RNA binding and replication by the poliovirus RNA polymerase


Material Information

RNA binding and replication by the poliovirus RNA polymerase
Physical Description:
x, 115 leaves : ill. ; 29 cm.
Oberste, Mark Steven, 1957-
Publication Date:


Subjects / Keywords:
Polioviruses   ( mesh )
RNA Nucleotidyltransferases   ( mesh )
Virus Replication   ( mesh )
Immunology and Medical Microbiology thesis Ph.D   ( mesh )
Dissertations, Academic -- Immunology and Medical Microbiology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1988.
Bibliography: leaves 101-114.
Statement of Responsibility:
by Mark Steven Oberste.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001071821
oclc - 20429294
notis - AFF6302
System ID:

This item is only available as the following downloads:

Full Text








I wish to thank Bert Flanegan for allowing me the opportunity to

work in his laboratory and for instilling in me the principles of good

science. He has been an excellent mentor, colleague, and friend, and I

shall always hold him in high regard. I want to thank the members of my

committee, Drs. Dick Moyer, Bill Hauswirth, Ward Wakeland, and Al

Lewin, for all their helpful critiques and suggestions throughout the

course of my work. I also want to thank Vince Racaniello for

participating in my research as my outside examiner. I thank my

parents, Marc and Judy Oberste, for their persistence in raising a

precocious only child and for always encouraging excellence.

I would also like to thank the current and former students of the

Department of Immunology and Medical Microbiology for friendship and

(mostly) good times over the past six years. In particular, I want to

thank Randy Horwitz, Paul Kroeger, Michael Siegel, Erv Faulmann, and

the members of the Flanegan lab, Greg Tobin, Carol Ward, and Phil

Collis for ideas, inspiration, and lots of fun. Greg deserves special

thanks for his help with experiments in the waning months of my

research. As always, none of this would have been possible without

outstanding technical assistance from Joan Morasco, Brian O'Donnell,

and Mike Duke, as well as outstandingly clean glassware from Greg Brown

et al. I am doubly indebted to Joan, Brian, and Phil for giving me

places to live in the waning days of my graduate career. I am grateful

to David Jewell for helping me get started with Turbo Pascal and


Finally, I want to extend my extra special thanks to my wife, Anna

Oberste, for hanging in there through the darkest hours. She is the

best friend I could ever have. She was indispensable in helping me to

think through many difficult technical problems with insightful ideas,

constructive criticisms, and good suggestions. I truly never would have

made it without her.


ACKNOWLEDGEMENTS ..................................................... ii

ABBREVIATIONS ........................................................vi

ABSTRACT. ............................................................. ix


1 INTRODUCTION ................................................... 1

Enzymes ...........................
Radiolabeled Compounds............
Ribonucleic Acid Homopolymers.....
Cell Culture ......................

Purification and Labeling of Poliovirion RNA.............
Purification of Poliovirus RNA Polymerase
and HeLa Host Factor ..................................
Poliovirus Polymerase-RNA Binding Reactions..............
Computer Methods.........................................
Synthesis of Poliovirus-Specific RNA
Using Phage RNA Polymerases...........................
Poliovirus RNA Polymerase RNA Synthesis Reaction.........
Poly(U) Isolation and Nearest Neighbor Analysis..........

INTO A TRANSCRIPTION VECTOR ...........................
Construction of a Full-Length Poliovirus cDNA Clone......
Construction of Subgenomic Clones ........................
Construction of Clones Whose In Vitro Transcripts
Terminate in Poly(A) ..................................
Discussion ...............................................

Binding of Polymerase to Poliovirion RNA.................
Optimal Conditions for RNA Binding .......................
Specificity of Polymerase Binding to Ribohomopolymers....
Competitive Binding Curves Using Virion RNA
and Ribohomopolymers ..................................
Computer Scan of Poliovirus RNA for G+U-Rich Sequences...
Calculation of Ka for Polymerase Binding to Virion RNA...

..... 13
..... 13
..... 13
..... 14
..... 14
..... 18
..... 18
..... 19

..... 19
..... 20
..... 21

..... 21
..... 22
..... 23

..... 24
..... 24
..... 25
..... 25

..... 28
..... 29

..... 39
..... 39
..... 39
..... 41
..... 42

..... 42
..... 43
..... 44

. . .
. . .
. ......
. .
. .
. .
. ......

Calculation of Ka for Polymerase Binding to Transcript RNAs...45
Discussion ....................................................47
BY POLIOVIRUS RNA POLYMERASE ............................... 68
Introduction .................................................. 68
Product RNA Is Covalently Linked to the Template.............. 69
Product RNA Contains Poly(U) .................................. 71
Template Activity of RNAs which Lack the 3' End
of the Genome .............................................. 72
Structural Features at the 3' End of the Poliovirus Genome....73
Sequences at the 3' End Are Required for Initiation
of RNA Synthesis ........................................... 74
Discussion .................................................... 75

6 CONCLUSIONS AND FUTURE STUDIES ................................ 91
Conclusions ................................................... 91
Future Studies ................................................ 93

APPENDIX ............................................................. 96

REFERENCES ........................................................... 101

BIOGRAPHICAL SKETCH ................................................. 115


A adenine

Ap ampicillin

ATP adenosine triphosphate

BAP bacterial alkaline phosphatase

BMV brome mosaic virus

C cytidine

C inhibitor concentration which gives half-maximal binding

cDNA complementary DNA

Ci Curie

CIP calf intestinal phosphatase

cpm counts per minute

CRC poliovirus crude replication complex

CTP cytidine triphosphate

dATP deoxyadenosine triphosphate

dCTP deoxycytidine triphosphate

dGTP deoxyguanosine triphosphate

DNA deoxyribonucleic acid

DNase I deoxyribonuclease I

DTT dithiothreitol

EDTA ethylenediaminetetraacetic acid

EMCV encephalomyocarditis virus



























foot-and-mouth disease virus


guanosine triphosphate


N-2-hydroxyethylpiperizine-N'-2'-ethanesulfonic acid

HeLa cell host factor

association constant

Luria-Bertani broth

multiple cloning site


oligoadenylic acid

oligocytidylic acid

oligoguanylic acid

oligouridylic acid

poly(A) Sepharose

cytidine 5'-3'-bis-phosphate

polyadenylic acid

polycytidylic acid

polyguanylic acid

polyuridylic acid

ribonucleic acid


RNA-qualified deoxyribonuclease

sodium dodecyl sulfate


thymidine triphosphate

TUT terminal uridylyl transferase

TYMV turnip yellow mosaic virus

U uridine

UTP uridine triphosphate

VPg genome-linked virion protein

vRNA poliovirion RNA


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Mark Steven Oberste

December, 1988

Chairman: James B. Flanegan
Major Department: Immunology and Medical Microbiology

RNA binding and RNA synthesis by the poliovirus RNA-dependent RNA

polymerase were studied in vitro using purified polymerase. Templates

for binding and RNA synthesis studies were natural RNAs, homopolymeric

RNAs, or subgenomic poliovirus-specific RNAs synthesized in vitro from

cDNA clones using SP6 or T7 RNA polymerases. The binding of the

purified polymerase to poliovirion and other RNAs was studied using a

protein-RNA nitrocellulose filter binding assay. A cellular poly(A)-

binding protein was found in the viral polymerase preparations, but was

easily separated from the polymerase by chromatography on poly(A)

Sepharose. The binding of purified polymerase (fraction 5-PAS) to 32p

labeled ribohomopolymeric RNAs was examined, and the order of binding

observed was poly(G) >>> poly(U) > poly(C) > poly(A). This suggested

that the polymerase binding site may be rich in G and/or U residues.

The Ka for polymerase binding to poliovirion RNA and to a full-length

negative strand transcript was about 1 x 109 M-1. The polymerase binds

to subgenomic RNAs which contain the 3' end of the genome with a Ka

similar to that for virion RNA, but binds less well to 18S rRNA, globin

mRNA, and subgenomic RNAs which lack portions of the 3' noncoding

region (Kas about 2 x 108 M-1). The polymerase copies subgenomic RNAs

which contain the 3' end of the genome into dimer-size products which

have a poly(U) tract equal in length to the poly(A) tract in the

template. This, coupled with the observation that labeled template can

be chased into dimer-length product, demonstrates conclusively that RNA

synthesis by poliovirus polymerase is template-primed in vitro. Using

subgenomic RNAs which lack defined regions of the genome as templates,

it was determined that the purified polymerase requires sequences near

the 3' end to initiate RNA synthesis. In particular, removal of the

last four nucleotides of the 3' noncoding region, GGAG, results in a

total loss of binding and a three-fold reduction in template activity.

The results indicate that the poliovirus polymerase recognizes specific

structures during initiation of RNA synthesis in vitro.


The Picornaviridae are a family of small, nonenveloped viruses

which comprise a large number of important human and animal pathogens.

The picornaviruses are classified into four subgroups according to

physical properties of the virions. These groups are the enteroviruses

(poliovirus, Coxsackieviruses A and B, Theiler's murine encephalitis

virus [TMEV], swine vesicular virus, human echoviruses, and hepatitis A

virus), rhinoviruses (human and bovine rhinoviruses), aphthoviruses

(foot-and-mouth disease viruses [FMDV]), and cardioviruses

(encephalomyocarditis virus [EMCV], mengovirus, Maus-Elberfeld virus ,

and Columbia SK virus). Palmenberg (1986a) has recommended subdivision

of the family into four genera based on sequence homology, rather than

physical properties. In this scheme, FMDV and hepatitis A are each in

their own group and the murine viruses (EMCV, mengovirus, and TMEV) are

classified together, as are most of the human viruses (poliovirus,

coxsackievirus, and rhinovirus).

The typical picornavirus genome is 7000-9000 nucleotides of

positive-sense single-stranded RNA. The polyprotein translation

product is processed by two virus-encoded endoproteases (reviewed by

Palmenberg, 1986b) to yield the individual virus-specific proteins. The

RNA is characterized by a covalently-linked viral protein, VPg, at the

5' end and a poly(A) tract of variable length at the 3' end. The long



open reading frame is divided into three regions based on function. The

5' region, Pl, encodes the four capsid proteins, 1A, IB, 1C, and ID.

The center region, P2, encodes three viral peptides, 2A (protease) and

2B and 2C, whose functions are unknown. Protein 2A is involved in the

first processing step of the capsid precursor and has also been

implicated in the shut-off of host cell functions during infection. The

replication proteins, 3A (unknown function), 3B (VPg), 3Cpro (major

protease), and 3DPol (RNA polymerase) are contained in P3. The coding

region is flanked by 5' and 3' nontranslated sequences. In addition,

some of the viruses (for example, EMC, mengovirus, FMDV) have a leader

peptide at the 5' end of P1 and a poly(C) tract within the 5' noncoding

region. In poliovirus, translation begins at the AUG at nucleotide 743

and terminates at a UAG at base 7370, yielding a polyprotein of 2209

amino acids.

As would be expected for closely related viruses, there is a high

degree of sequence similarity within the coding region of the different

picornaviruses, at both the nucleotide and amino acid levels. There is

also a great deal of nucleotide sequence similarity in the 5' and 3'

noncoding regions, largely in the form of conserved secondary

structures (Palmenberg, 1986a; 1986b; 1988).

The virion RNA of poliovirus type 1 (Mahoney strain) contains 7440

nucleotides and a poly(A) tract of 75 to 100 residues at its 3' end

(Armstrong et al., 1972; Yogo and Wimmer, 1972; Spector and Baltimore,

1975a), most or all of which is coded by a poly(U) sequence in the

negative strand RNA (Yogo and Wimmer, 1973; Yogo et al., 1974; Yogo and

Wimmer, 1975; Spector and Baltimore, 1975c). VPg is linked via a


uridylyl-0-tyrosyl bond to the 5' end of virion RNA and the positive-

and negative-strands of dsRNA and the replicative intermediate

(Flanegan et al., 1977; Nomoto et al., 1977a; Pettersson et al., 1978).

The only viral RNA lacking VPg is the mRNA, found in polysomes (Nomoto

et al., 1977b; Pettersson et al., 1977). VPg is specifically removed

from viral mRNA by an unlinking enzyme found in both poliovirus-

infected and uninfected HeLa cells (Ambros et al., 1978; Ambros and

Baltimore, 1980). VPg is not required for infectivity (Hewlett et al.,

1976). When the poly(A) tract is removed from purified virion RNA by

RNase H in the presence of poly(dT), the RNA is no longer infectious

(Spector and Baltimore, 1974). Since the deadenylated RNA retains the

ability to direct protein in an in vitro translation system, it is

likely that the poly(A) tract is required for replication (Spector and

Baltimore, 1975c; Spector et al., 1975).

Poliovirus replicates in the cytoplasm of infected cells (Franklin

and Baltimore, 1962). Replication of viral RNA requires the virus-

specific RNA-dependent RNA polymerase (replicase), which was originally

purified as a component of the endogenous RNA replication complex

(Baltimore et al., 1963; Lundquist et al., 1974; Flanegan and

Baltimore, 1977). The RNA component of this complex, the replicative

intermediate, contains one complete negative strand and about six

nascent chains of positive-strand RNA (Girard and Baltimore, 1966).

Viral replication produces about ten times as many positive strands as

negative strands. The rationale for such unbalanced RNA synthesis is

that positive strands must function as template, messenger RNA, and

nascent virion RNA, while minus strands function only as template


(Baltimore, D., 1969). The mechanism of regulation of positive strand

function is unknown, but may involve ribosome binding, binding to viral

replicase, packaging signals, or other factors.

A single virus-specific protein (3DPol) copurifies with polymerase

activity isolated from cytoplasmic extracts of infected cells (Van Dyke

and Flanegan, 1980). In vitro, the purified polymerase will copy a

number of homopolymeric and heteropolymeric RNA templates using an

oligoribonucleotide primer (Flanegan and Van Dyke, 1979; Tuschall et

al., 1982). These templates include several natural and synthetic

polyadenylated RNAs, as well as synthetic ribohomopolymers. All

homopolymer templates are not copied with equal efficiency, however.

