002. Transcriptional control of the
RNA-dependent RNA polymerase of vesicular stomatitis virus. Biochim. Biophys. Acta 1577, 337 – 353.
Bilsel, P.A., Nichol, S.T., 1990. Polymerase errors accumulating during
natural evolution of the glycoprotein gene of vesicular stomatitis virus
Indiana serotype isolates. J. Virol. 64, 4873 – 4883.
Blanchard, L., Tabouriech, N., Blackedge, M., Timmins, P., Burmeister,
W.P., Ruigrok, R.W., Marion, D., 2003. Structure and dynamics of the
nucleocapsid-binding domain of the Sendai virus phosphoprotein in
solution. Virology, in press.
Blumberg, B.M., Leppert, M., Kolakofsky, D., 1981. Interaction of VSV
leader RNA and nucleocapsid protein may control VSV genome replication. Cell 23, 837 – 845.
Blumberg, B.M., Giorgi, C., Kolakofsky, D., 1983. N protein of vesicular
stomatitis virus selectively encapsidates leader RNA in vitro. Cell 32,
559 – 567.
Bowman, M.C., Smallwood, S., Moyer, S.A., 1999. Dissection of individual functions of the Sendai virus phosphoprotein in transcription.
J. Virol. 73, 6474 – 6483.
Bressanelli, S., Tomei, L., Roussel, A., Incitti, I., Vitale, R.L., Mathieu, M.,
De Francesco, R., Rey, F.A., 1999. Crystal structure of the RNA-dependent RNA polymerase of hepatitis C virus. Proc. Natl. Acad. Sci.
U.S.A. 96, 13034 – 13039.
Calain, P., Roux, L., 1993. The rule of six, a basic feature for efficient
replication of Sendai virus defective interfering RNA. J. Virol. 67,
4822 – 4830.
Castaneda, S.J., Wong, T.C., 1990. Leader sequence distinguishes between translatable and encapsidated measles virus RNAs. J. Virol.
64, 222 – 230.
Chuang, J.L., Perrault, J., 1997. Initiation of vesicular stomatitis virus
mutant polR1 transcription internally at the N gene in vitro. J. Virol.
71, 1466 – 1475.
Collins, P.L., Olmsted, R.A., Spriggs, M.K., Johnson, P.R., Buckler-White,
A.J., 1987. Gene overlap and site-specific attenuation of transcription of
the viral polymerase L gene of human respiratory syncytial virus. Proc.
Natl. Acad. Sci. U.S.A. 84, 5134 – 5138.
Colonno, R.J., Banerjee, A.K., 1976. A unique RNA species involved in
initiation of vesicular stomatitis RNA transcription in vitro. Cell 8,
197 – 204.
Cramer, P., Bushnell, D.A., Fu, J., Gnatt, A.L., Maier-Davis, B., Thompson, N.E., Burgess, R.R., Edwards, A.M., David, P.R., Kornberg, R.D.,
2000. Architecture of RNA polymerase II and implications for the transcription mechanism. Science 288, 640 – 649.
Cramer, P., Bushnell, D.A., Kornberg, R.D., 2001. Structural basis of transcription: RNA polymerase II at 2.8 angstrom resolution. Science 292,
1863 – 1876.
Curran, J., 1996. Reexamination of the Sendai virus P protein domains
required for RNA synthesis: a possible supplemental role for the P
protein. Virology 221, 130 – 140.
Curran, J., 1998. A role for the Sendai virus P protein trimer in RNA
synthesis. J. Virol. 72, 4274 – 4280.
Curran, J., Kolakofsky, D., 1988. Scanning independent ribosomal initiation of the Sendai virus X protein. EMBO J. 7, 2869 – 2874.
Curran, J., Pelet, T., Kolakofsky, D., 1994. An acidic-like domain of the
Sendai virus P protein is required for RNA synthesis and encapsidation.
Virology 202.
Curran, J., Boeck, R., Lin-Marq, N., Lupas, A., Kolakofsky, D., 1995a.
Paramyxovirus phosphoproteins form homotrimers as determined by
an epitope dilution assay, via predicted coiled coils. Virology 214,
139 – 149.
Curran, J., Marq, J.-B., Kolakofsky, D., 1995b. An N-terminal domain of the
Sendai paramyxovirus P protein actsw as a chaperone for the NP protein
during the nascent chain assembly step of genome replication. J. Virol.
69, 849 – 855.
Egelman, E.H., Wu, S.-S., Amrein, M., Portner, A., Murti, G., 1989. The
Sendai virus nucleocapsid exists in at least four different helical states.
