Note: Descriptions are shown in the official language in which they were submitted.
~09'/13~711 PCT/~S~'tll~6U~
2~ 03~51
DESCRIPrION
Requlation of Nucleic Acid ~ranslation
ackqround of the Invent on
The invention relates to the regulation of RNA
translation.
Antisense RNA is RNA whose sequence is complementary
to that of a particular RNA molecule (see, e.q., Kimelman
et al., Cell 59:687, 1989; Melton, ~ tl. Acad. Sci.
USA, 82:144, 1985). In vlvo, antisense RNA corresponding
to a particular gene is usually produced ~y an artificial
gene which has been engineered to transcribe the normally
untranscri~ed strand of the chosen gene. Such an engin-
eered gene is easily generated by reversing the orienta-
tion of the transcribed DNA in the normal gene.
Antisense RNA blocks the production of the polypep-
tide encoded by its complementary sense RNA. This inhibi-
tion of translation is thought to occur because an RNA-
RNA duplex is formed which cannot be translated. Anti-
sense RNA has been used to control production of manganous
superoxide dismutase in human embryonic kidney cells
(Wong, et al., Cell 5~:923, 1989), amyloid ~ protein pre-
cursor in human fibroblasts (Saitoh, et al., Cell 58:615,1989), and ribulose biphosphate carboxylase in tobacco
plants (Rodermal et al., Cell 55:673). In Xeno~us oocytes
antisense RNA îs thought to cause modification of the RNA
molecules to which it hy~ridizes and this modification is
thought to cause rapid degradation of the RNA (Kimelman
et al., suDra). Antisense RNA and endogenous RNAse H have
been used to block cyclin production in Xenopus oocytes
cell extracts (Minshull et al., Cell 56:947, 1989).
Summary of the InventiQn
The invention features an RNA molecule, termed a
responsive RNA molecule which, when present in a cell,
responds to the presence of other nucleic acids. By
3STITUTE ~iHE~ T
~09~tl3~7ll PcT/~s9~t(~ub
2 1 '~ r~
"responds" is meant that the responsive RNA molecule will
be translated to form one or more polypeptides in the
presence of certain nucleic acids (which can hybridize to
the responsive RNA) and will not be significantly trans-
S lated to form these polvpeptides in the absence of suchnucleic acids. Such a responsive RNA molecule will gener-
ally encode one or more polypeptide molecules, the pro-
duction of which depends on translation of that responsive
RNA molecule. Generally, translation of the responsive
RNA molecule, and thus production of polypeptide, will not
occur in any particular cell unless a specific nucleic
acid, termed a signal nucleic acid, is also present within
that cell.
A responsive RNA can be used to kill or lnjure spe-
cific cells within a population of cells. For example, aresponsive RNA may encode a toxin molecule which is pro-
duced from the responsive RNA only when the responsive RNA
molecule within a given cell is exposed to a signal
nucleic acid indicative of a condition (e.q., infection
with a harmful virus such as HIV-I) requiring that the
cell be killed. More specifically, the responsive RNA
molecule may encode a cytotoxic protein such as cholera
toxin, diptheria toxin, ricin and the hok, gef, RelF or
flm gene products of E. coli, and translation of the
responsive RNA molacule and production of cytotoxic
protein occurs only when the respo~sive RNA molecule is
present within a cell which is infected with HIV-I. Here,
an RNA molecule specific to HIV-I serves as the signal
nucleic acid and interacts with the responsive RNA mole-
cule to allow translation of the toxin-encoding sequences
of the responsive RNA molecule.
A responsive RNA molecule is produced by designing a
polypeptide-encoding RNA which, in the absence of a signal
nucleic acid, has a structure which prevents translation.
One type of responsive RNA molecule can fold to form a
base-paired domain, e.~., which, when sufficiently stable,
prevents translation by preventing the translational
-
SUIE3STITUTE SHEET
~O9~/130711 PCT/-S9'/~6Ul
machinery of a cell from reading the nucleotide sequence
of the RNA. A specific example of a responsive RNA mole-
cule of this type has a domain which encodes the desired
polypeptide (or "protein-coding region") and a regulatory
domain (i.e., a domain which includes regulatory elements
including an inhibitor region, inverted repe~ts and nucle-
ation regions). The regul3tory domain may be located
anywhere in the responsive RNA molecule so long as the
sequence of the elements of the regulatory domain are
selected so as not to interfere with the activity of the
coded polypeptide. The inhibitor region is complementary
in sequence to both a substrate region (which can include
portions of either the protein-coding region and/or a
leader region which is the non-translated ~NA 5' of the
protein-coding region) and to a region of the signal
nucleic acid referred to as an anti-inhibitor region. In
the absence of the signal nucleic acid, the inhibitory
region of the responsive RNA molecule hybridizes to the
substrate region of responsive RNA molecule forming an
intramolecular base-paired domain which prevents or
reduces translation. When the signal nucleic acid is
present, the anti-inhibitor region competes with the
substrate region for binding to the inhibitor region.
Formation of an intermolecular base-paired domain between
the anti-inhibitor region of the signal nucleic acid and
the inhibitor region of the responsive RNA prevents for-
mation of a base-paired region with the protein-coding
region; under these circumstances the protein-coding
region(s) can be translated.
A second type of responsive RNA molecule has an
intervening sequence or "intron", the presence of which
prevents translation of one or more "exons". Introns do
not code for the desired polypeptides. Segments of the
RNA which code for desired polypeptides are called "exons"
as are non-coding sequences (~ g~, the leader region,
secretory signal sequences, polyA tails, and the like)
that remain after the splicing reaction. This second type
~5l.1BSTlTUTE SHEFT
'/130,~) PC-r/L'S9'/~)~J61~.`
21~%~1 .
of responsive RNA molecule is designed so that it can
undergo a splicing reaction under desired conditions
(e.q., in the presence of a specific RNA molecule) which
removes the intron and joins the two flanking portions of
the RNA molecule, thus formin~ a molecuIe which is the
proper template for the active polypeptide. It is the
regulation of this splicing reaction which in turn regu-
lates translation. This second type of responsive RNA
molecule is similar to the first type of responsive RNA
molecule in that it has an inhibitor region which is
complementary in sequence to both the anti-inhibitor
region of a signal nucleic acid and to a substrate region
within the responsive RNA molecule. In this second type
of responsive RNA, the substrate region is not necessarily
part of an exon, but rather contains a region which is
essential to the self-splicing reaction. When the sub-
strate region is base--paired to the inhibitor region, the
self-splicing reaction cannot occur, thus translation is
prevented. In contrast, when a signal nucleic acid is
present, its anti-inhibitor region hybridizes to the
inhibitor region of the responsive RNA forming an inter-
molecular base-paired domain, which prevents intramole-
cular base-pairing between the inhibitor region and the
substrate region. Under these circumstances, the sub-
strate region is free to participate in the splicingreaction, the intron is removed, and translation of
properly joined exons can occur.
Thus, in a first aspect the invention features a
- responsive RNA molecule which encodes, in one or more
pro~ein-coding regions, a polypeptide, and which includes
a regulatory domain, a substrate region, and a ribosome
recognition sequence, e.a., a ribosome binding site, a
translation initiation site, and all non-coding regions
necessary for the translation of an RNA. This responsive
RNA molecule has an inhibitor region in the regulatory
domain which is complementary to both a substrate region
of the responsive RNA molecule and to an anti-inhibitor
~;LJBSrllnVTE~ SHE~ET
PC~ 59'//~U~U.~
;
~ 1 ~3 i" 2 ~i ~
region of a signal nucleic acid such that, in the absence
of the signal nucleic acid, the inhibitor and substrate
regions form a base-paired domain which reduces the level
of translation of the responsive RNA molecule compared to
that level observed in the presence of a signal nucleic
acid.
The "regulatory domain" is a region of the responsive
RNA molecule which will regulate the level of translation
of the responsive RNA molecule dependent upon the presence
of the signal nucleic acid. The regulatory region
includes the inhibitor region, inverted repeats and nucle-
a~ion regions. A "ribosome recognition sequence" is a
region of an RNA molecule that is required in order for
translation to begin at a given initiation codon (typi-
cally AUG). Such a site is recognized by a ribosome andbound by the ribosome prior to the initiation of transla-
tion of the RNA. In procaryotes, the ribosome recognition
sequence is a ribosome binding site and includes a purine-
rich sequence centered about 10 nucleotides 5' to the
initiation codon (Shine and Dalgarno, Proc. Natl. Acad.