The order of preference is poly(A):oligo(U) = poly(C):oligo(G) -

poly(I):oligo(C) > poly(U):oligo(A); poly(G):oligo(C) is not copied at

all (Tuschall et al., 1982; Ward et al., 1988). The mechanism of

template specificity in vivo is unknown. Specificity may be due to

interaction of the replication complex with specific viral RNA

sequences or structures, but partitioning of the viral replication

complex away from cellular RNAs may also play a role (Kuhn and Wimmer,

1987). The crude replication complex cosediments with smooth membranes

on sucrose gradients. This smooth membrane fraction is devoid of

polysomes, implying an absence of cellular RNA (Caliguiri and Tamm,


The replicase requires an oligonucleotide primer to copy RNA

templates in vitro (Flanegan and Van Dyke, 1979), but a HeLa cell

protein known as "host factor" (HF) can substitute for the primer for

polyadenylated heteropolymeric templates (Dasgupta et al., 1980). In


addition, HF stimulates, but is not required for synthesis of product

RNA from a poliovirion RNA template with an oligo(U) primer (Young et

al., 1987). Host factor has been purified from uninfected cells by

different methods in different laboratories. Its activities in vitro

are reported to be protein kinase (Morrow et al., 1985), terminal

uridylyl transferase (TUT) (Andrews, 1985; Andrews et al., 1985;

Andrews and Baltimore, 1986), and RNA endonuclease (Hey et al., 1987).

It is not known whether any or all of these activities reside in the

same protein, but our laboratory has recently separated the TUT

activity from endonuclease activity (D.C. Young and J.B. Flanegan,

unpublished results).

Highly purified polymerase produces template-sized product in the

presence of an oligo(U) primer and product twice the size of the

template when HeLa cell host factor is substituted for oligo(U) (Young

et al., 1985; Young et al., 1986; Hey et al., 1986). The precise role

of host factor in the generation of dimer-sized product remains

unclear. The TUT activity in some HF preparations may be involved. TUT

adds about five uridylate residues to the 3' end of several different

polyadenylated RNA templates. This short oligo(U) sequence could

potentially form a hairpin on a polyadenylated template, providing a

primer for RNA synthesis (Flanegan et al., 1987). Alternatively, HF

could bind to the 3' end of the template and stabilize non-base-paired

hairpin formation, allowing initiation by the polymerase. In this case,

one would have to invoke template slippage, addition of adenylate

residues by a cellular poly(A) polymerase, or some other mechanism to

prevent the loss of genetic information at the ends of the replicating


RNA molecules (Spector and Baltimore, 1975b). Hey et al. (1987) have

suggested that endonuclease activity associated with a crude host

factor preparation may incise the template RNA at the base of a pre-

existing internal hairpin, creating a primer to be elongated by the

polymerase. They suggest that this type of host factor activity may be

an in vitro artifact, since its activity is inconsistent with RNA

replication in vivo. Their discovery of poly(A)-deficient dsRNA led

them to propose that the endonuclease activity may function in vivo,

but for no useful purpose (Richards et al., 1987). Whatever its

mechanism, host factor has been shown to be associated with poliovirus

polymerase in infected cells and in vitro, strongly implying a role in

viral replication in vivo (Dasgupta, 1983b). Curiously, HF does not

promote RNA synthesis by the polymerase with a poly(A) template in the

absence of an oligo(U) primer, even though poliovirion RNA and several

other active RNA templates terminate in a poly(A) tail. This suggests

that additional sequences or structures are required for replication by

the poliovirus polymerase.

Two models have been proposed for the initiation of poliovirus RNA

synthesis in infected cells. These alternative mechanisms are based

upon work with two different in vitro poliovirus replication systems,

one using membranous crude replication complexes (CRC) (Takegami et

al., 1983b) and the other, highly purified components from soluble cell

extracts (Flanegan and Van Dyke, 1979). In addition, the CRC system

synthesizes plus-strand RNA from a minus-RNA template while the soluble

system has been used to synthesize minus-strand product from a virion

RNA (plus-strand) template, making comparison of results difficult


(Kuhn and Wimmer, 1987). Also, different laboratories using soluble

components often purify their polymerase or host factor by different

methods (Andrews et al., 1985; Hey et al., 1986; Young et al., 1986;

Lubinski et al., 1987), further confusing matters. Any replication

model must account for the fact that both the positive strand 5'-and

3'-terminal sequences and the poly(A) tail are required for infectivity

and RNA replication, and therefore all genetic information must be


The benefit of the CRC system used by Wimmer and colleagues is

that it is membrane-bound and is thought to contain most of the viral

and cellular proteins required for replication. Full-length RNA

synthesized in CRC is indistinguishable from virion RNA. The

disadvantage is that the protein components of the CRC are undefined

and only preinitiated templates are copied, preventing the addition of

exogenous components which might shed light on the replication

mechanism. Based on studies using CRC, it was suggested that a

uridylylated form of VPg acts as a primer for RNA synthesis (Toyoda et

al., 1987). The evidence for this model is the presence of VPg-pUpU in

poliovirus-infected cells (Crawford and Baltimore, 1983) and the in

vitro synthesis of VPg-pUpU in CRC (Takegami et al., 1983a; Takeda et

al., 1986). There is some evidence that preformed VPg-pUpU can be

chased into longer RNA in vitro (Takeda et al., 1986), but these data

remain equivocal. The VPg-linked U residues would presumably base-pair

with the terminal adenylates found at the 3' end of both positive and

negative strands. This short paired region may be stabilized by HF or

other proteins until elongation can proceed. According to this model,

VPg or a VPg precursor peptide is uridylylated, perhaps by 3DPo,

yielding VPg-pUpU which then primes RNA synthesis.

The soluble component system, first developed by Flanegan and

Baltimore (1977) permits manipulation of exogenous templates and enzyme

preparations, allowing the characterization of the components required

for poliovirus replication in vitro. Young et al. (1985) proposed a

template-priming model based upon the in vitro synthesis of dimer-sized

product RNA in the host factor-primed replication reaction described

above. A similar model was also proposed by Andrews et al. (1985) after

finding that TUT activity was associated with some HF preparations. In

this model, TUT adds a short oligo(U) sequence to the 3' end of the

positive strand. The oligo(U) then base-pairs with the poly(A)

sequence, forming a 3'-terminal hairpin which acts as primer for the

polymerase. Since the minus strand terminates in ...pApA, the same

mechanism could be involved in priming of new positive strand

synthesis. Recent studies in our laboratory show that synthetic VPg can

convert dimeric product RNA into template size by forming a covalent

bond with the poly(U) 5' phosphate via a phospho-0-tyrosyl bond (Tobin

and Flanegan, 1988). This appears to be the result of a trans-

esterification reaction in which the hydroxyl group of tyrosine acts as

a nucleophile. This reaction results in the cleavage of a

phosphodiester bond in the terminal poly(A)-poly(U) hairpin and the

formation of a new phosphodiester bond between VPg and the product RNA.

The best studied RNA-dependent RNA polymerase is that encoded by

bacteriophage Q8 in (reviewed by Blumenthal and Carmichael, 1979 and

Blumenthal, 1982). The active replicase is a complex of one viral

protein and three cellular proteins. An additional cellular protein,

known as host factor, is required for in vitro replication of QP RNA.

The QO8 replicase binds to QP RNA about tenfold more tightly than to

nonhomologous RNAs (Silverman, 1973b). The replicase interacts with

two internal binding sites and with a C-rich sequence at the 3' end of

QO RNA, apparently relying more upon RNA structure than sequence for

recognition (Senear and Steitz, 1976). The Qfi replicase-host factor

complex exhibits a high degree of specificity for QO RNA templates

(Silverman, 1973b). Under certain conditions, however, the template

specificity of the replicase is significantly reduced. In the presence

of an oligonucleotide primer, the specificity of initiation is entirely

bypassed and the replicase will efficiently copy nonhomologous RNAs.

This is similar to the elongation activity observed with the poliovirus

polymerase on different templates in the presence of oligonucleotide


Like QP replicase and other polymerases, the poliovirus RNA

polymerase must bind to its template before initiation and elongation

can occur. In addition, the polymerase must distinguish between viral

RNA templates and cellular RNAs in vivo, by differential template

binding as in Q8, by compartmentalization of the replication complex

away from cellular RNAs, or by some other mechanism. The polymerase

must also distinguish between viral (+) and (-) RNA strands, perhaps by

a similar mechanism, to produce the 10:1 positive- to negative-strand


All picornavirus genomes so far sequenced have a putatively

stable, base-paired hairpin of approximately 25 bases at or near the 5'


end of the virion RNA (3' end of the negative strand). A deletion of a

single nucleotide at the base of this hairpin results in temperature-

sensitive replication of polioviral RNA (Racaniello and Meriam, 1986).

A nearby second-site mutation restores the hairpin structure, but not

the sequence, and results in normal replication at both permissive and

restrictive temperatures. In all poliovirus serotypes, as well as in

several related picornaviruses, a putative stem-loop of 42 bases is

found in the 3' noncoding sequence 24 bases from the poly(A) tract in

the positive strand (lizuka et al., 1987). Sarnow et al. (1986) have

constructed insertion mutants within this region, including an eight-

base insertion which results in a temperature-sensitive phenotype. Two-

and ten- base insertions at the same site have no effect, strongly

suggesting that specific RNA folding is required and insertion of

eight, but not two or ten bases, interferes with the formation of a

required structure. More recently, they have isolated a second-site

revertant whose mutation appears to map to the polymerase region of the

genome (Jacobson et al., 1988). This mutation may allow the altered

polymerase protein to associate more efficiently with the altered

sequence and structure, leading to normal replication.

All positive-sense RNA viruses face a problem early in their

replication cycle: the same RNA molecule must act both as mRNA for

direction of viral protein synthesis and as template for viral

replication. The ribosome must track along the RNA in the 5' to 3'

direction while polymerase must travel 3' to 5' along the template.

Phage QP solves this problem by utilizing the viral replicase as a

specific inhibitor of ribosome binding to viral RNA (Weber et al.,


1972). The replicase is unable to dislodge ribosomes from preformed

polysomes, but can bind to free RNA preventing polysome formation.

Detailed studies showed that the replicase and ribosome bind to

overlapping sites on the QOf RNA, mutually excluding one another.

Only limited information is available on sequence recognition signals

for RNA-dependent RNA polymerases of eukaryotic viruses. A 3'-terminal

134-nucleotide sequence with a tRNA-like structure in brome mosaic

virus RNA was recently shown to have a very important role in template

recognition by its viral polymerase (Miller et al., 1986). Genetic

studies with site-directed mutants of poliovirus suggest that 3'-

terminal sequences in both the positive- and negative-strand RNAs are

required for RNA replication (Racaniello and Meriam, 1986; Sarnow et

al., 1986). Kaplan and Racaniello (1988) have recently shown that

poliovirus-specific RNA transcripts lacking 386 nucleotides at the 3'

end is not replicated in transfected HeLa cells, supporting the idea

that these sequences play a major role in replication.

Much is known about the sequence specificity of DNA-dependent RNA

polymerases, but very little about RNA-dependent polymerases in animal

systems. Nucleic acid binding proteins such as transcription initiation

factors and ribosomal proteins have also been studied in great detail

(reviewed by Vogel, 1977). The stability and ease of manipulation of

DNA sequence and structure have facilitated the mapping of specific

protein binding sites at the nucleotide level in several systems (for

example, see Nagata et al., 1983; Milman and Hwang, 1987). Due to

recombinant DNA technology, these methods can be applied to the study

of poliovirus RNA replication. Racaniello and Baltimore (1981a)


constructed a complete cDNA copy of the poliovirus genome, cloned into

the Pst I site of pBR322. Plasmid DNA containing the complete viral

sequence produces infectious virus when transfected into susceptible

cells (Racaniello and Baltimore, 1981b). More recently, we (Flanegan et

al., 1987; Oberste et al., 1988) and others (Kaplan et al., 1985; van

der Werf et al., 1986; Ypma-Wong and Semler, 1987) have cloned the

poliovirus cDNA into transcription vectors and produced poliovirus-

specific RNA in vitro which is of higher specific infectivity than

poliovirus-specific cDNA. These clones allow the manipulation of viral

sequences in DNA form and conversion back into RNA by specific

transcription of the insert sequences. We are now in a position to

modify the primary sequence of poliovirus to study the sequence and

structure requirements for viral replication.

The focus of the work described here is (1) to construct a set of

poliovirus cDNA clones to allow easy manipulation of the poliovirus

genome, (2) to determine whether polymerase binding specificity plays

a role in template selection in vitro, and (3) to determine whether

specific viral RNA sequences or structures are required for efficient

replication in vitro.



Restriction enzymes, SP6 RNA polymerase, and T7 RNA polymerase

were obtained from Bethesda Research Laboratories (BRL, Gaithersburg,

MD), New England Biolabs (Biolabs, Beverly, MA), and Promega Corp.

(Madison, WI). BRL also supplied T4 RNA ligase, bacterial alkaline

phosphatase (BAP), mung bean nuclease, nuclease Bal 31, Klenow fragment

of E. coli DNA polymerase I, T4 DNA polymerase, T4 polynucleotide

kinase, and Sl nuclease. T4 DNA ligase was purchased from BRL or

International Biotechnologies, Inc. (New Haven, CT), calf intestinal

phosphatase from Sigma Chemical Co. (St. Louis, MO), and E. coli DNA

polymerase I from Worthington Biochemical Corp. (Freehold, NJ).

Pancreatic DNase I was from Boehringer Mannheim Biochemicals

(Indianapolis, IN) and RQ1TM DNase I from Promega. E. coli poly(A)

polymerase was purchased from Pharmacia Fine Chemicals (Piscataway,

NJ). Lysozyme was from Miles Laboratories (Naperville, IL). SequenaseTM

DNA sequencing kit was purchased from US Biochemical Corp. (Cleveland,

OH). All enzymes were used as recommended by the manufacturer.

Radiolabeled Compounds

[3H]UTP (50-60 Ci/mmole), [3H]ATP (50 Ci/mmole), [a-35SJdATP (>600

Ci/mmole), [a-32P]dATP (800 Ci/mmole), [a-32P]UTP (>410 Ci/mmole), 5'-


[32P]pCp (>3000 Ci/mmole), [y-32P]ATP (>2000 Ci/mmole), and [a-32P]dCTP

(>410 Ci/mmole) were obtained from Amersham (Arlington Heights, IL.