J. Virol. 63, 2233 – 2243.
Emerson, S.U., 1982. Reconstitution studies detect a single polymerase
entry site on the vesicular stomatitis virus genome. Cell 31, 635 – 642.
Fearns, R., Collins, P.L., 1999. Model for polymerase access to the overlapped L gene of respiratory syncytial virus. J. Virol. 73, 388 – 397.
Gnatt, A.L., Cramer, P., Fu, J., Bushnell, D.A., Kornberg, R.D., 2001.
Structural basis of transcription: an RNA polymerase II elongation
complex at 3.3 A resolution. Science 292, 1876 – 1882.
Gubbay, O., Curran, J., Kolakofsky, D., 2001. Sendai virus genome synthesis and assembly are coupled: a possible mechanism to promote viral
RNA polymerase processivity. J. Gen. Virol. 82, 2895 – 2903.
Gupta, A.K., Shaji, D., Banerjee, A.K., 2003. Identification of a novel
tripartite complex involved in replication of vesicular stomatitis virus
genome RNA. J. Virol. 77, 732 – 738.
Hansen, J.L., Long, A.M., Schultz, S.C., 1997. Structure of the RNAdependent RNA polymerase of poliovirus. Structure 5, 1109 – 1122.
Hausmann, S., Jacques, J.P., Kolakofsky, D., 1996. Paramyxovirus RNA
editing and the requirement for hexamer genome length. RNA 2,
1033 – 1045.
Hausmann, S., Garcin, D., Delenda, C., Kolakofsky, D., 1999a. The versatility
of paramyxovirus RNA polymerase stuttering. J. Virol. 73, 5568 – 5576.
Hausmann, S., Garcin, D., Morel, A.S., Kolakofsky, D., 1999b. Two nucleotides immediately upstream of the essential A6G3 slippery sequence modulate the pattern of G insertions during Sendai virus
mRNA editing. J. Virol. 73, 343 – 351.
Helfman, W.B., Perrault, J., 1988. Altered ATP utilization by the poIR
mutants of vesicular stomatitis virus maps to the N-RNA template.
Virology 167, 311 – 313.
Horikami, S.M., Curran, J., Kolakofsky, D., Moyer, S.A., 1992. Complexes
of Sendai virus NP – P and P – L proteins are required for defective interfering particle genome replication in vitro. J. Virol. 66, 4901 – 4908.
Iseni, F., Barge, A., Baudin, F., Blondel, D., Ruigrok, R.W., 1998. Characterization of rabies virus nucleocapsids and recombinant nucleocapsid-like structures. J. Gen. Virol. 79, 2909 – 2919.
Iverson, L.E., Rose, J.K., 1981. Localized attenuation and discontinuous
synthesis during vesicular stomatitis virus transcription. Cell 23,
477 – 484.
Johansson, K., Bourhis, J.-M., Campanacci, V., Cambillau, C., Canard, B.,
Longhi, S., 2003. Crystal structure of the measles virus phosphoprotein
domain responsible for the induced folding of the C-terminal domain of
the nucleoprotein. J. Biol. Chem. 278, 44567 – 44573..
Karlin, D., Longhi, S., Receveur, V., Canard, B., 2002. The N-terminal
domain of the phosphoprotein of Morbilliviruses belongs to the natively
unfolded class of proteins. Virology 296, 251 – 262.
Kolakofsky, D., Blumberg, B.M., 1982. A model for the control of nonsegmented negative strand virus genome replication. In: Mahy,
A.C.M.a.G.K.D. (Ed.), Virus Persistence. Cambridge Univ. Press, Cambridge, pp. 203 – 213.
D. Kolakofsky et al. / Virology 318 (2004) 463–473
Korzheva, N., Mustaev, A., Kozlov, M., Malhotra, A., Nikiforov, V., Goldfarb, A., Darst, S.A., 2000. A structural model of transcription elongation. Science 289, 619 – 625.
Lamb, R.A., Kolakofsky, D., 2001. Paramyxoviridae: the viruses and their
replication. In: Knipe, D.M., Howley, P.M. (Eds.), Fields Virology,
vol. 1. Lippincott, Williams & Wilkins, Philadelphia, pp. 1305 – 1340.
Landick, R., 1999. Shifting RNA polymerase into overdrive. Science 284,
598 – 599.
Landick, R., 2001. RNA polymerase clamps down. Cell 105, 567 – 570.
Le Mercier, P., Garcin, D., Hausmann, S., Kolakofsky, D., 2002. Ambisense Sendai viruses are inherently unstable but are useful to study viral
RNA synthesis. J. Virol. 76, 5492 – 5502.