Sci. USA 71:1342, 1974~. For eucaryotes, the sequence A/G
NNAUGG described by Kozak (Kozak, J. Cell Biol. 108:229,
1989) is the minimal ribosome recognition sequence
required for initiation of translation. This sequence
includes the AUG initiation codon.
The "signal nucleic acid" is a nucleic acid (e.a., a
viral RNA) which is indicative of a condition under which
it is desirable to produce the polypeptide encoded by the
responsive RNA molecule.
A "base-paired" domain is a region over which the
nucleotides of two regions of nucleic acid are hydrogen
bonded to each other. The term includes bonding of less
than all contiguous nucleotides of such regions.
The "substrate region" is a region of the responsive
RNA molecule which when base-paired reduces the level of
translation of one or more of the protein-coding regions
in the responsive RNA molecule.
~;UB~;TtTlJ~E: SHEET
~097/13(371) 21~G~ PC'r/~S'J~ ()6U.
The ~inhibitor region" is a region of the responsive
RNA molecule whic~l when base-paired to the substrate
region reduces the level of translation of one or more
protein-coding regions in the responsive RNA molecule.
The "anti-inhibitor region" is a region of the signal
nuclei^ acid which when base-paired to the inhibitor~;
region increases th~ level of translation of one or more
protein-coding regions of the responsive RNA molecule
compared to that observed in the absence of the si~nal
nucleic acid molecule. These three regions interact to
regulate the level of translation af the responsive RNA
molecule and are selected to ensure appropriate levels of
polypeptide production dependen~ upon the presence o~ the
signal nucleic acid.
By "appropriate level" is meant that in the absence
of the signal nucleic acid the level of polypeptide is
sufficiently low to have little or no effect on the physi-
ology of the cell, and in the presence of the signal
nucleic acid the level of polypeptide is sufficiently high
to reduce viability of the cell. The level of translation
of the responsive RNA can be de~ermined by standard proce-
dures. Generally, a low level of translation is one in
which less than 0.1% of the polypeptide produced by a cell
is polypeptide encoded by the responsiYe RNA molecule.
In preferred embodi~ents, the substrate region is
part of an exon or a leader region or overlaps the junc-
tion between the two (which includes the ribosome recog-
nition sequence, and the initiation codon), or includes a
rPgion necessarv for the self-splicing reaction. In a
procaryotic system, it is preferred ~o have the substra~e
reqion include the ribosome binding site and t~e initi2-
tion codon, (e.~., by overlapping the junction between the
leader region and the protein-coding region). The pro-
caryoti- ribosome binds a. the ribosome binding si~e
3~ unless this si~e is occlu~ed. Once bound, the ribosome
will transla~e the exon and, in the process, unwind the
proposed substrate-inhibito region. Occluding the -ibc-
8~
W09~/l3070 PCT/I;S92/0060~
2 :L ~
some binding site and the initiation codon with the
substrate-inhibitor hybrid region may eliminate the
translation of a procaryotic message in this type of
model. (See Figs. lD and lG.)
For eucaryotes, the 40S subunit of the eucaryotic
ribosome binds at the 5'-end of a capped mRNA and "scans"
down the message in search of the first initiation codon
(see ~enerally Kozak, J. Cell. ~iol. ~ 229, 1989). In
this process, all but extremely stable hybrids (i.e.,
those having a free energy of formation of <-50 kcal/mol)
are unwound and scanned through (~ozak, Proc._Natl. Acad.
Sci. USA 83:2850, 1986). Thus, to inhibit scanning of the
40S subunit to the translation initiation site, the inhi-
bitor region must form an ex~ensive hybrid with the sub-
strate region (which may include the ribosome recognition
sequence and~or the initiation codon) in which the base-
paired region has a free energy of formation that is -50
kcal/mol or lower. Thus, it is preferred that the inhi-
bitor region be located downstream (3') of the ribosome
r~cognition sequence (in the exon or perhaps nearer the 3'
end of the message) so that the interaction between ths
inhibitor region and the anti-inhibitor signal RNA (which
would have a similar if not lower free energy of forma-
tion) would not also prevent movemen~ of the 40S ribosomal
subunit to the initiation site (see, Figs. lF, lG, and
lH). Accordingly, in a eucaryotic system, having the
self-splicing intron interrupt the protein-coding region
is preferred.
As used herein an "intron" is a domain of the respon-
sive RNA molecule which is separate from the exons.Preferably the intron is an RNA molecule having catalytic
activity including RNA cleavage and ligation activity. It
is preferred that such an intron be able to self splice
and thus is chosen from a group I or group II intron, such
as that present in Tetrahymena thermophila.
In more preferred embodiments, the responsive RNA
molecule is puri~ied, and the responsive RNA encodes a
8UB8TITUTE SHEE~
w~2~1307~) PCT/~'S92~00603
21 ~92~1
polypeptide which modifies cell viahility, cell prolifer-
ation, transcription of DNA, translation of RNA, or repli-
cation of DNA, e.q., the responsive RNA molecule encodes
a polypeptide which has diphtheria toxin activity or ribo-
nuclease activity.
"Purified RNA" is RNA isolated from one or more com-
ponents of the environment in which i~ naturally occurs.
For examplP, the RNA is presen~ in a cell in which it does
not naturally occur~ Preferably it is provided as a
homogeneous solution of nucleic acid.
In other preferred embodiments, the substrate region
includes the 5'-splice junction of the intron; the intron
reduces the level of translation of the exons compared to
the level of translation in the absence of the intron; the
intron is located between the ribosome recognition
sequence and a 5'-most exon or between two exons. Even
more preferably, the intron overlaps at its 5'-end a
5'-splice junction, and at its 3'-end a 3'-splice junc-
tion; the intron catalyzes two cleavage reactions, one
within the 5'-splice junction and one within the 3'-splice
junction; the intron i5 a sel~-splicing intron; the sub-
strate region includes the s~-splice junction; and the
inhibitor region interferes with ~he cleavage reaction
within the 5'-splice junction~
A "5'-splice junction" refers to the sequence over-
lapping or abutting the s~-end of an intron which is
required for a splicing reaction. A "3'-splice junction"
refers to the sequence at the 3~-end of an in;ron which is
required for a splicing reac~ion. Such splice junctions
overlap the ends of a self-splicing intron such as those
bordering the intervening sequence of Tetrahymena
thermophilaO
A "self-splicing intron" is a piece of RNA which
contains all of the sequences required except for the
necessary abutting splice junction sequences for the
intron to excise itself from a larger piece of RNA and to
- join the two pieces of RNA that flanked ~he intron prior
8UB8TlTl.Jll~ SHEET - -
Wo92/l307~ ~CT/~'S92t00603
9 2~30~ )1
to the excision reaction. That is, the intron is able to
cleave and ligate two portions of an RNA molecule.
In yet more preferred embodiments, the signal nucleic
acid is single stranded, e.q., it is viral RNA.
Examples of responsive RNA include TetrahYmena RNA
which has been modified, for example, by nucleotide
changes at positions -14, -19, -21, -22, -23 and/or -24
relative to the 5'-splice site.
In a related aspect the invention features a method
for interfering with the growth of a cell harboring a
signal nucleic acid by introducing a responsive RNA
molecule as described above into the cell.
In a related aspect the invention features a DNA
molecule encoding the above responsive RNA molecules.
Other features and advantages of the invention will
be apparent from the following description of the pre-
ferred embodimen~s thereof, and from the claims.
Description of the Preferred Embodiments
The drawings are first briefly described.
Brief Description of the Drawinas
Figs. 1, lA and lB are schematic drawings of a
responsive RNA molecule. The thin line represents the
leader region, the thick line repres~nts a protein-coding
region, a series o~ shor~ vertical lines indicates a base-
paired domain, and the boxes above and below these linesindicate various features of the RNA. Specifically, in
Fig. lA the responsive RNA is drawn so as to depict intra-
- molecular base-pairing which prevents translation; and inFig. lB the responsive RNA molecule is depicted as hybri-
dized to a signal nucleic acid.