Cloning vectors pGEM-1 and pGEM-2 were from Promega. Poliovirus

cDNA clone pEV104 was a gift of Dr. Bert Semler, UC Irvine (Semler et

al., 1984). Oligonucleotides used in cloning experiments were

synthesized on an Applied Biosystems model 380A or 380B automated DNA

synthesizer, using phosphoramadite chemistry. When necessary,

oligonucleotides were further purified by preparative gel

electrophoresis in 20% acrylamide gels containing 7 M urea (Lloyd et

al., 1986).


All plasmids were propagated in E. coli strain DH5a (Hanahan,

1983) in LB medium (Lennox L broth base, Gibco, Madison, WI.)

supplemented with 50 pg/ml ampicillin. Large-scale DNA isolation was by

the method of Wang and Becherer (1983).

Plasmids containing poliovirus cDNA sequences were constructed by

standard methods (Maniatis et al., 1982). The following is a

description of the general cloning procedure; a detailed account of the

construction of each clone is given in chapter 3. Polio-specific DNA

fragments were purified from agarose gels using the low-melting-point-

agarose method. Thirty micrograms of parental plasmid DNA was digested

with appropriate restriction enzyme(s) in a volume of 300 il and the

fragment of interest was isolated by the low-melting-point agarose

method (Maniatis et al., 1982). After addition of 30 pl 10X DNA gel

loading dye, the digest reaction was electrophoresed through an agarose


gel in 0.5X TEA buffer, containing 0.5 pg/ml ethidium bromide in both

the gel and running buffer. Once the vector- and polio-specific bands

had resolved, as visualized under UV illumination, a trough (one cm by

width of the band) was cut just ahead of the fragment. A solution of

molten low melting point (IMP) agarose (BRL) in 0.5X TEA was poured

into the trough and allowed to gel before continuing electrophoresis.

The fragment was run to the center of the LMP agarose and cut out with

a sterile spatula. It was weighed to estimate its volume and placed in

a sterile tube with 0.1 volume 1 M Tris-HCl, pH 7.0 and 0.1 volume 3 M

sodium acetate, pH 6.0. The gel slice was melted at 65*C for 20 min and

the DNA extracted with 65*C phenol, room temperature phenol, phenol-

chloroform (phenol-chloroform-isoamyl alcohol [25:24:1]), and

chloroform (chloroform-isoamyl alcohol [24:1]). The DNA was recovered

by precipitation with 2 volumes of cold ethanol, incubation at -20C

for at least 30 min, and centrifugation at 12,000 x g for 15 min at

4C. The pellet was dried in vacuo and resuspended in sterile water at

a concentration of 100 pg/ml.

Vector DNA (pGEM-1) was digested with Eco RI and dephosphorylated

with BAP (20 units per pg DNA at 65C for 60 min). The

dephosphorylation step was necessary to reduce the number of insert-

negative clones following ligation. Digested and dephosphorylated DNA

was extracted with phenol-chloroform and chloroform, ethanol-

precipitated, and resuspended in sterile water at 100 pg/ml.

Ligations were carried out by a modification of a published

procedure (Rodriguez and Tait, 1983). Insert DNA was combined with 100

ng vector at a 2:1 molar insert:vector ratio in a 25 1l reaction as


described in Materials and Methods. Reactions were incubated for 16 h

at 4*C. E. coli DH5a was transformed with the ligated DNA by combining

the ligation mixture with 100 1l competent cells (Rodriguez and Tait,

1983). The transformation reaction was incubated on ice for 30 min,

then at 37*C for 5 min, prior to addition of 0.9 ml LB. Incubation was

continued at 37*C for 1-3 h and 100 pl of the mixture plated on LB-Ap

plates (LB + 15 mg/ml bacto-agar [Difco] + 50 pg/ml ampicillin). The

plates were incubated overnight at 37*C and screened by a modification

of the mini-prep plasmid isolation of Maniatis et al. (1982).

Individual colonies were picked and plated on an LB-Ap master plate. A

sample of each colony was inoculated into 1 ml LB-Ap in a 1.5 ml

microfuge tube and grown overnight at 37*C with shaking. Cells were

pelleted by centrifugation at 12,000 x g for one minute and resuspended

in 100 pA STET (50 mM Tris-HCl, pH 8.0, 8% [w/v] sucrose, 1 mM EDTA,

and 0.5% [v/v] Triton X-100). Five microliters of 10 mg/ml lysozyme

(Miles Laboratories) were added to each tube and the tubes were

immersed in a boiling water bath for exactly 40 sec. Cellular debris

from the lysed bacteria was removed by centrifugation at 12,000 x g for

15 min. The supernatant was removed, extracted with phenol-chloroform

and chloroform, and the nucleic acids precipitated with two volumes of

cold ethanol. The pellets were resuspended in 40 1l of water and an 8

1l aliquot was analyzed by digestion with one or more restriction

enzymes and agarose gel electrophoresis. Since the sequence of the

poliovirus genome is known, the size of the digest products is

indicative of a particular insert. Such digestion can also be used to

determine the orientation of the inserted fragment relative to the


vector sequences. This basic methodology was used to construct all


Subgenomic polio-specific subclones were created by one of two

methods (see chapter 3 for details). Subclones containing one or both

ends of the polio genome were constructed by simple deletion of

internal sequences, followed by religation. Subclones containing only

internal sequences were constructed by digesting pOF2612 with one or

more restriction enzymes and lighting to compatible ends of digested

and dephosphorylated pGEM-1.

For the synthesis of RNA which terminates in poly(A), the Eco RI

site of pOFl211 was replaced with an Mlu I site using a synthetic

oligonucleotide linker in a multistep procedure. The DNA was phenol-

extracted and ethanol-precipitated after each step to remove protein

and byproducts which might interfere with subsequent steps. Clone

pOFl211 was linearized with Eco RI and the overhanging nucleotides were

removed by treatment with mung bean nuclease to leave ...AAAAAG-3'OH.

The 3'-terminal G residue was removed by treatment with T4 DNA

polymerase in the presence of dATP. This reaction produces ...AAAAA-

3'OH in the top strand and a 5'-overhanging C residue in the lower

strand. The extra C was removed by a second treatment with mung bean

nuclease. An oligonucleotide linker containing the Mlu I recognition

site (AAACGCGTTT) was ligated to the plasmid with T4 DNA ligase. After

ligation, excess linker was digested away with Mlu I, the DNA was

diluted to 1 pg/ml, and the plasmid was recircularized with T4 DNA

ligase to yield pOFl211-Mlu.


A nested set of clones having deletions of the 5' portion of the

3' noncoding region were constructed from pOFl211-Mlu by the Bal 31

nuclease method (Maniatis et al., 1982). DNA was extracted and

precipitated as described above after each step. The parental clone was

digested within the insert with Sty I (polio base 7210) and treated

with Bal 31 nuclease to remove unwanted nucleotides. The reaction

conditions were adjusted give an average digestion rate of about 28 bp

per minute using the method of Maniatis et al. (1982). To obtain a

useful range of deletions, time points were taken at five, six, seven,

eight, nine, and ten minutes. The DNA was cleaved in the MCS with Hind

III and the ends repaired with the Klenow fragment of E. coli DNA

polymerase I prior to blunt-end ligation with T4 DNA ligase. Clones

were characterized by sequencing of double-stranded mini-prep DNA using

the primer-extension, dideoxy chain-termination method first described

by Sanger et al. (1977) and modified by Tabor and Richardson (1987).

Ribonucleic Acid Homopolymers

Poly(A), poly(C), poly(G), and poly(U) were purchased from

Pharmacia P-L Biochemicals (Piscataway, NJ). For direct RNA-binding

assays, homopolymers were dephosphorylated with BAP and 5' end labeled

with T4 polynucleotide kinase and [7-32P]ATP, as described (Maniatis et

al., 1982). Specific activity of labeled homopolymers was in the range

1-10 x 106 cpm/pg.

Cell Culture

HeLa cells were maintained in suspension culture and infected with

poliovirus type 1 (Mahoney strain) as previously described (Villa-

Komaroff, et al., 1974).


Purification and Labeling of Poliovirion RNA

Poliovirion RNA (vRNA) was isolated from infected cells as

described (Young et al., 1986). For binding experiments, vRNA was

labeled at the 3' end using T4 RNA ligase and 5'-[32PlpCp by the method

of Romaniuk and Uhlenbeck (1983), except that a 30 pl reaction

contained 8 pg vRNA, 40 pCi 5'-[32P]pCp, and 8 pg ligase. The reaction

was Incubated for 16 h at 4*C, phenol-extracted, and ethanol-

precipitated three times. The specific activity of labeled vRNA was

typically about 1 x 106 cpm/pg (2.5 x 106 cpm/pmole).

Purification of Poliovirus RNA Polymerase and HeLa Host Factor

Fraction 4-HA polymerase was purified from poliovirus-infected

cells as described (Young et al., 1986). Polymerase used in some

experiments was further purified by chromatography on a 1.8 x 1.0 cm

column of poly(A)-Sepharose (AGPOLY(A)TM type 6, Pharmacia Fine

Chemicals). The sodium phosphate content of pooled fraction 4-HA

polymerase was diluted to 35 mM with poly(A)-Sepharose (PAS) binding

buffer (50 mM Tris-HCl, pH 7.5, 2 mM dithiothreitol, 10 pg/ml

ovalbumin, 20% glycerol). The polymerase was loaded onto a poly(A)-

Sepharose column equilibrated with PAS binding buffer, washed with 20

ml PAS binding buffer, and eluted with a 60 ml linear 0-0.5 M KC1

gradient in 1 ml fractions. KC1 concentration of selected fractions

was calculated from the conductivity measured with a Radiometer

conductivity meter. Polymerase elongation activity was assayed using a

poly(A) template and oligo(U) primer as described (Flanegan and

Baltimore, 1977). RNA binding activity was measured as described below

and fractions containing both binding and elongation activities were


pooled and stored at -70*C. This polymerase preparation was designated

fraction 5-PAS and was stable for at least six months at -70C.

A control extract preparation was isolated from uninfected cells

following the same protocol used to isolate the polymerase. A mock-

infected HeLa cell suspension culture was prepared by treating

uninfected HeLa cells with 5 pg/ml actinomycin D (Sigma Chemical Co.,

St. Louis, Mo.) and incubating at 37*C for six hours. A cytoplasmic

extract was prepared and carried through the fraction 4-HA step as

previously described for the purification of poliovirus polymerase from

infected cells (Young et al., 1986). Control extract (fraction 4-HA)

was chromatographed as above on a separate poly(A)-Sepharose column.

RNA binding activity was measured as described below. Since RNA-

dependent RNA polymerase activity was not present in the uninfected

extracts, the fractions which normally contain polymerase activity were

selected at each purification step for the isolation of the control


HeLa cell host factor was purified from uninfected HeLa cells

using protocol II of Young et al. (1985).

Poliovirus Polymerase-RNA Binding Reactions

Protein-RNA filter binding assays were performed by a modification

of the procedure of Zimmern and Butler (1977). Standard binding

reactions contained 200 fmoles 32P-labeled vRNA substrate or a varying

amount of 32P-labeled homopolymer, and fraction 4-HA or fraction 5-PAS

polymerase in 50 mM Hepes-KOH, pH 8.0, 7 mM MgCl2 in a total volume of

100 p1. After incubation at 30*C for 60 min, reaction solutions were

diluted to 1 ml with TE (10 mM Tris-HCl, pH 7.5, 1 mM EDTA) and


filtered through 0.45 Am nitrocellulose filter disks (Schleicher and

Schuell BA 85) which were previously boiled for 30 min in water and

rinsed extensively. The filters were washed with 6 ml of TE and counted

in ReadySolvTM (Beckman) in a Beckman liquid scintillation counter.

Binding is expressed as fmoles labeled RNA specifically bound,

calculated from the specific activity of the labeled RNA and corrected

for isotope half-life. Binding in the absence of added protein was

typically 0.5-2.0% of the added radioactivity and was subtracted from

the experimental values.

Computer Methods

DNA and RNA sequences were analyzed using an IBM-compatible

personal computer and the following software packages: MicroGenie,

Beckman Instruments, Palo Alto, CA (Queen and Korn, 1984); PCFold

(Zuker and Stiegler, 1981); Molecule (Bionet, Palo Alto, CA); IG Suite

(Intelligenetics, Palo Alto, CA). RNAScan was written using Turbo

Pascal, version 3.0 (Borland International, Scotts Valley, CA). The

source code of RNAScan is listed in the Appendix.

Synthesis of Poliovirus-Specific RNA Using Phage RNA Polymerases

Poliovirus-specific RNA was synthesized from cloned cDNA using

purified DNA-dependent RNA polymerase from SP6 or T7 bacteriophages, as

indicated in the text (Chamberlin et al., 1983; Melton et al., 1984).

For synthesis of large amounts of RNA for binding or replication

studies, the reaction contained 40 mM Tris-HCl, 10 mM MgC12, 10 mM DTT,

2 mM spermidine, 0.5-1 mM each nucleoside triphosphate, 0.5-2.0 Mg

linear DNA template, and 60 units SP6 polymerase or 15 units T7

polymerase. Reactions were stopped by incubating with 1 unit RQ1TM


DNase I per pg template for 15 min at 37C, followed by phenol-

extraction and ethanol-precipitation. The RNA transcript was separated

from free NTPs by the spun-column method (Maniatis et al., 1982). When

needed for tracer, 10-20 pCi of an appropriate radiolabeled nucleoside

triphosphate were added to the reaction. The amount of RNA polymerase

which gives optimal incorporation was determined empirically.

Typically, about 20 pg of RNA was synthesized from 2 pg of template in

a two-hour reaction at 37*C. Synthetic RNA templates are referred to in

the text by the template used in their synthesis. For example, RNA

synthesized using an Eco RI digest of pOF1213 as template is referred

to as 1213 RNA.

Poliovirus RNA Polymerase RNA Synthesis Reaction

Unless specified, the conditions used for poliovirus RNA

polymerase synthesis reactions were those which yield the maximum

elongation rate in vitro (50 mM Hepes-NaOH, pH 8.0, 7 mM MgCl2, and 10

mM DTT) (Flanegan and Van Dyke, 1979). The reactions also contained

0.5-2.0 pg synthetic RNA template, 0.5 mM ATP, CTP and GTP, 5 pM cold

UTP, 10-50 pCi [a-32P]UTP, poliovirus RNA polymerase, and host factor.