Le Mercier, P., Garcin, D., Garcia, E., Kolakofsky, D., 2003. Competition
between the Sendai virus N mRNA start site and the genome 3V-end
promoter for viral RNA polymerase. J. Virol. 77, 9147 – 9155.
Leppert, M., Rittenhouse, L., Perrault, J., Summers, D.F., Kolakofsky, D.,
1979. Plus and minus strand leader RNAs in negative strand virusinfected cells. Cell 18, 735 – 747.
Lis, J., 1998. Promoter-associated pausing in promoter architecture and
postinitiation transcriptional regulation. Cold Spring Harbor Symp.
Quant. Biol. 63, 347 – 356.
Longhi, S., Receveur-Brechot, V., Karlin, D., Johansson, K., Darbon, H.,
Bhella, D., Yeo, R., Finet, S., Canard, B., 2003. The C-terminal domain
of the measles virus nucleoprotein is intrinsically disordered and folds
upon binding to the C-terminal moiety of the phosphoprotein. J. Biol.
Chem. 278, 18638 – 18648.
Marion, D., Tarbouriech, N., Ruigrok, R.W., Burmeister, W.P., Blanchard, L., 2001. Assignment of the 1 H, 15 N and 13 C resonances
of the nucleocapsid-binding domain of the Sendai virus phosphoprotein.
J. Biomol. NMR 21, 75 – 76.
Masters, P.S., Banerjee, A.K., 1988. Complex formation with vesicular
stomatitis virus phosphoprotein NS prevents binding of nucleocapsid
protein N to nonspecific RNA. J. Virol. 62, 2658 – 2664.
Mavrakis, M., Iseni, F., Mazza, C., Schoehn, G., Ebel, C., Gentzel, M.,
Franz, T., Ruigrok, R.W., 2003. Isolation and characterisation of the
rabies virus N degrees – P complex produced in insect cells. Virology
305, 406 – 414.
McGeoch, D.J., 1979. Structure of the gene N:gene NS intercistronic junction in the genome of vesicular stomatitis virus. Cell 17, 673 – 681.
Melcher, K., 2000. The strength of acidic activation domains correlates
with their affinity for both transcriptional and non-transcriptional proteins. J. Mol. Biol. 301, 1097 – 1112.
Nudler, E., Mustaev, A., Lukhtanov, E., Goldfarb, A., 1997. The RNA –
DNA hybrid maintains the register of transcription by preventing backtracking of RNA polymerase. Cell 89, 33 – 41.
Ollis, D.L., Brick, P., Hamlin, R., Xuong, N.G., Steitz, T.A., 1985. Structure of large fragment of Escherichia coli DNA polymerase I complexed with dTMP. Nature 313, 762 – 766.
Pattnaik, A.K., Hwang, L., Li, T., Englund, N., Mathur, M., Das, T., Banerjee, A.K., 1997. Phosphorylation within the amino-terminal acidic domain I of the phosphoprotein of vesicular stomatitis virus is required for
transcription but not for replication. J. Virol. 71, 8167 – 8175.
Peluso, R.W., Moyer, S.A., 1988. Viral proteins required for the in vitro
replication of vesicular stomatitis virus defective interfering particle
genome RNA. Virology 162, 369 – 376.
Perrault, J., Clinton, G.M., McClure, M.A., 1983. RNP template of vesicular stomatitis virus regulates transcription and replication functions.
Cell 35, 175 – 185.
Poch, O., Sauvaget, I., Delarue, M., Tordo, N., 1989. Identification of four
conserved motifs among the RNA-dependent polymerase encoding elements. EMBO J. 8, 3867 – 3874.
Price, D.H., 2000. P-TEFb, a cyclin-dependent kinase controlling elongation by RNA polymerase II. Mol. Cell. Biol. 20, 2629 – 2634.
Roberts, J.W., Yarnell, W., Bartlett, E., Guo, J., Marr, M., Ko, D.C.,
Sun, H., Roberts, C.W., 1998. Antitermination by bacteriophage
lambda Q protein. Cold Spring Harbor Symp. Quant. Biol. 63,
319 – 325.
473
Rose, J.K., 1980. Complete intergenic and flanking gene sequences from
the genome of vesicular stomatitis virus. Cell 19, 415 – 421.
Roux, L., Kolakofsky, D., 1974. Protein kinase associated with Sendai
virions. J. Virol. 13, 545 – 547.