- Fig. lC depicts a second variation of this type of
responsive RNA molecule; in Fig. lD the responsive RNA
molecule is drawn to depict the intramolecular base-
pairing that prevents translation; and in Fig. lE the
:
:
$UE~8TITUTE SHEET
W0~2/13070 PcT/~ls92/006~3
2~o~
responsive RNA molecule is hybridized to a signal nucleic
acid.
Fig. lF depicts a third variation of a responsive RNA
molecule; ln Fig. lG this responsive RNA molecule is drawn
to show the intramolecular ba~e-pairing which prevents
translation; and in Fig. lH the responsive RNA molecule is
hybridized to a signal nucleic acid.
Figs. 2, 2A, 2B, and 2C are schematic drawings of a
responsive RNA molecule which includes a self-splicing
intron. The thin line represents a leader region, the
broken line represents a self-splicing intron, the thick
line represents an exon, a series of short vertical lines
indicates a base-paired domain, and the boxes above and
below these lines represent various features of the RNA.
Speci~ically, in Fig. 2A the responsive RNA molecule is
drawn so as to depict the intramolecular base-pairing
which prevents self-splicing; in Fig. 2B the responsive
RNA molecule is depicted as hybridized to a signal nucleic
acid; and Fig. 2C depicts the spliced molecule produced by
the self-splicing reaction.
Fiqs. 2D, 2E, 2F and 2G depict a variation of the
type of responsive RNA molecule shown in Figs. 2-2C. In
Fig. 2D, a self-splicing intron separates the polypeptide-
coding sequence in the responsive RNA ~olecule; in
Fig. 2E, the responsive RNA molecule is drawn to depict
the intramolecular base-pairing which prevents self-
splicing; in Fig. 2F, the responsive RNA molecule is
hybridized to a signal nucleic acid; and in Fig. 2G, the
spliced molecule produced by the self-splicing reaction is
depicted.
Fig. 3 depicts P(l) and P(-l) stem-loop structures at
or just upstream of the 5' exon-intervening sequence (IVS)
junction of Tetrahymena thermoE~hila. The IVS (uppercase)
contains the internal guide sequence (boxed) which can
hybridize with the end of the 5' exon (lowercase) to form
the P(1) stem-loop, the conformation required at the
5'-splice site (shown by filled-in triangle) for self-
8U~ TITUTE 8HEET
W~92/l307~ PCT/~S92/00603
~13~
11
splicing. The alternative structure P(-l), which does not
support self-splicing, is formed by hybridization between
a portion of the P(l) stem (boldface) with an upstream 5'
exon sequence (overlined). The sequence shown at the top
is that for RNA from the parent plasmid pTETBLU. The
lower three RNA structures represent modified P(-1) stem-
loops from three mutant plasmids that were made by
sequence changes (shaded) in the 5~ exon. Calculated free
energies at 37'C for each of these structures are given.
Fig. 4 is a photograph of a polyacrylamide gel show-
ing the results o~ 'n itro transcription reactions car-
ried out in the presence of [~ZP]CTP using the parent
plasmid (pTETBLU) and the three splicing mutants (pTET14,
pTET1419, pTET21-24) as templates. Each set of three
lanes represents the transcription products before (0) and
after (15 or 60 min) the change to splicing conditions.
The template used is given above each set of lanes, and
the restriction enzyme used to linearize the template is
shown at the top. For this and the following two figures,
FL denotes the full-length precursor ~NA and LE indicates
ligated exons. The positions of linear IVS RNA (L-IVS~
and circular IVS RNA (C-IVS) are also indicated, as are
the shortened forms of L-IVS in which 15 or 19 nt have
been removed from the 5' end by the circularization reac-
tion (L-15 and L-19, respectively). An asterisk denotes
an RNA thought to be the produc~ of 3'-splice site hydrol-
ysis (l.e., a 5' exon-IVS fragment). An as yet unidenti-
fied small RNA product is also indicated (<).
Fig. 5 is a photograph of a polyacrylamide gel show-
ing the results of experiments in which gel-purified
pTET1419 RNA was incubated under splicing conditions in
the absence (0) or presence of the given concentrations of
either of two signal RNAs (4S or 4S3) for 15 or 60 min.
The resultant products were analyzed on a 4% denaturing
polyacrylamide gel and are indicated in Fig. 5.
Fig. 6 is a photograph of a polyacrylamide gel show-
ing the results of experiments in which gel-purified
- 81JB8TITUTE SHEET
W092/l3070 PCT/~S92/006~3
~ 2 ~ ~ .
12
pTETBLU RNA or pTET21-24 RNA (10 nM) was incubated in
splicing buffer at 4 or 37nC. Where indicated, Mg2 was
added to S mM to initiate the splicing reaction. For
pTET21-24, splicing was initiated in the absence or pres-
S ence of either of two signal RNAs specific for the pTET21-
24 sequence (8S4 or 12S). When present, the concentration
of the signal RNA is 1 ~M. The resulting products, ana-
lyzed on a 4~ denaturing polyacrylamide gel, are labeled
as in Fig. 4. Templates used for transcription were lin-
earized with either EcoRI or ~HI as indicated at thetop. An additional product seen when the ~_RI~runoff
precursor is incubated under splicing conditions in the
presence of signal RNA is indicated with a dot. A short
RNA product (<) seen when pTET21-24 is incubated under
splicing conditions in the absence of a signal RNA is
marked with an arrowhead. This same RNA product is also
visualized in Fig. 4.
Responsive RNA Molecules
Responsive RNA molecules are generally described
above. Below are presented specific examples to illus-
trate these molecules to those of ordinary skill in the
art. These examples are not limiting to this invention.
Example 1: Responsive RNA Molecules Without Introns
A first type of responsive RNA molecule is illus-
trated in Fig. 1. One por~ion of this molecule, theprotein-coding region encodes a polypeptide ~hose produc-
tion is desired only in the presence of a signal nucleic
acid. Another portion of the molecule, the regulatory
domain, includes an inhibitor region which is complemen-
tary in sequence to a substrate region within the protein-
coding region. The inhibitor region can base pair with
the substrate region to form a base-paired domain which
blocks translation of the protein-coding region. The
substrate region can be a part of the protein-coding
8U~8TITUTE SHEET
W092/13070 PC~ S92/0~603
13
region, part of the leader region, or overlap the junction
between the two.
Referring to Fig. 1, responsive RNA molecule 10 has
a 5'-end 12, and a 3'-end 14. Adjacent to 5'-end 12 is a
leader region 2S and regulatory domain 16; adjacent to
3'-end 14 is a protein-coding region 18. Within regula-
tory domai~ 16 is an inhibitor region 20; within protein-
coding region 18 is a substrate region 22. At the 5' of
protein-coding region 18 is a ribosome racognition
sequence 21 and an initiation codon 23.
R~ferring to Fig. lA, inhibitor region 20 hybridizes
to substrate region 22 to form a base-paired domain 28.
Such base-pairing within responsive ~NA molecule 10 inhi-
bits translation of the protein-coding region of the
responsive ~NA molecule.
The inhibition of translation is relieved by the
presence of a signal nucleic acid, a region of which,
referred to a~ the anti-inhibitor, is complementary to the
inhibitor region of the responsive RNAo The anti~
inhibitor region of the signal nucleic acid competes with
the substrate region of the responsive RNA molecule for
hybridiza~ion (base pairing) with the inhibitor region of
the responsive RNA molecule. Under these circumstances
there is no base pair formation with the substrate region,
translation occurs and the desired polypeptide is
produced.
For example, referring to Fig. lB, signal nucleic
acid 30 has a 3'-end 32, a 5'-end 3~, and an anti-
inhibitor region 36 complementary in sequence to inhibitor
region 20 of responsive RNA molecule 10. Hybridization of
anti-inhibitor region 36 with inhibitor region 20 forms
base-paired dom~in 38 and prevents hybridization of inhi-
bitor region 20 to substrate region 22. Under these cir-
cumstances, translation of protein-coding region 18
occurs.
In a variation of this type of responsive RNA mole-
cule, the substrate .region is not entirely contained
.
~3U~18TITUTE g~
WO92/13070 PCr/~'S92/00603
2 1 ~
14
within the protein-coding region but extends upstream of
the protein-coding region into the leader region. Spe-
cifically, the responsive RNA molecule depicted in Fig. lC
has a substrate region 22 which includes the ribosome
recognition sequence 21 and the initiation codon 23.