Incubation was at 30C for the period indicated in the text. Labeled

product RNA was quantitated by precipitation of an aliquot of each

reaction in 7% trichloroacetic acid and 1% sodium pyrophosphate,

collection on membrane filters, and liquid scintillation counting.

Samples were phenol-extracted and ethanol-precipitated prior to

analysis of product RNA by gel electrophoresis. In general, large RNAs

were analyzed on agarose gels containing methylmercuric hydroxide,


while small RNAs were treated with glyoxal and fractionated on

polyacrylamide gels containing 7-10 M urea (Maniatis et al., 1982).

Poly(U) Isolation and Nearest Neighbor Analysis

Poly(U) was isolated from product RNA synthesized by fraction 4-HA

polymerase and phosphocellulose-purified host factor. Gel-purified

dimer-sized product RNA was digested for 3 hours with RNases TI, Ul,

and U2 in 10 mM Tris-HCl, pH 7.5, 1 mM EDTA, phenol-chloroform

extracted and ethanol-precipitated. Poly(U) was separated from shorter

oligomers by polyacrylamide gel electrophoresis, visualized by

autoradiography, and eluted from the gel for nearest neighbor analysis.

The eluted RNA was digested to mononucleotides with RNases TI, T2, and

A. The mononucleotides were resolved by high-voltage paper ionophoresis

on Whatman 3MM paper at pH 3.5 and analyzed by autoradiography as

previously described (Barrell, 1971; Rose, 1975). Spots corresponding

to each of the mononucleotides were cut from the paper and counted in 5

ml Aquasol-2 scintillation fluid (Ward et al., 1988).



Poliovirus-specific cDNA has been cloned and the sequence of the

entire genome determined (Racaniello and Baltimore, 1981a and 1981b;

van der Werf et al., 1981; Toyoda et al., 1984). Although a number of

poliovirus cDNA clones are directly infectious in susceptible cells

(Racaniello and Baltimore, 1981a; Semler et al., 1984; Kean et al.,

1986; Kuhn et al., 1987), their specific infectivity is several orders

of magnitude below that of purified poliovirion RNA. The development of

transcription vectors has alleviated this problem by providing a

method to synthesize genome-length RNAs in vitro using bacteriophage

RNA polymerase promoters (Melton et al., 1984). The sequence of

interest is cloned into a site adjacent to and downstream from a

bacteriophage RNA polymerase promoter. The clone is cleaved with a

restriction enzyme at the 3' side of the insert. Transcription

initiates 5' of the insert and results in synthesis of a run-off

transcript which includes the entire insert. A number of laboratories

have synthesized infectious RNAs of mammalian, plant, and insect

viruses from cDNA clones constructed in such vectors. These viruses

include poliovirus (Kaplan et al., 1985; van der Werf et al., 1986),

human rhinovirus 14 (Mizutani and Colonno, 1985), brome mosaic virus

(Ahlquist et al., 1984), tobacco mosaic virus (Dawson et al., 1986;



Meshi et al., 1986), cucumber mosaic virus satellite RNA (Masuta et

al., 1987), and black beetle virus (Dasmahapatra et al., 1986). We have

inserted full-length poliovirus cDNA into the transcription vector,

pGEM-1 (Promega-Biotec, Madison, Wi.), and used this clone to study the

replication of poliovirus in vitro (Oberste et al., 1988).

Construction of a Full-Length Poliovirus cDNA Clone

The poliovirus genome, in the form of cDNA, was cloned into the

Eco RI site of pGEM-1 using standard molecular cloning procedures, as

described in Materials and Methods (chapter 2). The sequence of the

pGEM-1 multiple cloning site (MCS) and adjacent phage RNA polymerase

promoters is shown in figure 3-1. Poliovirus cDNA derived from pEV104

(Semler et al., 1984) was ligated to Eco RI-digested pGEM-1 and the

ligated product was used to transform E. coli strain DH5a. Polio-

specific clones were identified among the ampicillin-resistant colonies

by colony hybridization (Grunstein and Hogness, 1975) using 32P-labeled

pEV104 insert as a probe. Clone pOF2612 (figure 3-2) contains the

entire polio sequence in what I have arbitrarily termed the positive

orientation (5' end of the polio positive strand adjacent to the SP6

promoter). Clone pOF2612 served as the starting point for the

construction of subgenomic clones.

Construction of Subgenomic Clones

A number of subgenomic clones, representing the entire genome,

have been constructed in pGEM-1 (table 3-1). The clones are designated

pOFl2xx, where 12 indicates their derivation from pOF2612 and xx is the

specific clone identifier. Odd-numbered clones contain the insert in

the positive orientation (same as pOF2612), whereas even-numbered


clones contain the insert in the negative orientation. Clones which

contain the same insert in opposite orientations are numbered

consecutively (eg., 1263 and 1264). These subgenomic clones were

constructed by one of two general methods, as described in chapter 2.

Clones containing only the 5' end or only internal sequences were

constructed by insertion of poliovirus-specific restriction fragments,

derived from pOF2612, into the pGEM-1 multiple cloning site (figure 3-

4). Clones containing both ends or only the 3' end of the genome were

constructed by deletion of internal or 5'-proximal sequences from

pOF2612 using one or more restriction enzymes and religation of

remaining sequences (figure 3-5; except pOF1215, see below). This

second strategy took advantage of the fact that only sequences which

contain both an origin of replication and a P-lactamase (ampicillin

resistance) gene will survive in bacteria grown on ampicillin-

containing media. For reference, an abbreviated restriction map of the

poliovirus genome is shown in figure 3-3. All clones which retain the

3' end also contain a poly(A) tract of 83 bases, inherited from pEV104.

For construction of internal clones, restriction fragments from

Bam HI, Hinc II, or Pst I digests of pOF2612 were inserted into pGEM-1

which had been digested with the same enzyme (figure 3-4). Insertion of

Bam HI fragments into pGEM-1 yielded three clones, 1261, 1263, and

1264. Clone 1261 contains polio bases 670 to 2105, while 1263 and 1264

both contain bases 2099 to 4605, but in opposite orientations. One Hinc

II clone was obtained, pOF1265, containing bases 5240 to 6775. The Pst

I digest resulted in three clones, 1267 and 1269/1270, containing bases


2243 to 3422, and 1 to 1814, respectively. Again, 1269 and 1270 have

the same insert in opposite orientations.

Clone pOF2612 was digested with Asu II at bases 866 and 6011 and

religated to yield pOF1203 (figure 3-5). This construct lacks

poliovirus nucleotides 867 to 6011, an in-frame deletion of 5145

nucleotides, but retains both ends of the genome and the translation

initiation site. A 4.6 kb internal deletion (dl 389-4911) was generated

by digestion with Nco I and religation to form pOF1207 (figure 3-5).

This deletion removes the translation start site at base 743, but

retains both ends as well as the entire P3 region. Clone pOFl201 was

created by digestion of pOF2612 with Kpn I at base 67 and Sty I at base

7209 (figure 3-5). The digested DNA was treated with Sl nuclease and

the Klenow fragment of E. coli DNA polymerase I in the presence of

dNTPs to produce blunt ends prior to ligation. This clone contains only

66 bases at the 5' end of the 5' noncoding region, 230 bases at the 3'

end of 3D, and the entire 3' noncoding region.

pOFl211 was constructed by cleaving 2612 at polio base 4911 (Nco

I) and in the MCS (Sal I) (figure 3-6). Blunt ends were generated by

filling in the overhangs with Klenow fragment and dNTPs prior to

ligation. Clone 1211 contains the same 3'-proximal bases as pOF1207.

The full-length clone was digested with Asu II (polio base 6011) and

Sal I (MCS), treated with Klenow fragment as above, and ligated to

produce pOF1209 (figure 3-6). This clone contains the 3'-proximal

portion of pOF1203. Clone pOF1205 was constructed by cutting with Hind

III in the MCS and at polio base 6515 (figure 3-6). The religated clone

contains the 3' 925 bases of the genome. Subgenomic clone pOF1213 was


made by digesting 2612 with Sty I (polio base 7209) and Sal I (MCS),

treatment with Klenow fragment, and religation. pOF1213 retains only

161 bases of 3D plus the 3' noncoding region (figure 3-6). Clone

pOF1215 was generated by inserting a poliovirus-specific Taq I-Eco RI

restriction fragment into pGEM-1. The 147 bp Taq I-Eco RI fragment of

pEV104 was gel-purified and ligated to pGEM-1 which had been digested

with Acc I and Eco RI (Taq I and Acc I generate compatible ends). This

clone contains only 62 nucleotides of the poliovirus heteropolymeric

sequence, beginning at nucleotide 7379 (figure 3-8).

Construction of Clones Whose In Vitro Transcripts Terminate in Poly(A)

The Eco RI site of pOFl211 was replaced with an Mlu I site to

facilitate synthesis of RNA which terminates in poly(A) (figure 3-7).

pOFl211 was cleaved with Eco RI (figure 3-7A1) and the overhanging

bases were removed by treatment with mung bean nuclease (figure 3-7A2).

Following phenol-extraction and ethanol-precipitation, the DNA was

treated with T4 DNA polymerase in the presence of dATP to remove the

3'-terminal G residue (figure 3-7A3). This was followed by a second

treatment with mung bean nuclease to remove the 5'-overhanging C

residue (figure 3-7A4). Finally, a synthetic Mlu I linker was inserted

and the DNA ligated to form pOFl211-Mlu (figure 3-7A5). When digested

with Eco RI and treated with mung bean nuclease, this clone directs

synthesis of an RNA which terminates in poly(A) (figure 3-7B).

pOFl211-Mlu was the starting point for the generation of deletions in

the 3' noncoding region.

Clones of the series p3NCxx were constructed by deletion of

sequences in the 3' noncoding region of pOFl211-Mlu using nuclease Bal


31 as described in Chapter 2. The xx in the clone designation denotes

the number of 3' noncoding region bases remaining, counted from the 3'

end. That is, base 1 is the G residue immediately 5' of the poly(A)

tract. Clones were constructed with progressively larger deletions

(figure 3-8), covering all potentially important structures in the 3'

noncoding region (figure 5-10). Clones p3NCOO contain a poly(A) tract

of 60 nucleotides, while the other p3NCxx clones have 83-nucleotide

poly(A) sequences.


Full-length and subgenomic poliovirus cDNA clones have been

constructed in a transcription vector. This allows us to easily

manipulate the sequence and structure of the poliovirus genome in DNA

form to study the molecular genetics of the virus on a fine scale not

possible using naturally occurring genome variants.

Clones which lack specific regions of the genome have been

constructed to enable us to determine whether defined regions of the

genome interact specifically with the polymerase in vitro and whether

particular sequences or structures are required for initiation of RNA

synthesis by the viral polymerase. Polymerase-RNA binding studies are

described in Chapter 4 and replication studies in Chapter 5.

SP6 Promoter

SP6 transcription start

I I Il iL I
Hind III Pst I Sal I Xba I
Acc I
Hinc II

T7 transcription start

Bam HI Sac I Eco RI
Sma I
Ava I

T7 Promoter


Figure 3-1. Sequence and restriction map of the multiple cloning site
(MCS) and bacteriophage RNA polymerase promoters in the cloning vector,
pGEM-1 (Promega Biotec). The sequence shown is the sense strand for SP6
RNA polymerase and the anti-sense strand for T7 RNA polymerase. SP6 and
T7 promoters are underlined and transcription initiation sites are
indicated by arrows.


Construction of pOF2612


pEV1 04

Eco RI Eco RI

Digest with Eco RI
Isolate 7.5 kb fragment

Eco RI
SP6 T7


Digest with Eco RI
Treat with Phosphatase



5' SP6 T7
Eco RI Eco RI

Figure 3-2. Construction of a transcription clone containing the entire
poliovirus genome. Thick lines indicate poliovirus-specific sequences
and thin lines indicate vector sequences. Location and transcription
direction of phage RNA polymerase promoters are shown as arrowheads.
Full-length clone pOF2612 was constructed as described in Chapter 2.

N 0
wd 0

(A -4 i
o4 *
-) 1) (U
m0 0

C~ .,-I4-
*> *0 0

o bo 0
0 0)

'-Cd u 4

*4 0 4J

Mt cc

4-) V) Cd


o 0) 4


04 bo
Cd ,4 41)

0 *U

. 4-I ,

0) 4i
c c

,CC 0 4
Wf o *

*- 0)CU

w (Dl
a> 5-i
0-) 0 ) f

O3 01
bo0 0Ur-
0r ZU 0)
r41 W Cd


- o



- o



- O

- o

- o


743 7370 7440
5' VPg---- Poliovirion RNA --poly(A) 3'


1261 --:




Figure 3-4. Construction of subgenomic clones containing sequences
which are internal to the poliovirus genome. Open bars indicate the
coding region and thin lines indicate the 5' and 3' noncoding regions.
Numbers above bars are nucleotide positions, relative to virion RNA.
Clones were constructed as described in Chapter 2.

Poliovirion RNA

7370 7440
IIoly(A) 3
i -poly(A) 3'

1203 -----E

1207 --


dl (389-4911)

dl (63-7213)


Figure 3-5. Construction of subgenomic clones which contain both ends
of the poliovirus genome. Open bars indicate the coding region and thin
lines indicate the 5' and 3' noncoding regions. Extent of deletion is
indicated in the center of each figure. Clones were constructed as
described in Chapter 2.

5' VPg-----


I |-poly(A)

1201 -

I 1-poly(A)

743 7370 7440
5' VPg----- Poliovirion RNA -poly(A) 3'

1211 -poly(A)

1209 | Z-poly(A)

1205 | -poly(A)

1213 E--poly(A)

Figure 3-6. Construction of subgenomic clones which contain the 3' end
of the poliovirus genome. Open bars indicate the coding region and thin
lines indicate the 5' and 3' noncoding regions. Numbers above bars are
nucleotide positions, relative to virion RNA. Clones were constructed
as described in Chapter 2.