Ryan, K.W., Murti, K.G., Portner, A., 1990. Localization of P protein
binding sites on the Sendai virus nucleocapsid. J. Gen. Virol. 71,
997 – 1000.
Sanchez, A., Kiley, M.P., Holloway, B.P., Auperin, D.D., 1993. Sequence
analysis of the Ebola virus genome: organization, genetic elements,
and comparison with the genome of Marburg virus. Virus Res. 29,
215 – 240.
Santangelo, T.J., Mooney, R.A., Landick, R., Roberts, J.W., 2003. RNA
polymerase mutations that impair conversion to a termination-resistant
complex by Q antiterminator proteins. Genes Dev. 17, 1281 – 1292.
Schoehn, G., Iseni, F., Mavrakis, M., Blondel, D., Ruigrok, R.W., 2001.
Structure of recombinant rabies virus nucleoprotein-RNA complex
and identification of the phosphoprotein binding site. J. Virol. 75,
490 – 498.
Schubert, M., Keene, J.D., Herman, R.C., Lazzarini, R.A., 1980. Site on
the vesicular stomatitis virus genome specifying polyadenylation and
the end of the L gene mRNA. J. Virol. 34, 550 – 559.
Shilatifard, A., Conaway, R.C., Conaway, J.W., 2003. The RNA polymerase II elongation complex. Annu. Rev. Biochem 72, 693 – 715.
Smallwood, S., Ryan, K.W., Moyer, S.A., 1994. Deletion analysis defines a
carboxy-proximal region of Sendai virus P protein that binds to the
polymerase L protein. Virology 202, 154 – 163.
Smallwood, S., Cevik, B., Moyer, S.A., 2002. Intragenic complementation
and oligomerization of the L subunit of the Sendai virus RNA polymerase. Virology 304, 235 – 245.
Stillman, E.A., Whitt, M.A., 1998. The length and sequence composition of
vesicular stomatitis virus intergenic regions affect mRNA levels and the
site of transcript initiation. J. Virol. 72, 5565 – 5572.
Stillman, E.A., Whitt, M.A., 1999. Transcript initiation and 5V-end modifications are separable events during vesicular stomatitis virus transcription. J. Virol. 73, 7199 – 7209.
Tao, Y., Farsetta, D.L., Nibert, M.L., Harrison, S.C., 2002. RNA synthesis
in a cage-structural studies of reovirus polymerase lambda3. Cell 111,
733 – 745.
Tarbouriech, N., Curran, J., Ebel, C., Ruigrok, R.W., Burmeister, W.P.,
2000a. On the domain structure and the polymerization state of the
Sendai virus P protein. Virology 266, 99 – 109.
Tarbouriech, N., Curran, J., Ruigrok, R.W., Burmeister, W.P., 2000b. Tetrameric coiled coil domain of Sendai virus phosphoprotein. Nat. Struct.
Biol. 7, 777 – 781.
Teninges, D., Bras, F., Dezelee, S., 1993. Genome organization of the
sigma rhabdovirus: six genes and a gene overlap. Virology 193,
1018 – 1023.
Tsurudome, M., Bando, H., Kawano, M., Matsumura, H., Komada, H.,
Nishio, M., Ito, Y., 1991. Transcripts of Simian Virus-41 (SV41) matrix
gene are exclusively dicistronic with the fusion gene which is also
transcribed as a monocistron. Virology 184, 93 – 100.
Vidal, S., Kolakofsky, D., 1989. Modified model for the switch from Sendai virus transcription to replication. J. Virol. 63, 1951 – 1958.
Vidal, S., Curran, J., Orvell, C., Kolakofsky, D., 1988. Mapping of monoclonal antibodies to the Sendai virus P protein and the location of its
phosphate. J. Virol. 62, 2200 – 2203.
Vidal, S., Curran, J., Kolakofsky, D., 1990. A stuttering model for paramyxovirus P mRNA editing. EMBO J. 9, 2017 – 2022.
Whelan, S.P., Wertz, G.W., 2002. Transcription and replication initiate at
separate sites on the vesicular stomatitis virus genome. Proc. Natl.
Acad. Sci. U.S.A. 99, 9178 – 9183.
Whelan, S.P., Barr, J.N., Wertz, G.W., 2000. Identification of a minimal size
requirement for termination of vesicular stomatitis virus mRNA: implications for the mechanism of transcription. J. Virol. 74, 8268 – 8276.
Zhang, G., Campbell, E.A., Minakhin, L., Richter, C., Severinov, K., Darst,
S.A., 1999. Crystal structure of Thermus aquaticus core RNA polymerase at 3.3 A resolution. Cell 98, 811 – 824.