Referring to Fig. lD, substrate region 22 base-pairs to
inhibitor region 20 forming intramolecular base-paired
region 28. In a procaryotic system, this con~iguration
physically blocks a ribosome from interacting with the
ribosome binding site and the initiation site, and trans-
lation is inhibited. Referring to Fig. lE, the anti-
inhibitor region 36 of the signal nucleic acid 30 is
hybridized to the inhibitor region 20 to form base-paired
region 38. In this configuration, a procaryotic ribosome
initiates translation and the desired polypeptide is
produced.
In another variation of this type of responsive RNA
molecule, the inhibitor region is located downstream of
the substrate region. The inhibitor region can be within
the protein-coding region itself, as diagrammed in this
figure, or located in a region 3' of the protein-coding
region. In Fig. lF, the responsive RNA molecule is
depicted as having a substrate region 22 that includes
ribosome recognition sequence 21 and initiation codon 23
and has an inhibitor region 20 located 3' of the substrate
region. Referring to ~ig. lG, the inhibitor region, 20,
base-pairs with the substrate region 22 forming intramole-
cular base-paired region 28. In this configuration, a
scanning eucaryotic ribosomal subunit cannot invade or
bind to the base-paired domain to initiate translation
provided this basepairing interaction is sufficiently
strong. In Fig. lH, the anti-inhibitor region 36 of sig-
nal nucleic acid 30 is hybridized to inhibitor region 20
~orming base-paired region 38. A eucaryotic ribosome can
scan to the proper initiation codon (provided there are no
other upstream initiation codons) and initiate transla-
tion. Translation of the polypeptide occurs, with dis-
WO 92/13070 PC r/~;s~2/no603
~ 1 ~ ? ~
ruption of base-paired region 38 by the translating
ribosome.
Since the inhibitor region of the responsive RNA must
be complementary to both the substrate region of the
responsive RNA, and the anti-inhibitor region of the tar-
get nucleic acid, the sequences of these three regions
must be chosen to allow suitable regulation of translation
of the responsive RNA. This does not mean that the
sequence of the substrate region must be identical to the
sequence of the anti-inhibitor region. Neither of the two
base-paired domains which can form need to be perfectly
base-paired (l.e., all contiguous bases along the domains
are base-paired), nor do they have to be the same length.
There is flexibility in the selection of the anti-
inhibitor region so long as the region is specific enoughto indicate when translation must occur. For example, if
the signal to which the responsive RN~ responds is the
presence of HIV-I within a cell, any specific nucleic acid
sequence of HIV-I could be chosen, and of course, one is
limited in selecting a nucleic acid seguence present in
HIV-I. The sequence of the substrate region is chosen to
create a responsive RNA molecule which produces a biologi-
cally active polypeptide. Since the substrate region may
include portions of a protein-coding region, any modifica-
tion of its sequence ~ust preserve a significant amount ofthe activity of the encoded polypeptide. The degeneracy
of the genetic code allows for changes in the sequence of
the protein-coding region which do not affect the sequence
of the encoded polypeptide. Because guanosine can base-
pair with uridine as well as with cytosine there is addi-
tional flexibility in the sequences which can be used.
In addition, since conservative amino acid changes at one
or more positions in proteins often do not eliminate
activity of the protein the number of useful sequences is
increased substantially.
The base-paired domain formed by hybridization of the
inhibitor region to the substrate region must be stable
- - 8UB8TITUTE 8HE~ET
W092/l3070 PCT/US92/00603
~t~ 2 ~ 1
16
enough so that it will not be disrupted by nucleic acids
other than the signal nucleic acid, which may also be
present within the cell. For example, if the inhibitor
region and the substrate region are complementary over
only four contiguous nucleotides, any sinyle strandad
nucleic acid that includes that four base sequence could
compete with the substrate region for hybridization to the
inhibitor region, and if the nucleic acid including this
sequence was present at a high enough concentration inhi-
bition of transla~ion would be relieved. Generally, thebase-paired domain formed by the hybridization of the
substrate region to the inhibitor region should include at
least 12, and preferably 15, contiguous nucleotides in
order for the molecule to respond to only the signal
nucleic acid.
The responsive RNA molecule can include a region that
will allow the signal nucleic acid to more readily hybri-
dize to the inhibitor region. This additional region is
called a nucleation region and consists of a number of
nucleotides immediately adjacent to the inhibitor region
and complementary to the sequence of the signal nucleic
; - acid such that the nucleation region and the inhibitor
together form a region of extended complemantarity with
the signal nucleic acid. The nucleation region provides
a single stranded region tha~ is readily available for
hybridization to the signal nucleic acid. Base-pair
formation over this region will tend to ~avor displacement
of the substrate region from the inhibitor r~gion by posi-
tioninq the anti-inhibitor region correctly for hybridiza-
tion to the inhibitor region. In addition, such a nuclea-
tion region will increase the stability of the base-
paired region formed with a signal nucleic acid.
The regulatory domain may also include a region that
will disfavor hybridization of non-spe~ific nucleic acids
(i.e., nucleic acids other than the signal nucleic acid)
to the region immediately adjacent to the inhibitor
8VB8TITUTE SIHEET
~VOg~/l3070 PCT/~S9~/00603
2 1 ~
17
domain. This region is re~erred to as an inverted repeat
and can fold to forma hairpin structure.
The detailed nature of the inhibitor region, the sub-
strate region, and the anti-inhibitor region will depend,
in part, on how ~ightly translation is to be regulated.
The more stable the intramolecular base-paired domain
formed by hybridization of the inhibitor region to the
substrate region, the more translation will be inhibited.
For RNA-RNA duplexes, the stability of a base-paired
domain depends on the number of nucleotides actually base-
paired within a contiguous region of nucleotides, the
number of mismatches within a generally base-paired
domain, and t~e nucleotide composition of the base-paired
domain. Intramolecular base-pair formation depends on the
distance between the two regions to be base paired. For
example, when there are too few nucleotides between the
two regions, torsional-type constraints can prevent base
pair formation. Those in ~he art are well aware of how
these param~ters can be adjusted in order to make a more
or less stable base-paired domain. The stability of the
intramolecular base-paired domain can be adjusted depen-
dent upon the level of translation that is desired at any
given level of signal nucleic acid. The level of transla-
tion depends on the proportion of respon~ive RNA molecules
in which the inhibitor region is hybridized to th2 sub-
strate region. This proportion, in the presence of the
signal nucleic acid, depends on the proportion of the
responsive RNA molecules in which the inhibitor region is
hybridized to the anti-inhibitor region of the signal
nucleic acid. Those in the art will appreciate that the
amount of each duplex which forms depends on the relative
stability of the two duplexes as well as the amount of
signal nucleic acid and responsive RNA present in a given
cell. If a highly toxic molecule is encoded by the
responsive RNA then a high degree of regulation is
required. For example, if the active subunit of cholera
toxin is encoded, only a few molecules are required to
8lJBSTlTUTE SHEE~T -
w(~ 92/1307f~ PCT/I,'S92/00603
o ~
18
kill a cell. In this case translation must be completely
inhibited in the absence of signal nucleic acid. This is
best ensured by having almost complete complementarity of
the substrate and the inhibitor regions, ~g~, 85% comple-
mentarity of a 20 nucleotide region. Expression occursonly when a highly complementary signal RNA is present,
having e.~, 100% complementarity to the inhibitor region
over a 25 nucleotide region.
The inhibitor region may be on the 5'-side or the
3'-side of the protein-coding region or within the
protein-coding region itself. If the responsive RNA
molecule is subject to exonucleolytic degradation, this
should be taken into account when designing the molecule.
Thus, if the molecule is degraded beginning at the 3'-end
it would be best to locate the inhibitor region at the
5'-end of the molecule in order to prevent formation of a
molecule containing all of the sequences required for
translation but lacking an inhibitox region.
Example_2: Responsive RNA Molecules_ ith Self-Splicinq
Introns
A second type of responsive RNA molecule lncludes a
self-splicing intron which prevents production of the
desired polypeptide. The intron can be removed by a
splicing reaction, and the spliced molecule serves as a
template for the production of the desired polypeptide.