Table 3-1. Clones containing subgenomic poliovirus cDNA sequences
inserted into pGEM-1.

type clone polio bases insert sizea

both ends 1201 1-62, 7214-7440 298
1203 1-866, 6012-7440 2294
1207 1-388, 4912-7440 2917

5' end 1269/70 1-1814 1814

3' endb 1205 6516-7440 925
1209 6012-7440 1429
1211 4912-7440 2529
1211-Mlu 4912-7440 2529
1213 7210-7440 231
1215 7379-7440 145

internal 1261 670-2105 1436
1263/64 2099-4605 2507
1265 5240-6775 1536
1267 2243-3422 1180

a Size of insert does not include the 83 base poly(A) tail in
clones containing the 3' end of the genome. Synthetic RNAs are longer
than the inserts due to vector sequences at either end of the
b Clones constructed by Bal 31 nuclease digestion are listed in
figure 3-8.


1. Eco RI



1. Mlu I


2. mung bean nuclease


2. mung bean nuclease



3. T4 DNA polymerase + dATP



3. transcription with SP6 RNA




4. mung bean nuclease



5. Mlu I linker: AAacgcgtTT


Figure 3-7. Construction of pOFl211-Mlu. Restriction enzyme recognition
sites are indicated by lower-case letters. (A) Removal of the Eco RI
site from pOFl211 and insertion of an Mlu I site. pOFl211 was digested
with Eco RI (Al) and the 3' overhang was removed with mung bean
nuclease (A2). The 3' G and 5' C residues were removed by treatment
with T4 DNA polymerase in the presence of dATP (A3) and a second mung
bean nuclease digestion (A4). An Mlu I was added by lighting a 10-bp
Mlu I linker to the plasmid (AS). (B) Synthesis of pOFl211-Mlu RNA
which terminates in poly(A). pOFl211-Mlu may be digested with Mlu I and
treated with mung bean nuclease to remove the overhang, leaving a
poly(A) tract and free OH at the 3' end.

0 0 0 0 0 0 0 0 i 41 0

0 0U 0
S P ,4 0 4 4)
rul rl ul rl ul u3 ( ,-- ,a 0 w
o a

110 0-0
S ,. -i L a4

,- 4) 0- .
O3 U U .) 0 04
0 o


r,.) m ,. in ) u
Uo U U 0 4-1 0-,

o Z 0
U U U 3U I) t W

0 0 4 C

Su ,0 ) 1- 0
0 0 0 0 M *M

,3 U U, r0 V
u 0 0) U)
u u u .-4 4 L$ )
U 0 0 0 0
0 P 0 l r-q
p,.4 bO 0

41) 0 0

0U 0 O) "U

0u U)
S4--1 ,C 0 4
S) 4-> U

u ) *0 0 u

0 L 4 J
u Zr> M r--4 :U
r. 0 a u
O (C 0 0 >

0 0 0 ) Z u

tko o

PC 0 op

co mn l- C4 C14 04 0 0 oo t0 C 0<
u U U u u U U u .. A C3 0<
z z z z N a & 0 41 "u u



Little is known about the physical interactions of eukaryotic RNA-

dependent RNA polymerases with their templates, but the binding of

bacteriophage QO replicase to QP and other RNAs has been studied in

great detail (reviewed by Blumenthal and Carmichael, 1979, and

Blumenthal, 1982). Q6 replicase binds to homologous RNAs at three

sites, two internal and one at the 3' end of the genome (Senear and

Steitz, 1976). In addition, the replicase does not bind to RNAs which

it cannot copy, suggesting that efficient enzyme-RNA binding is a

prerequisite for RNA synthesis. We have developed a rapid and sensitive

filter binding assay to measure the binding of poliovirus polymerase to

poliovirion and other RNAs.

Binding of Polymerase to Poliovirion RNA

In our initial experiments, we found that fraction 3 and fraction

4-HA polymerase preparations (Young et al., 1986) bound 32P-labeled

poliovirion RNA in the filter binding assay. To determine if the

observed binding activity was virus specific, a control extact

preparation (fraction 4-HA) was prepared from uninfected HeLa cells

using the same purification protocol. Unexpectedly, the control

extract preparation also bound virion RNA, although at a reduced level.

We used ribohomopolymers as competitive inhibitors to determine if



there were any differences in the specificities of the RNA binding

activities in the polymerase and control extract preparations.

Increasing amounts of unlabeled poly(A), poly(C), poly(G), or poly(U)

were added with 200 fmole of 32P-labeled virion RNA during the 60 min

binding reaction, and the resulting amount of protein-bound RNA was

determined using the filter binding assay. With the polymerase, the

relative order of inhibition observed was poly(G) >>> poly(U), poly(A)

> poly(C) (figure 4-1). With the control extract, however, poly(A) was

a very strong inhibitor, and the other ribohomopolymers had little or

no effect (data not shown). Thus, the cellular binding activity was

very specific for poly(A) and was distinct from the binding activity

observed with the polymerase.

Based on the above results, we reasoned that it should be possible

to separate the polymerase from a contaminating cellular poly(A)-

binding protein by chromatography on a poly(A)-Sepharose column. The

polymerase and control extract preparations (fraction 4-HA) were

chromatographed on separate columns. The peak of binding activity in

the control extract eluted from the column with 120 mM KC1 (figure 4-

2). In contrast, the peak of RNA binding activity in the polymerase

eluted from the column at 70 mM KC1 along with the peak of polymerase

elongation activity (figure 4-2). The trailing shoulder of RNA binding

activity in the polymerase appeared to represent the contaminating

cellular binding activity. The peak column fractions (i.e., fractions

8-10) which contained both polymerase elongation activity and RNA

binding activity and which did not overlap with the fractions

containing cellular RNA binding activity were pooled for further


studies. This added step of purification resulted in a viral

polymerase preparation (fraction 5-PAS) which exhibited only virus-

specific binding to poliovirion RNA.

Optimal Conditions for RNA Binding

The conditions for RNA binding by the purified polymerase

(fraction 5-PAS) were optimized with respect to time, temperature, pH,

KCI and MgC12 concentration. Under standard reaction conditions (see

Chapter 2), maximum binding to virion RNA was observed in less than 5

min. Thus, 60 min. binding reactions were adequate to reach binding

equilibrium. Changes in the temperature from 10-42*C and changes in pH

from 6.5 to 8.5 had little or no effect on polymerase binding.

Increasing the KCl concentration from 0-200 mM was found to inhibit

polymerase binding activity (figure 4-3). At 200 mM KC1, the binding

activity was reduced by about 50% (figure 4-3). This concentration of

KCI, however, will totally inhibit polymerase elongation activity on

virion RNA (Flanegan and Van Dyke, 1979). This suggests that the

inhibition of polymerase activity by added KC1 can be explained only in

part by decreased template binding. The requirement for Mg+2 in the

RNA binding assay was determined by increasing the MgC12 concentration

from 0-16 mM (figure 4-4). Significant binding activity was observed

in the absence of any added MgC12 with a clear optimum at 7 mM. This

concentration of Mg+2 also gave maximum elongation rates on virion RNA

in vitro (Van Dyke et al., 1978). Addition of EDTA to 1 mM in the

absence of added Mg+2 had no effect on binding (data not shown). The

addition of 10 mM dithiothreitol to a standard binding reaction had no

effect, and as expected the addition of 0.1% SDS blocked all binding.


Specificity of Polymerase Binding to Ribohomopolymers

The binding specificity of fraction 5-PAS polymerase to labeled

ribohomopolymers was determined using the filter binding assay.

Increasing amounts of each labeled ribohomopolymer were added to the

polymerase under the optimal binding conditions determined above. The

polymerase showed a high degree of specificity for binding poly(G)

(figure 4-5A). A reduced but significant amount of binding was also

observed with poly(U) and poly (C) (Fig 4-5B). No detectable binding

was observed with poly(A) at the concentrations tested (Fig 4-5B).

Thus, the overall order of binding was poly(G) >>> poly(U) > poly(C) >

poly(A). The lack of binding with poly(A) indicated that essentially

all of the contaminating cellular poly(A)-binding protein was removed

in the purification of the fraction 5-PAS polymerase.

Competitive Binding Curves Using Virion RNA and Ribohomopolymers

Ribohomopolymers were used as competitor RNAs for the binding of

purified polymerase to 32P-labeled virion RNA. Competitive binding

data were analyzed using equation (5) described by Lin and Riggs

(1972), in which the parameter 0 is defined as the ratio of the amount

of labeled nucleic acid bound to the filter in the presence of

competitor to the amount bound in the absence of competitor, i.e.,

cpm [32P]vRNA bound plus competitor RNA
e = (1)
cpm [32P]vRNA bound

This term is useful because of the automatic correction for any changes

in the retention efficiency or specific activity of the labeled RNA.

In addition, the concentration of competitor RNA which reduces 0 to 0.5


(i.e., C3), is inversely proportional to Ka, the apparent equilibrium

association constant for polymerase binding to competitor RNA.

Using poly(A), poly(C), poly(G), and poly(U) as competitor RNAs,

we determined 0 as a function of competitor RNA concentration in the

binding reaction (figure 4-6). Values of C were determined for each

competitor RNA at 6 = 0.5 in figure 4-6 (Table 1). A relative Ka for

each ribohomopolymer was calculated assuming a value of 1 for poly(G)

(Table 1). Overall, the relative Ka values varied as much as 200-fold

and were consistent with the relative order of binding observed with

the individual ribohomopolymers (figure 4-5).

Computer Scan of Poliovirus RNA for G- and/or U-Rich Sequences

Since the homopolymer binding and template activity data (see

Chapter 5) suggest a role for G- and/or U-rich sequences in poliovirus

replication, I have developed a computer program to search for locally

G- and/or U-rich regions in an RNA sequence. The complete Turbo Pascal

source code for "RNAScan" is listed in the Appendix. Briefly, the

program reads the input sequence from a file and assigns a score of 1

for each G or U residue and a score of 0 for each A or C. These scores

are then summed in a sliding window, the size of which is user-

determined. As currently implemented, RNAScan data is formatted for

output to a dedicated graphics program, but internal viewing and

printing procedures may be added in the future.

The 3' noncoding region of poliovirus type 1 (Mahoney strain) was

scanned for G+U content as a demonstration of RNAScan's potential

(figure 4-7). This portion of the poliovirus genome is relatively G+U-

rich at 60% G+U versus 47% for the genome as a whole.


Calculation of Ka for Polymerase Binding to Virion RNA

The absolute value of Ka for the binding of polymerase to

poliovirion RNA was calculated by the method of Riggs et al. (1970). If

one assumes a stoichiometry of one polymerase molecule per RNA

molecule, the binding of poliovirus polymerase to RNA may be described

at equilibrium as:

Pol + RNA = Pol-RNA (2)

[Pol-RNA] [Pol-RNA]
Ka (3)
[Pol (Pol.RNA)][RNA (Pol.RNA)] [Polf][RNAf]

where Pol is total polymerase, RNA is total RNA, and Pol-RNA is

polymerase-RNA complex. Polf and RNAf are free polymerase and free RNA,

respectively (equations 1 and 2, Riggs et al., 1970). At an RNA

concentration which gives one-half saturation of polymerase, equation

(3) may be rearranged to:

[RNA]; = Ka-1 + 1[Pol] (4)

where [RNA] is the RNA concentration which gives one-half saturation

of Pol (equation 3, Riggs et al., 1970). It can also be shown that at

one-half RNA saturation,

[Pol]0 Ka-1 + [RNA] (5)

where [Pol] is the concentration which gives one-half saturation of

RNA. When polymerase and RNA concentrations are known, Ka may be easily

calculated. The concentration of RNA is easily measured

spectrophotometrically or by other means, but the polymerase

concentration must be estimated. The estimate must be reasonably

accurate if the value of Ka is to be useful.


Polymerase concentration was estimated using method II of Riggs et

al. (1970). Binding of fraction 5-PAS polymerase was titrated against

1.33 x 10-12 M 32P-labeled virion RNA (figure 4-8A). If we assume that

all RNA is in polymerase-RNA complexes at plateau, we may derive a

"conversion factor" which measures the efficiency of the binding assay.

At plateau, 5.25 x 10-17 moles were retained on the filters but 1.33 x

10-16 moles were in the complex. This gives a conversion factor of 2.53

[(13.3 x 10-17)/(5.25 x 10-17)]. For the second part of the experiment,

[32P]vRNA was titrated against a fixed concentration of polymerase

(figure 4-8B). At plateau, all polymerase was saturated with RNA, ie.

Pol*RNA = Pol. In figure 4-8B, 3.4 x 10-14 moles of RNA were bound at

plateau. Therefore, [Pol] (3.4 x 10-14 moles) x 2.53 / (30 x 10-6

liters) = 2.8 x 10-9 M.

Using figure 4-8A and equation (5), Ka 1.7 x 109 M-1. This value

agrees closely with that derived from figure 4-8B and equation (4), 0.9

x 109 M-1. A similar value for Ka was derived from an RNA titration

experiment using 6.0 x 10-10 M polymerase (data not shown). Thus, the

Ka for polymerase binding to virion RNA is about 1 x 109 M-1.

Calculation of Ka for Polymerase Binding to Transcript RNAs

Once the polymerase concentration was determined, similar methods

were used to calculate Ka for binding of polymerase to labeled negative

strand RNA. Full-length negative strand was synthesized using T7 RNA

polymerase and Sac I-digested pOF2612 template as described in Chapter

2. Since the polymerase concentration was already known from the

experiments described above, it was most convenient to titrate RNA

against a fixed polymerase concentration (figure 4-9) and calculate Ka


using equation (4). There were two plateaus in the negative strand

titration experiment. At plateau 1, Kal 1.2 x 109 M-1 and at plateau

2, Ka2 9.5 x 107 M-1.

The association constants for binding of polymerase to subgenomic

poliovirus-specific transcripts were measured using the competitive

binding assay as described in Chapter 2. Since C, the concentration of

inhibitor which gives half-maximal binding, is proportional to Ka,

measurement of C for both the RNA in question and for virion RNA can

be used to derive Ka for that RNA once the Ka for virion RNA is known.

In this type of experiment, values for CQ are internally comparable

only. That is, C; for the test RNA can be compared to that of virion

RNA in the same experiment, but not to a C from another experiment.

Transcript RNAs were synthesized using [3H]ATP to allow accurate

quantitation of RNA concentration. Tritiated transcripts made from

clones 1205, 1209, 1213, and 3NC84, along with unlabeled poliovirion

RNA, were used as competitors in polymerase-[32P]vRNA binding reactions

(figure 4-10). Competitor RNA concentration was varied from 0.3 nM to

30 nM in the presence of 0.5 nM labeled virion RNA. The C for virion

RNA was 1.2 nM, while the values for 1205, 1209, 1213, and NC84 RNAs

were 1.5 nM, 2.8 nM, 2.0 nM, and 7.0 nM, respectively. Therefore, using

Ka(vRNA) 1 x 109 M-1, the Ka values for the transcripts are: 1205,

8.Ox 108 M-1; 1209, 4.2 x 108 M-1; 1213, 6.0 x 108 M-1, 3NC84, 1.7 x

108 M-1 (table 4-2). The values of Ka for 1205, 1209, and 1213 are

within a factor of 2-3 of that of virion RNA and this is probably

within experimental error. The value of Ka for NC84 is about six-fold

less than that for virion RNA and appears to be a significant


difference. Values of C; and Ka for all transcript RNAs are summarized

in table 4-2.

Competitive inhibition was used to assess the binding of

polymerase to RNAs from the 3' end deletion clones (p3NCxx). In this

experiment, C, for virion RNA was 1.2 nM while C; for the transcripts

varied from 2.1 (NC54) to >100 (NCOO) (figure 4-11 and table 4-2).

Since the experimental error is a factor of 2 to 3, C4 for NC54 is

considered to be the same as that for virion RNA and while C4 values

for the other RNAs are considered significantly different (table 4-2).

Polymerase binding to 18S ribosomal RNA and to rabbit globin mRNA was

also measured by competitive inhibition (figure 4-12 and table 4-2).

The Ka derived for 18S rRNA was 2.5 x 108 M"1 (C4 = 10 nM) and that for

globin mRNA was 1.7 x 108 M-1 (C0 15 nM). These values are within

experimental error of one another, but about five-fold less than that

of virion RNA.


We have developed a filter-binding assay for quantitating the

binding of purified poliovirus RNA polymerase to viral and nonviral

RNAs. The binding of labeled RNA to nitrocellulose filters was

dependent upon the presence of purified polymerase and was inhibited by

conditions which also inhibit RNA synthesis, such as high salt

concentration and ionic detergent. The optimum reaction conditions for

polymerase binding to virion RNA were similar to those which give

maximum elongation rates in vitro. RNA binding activity copurified

with the elongation activity of the polymerase, and only virus-specific


RNA binding activity was observed after the final step of purification

on a poly(A) Sepharose column (fraction 5-PAS).

We were surprised to find a cellular poly(A) binding protein

associated with the fraction 4-HA polymerase. We have not yet

determined if this protein directly interacts with the viral polymerase

or if the addition of this protein stimulates polymerase activity on

virion RNA in the oligo(U)-primed or host factor-dependent reactions.

This is an interesting possibility since the results of this study

showed that the polymerase had a very low binding affinity for poly(A).

Since negative strand RNA synthesis initiates at the 3'-end of the

poly(A) sequence in virion RNA, one can hypothesize that the polymerase

forms a complex with the poly(A) binding protein and that this mediates

its binding to virion RNA. Direct binding of the polymerase to a short

3' terminal oligo(U) primer synthesized by a terminal uridylyl

transferase host factor (Andrews, et al., 1985) is another


In the binding studies with ribohomopolymers, the polymerase

exhibited a high degree of specificity for binding poly(G) relative to

poly(C) and for binding poly(U) relative to poly(A). These results

were consistent with the data from competitive binding experiments with

labeled virion RNA and unlabeled ribohomopolymers. The value of the

association constant, Ka, for polymerase binding to poly(G) was 36

times the value for binding to poly(C). Likewise, the Ka for binding

poly(U) was 17 times the value for binding to poly(A). In previous

studies in this laboratory, where polymerase activity was measured

using ribohomopolymeric template:primers, there were high levels of


activity with poly(A):oligo(U) and poly(C):oligo(G), a five-fold lower

level of activity with poly(U):oligo(A) and essentially no activity

with poly(G):oligo(C) (Tuschall et al., 1982; Ward et al., 1988).

Taken together, these data suggest that preferential binding of the

polymerase to the primer correlates with high levels of RNA synthesis,

whereas tight binding to the template correlates with very low levels

of synthesis. A high binding affinity between the polymerase and the

template may slow the movement of the polymerase on the template and

inhibit RNA synthesis. A similar idea was previously proposed for AMV

reverse transcriptase based on a correlation between low levels of DNA

synthesis and high binding affinities for template:primers (Parnaik and

Das, 1983).

Based on our results with the ribohomopolymers, it is tempting to

speculate that polymerase binding sites in viral RNA may contain G-

and/or U-rich sequences. In a computer search of the viral genome, we

found several very G- and/or U-rich sequences (figure 4-7). Of

particular interest are sequences (about 75% G+U) which exist on one

side of the stem in hairpin structures which have been proposed to

exist near the 3' ends of both positive (figure 4-7; lizuka et al.,

1987) and negative strand RNA (data not shown; Racaniello and Meriam,


The association constants for the binding of polymerase to

poliovirion and negative strand RNAs were comparable to published KaS

for other RNA binding proteins, including the binding of phage R17 and

fr coat proteins to their homologous RNAs (Wu et al., 1988). The

finding that polymerase binds with equal affinity to both positive and

negative strands in vitro argues against the involvement of

differential binding in the regulation of strand selection in infected

cells. This is consistent with the compartmentalization model of Kuhn

and Wimmer (1988). Differential binding cannot be ruled out, however.

If the polymerase binds near the 3' end, the presence of poly(A) at the

3' end of the positive strand may cause polymerase to fall off the

template before reaching the 3' hydroxyl group. In addition, viral or

cellular factors not present in fraction 5-PAS polymerase may play an

important role in binding specificity in vivo, similar to the control

of OfQ replicase template specificity by E. coli host factor. Partially

purified HeLa cell host factor preparations do, in fact, bind to

poliovirion RNA (data not shown) but the impurity of the preparations

precludes assignment of RNA-binding activity to host factor itself. It

will be important to obtain highly purified host factor preparations to

see whether host factor influences the binding specificity of

poliovirus polymerase in vitro.

The presence of two plateaus in the negative strand titration

indicates the presence of two RNA binding sites of different affinity.

The functions) and genome locations of these sites are unknown. One

may speculate that the two sites interact in some way as in the Qfi

system, however there are no data to support this. Construction of

appropriate subgenomic clones and measurement of polymerase binding to

subgenomic negative strand RNAs should allow mapping of both binding

sites and perhaps the assignment of function.

Since polymerase binds with approximately equal efficiency to

virion RNA and to subgenomic RNAs (> 350 nucleotides long) which


contain the 3' end, a binding site(s) can be mapped near the 3' end of

the positive strand. However, interaction of polymerase with other

regions of the RNA cannot be ruled out as the native conformation of

virion RNA may bring 3' and 5' sites into close proximity, in a manner

similar to OQW RNA. In the QP system, polymerase binds to different

genome regions under different ionic conditions (Meyer et al., 1981).

It may be possible to further dissect poliovirus polymerase binding to

subgenomic RNAs by altering the reaction conditions.

The differential binding of RNAs which contain only a portion of

the 3' noncoding region (figure 4-11) suggests that structures in this

region are important for binding. In particular, the very low affinity

for NCOO RNA shows that heteropolymeric polyadenylated RNA is not

sufficient for binding, but that there is recognition of specific

sequences or structures in the 3' noncoding region. (RNA from 3NCOO

contains 11 vector-derived nucleotides with a poly(A) tail at the 3'

end). The Kas for the remaining RNAs vary within a factor of 3 to 4,

with NC84 and NC42 having the highest affinities at about 1 x 108 M-1

to 2 x 108 M-1. Transcript NC04 has GGAG plus a longer poly(A) tract

(by about 20 residues) than NCOO. The additional poly(A) sequence would

not be expected to change the binding, but the G-rich GGAG adjacent to

the poly(A) appears to be very important. This sequence is highly

conserved in human enteroviruses and it is interesting to speculate

that this is an important polymerase binding determinant. One could

envision a mechanism in which the polymerase first binds at the GGAG

site and then translocates to the 3' end by sliding along the RNA until

it reaches a 3' hydroxyl group. Such a mechanism would not require the


poly(A) tract to be a specific length and may explain why poly(A)

tracts of heterogeneous length are tolerated in vivo. The values for

the 3NC RNAs (except NCOO) are very similar to one another, but differ

significantly from that of virion RNA. This difference, and the fact

that longer 3'-terminal RNA transcripts bind with Kas within a factor

of two of virion RNA, suggests that sequence and/or structural

information immediately 5' to the 3' noncoding region is also involved

in polymerase binding. Perhaps the 3' noncoding region contains only a

portion of a larger recognition site which the larger clones retain.

Site-directed mutagenesis of the 3' noncoding region, particularly in

the GGAG sequence, could be used to map the specific sequences and/or

structures required for binding on a very fine scale.

The polymerase binds to 18S rRNA and to globin mRNA equally well

(Ka about 2 x 108 M-1), but with a five-fold lower affinity than to

virion RNA (Ka 1 x 109 M-1) (figure 4-12 and table 4-2). Ribosomal

RNAs are highly structured and interact with ribosomal proteins in the

assembly of mature ribosomes, so it is not surprising that another

protein which interacts with RNA (polymerase) also binds 18S rRNA. The

structure of globin mRNA is not well-known, but a preliminary computer

analysis of the sequence has revealed short sequences which have

homology to the poliovirus 3' noncoding region and which have the

potential to form secondary structures (data not shown). Polymerase

shows a relatively high affinity (Ka) for most RNAs, except for

poly(A), for which it has a very low affinity. This is interesting

since RNA synthesis must initiate at the end of a poly(A) tract. Virion

RNA is bound with the highest affinity. Negative strand and 3'


transcripts which were 350 nucleotides or longer were bound with about

the same affinity as was virion RNA. That the GGAG sequence appears to

be an important binding determinant is consistent with the homopolymer

binding studies, which predicted a G-rich binding site.

I--_I poly(A) A -
30- O poly(C) O D
0 poly(G) -
E25- poly(U) 0- 0

o n
< 15
J 5-

0 7 I I i
0 10 20 30 40 50
Competitor RNA concentration (upg/ml)

Figure 4-1. Inhibition of poliovirus polymerase (fraction 4-HA) binding
to poliovirion RNA by ribohomopolymers. The filter binding assay was
used to measure polymerase binding to 32P-labeled poliovirion RNA in the
presence of the indicated concentration of unlabeled ribohomopolymer.
The conditions were as described in Chapter 2 except MgC12 was not added
to the binding reaction.

35 35

1 30- 0 -30
0 0
S 25- 25 0 -

0 2 20- -20 Q o
c x 2.. x
E 15 A 0.2 15 l E
0 -o.2
z 10. -10.
> 50

A 0"/ A^-' \\ A
0 5 o ., L
^ 0 o V9 0
0 5 10 15 20 25 30
Fraction number

Figure 4-2. Purification of poliovirus polymerase on poly(A) Sepharose
column. Fraction 4-HA viral polymerase and mock polymerase were
isolated and chromatographed on separate poly(A) sepharose columns using
a linear KC1 gradient (dashed lines) as described in Chapter 2. The
separate elution profiles were superimposed for clarity. The viral
polymerase fractions were assayed for both elongation activity (open
squares) and binding activity (closed squares). The mock polymerase
fractions were assayed for binding activity (closed triangles) only.
The binding reactions contained P-labeled virion RNA as described in
Chapter 2. The activities recovered in the flow-through are indicated in
fraction 0.


E 60-



0 20-

0 50 100 150 200

[KCI] (mM)

Figure 4-3. Effect of KCI concentration on poliovirus polymerase
binding to virion RNA. The binding of purified polymerase (fraction 5-
PAS) to labeled virion RNA was measured using the filter binding assay.
KCI was added at the indicated concentration during the 60 min binding


25 -
o *


: 10>0\
S- ------

0 4 8 12 16
[MgCI2] (mM)

Figure 4-4. Effect of MgCl2 concentration on polymerase binding to
virion RNA. The filter binding assay was used to measure the binding of
fraction 5-PAS polymerase to labeled virion RNA with the indicated
concentration of MgC12 added during the 60 min binding reaction.

- 120 12
c 0-D poly(A)

S100- a-- poly(C) 10-
:- zA-A poly(G)
80 A-A poly(U)
80- 8-
o 60 6-

o 40 4-
^ 20 2-
N 10

0 1 2 3 4 5 0 1 2 3 4 5
[Ribohomopolymer] (pg/ml) [Ribohomopolymer] (Acg/ml)

Figure 4-5. Binding of polymerase to 32P-labeled ribohomopolymers. (A)-
The filter binding assay was used to determine the amount of labeled
ribohomopolymer bound to fraction 5-PAS polymerase in reactions
containing the indicated concentration of each ribohomopolymer. (B)-
Data for poly(U), poly(C), and poly(A) from panel A plotted on an
expanded scale.

1.0 -
poly(A) 03-0
SA\ poly(C) E- 0
-0 0.8- 0 poly(G) A- A
c A \ poly(U) A A

< 0.6- \
z \0
> -------------- ------\------------ ---------------A

S 0.4- A 0

C Al

0.0 .. ...... ...... ... ... ...
0.01 0.1 1 10 100
Competitor RNA concentration (/g/ml)

Figure 4-6. Inhibition of polymerase binding to virion RNA by
ribohomopolymeric competitor RNAs. The binding of fraction 5-PAS
polymerase to 5 fg/ml 32P-labeled virion RNA was measured in the
presence of the indicated concentration of each unlabeled
ribohomopolymer using the filter binding assay. The data are presented
as the fraction of labeled RNA bound using the term 0 as defined in
equation (1) in the text. A logarithmic scale was used for the
competitor RNA concentrations. The horizontal dashed line indicates 0 -
0.5. The values of C3 in Table 4-1 were determined at 0 0.5 for each

Table 4-1. Relative association

constants for polymerase binding to

Ribohomopolymer CG (pg/ml) a Relative Ka b

poly(G) 0.25 1.0

poly(U) 3.0 8.3 x 10-2

poly(C) 9.0 2.8 x 10-2

poly(A) 50.0 5 x 10-3

a Concentration of ribohomopolymeric competitor RNA required to reduce
the fraction of labeled virion RNA bound to 0.5 in figure 4-6.
b The relative association constants (Ka) are inversely proportional to
the values of C; and were calculated assuming a value of 1 for poly(G).