A signal nucleic acid regulates translation of this type
of responsive RNA molecule, but the regulation is achieved
indirectly by using the signal nucleic acid to regulate
the splicing reaction. In order for this type of regula-
tion to work the responsive RNA molecllle must, in the
absence of the signal nucleic acid, fold so as to form an
intramolecular base-paired domain which prevents splicing.
In the presence of the signal nucleic acid an alternative
intermolecular base paired domain forms and splicing
occurs.
- - 8UB8TITUTE SHE~ET
w092/l307~ PCT/~592/00603
19 2i~
An example of thls second type of responsive molecule
is illustrated in Fig. 2. This molecule has an intron
located between the ribosome recognition sequence and the
initiation codon of a single protein-coding region which
encodes a desired polypeptide. This intron prevents
translation because it places the ribosome recognition
sequence too far away from the initiation codon~ In this
example, the intron is a self-splicing intron derived from
the pre-rRNA of Tetrahymena. Introns of this type can
fold into a structure which causes two cleavage reactions,
one on either side of ~he intron, and a ligation reaction
which joins ~he portions of the ~JA molecule flanking the
intron. An essential step in the self-splicing of such
introns is hybridization of a region of the intron,
referred to as the 5'-splice junction, to a second region
of the in~ron, referred to as an internal guide sequence.
Thus, one way the self-splicing activity of the intron can
be regulated is by preventing hybridization of the
5'-splice junction to the internal guide sequence. The
responsive RN~ molecule depicted in Fig. 2 has a regula-
tory domain which is distinct from the intron and the
protsin-coding region. This regulatory domain has an
inhibitor region which is complementary to the substrate
region which in this molecule includes the 5'-splice
junction of the self-splicing intron. Intramolecular base
pair formation between the inhibitor region and the sub-
strate region prevents hybridization of the 5'-splice
junction to the internal guide sequence, and splicing is
prevented. This responsive RNA molecule is designed so
that the inhibitor region is also complementary to the
anti-inhibitor region of the signal nucleic acid. Thus,
in the presence of the signal nucleic acid, the inhibitor
region hybridizes to thP anti-inhibitor region freeing the
5'-splice junction for participation in the self-splicing
reaction.
Referring to Fig. 2, responsive RNA molecule 40 has
a 5'-end 42, and a 3~-end 44. Adjacent to 5~-end 42 is a
- 8UB8TiTlJTE ~;HE~Fr
WO97/l307~) PCT/US92/00603
3 1
leader region 49 adjacent to which is a self-splicing
intron ~8, and then polypeptide-encoding exon 50. Regu-
latory domain 46 lies within leader region 49. Self-
splicing intron 48 thus l.ies between regulatory domain 46
and exon 50, and is flanked on its 5' side by a ribosome
recognition sequence 56, and on its 3' side by an AUG
codon 66. An inhibitor region 52 within reyulatory domain
46 is complementary to a substrate region 54 at the junc-
tion between leader region 49 and self-splicing intron 48.
Within the regula'eory do~ain, on the 3'-side of the inhi-
bitor region, is a nucleation region 45 which is contigu-
ous with the inhibitor re~ion 52 and complementary to a
region of the signal nucleic acid immediately adjacent to
the anti-inhibitor region referred to as the anti-
inhibitor extension. The regulatory region may alsoinclude an inverted repeat 47 on the 5~-side of the inhi-
bitor region. Substrate region 54 includes ribosome
recognition sequence 56, a 5'-splice junction 58, and a
stabilizer region 60. Self-splicing intron 48 is over-
lapped by a 5'-splice junction 58, and a 3'-splice junc-
tion 64 adjacent to AUG codon 66, and includes an internal
guide sequence 62.
Referring to Fig. 2A, when inhibi~or region 52 hybri-
d~zes to substrate region 54 a base-paired domain 70 forms
preventing 5'-splice junction 58 ~rom interacting with
internal guide sequence 62. The inverted repeat can fold
so as to create a stabiliæer hairpin 63.
In the presence of a signal nucleic acid, an inter-
molecular bas~-paired domain forms between the anti-
inhi~itor and an~i-inhibitor extension regions of the
signal nucleic acid and the inhibitor and nucleation
regions of the responsive RNA molecule. This interaction
frees 5'-splice junction 58 allowing it to interact with
internal guide sequence 62. Under ~hese circumstances, a
self-splicing reaction occurs. Thus, referring to
Fig. 2B, signal nucleic acid 71 having a 3'-end 72 and a
5'-end 73 includes an anti-inhibitor region 74 and an
SIIB8TITUTE SHEET
W~92/13070 PCT/VSg~/00603
C~
21
anti-inhibitor extension 77 which hybridize to inhibitor
region 52 and nucleation region 45 forming base-paired
domain 15.
The self-splicing reaction removes all of the self-
splicing intron. The spliced molecule now can produce theencoded polypeptide from exon 50 because the ribosome
recognition sequence is now in close juxtaposition to the
initiation codon of the polypeptide encoding exon allowing
utilization of the initiation sequence as the first codon
of a polypeptide.
Referring to Fig. 2C, spliced molecule 90 includes
5'-end 42, 3'-end 44, leader region 49, exon 50, ribosome
recognition sequence 56, ini~iation codon 66, and fused
splice junction g5 containing a small por~ion of 5' splice
junction 58 and a small portion of 3' splice junction 64.
Any intron known to have self-splicing activity can
be adapted for use as a responsi~e RNA molecule. Suitable
self-splicing RNA can be derived form the nuclear pre-
rRNA of Tetrahvmena, the mitochondrial pre-rRNA of
Saccharomvces and Neurospora, the introns of Arqobacterium
or Azoarcus, and the mitochondrial pre-mRNA of Saccharo-
myces or other equivalent group I self-splicing ~NAs.
Group II introns can also be used in this invention, or
any RNA which has at least ~NA cleavage activity. ~NA
Z5 ligase activity can be provided by other RNA molecules or
their equivalent.
Once a sel~-splicing RNA has ~een selected it must be
correctly positioned between the ribosome recognition
sequence site and the start codon of the polypeptide
encoded so that after the self-splicing reaction has
occurred the ribosome recognition sequence is positioned
correctly relative to the start codon. In eucaryotes
translation generally begins at the most 5' AUG of a
capped RNA providing that the sequence surroun~ing the AUG
conforms to A/G~NAUGG. Accordingly, the responsive RNA
molecule must be designed so that this sequence appears
only after splicing has occurred. Moreover, an AUG or
.
81JB8TITIJTE 8HEE~ -
WO9t/13Q7n PCT/~'S92/00603
2 ~ ~ 0 '~
22
other codon in a favorable sequence context can be
included in the intron so that it is recognized and used
as the 5' most translation initiation site. The inhibi-
tory effect of this upstream AUG on translation initiation
at the downstream site will be relieved only upon removal
of the intron by self-splicing, thus ensuring that no
scanning ribosomal subunits reach the downstream initia-
tion site from which translation of the toxic protein
would occur.
In a variation on this ~ype o~ responsive RNA mole-
cule the self-splicing intron is placed 50 as to interrupt
a polypeptide-coding sequence. As illustrated in Fig. 2D,
this molecule has an intron located between two exons that
together encode the desired polypeptide. I~ the intron
includes a stop codon, translation will be blocked. Even
if the intron does not encode a stop codon, translation of
the intron may be out-of-frame with the downstream exon
and/or will add amino acids to the polypeptide that will
likely destroy activity. Removal of the intron results in
the fusion o~ the two exons and ~ormation of a translat-
able nucleotide sequence coding for a polypeptide having
the desired activity.
Referring to Fig. 2D, responsive RNA molecule 40 has
a 5'-end 42 and a 3'-end 44. The polypeptide is encoded
in two regions, 50 and 51, separa~ed by self-splicing
intron 48. Intron 48 is overlapped by a s~-splice junc-
tion 58, and a 3'-splice junction 6~ and includes internal
guide sequence 62. The protein-coding region 50 is pre-
ceded by a ribosome recognition sequence 56 and a transla-
tional initiation codon ~6. An inhibitor region 52 lieswithin exon 50 and is complementary to substrate region 5~
which overlaps the 3'-end of region 50 and the 5'-splice
junction 58 and includes stabilizer region 60. Flanking
the inhibitor region on its 5~ side is nucleation region
45 that is contiguous with the inhibitor region and is
complementary to regions in the signal nucleic acid imme-
diately adjacent to the anti-inhibitor region.