= 40-


0 -2. .
7370 7380 7390 7400 7410 7420 7430 7440
Nucleotide position

Figure 4-7. Computer analysis of G+U content in the 3' noncoding region
of poliovirus type 1 (Mahoney strain) by the program "RNAScan." G+U
content is expressed as a function of nucleotide position, relative to
virion RNA. Turbo Pascal source code is listed in the Appendix.

-10 50

0 8 40
cn =E

0 6 3_0 30_

4 20 -
0 Z

2 10

- O 0 l. I .I. 0 ... -. .

0 10 20 30 40 50 60 70 0 5 10 15
Polymerose added (Al) RNA Concentration (nM)

Figure 4-8. Determination of polymerase concentration and measurement of
Ka for binding of polymerase to poliovirion RNA. Binding reactions were
as described in Chapter 2. The dashed line indicates the plateau value.
The horizontal dotted lines indicate the half-plateau values, while the
vertical dotted lines indicate [Pol] in (A) and [RNA]; in (B). (A)
Titration of polymerase against a fixed virion RNA concentration.
Increasing amounts of polymerase were added to binding reactions
containing 1.33 x 10-12 M 32P-labeled vRNA. (B) Titration of RNA against
a fixed polymerase concentration. [32P]vRNA was added in increasing
amounts to 30 pl polymerase in a 100 pl reaction.


E 0
S20 -

o 15

< 10
q -
r) 5
0, L ---.--I----------

0 5 10 15 20

[RNA] (nM)

Figure 4-9. Titration of negative strand RNA against fixed polymerase
concentration. Concentration of labeled synthetic negative strand was
increased fixed fixed at of 1.4 x 10-10 M. RNA binding reactions were as
described in Chapter 2. The dashed lines indicates the plateau values.
The horizontal dotted lines indicates the half-plateau values, while the
vertical dotted lines indicate values for [RNA];.

A 1205
S 0.8 D 1209
:3 v 1213
o O NC84

0.6 -

N 0.4

0 .0 .... .... ....
0.1 1 10 100

Competitor concentration (nM)

Figure 4-10. Inhibition of polymerase binding to virion RNA by
poliovirus-specific subgenomic RNAs. The binding of fraction 5-PAS
polymerase to 32P-labeled virion RNA was measured in the presence of the
indicated concentration of unlabeled virion RNA or each unlabeled
subgenomic RNA using the filter binding assay. The polymerase
concentration was 4.2 x 10-10 M. The RNAs used were virion RNA (closed
circles), and transcripts from clones 1205 (open triangles), 1209 (open
squares), 1213 (open inverted triangles), and 3NC84 (open diamonds). The
data are presented as the fraction of labeled RNA bound using the term 0
as defined in equation (1) in the text. A logarithmic scale was used for
the competitor RNA concentrations. The horizontal dashed line indicates
0 = 0.5. The values of CG in Table 4-2 were determined at 0 = 0.5 for
each competitor.

1.0 1


< 0.6



0.1 1 10 100

Competitor concentration (nM)

Figure 4-11. Inhibition of polymerase binding to virion RNA by
poliovirus-specific subgenomic RNAs transcribed from 3NCxx clones. The
binding of fraction 5-PAS polymerase to 32P-labeled virion RNA was
measured in the presence of the indicated concentration of unlabeled
virion RNA or each unlabeled subgenomic RNA using the filter binding
assay. The polymerase concentration was 4.2 x 10-10 M. The RNAs used
were virion RNA (closed circles), and transcripts from clones 3NC84
(open triangles), 1215 (open squares), 3NC42 (open inverted triangles),
3NC26 (open diamonds), 3NC24 (closed triangles), 3NC04 (closed squares),
and 3NCOO (closed inverted triangles). The data are presented as the
fraction of labeled RNA bound using the term 0 as defined in equation
(1) in the text. A logarithmic scale was used for the competitor RNA
concentrations. The horizontal dashed line indicates 0 = 0.5. The
values of C in Table 4-2 were determined at 0 0.5 for each




> -- ---- -- -- ---------------

< 0.4



RNAs. The binding of fraction 5-PAS polymerase to 32P-labeled virion RNA
was measured in the presence of the indicated concentration of unlabeled
virion RNA, 18S rRNA, or rabbit globin mRNA using the filter binding
assay. The polymerase concentration was 4.2 x 10 10 M. The RNAs used
were virion RNA (closed circles), and transcripts from 18S rRNA (open
triangles), and globin mRNA (open squares). The data are presented as
the fraction of labeled RNA bound using the term 0 as defined in
equation (1) in the text. A logarithmic scale was used for the
competitor RNA concentrations. The horizontal dashed line indicates 0 =
0.5. The values of C; in Table 4-2 were determined at 0 0.5 for each

Table 4-2. Summary of C] and Ka values for polymerase binding to
poliovirus-specific RNA transcripts and to nonpolio RNAs.

RNA C1 (nM)a Ka (M-1)

virion RNA

negative strand












18S rRNA

1.0 x 109

1.2 x 109,
9.5 x 107

8.0 x 108

4.3 x 108

6.0 x 108

1.7 x 108

5.7 x 108

1.0 x 108

7.1 x 107

4.8 x 107

7.1 x 107


2.5 x 108

1.7 x 108

globin mRNA

a Summary of data from figures 4-8, 4-9, 4-10, 4-11, and 4-12. In the
experiments using transcript RNAs as competitors, C, for unlabeled
virion RNA was 1.2 nM. In the experiment with 18S rRNA and globin mRNA
as competitors, C for unlabeled virion RNA was 2.5 nM.



The template-priming model for poliovirus replication makes

several testable predictions about the product RNA synthesized by

polymerase using a positive strand template (Young et al., 1985;

Andrews et al., 1985). First, if synthesis is template-primed, the

product strand will be covalently linked to the 3' end of the template

strand. Thus, all labeled reaction products will be template-length or

greater. Second, correct initiation at the 3' end will result in

labeled product RNA which contains a poly(U) tract equal in length to

the poly(A) tract in the template strand.

As discussed in Chapter 1, there is genetic evidence that

sequences near the ends of the poliovirus genome are required for

replication (Racaniello and Meriam, 1986; Sarnow, et al., 1986). More

recently, Kaplan and Racaniello (1988) reported that subgenomic RNAs

which lack the capsid region, but include the 3' end of the positive

strand, are replicated efficiently when transfected into HeLa cells.

RNAs which lacked the 3' end were not replicated, supporting the

proposal that 3' end sequences are required for replication. The 3' end

of brome mosaic virus RNA 4 has been shown to be required for in vitro

initiation of replication by the viral replicase (Miller et al.,



The nonstructural proteins of a number of RNA viruses exhibit a

high degree of amino acid sequence homology to their counterparts in

picornaviruses, suggesting a common origin for these diverse virus

families (Argos et al, 1984; Kamer and Argos, 1984; Domier et al.,

1987). The 3' ends of the genomic RNAs of several plant viruses contain

sequences which can be folded into three-dimensional structures which

strongly resemble the structure of tRNAs (Rietveld et al., 1983).

Viruses with this tRNA-like structure include the tymoviruses, the

tobamoviruses, the bromoviruses, and the cucumoviruses. These

structures are highly conserved among related viruses, even though the

sequences vary widely. A key feature of these structures is a tertiary

structure known as a pseudoknot. RNA pseudoknots have been shown to

exist in solution and are thought to stabilize long-distance base-

pairing, gaining energy by the coaxial stacking of adjacent helices

(Pleij et al., 1985; Puglisi et al., 1988). Weiner and Maizels (1987)

have proposed that tRNA-like structures such as these are genomicc

tags" which act as recognition signals for viral replicases and may be

remnants of ancient RNA genomes. The region of the BMV genome found by

Miller et al. (1986) to be required for in vitro initiation by the

viral replicase corresponds precisely to the BMV tRNA-like structure,

lending support to the genomic tag replication model.

Product RNA Is Covalently Linked to the Template

RNA was transcribed from clones 1209 and 1205 for use as template

in in vitro replication reactions (figure 5-1). In the presence of HeLa

cell host factor, poliovirus polymerase produced dimer-size products

using both synthetic templates (figure 5-2). RNA transcribed from clone


1211 was also copied efficiently into a dimer-size product (data not

shown). RNAs 1211, 1209, and 1205 contain 2612, 1512, and 1008

nucleotides of poliovirus sequence, respectively, including the poly(A)

tract (table 3-1). In all cases, the major reaction product was twice-

template length.

Clone 1213 RNA, corresponding to the 3' 314 nucleotides of the

genome, including poly(A) (figure 5-1), was efficiently copied by the

polymerase in the presence of either host factor or oligo(U) (figure 5-

3). The polymerase synthesized a product which was predominantly dimer-

sized in the presence of host factor (figure 5-3, lanes Al and Bl). In

the oligo(U)-primed reaction, the major product was monomer length

(figure 5-3, lanes A2 and B2). Dimer-sized product synthesized in the

presence of oligo(U) was probably due to contaminating host factor in

the polymerase. While clone 1213 product RNA could be analyzed on a 2%

agarose-CH3HgOH gel (figure 5-3A, lanes 1 and 2), the products were

better resolved if treated with glyoxal (Maniatis et al., 1982) and

fractionated on a 5% polyacrylamide-7 M urea gel (figure 5-3B, lanes 1

and 2). Product RNAs migrated aberrantly on polyacrylamide-urea gels if

the samples were not glyoxylated prior to loading, probably due to

incomplete denaturation (figure 5-3C, lanes 1 and 2).

To definitively determine whether RNA synthesis by the polymerase

is template-primed, clone 1213 RNA was used as template in a time

course experiment (figure 5-4). Suboptimal reaction conditions (pH 7.0,

3 mM MgC12) were used to slow the elongation rate to allow detection of

intermediate-length products, as previously described with poliovirion

RNA template (Van Dyke et al., 1982). Product RNAs synthesized during a


five-minute time course experiment were glyoxalated and run on a 5%

polyacrylamide-7 M urea gel (figure 5-4). At the earliest time points,

labeled RNA is template-size or slightly larger (figure 5-4, lanes 1-3,

8-10). As the reaction progressed, the products increased in size,

reaching twice-template length after about four minutes (figure 5-4,

lane 6 and 13). At no time was labeled product less than template

length. The elongation rate was determined by calculating the length of

the largest product at each time point as previously described (Van

Dyke et al., 1982). Under the conditions of this experiment, the

elongation rate was approximately 90 nucleotides per min. This was very

close to the value of 83 nucleotides per min determined by Van Dyke et

al. (1982) using a poliovirion RNA template under similar conditions.

Labeled template RNA was used in a chase experiment to determine

whether polymerase and host factor could chase a monomer template into

a dimer-sized molecule. Uniformly labeled template RNAs were

transcribed from clones 1205 and 1209 by spiking the transcription

reactions with 10 pCi [32P]UTP. Under standard reaction conditions,

these templates were chased into dimer-sized products by polymerase and

host factor in the presence of unlabeled ribonucleoside triphosphates

(figure 5-5).

Product RNA Contains Poly(U)

To assign physiological relevance to the above results, it was

important to demonstrate that initiation occurred correctly at the 3'

end of the template, rather than at an internal hairpin as was

suggested by Hey et al. (1987). Because the clones contain a poly(A)

tract of 83 bases, all products of correct initiation should contain a


poly(U) tract of approximately the same length. Product RNA was

synthesized using 1213 RNA template in a polymerase-host factor

reaction and fractionated on a 1.5% agarose-CH3HgOH gel (figure 5-6A,

lane 2). The dimer-sized product eluted from the gel was digested with

RNases Tl, Ul, and U2 and fractionated on a polyacrylamide gel in the

presence of 7 M urea (figure 5-6B). RNase-resistant material of 80-90

residues was eluted from the gel and subjected to nearest neighbor

analysis (figure 5-6C). The 80-90 nucleotide material was 94% UMP as

measured by scintillation counting of the spots cut from the high

voltage ionophoresis paper, indicating the presence of an 80-90

nucleotide poly(U) tract in the product RNA. Material of greater than

90 nucleotides was the product of partial digestion, as further

digestion reduced its size to 90 nucleotides or less (data not shown).

This material was also composed of essentially all UMP (data not


Template Activity of RNAs Which Lack the 3' End of the Genome

To determine whether sequences near the 3' end of the genome were

required for template activity, RNAs which lacked defined portions of

the 3' terminus were used as templates in polymerase-host factor

reactions (figure 5-7). First, the template activity of clone 1265 RNA

(figure 3-4) was compared to that of 1209 RNA. These two RNAs were

chosen because they are both about the same size (-1.5 kb) to minimize

any differences due to differences in the molar concentrations of the

templates. The RNAs were transcribed from the respective clones in the

presence of [3H]UTP to allow accurate quantitation of the template

concentration. Two micrograms of each RNA was used as template in a


standard polymerase-host factor reaction with [32P]UTP and the products

were fractionated on a 1.2% agarose-CH3HgOH gel (figure 5-7). The

templates remained intact, as judged by staining with ethidium bromide

(figure 5-7A). The autoradiogram showed that 1209 RNA was copied

efficiently (figure 5-7, lane 2), as previously determined (figure 5-

2), but that the 1265 RNA was copied much less efficiently (figure 5-7,

lane 1). Other templates which lacked the 3' end were also copied less

efficiently than 1209 RNA (data not shown). The template efficiency was

quantitated and defined as the ratio of fmoles of product RNA

synthesized to pmoles of template used (table 5-1). The yield of

product RNA was calculated measuring the acid-precipitable [32P]UMP

incorporated during the reaction and acid-precipitable [3H]template

remaining after the reaction. Since the sequence of the RNAs is known,

one can calculate the molar amount of both template and product. Clone

1209 RNA was copied most efficiently and assigned a relative efficiency

of 1 (table 5-1). Other RNAs derived from clone 1209 which contained

nucleotides 6012-7338, 6012-7213, and 6012-7116 were copied about two-

fold less efficiently (table 5-1). For the experiment shown in figure

5-7, 1265 RNA was copied about five-fold less efficiently than 1209

RNA. In other experiments with 1265 template, the range was two- to

five-fold, for an average relative efficiency of about 0.35 (table 5-


Structural Features at the 3' End of the Poliovirus Genome

The nucleotide sequence of the 3' noncoding region of poliovirus

type 1 (Mahoney strain) (figure 5-8A) has the potential to form stable

stem-loop structures (figure 5-8B; lizuka, et al., 1987). Although the


sequences in these regions vary considerably between different

picornaviruses, the structures are highly conserved. These sequences

may also be folded into more complex three-dimensional structures which

contain pseudoknots (figure 5-8C). Molecular modelling studies suggest

that this proposed three-dimensional structure may be stable (J.B.

Flanegan and M.S. Oberste, unpublished results). The tertiary folding

of the 3' end of the poliovirus positive strand strongly resembles that

of the 3' end of BMV RNA 3, but lacks arms C and D (figure 5-8D;

Rietveld et al., 1983). Phylogenetic data support this structural model

as the 3' ends of Coxsackieviruses Bl, B3, and B4 can all be folded

into structures similar to that of poliovirus (data not shown).

Involvement of 3'-Terminal Sequences in Initiation of RNA Synthesis

RNA transcripts were synthesized from p3NCxx clones (figure 3-7)

using [3H]ATP to allow accurate quantitation of RNA concentration.

These transcripts were used as templates to determine whether specific

sequences in the 3' noncoding region are required for initiation of RNA

synthesis. Since the amount of template used in these experiments was

too little to detect by staining (about 1 pmole each), the condition of

the transcripts was checked separately (figure 5-9B). Two micrograms of

each transcript was fractionated on an 8% polyacrylamide gel and

stained with ethidium bromide. All RNAs appeared intact and the correct

size (figure 5-9B). Product RNAs synthesized in standard reactions were

treated with glyoxal and fractionated on a 5% polyacrylamide-7 M urea

gel (figure 5-9A). The yield of product RNA was calculated measuring

the acid-precipitable [32P]UMP incorporated during the reaction and

acid-precipitable [3H]template remaining after the reaction. Since the

sequence of all the RNAs is known, one can calculate the molar amount

of both template and product. Template efficiency was expressed as

pmoles of product formed per pmole template (figure 5-10). RNA

transcripts from clones NC84, NC54, NC42, and NC26 were copied

efficiently into dimer-size products (figure 5-9A, lanes 1-4,

respectively; figure 5-10). Clone NC24 RNA lacks two G residues which

are present at the 5' end of the poliovirus sequence in clone NC26 RNA

(figure 3-7) and was a somewhat less efficient template (figure 5-9A,

lane 5; figure 5-10). In other experiments the difference between NC26

and NC24 was much less dramatic (data not shown). Clone NC04 and NCOO

RNAs, containing four 3' noncoding region nucleotides plus poly(A) and

poly(A) alone, respectively (figure 3-7), were very poor templates

(figure 5-9A, lanes 6 and 7; figure 5-10).


With all transcripts which contained an intact 3' noncoding

region, the polymerase-host factor reaction yielded a product which was

covalently linked to the template and twice-template length (figures 5-

2, 5-3, 5-7 and 5-9). This was consistent with previous results using

poliovirion RNA as template, but the smaller size of the RNA

transcripts allowed for the synthesis of a more homogeneous dimer-sized

product, better resolution of the product from the template RNA, and

more accurate measurement of the length of the product. In the presence

of the host factor used in this study, the polymerase synthesized

product RNA which increased in size from template length to twice-

template length (figures 5-4 and 5-5.) The presence of a poly(U) tract

in the product which was the same size as the poly(A) tract in the


template strand demonstrated that the polymerase initiated RNA

synthesis precisely at the 3' end of the template. These data provided

the clearest evidence yet support template-priming as a possible

mechanism for the initiation of poliovirus negative strand RNA

synthesis (figure 5-11). This is also supported by recent evidence

which indicates that the terminal protein, VPg, is active in a self-

catalyzed reaction which cleaves the dimer-size product RNA at the

poly(A)-poly(U) hairpin and results in linkage of VPg to the negative

strand (Tobin, 1988).

Experiments using NCxx RNAs as templates showed that specific

sequences in addition to a poly(A) tail were involved in the initiation

of RNA synthesis, supporting the template specificity hypothesis and

suggesting that a polymerase recognition element exists near the 3' end

of the genome. The requirement for specific sequences was not, however,

absolute. The sequences involved in initiation are summarized in the

context of the proposed 3' end structure in figure 5-12. Initiation of

RNA synthesis occurs efficiently as long as the 3' 24-26 nucleotides of

the noncoding region are present 5' to the poly(A) tract (figure 5-9A).

If the upstream sequence is reduced to four nucleotides or fewer,

synthesis is barely detectable (figure 5-9A, lanes 6 and 7; figure 5-

10). These results are most easily interpreted in terms of the proposed

structure (figure 5-12). As long as the UAAAU loop at the end of arm B

is maintained, RNA synthesis initiates efficiently (figure 5-9A, lanes

1-4; figure 5-10). If this region is deleted, synthesis is decreased

sharply. The sequence in the loop at the end of the stem is highly

conserved among picornaviruses, with the consensus among the human


viruses being UAAAU (Palmenberg, 1988). A similar conserved sequence,

UA(A/U)AA, is present about the same distance from the poly(A) tract in

EMCV, Mengovirus, and TMEV. At least two strains of hepatitis A virus,

the most isolated of the picornaviruses, also contain UAAAA in the loop

of a putative stem-loop structure.

743 7370 7440

5' VPg---- Poliovirion RNA -poly(A) 3'

1211 -poly(A)

1209 L-poly(A)

1205 1 -poly(A)

1213 [}-poly(A)

Figure 5-1. Map of subgenomic clones which contain the 3' end of the
poliovirus genome. Open bars indicate the coding region and thin lines
indicate the 5' and 3' noncoding regions. Numbers above bars are
nucleotide positions, relative to virion RNA.

ori -





1 2

Figure 5-2. Poliovirus-specific transcript RNAs are copied by the viral
polymerase in vitro. RNAs were transcribed from clones 1209 and 1205
and used as templates in polymerase-host factor reactions (lanes 2 and
4, respectively). 32P-labeled RNAs were transcribed for use as size
markers (lanes 1 and 3). Reaction products were phenol-extracted and
fractionated on a 1.2% agarose-CH3HgOH gel. Origin of migration (ori)
and sizes of labeled RNAs, in kilobases, are shown at left.

3 4


on- 1 2 1 2 1 2

404- 527-

309- 404-


Figure 5-3. Product RNA is covalently linked to the template. Clone
1213 RNA was copied in a standard polymerase reaction (Chapter 2), in
the presence of host factor (lanes Al, Bl, and Cl) or oligo(U) (lanes
A2, B2, and C2). The product was phenol-extracted, split into three
portions, and fractionated using three different gel systems: lanes Al
and A2, 2% agarose-CH3HgOH gel; lanes Bl and B2, glyoxalated, 5%
polyacrylamide-7 M urea gel; lanes Cl and C2, untreated, 5%
polyacrylamide-7 M urea gel. Origin of migration (ori) and sizes of
labeled RNAs, in nucleotides, are shown at left.

1 2 3 4 5 6

on -

527 -
404 -



Figure 5-4. Polymerase-host factor product RNA is template length or
larger. Clone 1213 RNA was incubated in a standard polymerase and host
factor reaction. Aliquots were removed at 0.5, 1.0, 1.5, 2.0, 3.0,
4.0, and 5.0 min (lanes 1-7, respectively; lanes 8-14 are a longer
exposure of lanes 1-7), phenol-extracted, glyoxalated, and fractionated
on a 5% polyacrylamide-7 M urea gel. Origin of migration (ori) and
sizes, in nucleotides, of DNA size markers are shown at left.

7 8 9 10 11

12 13 14
luMl ,

1 2 3 4
onri -




Figure 5-5. Subgenomic RNA transcripts are elongated into product RNA
twice the size of the template. 32P-labeled RNA was transcribed from
clones pOF1205 and pOF1209. The transcription reaction was terminated
with DNase I and 150 units CIP for 20 min at 37C, phenol-extracted,
and ethanol-precipitated. Transcripts were copied in a reaction
containing poliovirus polymerase, host factor, and 500 pM each
ribonucleoside triphosphate. Product RNAs were phenol-extracted and
electrophoresed through a 1% agarose-CH3HgOH gel. Lanes 1 and 3, 1205
template and product; lanes 2 and 4, 1209 template and product. Lanes 3
and 4 are a longer exposure of lanes 1 and 2. Origin of migration (ori)
and sizes of the RNAs, in kilobases, are indicated at left.





1 2










Figure 5-6. Isolation of poly(U) from poliovirus polymerase product
RNA, using clone 1213 RNA as template. (A) Autoradiogram of product
RNA fractionated by 1.5% agarose-CH3HgOH gel electrophoresis. Markers
indicate size in bases. Lane 1: 32P-labeled clone 1213 RNA (template-
size marker); lane 2: Clone 1213 product RNA. (B) RNase digestion and
fractionation of excised product RNA from gel (A). Material greater
than 90 bases long is the result of incomplete digestion. (C)- Nearest
neighbor analysis of 80-90 base material from gel (B). Migration is
from bottom to top.

12 1 2
on -



Figure 5-7. A transcript RNA which lacks the 3' end of the genome is
replicated inefficiently. Product RNA was fractionated on a 1.2 %
agarose-CH3HgOH gel. Lane 1, 1265 RNA; lane 2, 1209 RNA. Origin of
migration (ori) and sizes of template and product RNAs, in kilobases,
are indicated. (A) Ethidium bromide stain of the gel to determine
whether template RNA remained intact. (B) Autoradiogram of product RNA.

Table 5-1. Relative template efficiencies for transcript RNAs which
lack the 3' end of the genome.

Transcript RNAa


1209/Sca I

1209/Sty I

1209/Xmn I








fmole product per
pmole template











a Template RNAs were
polymerase. Unless a
contained the entire

transcribed from the indicated clone with SP6 RNA
restriction digest is indicated, the transcript
insert sequence.



U-G A.

(C) arm A (D) arm A
11111 11U1i1l i l 111111 I IIII 1ill ll
CUG 5' C arm D UAC GCGG / CC C U
GUG/A/5 \ ,G G / GU C

GU/G U/ CG\\U G C /A
U / U arm B UC\\C UG UA/C/ A
G A AG arm C U C m
UA UAG arm A

Figure 5-8. Structural features at the ends of the poliovirus genome.
(A) Sequence of the 3' end of the positive strand (3' noncoding region)
of poliovirus type 1 (Mahoney strain) (Koch and Koch, 1985). (B)
Putative stem-loop at the 3' end of the positive strand predicted by
PCFold (Zuker and Stiegler, 1981). (C) Proposed tertiary folding of the
3' end of the positive strand (J.B. Flanegan and M.S. Oberste,
unpublished results). (D) Tertiary folding of the 3' end of brome
mosaic virus RNA 3 (after Miller et al., 1986). The 3' end sequences
and structures are virtually identical in all four BMV RNAs.

ori -

404 -

309 -

242 -

217 -
201 -
190 -

180 -

160 -

122 -

90 -

67 -

1 2 3 4 5 6 7

ori -



Figure 5-9. Sequences in the 3' noncoding region are required for
initiation of RNA synthesis. (A) 3H-labeled RNA was transcribed from 3'
end deletion clones, NC84, NC54, NC42, NC26, NC24, NC04, and NCOO
(figure 3-8) and used as template in polymerase-host factor reactions.
Products were phenol-extracted, glyoxylated, and fractionated on a 5%
polyacrylamide-7 M urea gel. (B) Integrity of input template was
checked by fractionating 2 pg each RNA transcript on an 8%
polyacrylamide-7 M urea gel and visualizing the RNA by ethidium bromide
staining. Origin of migration (ori) and positions of DNA size markers
are indicated at left.

1 23 4 567

x 5.0

E 4.0

3 3.0



0 0.0
SNC84 NC42 NC26 NC24 NC04 NCOO

Template RNA

Figure 5-10. Quantitation of template efficiency of 3NCxx RNA
transcripts. The templates and products from the experiment in figure 5-
9 were quantitated from the acid-precipitable [3H]- and [32P]RNA present
at the end of the reaction. Template efficiency is expressed as pmoles
[32P]product per pmole [3H]template and presented as a histogram. Clone
NC54 template was omitted from the analysis since its vector-derived
sequences are different from those of the other clones.


SP6 promoter


pOF1213 DNA
linearized with Eco RI

SP6 RNA polymerase

- (A)83

poliovirus-specific RNA
(349 nucleotides)

Polymerase & Host Factor

(-700 nucleotides)

3' GCCUC(U) ~80

(~85 nucleotides)

Figure 5-11. Template-priming model for poliovirus replication. Eco RI-
linearized clone 1213 DNA is transcribed into RNA using SP6 RNA
polymerase. The transcript RNA contains 35 vector-derived nucleotides,
231 nucleotides from the 3DPol and 3' noncoding regions, and a poly(A)
tract of 83 nucleotides. In the presence of host factor, poliovirus
polymerase copies the 1213 template into a dimer-size, base-paired
product with a hairpin at the positive strand-negative strand junction.
Digestion of this product with RNases T1, Ul, and U2 yields a poly(U)
sequence of approximately 80 nucleotides with five nucleotides derived
from the 3' noncoding region (negative strand) at the 3' end.

II U)80

RNases T1, U 1, U2

arm A

NC42 ---u

uacuggguu agguuaagc NC54
a uaccucag UUUAAUUCGGAG-poly(A)-3'
I c I I I
u a 5' INC4 NCOO
u //

NC26 -aG/U U arm B

Figure 5-12. Terminal sequences required for recognition of template by
polymerase, in the context of the proposed tertiary folding as in
figure 5-8. Sequences which are required for initiation of in vitro RNA
synthesis are in upper case. Dispensable sequences are lower case.
Starting positions of Bal 31 deletion clones (figure 3-8) are
indicated. Clone NC84 (not shown) contains the entire 3' noncoding
sequence plus 13 nucleotides from the 3' end of the 3DPol region.

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E6F3KO97Y_HT1DOE INGEST_TIME 2012-02-20T21:38:21Z PACKAGE AA00009087_00001