8UB8TITIITE SHEE~T
W~92/13070 PCTt~'S92/00603
J ~J a.J _
23
Referring to Fig. 2E, when the inhibitor region 52
hybridiæes to substrate region 54 a base-paired domain 70
forms and thus prevents the 5'-splice junction 58 from
interacting with the internal guide sequence 62.
Referring to Fig. 2F, signal nucleic acid 71 having
a 3'-end 72 and a 5'-end 73 and including an anti-
inhibitor region 74 and an anti-inhibitor extension 77
hybridizes to the inhibitor region 52 and nucleation
region 45. The intermolecular base-paired domain 75 is
formed. Under these circumstances, the 5'-splice junction
58 is free to interac~ with the internal guide sequence 62
and self-splicing occurs.
Referring to Fig. 2G, the self-splicing reaction
removes all of the self-splicing in~ron 48 lea~ing the
fused spliced junction 95 which contains portions of the
5'-splice junction 58 and the 3'-splice junction 64.
Other strategies, for example, where the substrate
and/or inhibitor regions are contained within the intron,
may be used so that upon splicing these ele~ents are com-
pletely removed. When ~he substrate or inhibitor domainsremain in the protein-coding regions, their sequences must
be carefully chosen to preserve the biological activity of
the encoded protein. The degeneracy of the genetic code,
the possibility of guanosine-uridine ~ase-pairs and con-
servative amino acid changes that do not eliminate theprotein's activity will all be considered. Moreover, it
i5 known that many proteins contain regions not essential
to their inherent activity and that amino acid changes
and/or additions in these areas do not result in a drastic
loss of biological activity. The placement of the sub-
strate and/or inhibitor domains in such a region simpli-
fies the choice of the anti-inhibitor containing signal
RNA since changes to the protein-coding sequence might be
more easily tolerated.
The requirement that the inhibitor region be com-
plementary to both the anti-inhibitor region and the
substrate region places certain constraints on the
8UB8TITIJTE SHEET
WO92/13070 PCT/VS92/00603
21~2'j'1
sequences of these regions. First, as noted above, the
substrate region does not have to have the same sequence
as the anti-inhibitor region of the signal nucleic acid.
Since the anti-inhibitor region can be selec~ed but not
altered, the anti-inhibitor region must include a sequence
identical to the sequence of the 5'-splice junction. The
minimal 5'-splice junction in a ~et~hy~ena rRNA intron is
only four nucleotides long. Since any four nucleotide
sequence should occur with a probability of 1/64, many
potential anti-inhibitor regions will include the sequence
of the 5'-splice junction. It is very likely that many
different four-base sequences can serve as a 5~-splice
junction provided tha~ the sequence of the internal guide
region is adjusted to accommodate the changes in the
5'-splice junction (Zaug et al., Nature 324:430, 1986).
While it is suitable for the minimal 5~-splice junction to
be able to base pair With the internal guide sequence, a
complex with a single mis-match can be functional (Zaug
et al, Biochemistry 27:8924, 1988).
The base-paired domain formed by hybridization of the
inhibitor region and the substrate region must be more
stable than the base-pairing that occurs between the
5'-splice junction and the internal guide sequence during
a splicing reaction. This can b~ accomplished ky choosing
an inhibitor region and substra~e region that will hybri-
dize to form a base-paired domain longer than that formed
by hybridization of 5'-splice junction to the internal
guide sequence. The substrate region is designed to
include a stabilizer region that extends the homology
between the substra~e region and the inhibitor region
sequence beyond the 5~-splice junction. This stabilizer
region can be located just 3~ of the 5'-splice junction in
the case of self-splicing introns located between the
ribosome recognition sequence and the initiation codon.
This arrangement ensures that the stabilizer domain will
be removed as part of the splicing reaction and will not
interfere with the relationship between the ribosome
8U~8TITIJTE 8HEET
3V092/l3070 PCT/U592/0~603
3 1
recognition sequence and the initiation codon. The ribo-
some r~cognition sequence can also be included within the
region which base-pairs with the inhibitor region, but
there is no requirement that this be the case. In the
case of a self-splicing intron which is inserted between
exons which encode portions of the same polypeptide, the
stabilizer region should preferably be located within the
intron, i.e., on the 3~-side of the s~-splice junction so
that it will be removed along with the rest o~ the intron.
It is important that the inhibitor/substrate base-
paired domain be disrupted only by the signal nucleic acid
and not by other nucleic acids present in cell. As dis-
cussed above, for the first type of responsive RNA mole-
cule, this means that the intramolecular base-pair forma-
tion must be ext~nsiva enough to be disrupted only by a
unique nucleic acid. This requiremen~ can make it diffi-
cult for the signal nucleic acid to disrupt the intramole-
cular duplex. As outlined above, the inclusion of a
nucleation region adjacent to the inhibitor region will
favor hybridiza~ion of the inhibitor region to the anti-
inhibitor region.
Many arrangements of the regulatory domain of the
self-splicing intron and the exon will be useful. As
noted above, the self-splicing intron can be located
between two exons; under these circums~ances while the
most 5' exon of the unsplic~d molecule can be translated
a complete functional polypeptide cannot be produced. The
i~hibitor region can be located on the 5'- or the 3'- side
of the self-splicing intron or possibly within the intron
itself. Since RNA is synthesized in the 5' to 3' direc-
tion, it is preferred to locate the inhibitor region on
the 5'-side so that the inhibitor will be synthesized and
have an opportunity to hybridize to the 5'-splice junction
before the production of the internal guide seguence. The
inhibitor could be located on the 3'-side of the self-
splicing intron if folding of the RNA to form the splicing
8lJ~8TlTUTE SIHEE~
w092tl307() PCT/~'S92/00603
26
complex is slow compared to rate of synthesis of the inhi-
bitor region.
It is preferred that the self-splicing reaction be
specific and accurate; if the splice occurs at the wrong
location, the ribosome binding site will be positioned
incorrectly. In the case of a self-splicing intron
located between two exons, incorrect splicing may result
in an out-of-frame fusion of the polypeptide encoding
sequences. Self-splicing introns in which the distance
between the internal guide sequence and the 5'-splice
junction is relatively short tend to catalyze more accu-
rate splicing reactions. It is also important to ensure
that there are no sequences that will be recognized as
alternative 5'-splice junctions.
The above described responsive RNA molecules can be
prepared by any standard methodology. For example, the
RNA can be produced by a transcription of a DNA molecule,
eithex in vivo or n vitro. Generally, the RNA molecule
will be produced by construction of a plasmid or viral DNA
which includes sequences encoding the responsive molecule,
appropriate sequences for regulated transcription of the
responsive RNA molecule, and appropriate sequences for
replication of the DNA. In constructing the RNA molecule,
the general considerations are described above. From a
practical viewpoint, it is generally preferred to identify
an appropriate RNA molecule having enzymatic activity
which is able to cleave itself or other RNA molecules and
is preferably able to splice those two RNA molecules
together, e.q., a self-splicing RNA molecule. This RNA
molecule is then modified to change the 5~-splice junction
and the internal guide sequence as require~ within the
limitations described above so that the 5'-splice junction
is complamentary to part of the inhibitor region of the
responsive RNA molecule. This RNA molecule is then caused
to be ligated to RNA whi~h encodes the desired polypeptide
and to RNA which includes an appropriate regulatory
8UE~8TITUTE ~
WO9~/~3070 PCT/US92~00603
f~J .~ 1
27
domain. If required, nucleation sites and inverted
repeats can be designed into the requlatory domain.
The experiments discussed in the following Examples
3-7 describe preparation o~ responsive RNA molecules con-
taining inactive introns which can be reactivated by thepresence of specific signal RNAs. The responsive RNA
molecules were prepared from the self-splicing intron or
intervening sequence (IVS) in the rRNA of Tetra~n~ena
thermophila. For the IVS to self-splice requires the
proper folding of the core structure of the IVS RNA.
Included in this required conformation ls a base-paired
region known as P(1) that encompasses the 5'-splice site
(Fig. 3). In P(1), the internal guide sequence in the IVS
base pairs with the adjacent portion of the 5' exon to
form a stable stem-loop structure. The 5'-splice site is
located within this stem. The ability of the IVS RNA to
self-splice relies on the ability of the P(l) stem to
for~.
A na~ural sequence just upstrea~ of the 5'-splice
site can also form a hairpin structure with the exon
sequence immediately adjacen~ to ~he 5'-splice site
(Fig. 3). The stem-loop required for sel~-splicing, P(l),
and this alternative stem-loop, termed P(-l), are mutually
exclusive since the 5' exon sequence immediately ad~acent
to the splice site is included in both structures. The
alternative st m-loop struc~ure, P(-l), can be made more
stable by extending its stem region. See Woodson and
Cech, Biochemistry, 30.2042, 1991, reporting results of a
one-nucleotide change in the 5' exon (A to C change at
position -1~ relative to the s'-Splice site). In that
mutant, self-splicing was reported to be decreased. Con-
versely, RNAs containing mutations in the 5' exon which
either diminished the relative strength of P(-l) or abol-
ished it completely reportedly showed an increase in sel~-
splicing activity. Three mutants which contain sequencechanges in the 5' exon, which were predicted to strengthen
the alternative structure, P(-l), were made. In all three -
W092/13~70 PCT/US92/00603
2~2~
28
mutants, the level of ln vit~o self-splicing (as judged by
the formation of ligated exons) was decreased relative to
a parent construct in which the natural 5' exon sequence
is present. one mutant, in which the stem of P(-l) has
been lengthened by 5 additional base-pairs, exhibits no
detectable self-splicing activity n vitro.
Applicant demonstrated that self-splicing activity
can be recovered even in this strong, non-splicing mutant
by ths addition of signal ~NAs complementary to the
upstream 5' exon sequence (inhibitor region) involved in
the alternative structure. By binding to the 5' portion
of the P(-1) stem, these signal RNAs disrupted P(-l) and
left the sequence immediately adjacent to the 5~-splice
site in sinqle-stranded ~orm, fully capable of hybridizing
to the internal guide sequence in an active, self-splicing
conformation containing P(l).
Example 3: Plasmid Construction and DNA Preparation
The source of the IVS-containing fragment used to
prepare the responsive RNA molecules was plasmid pTTlA3T7
(obtained from Dr. A. Zaug; equivalent such plasmids are
readily constructed an~ this plasmid is used only for pur-
poses of illustration of ~he invention), which con~ains
the 482-bp ThaI fragment of Te~rahy~ena thermophila rDNA
inserted into the HindIII site of pT7-2 (U.S. ~iochemica1
Corporation, Cleveland, Ohio) on HindIII linkers~ This
fragment contains rDN~ sequence corresponding to 32 nt of
5' exon, the 413 nt IVS, and 37 nt of 3 -exon. The
HindIII fragment of pTTlA3T7 was isolated and inserted
into the HindIII site of pTZ19R (United States Biochemical
Corporation, Cleveland, OH) to generate a plasmid con-
taining the IVS and a small portion of the natural rDNA
sequence inserted into the first few codons of the lacZ'
gene, the ~-complementation fragment of the ~-galacto-
sidase gene. It has been reported previously by others
(Been and Cech, Cell 47:207, 1986: Price and Cech, Science
228:719,- 1985; Waring et al., Cell 40:371, 1985), that
8UB~TITUTE SHEI~
W~92/13Q7~ P~T/US92/00603
2~, 9~j2,
29
~-galactosidase activity in ~_ÇQli relies on the a~ility
of the IVS RN~ to excise itself and ligate the lacZ' cod-
ing region in frame so as to produce a translatable mRNA
product. In vitro mutagenesis was carried out on the
pTZ19R derivative containing the rDNA in~ert to generate
a clone in which the corresponding lacZ' RNA would self-
splice ana maintain the correct reading frame. In addi-
tion, a potentially useful SalI site was created in the
3'-exon and an in-frame AUG in the 3'-exon was destroyed
to ensure that it not be used as a translation start site.
The final DNA sequence and correct reading frame of the
3'-exon from the 3'-splice site (~) to ~he HindIII site
(underlined) in the vector sequence is shown below.
pTETBLU
T AAG GTA GCC AGC CGT CGA CAT CTA ATT AGT GAC GCA AGC
TT
pTETBL~ DNA was hen used as the parent for a series
of splicing mutants in which changes were made by in vitro
mutagenesis in the 5' exon sequence to improve the base-
pairing ability in the alternative P(-l) stem-loop struc-
- ture. Care was taken to maintain the correct reading
frame in the spliced RNA product and to avoid the creation
of translational start or stop codons. The resulting
sequence chang~s made in the 5' exon RNA and the RNA
alternative structures predicted to form are shown in
Fig. 3.
All site-specific mutations were generated using the
in vitro Mutagenesis Kit from United States Biochemical
Corporation. D~ oligos were made on an Applied Biosys-
tems 394 DNA/RNA Synthesizer using phosphoramidite chem-
istry and purified using OLIGOCLEANTM columns (United
States Biochemical Corporation) prior to use as mutagenic
oligo. Plasmids were maintained in strain MVll90 (E. coli
tsrl-recA~ 306::TNlO ~ ~lac-Pro! _thi-su~E (F' ~ro A+B+
lacI lacZ, Mi5 traD3"6). Each plasmid was verified ~y
DNA sequencing (Tabor and Richardson, Proc. Natl. ~cad.
Sci. USA 84:4767, 1987).
38TITIJTE 15;HI ET
WO92/1307~ PCT/~'S92/00603
Plasmids for use as ~ transcription templates
were purified by Qiagen (Qiagen Inc., Chatsworth, CA)
maxi-column preparation as described by ~he manufacturer
except that the final DNA preparation (400 ~1) was
extracted two times with an equal volume of phenol, once
with chloroform, and ethanol precipitated in the presence
of 0.25 M Tris-HCl, pH 7.5. The plasmids were linearized
by cleavage with either ~çQRI or ~HI to generate tem-
plates on which runoff T7 transcription will yield full-
length RNA of 548 or 527 nt, respectively. (The T7 pro-
moter sequence is located immediately upstream of the
polycloning site and within the coding sequence of
~-galactosidase.)
Example ~4: Signal RNAs
Short signal RNAs (11-26 nt) were chemically synthe-
sized on an Applie~ ~iosystems 380B DN~ synthesizer using
phosphoramidite chemistry. Prior to use, the signal RNAs
were desalted using a C18 SEP-PAC cartridge (Millipore
Corporation), gel-purified and quantified by absorbance at
260 nm. Signal RNAs were stored at -20-C in 1 mM EDTA, 10
mM Tris-HCl (pH 7.5). The sequences of the signal RNAs
specific for precursor RNA from PTET1419 and pTET21-24
(see FIG. 3) are given below:
pTET1419 4S 3' GCCGCUCUCAG 5'
4S3 3' GCCGCUCUCAGUGAU 5'
pTET21-24 8S4 3' CGCCCAUUUAAAUCUCUCAGUGAUA 5'
12S 3' CGGAAACGCCCAUUUAAAUCUCUCAG 5'
These signal RNAs are complementary to the upstream
exon sequence which forms the 5' side of the P(-1) stem in
the given construct. The underlined nucleotides corres-
pond to the portion of the signal sequence that will base
pair with 5' exon sequence involved in the P(-l) stem, the
remaining nucleotides base pair either with nucleotides at
the base of the stem or in the loop. For example, signal
RNA 4S3 will base pair with 4 nt 5' to the base of the
81J~8TITUTE ~
WO92/13070 PCT/US92/00603
31 2 J ~ {.~
stem in pTET1419 RNA, all the nucleotides included in the
5' side of the P(-l) stem and 3 nucleotides in the loop.
In pTET14 RNA (see FIG. 3), a U to C change at -14
relative to the 5'-splice site allows the formation of an
extra C-G base-pair to lengthen the P(-l) stem. This par-
ticular seyuence change was reported by Woodson and Cech
(Woodson and Cech, ~iochemîst~Y ~Q:2042, 1991) to decrease
self splicing activity of a short precursor RNA. pTET1419
RNA has a~ additional nucleotide change (G to A at -19)
which allows P(-l) to form a more stable stem by creating
an A-U base pair in place of a less stable G-U base pair.
Finally, pTET21-24 RNA has a very stable P(-l) stem gener-
ated by 4 additional nucleotide changes tat positions -21
to -24 relative to the splice site). Calculated free
energies at 37-C for these structures, based on the most
current values in the literature (Freier et al., Proc.
Natl. Acad. Sci. US~ 83:9373, 1986; Jaeger et al., Proc.
Natl. Acad. Sci. USA 86:7706, 1989), are also given in
Fig. 3. In all of these constructs, nucleotide changes
were made in the upstream 5' exon only, without altering
the IVS or the 13 nt at the 3' end of the 5' exon.
On templates linearized with EcoRI or BamHI, full-
length transcription from the T7 promoter yielded tran-
scripts of 548 and 527 nt, respectively. These differed
2~ only in the length of ~heir 3'-exon (92 vs. 71 nt), but
had equivalent length 5' exons (43 nt) and IVS RNA (413
nt). Correct ligation of the 3'-exon to the 5' exon with
excision of the IVS yielded an RNA of 135 nt for the EcoRI
runoff transcript and 114 nt for the corresponding BamHI
transcript. The appearance of ligated exons is an indi-
cation of the level of self-splicing supported by a par-
ticular IVS-containing construct.
ExamDle 5: Decreasinq Self-SPlicinq bv Increasinq
Stability of P(~
In vitro transcription was performed as follows.
Transcription reactions using T7 RNA polymerase were
8UB8TITUTE SHEET
W~9~/l3070 PCT/US92/00603
~i~i32~)~
32
carried out in transcription buffer (40 mM Tris-HCl, pH
7.S, 5 mM MgCl2, 10 mM dithiothreitol, 4 mM spermidine)
containing 500 ~M each NTP and -10 ~Ci [~32P]CTP. Indivi-
dual reactions (10 ~l total volume) contained 0.1 ~g line-
arized plasmid template and 20-30 U T7 RNA polymerase.
After 30 minutes at 30C, 2 ~1 o~ each sample was removed
and mixed with 2 ~l buffered formamide containing xylene
cyanol FF and bromphenol blue (formamide/dye mix). The
remainder of the sample was warmed to 37-C, and 2 ~l of lM
NaCl, 20 mM MgCl2, 1 mM GTP was added to adjust the reac-
tion conditions to better support splicing. After 15 or
60 minutes as noted, 2.5 ~l samples were removed and mixed
with 2.5 ~l of the formamide/dye mix. Samples were ana-
lyzed on denaturing gels containing 4% (19:1) acrylamide:
bisacrylamide and 7M urea in 0.4 X TBE (TBE is 89 mM Tris,
89 mM boric acid, 0.025 mM EDTA). Electrophoresis was
carried out at 30-60 watts using 0.4 X TBE as running
buffer. Gels were exposed to Kodak XOMAT XAR-5 film.
For gel purification of 32P-labelled, precursor RNAs,
transcription reactions were scaled up 2.5- to 10-fold and
incubated 1-2 hours at 37-C. In some cases, the concen-
tration of each NTP was increased to 2.5-3 mM in an
attempt to reduce self-splicing during the transcription
reaction and thereby maxi~ize the recovery of full-length
transcripts. An equal volu~e of formamide/dyes was added
to the completed reaction and t~e entire reaction was
loaded onto a denaturing gel as described above. After
visualization by autoradiography, the regi~n of the gel
containing the full-length ~ranscript was excised and
placed in 0.5-1 ml 0.5 M ammonium acetate, 1 mM EDTA.
After 12-16 hours at 4C, the eluent was removed and the
RNA precipitated by the addition of 2.5 volumes of etha-
nol. The ~inal RNA pellet was resuspended in 1 mM EDTA,
10 ~M Tris-HCl (pH 7.5) and stored at -20~C.
Transcription using the parent plasmid and ~he modi-
fied constructs as templates was carried out in the pres-
ence of-~32P]CTP to generate 32P-labelled transcripts that
811J138TlTlJTE $HEE~T
wo 92/13070 PCr/~'S9~/~0603
,~J i ~ U ,~
33
could be analyzed for their ability to self-splice
(Fig. 4, 0 min). Full-length ~ranscripts (FL), a slight
amount of IVS RNA (IVS), and additional "intermediate" RNA
products (~), were present for all templates A small
amount of an RNA product of the appropriate length to be
ligated exons (LE~ from the EcoRI run-off transcript (135
nt) as well as from the BamHI run-off transcript (114 nt)
was also visible, and indicated that a limited amount of
splicing could occur under these transcription conditions.
This faint band decreased in intensity with the order
pTETBLU~pTET14~pTET1419 and was not visible in pTET21-24.
From analysis of the resultant RNA products, it is
clear that transcription of the parent plasmid, pTETBLU,
generated transcripts capable of efficient self-splicing.
This is evidenced by an increased amount of ligated exons
15 and 60 minutes after adjusting the conditions to better
support splicing.
By comparison of the amount of ligated exon produced,
it is apparent that transcripts from pTET14 and pTET1419
were still capable of self~splicing, although less effi-
ciently than transcripts from the parent pTETBLU. Both
pTET14 and pTET1419 produced fewer ligated exons than
pTETBLU when shifted to splicing conditions, and of these
two mutants, pTET1419 was ~he least efficient. Under the
same conditions, however, transcripts from pTET21-24 did
not appear to self-splice. No ligated exons were visible
for pT~T21-24 precursors after conditions were altered to
support splicing. The relative observed ability of these
three mutant cnnstructs to self-splice, ~hen, follows the
order expected based on the increasing stability of the
P(-l) stem, i.e., there is a negative correlation between
the strength of the P(-1) stem and the RNA's ability to
self-splice. Moreover, the presence o~ the highly stabil-
ized P(-1) stem in pTET21-24 reduced 1 vitro splicing to
undetectable levels.
Under splicing conditions, a number of RNA products
in addition to the ligated exons were visualized. As
8U~3~TITUTE ~;HF~T
~092/1307() PCT/~'S92/00603
21~2~`~
3~
expected, splicing of the pTETBLU transcript generated a
significan~ amount of the exc~sed IVS RNA in its various
forms (circular and linear IVS and the shortened forms
lacking the 5' 15 or 19 nt). Some of these products were
visible for the mutant transcripts as well, even for
pTET21-24 where no ligated exons were visible. The pres-
ence of these IVS products may reflect the ability of
these mutant ~NAs, which are to various degrees misfolded
at the 5'-splice site due to a stronger than normal P(-1)
stem, to still support hydrolysis at their 3'-splice site
(See Woodson and Cech, Biochemistry 30:2042, 1991).
Although no released 3'-exon was visible, one RNA product
that was greatly enhanced in the mutant RNA lanes (indi-
cated with an asterisk in Fig. 4), was of the appropriate
size to represent the 5' exon-IVS RNA. This 5' exon-IVS
RNA would still be expected to undergo circularization
reactions, producing the linear IVS products (L-15 and
L-l9) seen on the gel. The short RNA indi~ated with an
arrowhead is unidentified. This RNA increased in inten-
sity after the switch to splicing conditions. It alsoseemed to increase in abundance as the ability of the
precursor RNA to self-splice decreased, and thus was most
prominent in the pTET21-24 RNA lanes.
It is clear from the lack of ligated exons in the
pTET21-24 lanes that this mutant was unable to undergo
correct ligation of the two exon products. The apparent
side reactions of the mutant IVS-containing RNAs (e.q.,
the formation of the RNA product labeled with the aster-
isk) when unable to undergo a correct splicing reaction
may be able to be used advantageously. For example, this
"self-destruc~ion" may be bene~icial for IVS-containing
mRNAs that encode toxins where rapid turnover of the
message would further diminish the possibility that a
toxin be produced in the absence of the proper signal.
- 8UB8TITUTE SHEEr
WO92/l307~ PCT/US92/0~60~
~ ~ ~ 'J ~J r~
Exam~le 6: Reactivity of $Plicinq Reaction bv Siqnal R~a
Gel-purified, full-length R~ precursors were sub-
jected to splicing conditions in the absence or presence
of signal RNAs to test the ability of short ~NAs comple-
mentary to the upstream 5' exon sequence to disrupt the
P(-1) structure and thereby allow the active P(l) struc-
ture to form.
Splicin~ reactions using gel-purified precursor RNAs
were carried out by incubating 0.1-0.25 pmole of 32p_
labelled transcription 10 ~1 splicing buffer (200 mM NaCl,
200 ~M GTP, 30 mM Tris-HCl, pH 7.5) in the presence of 0
to 1000-fold molar excess of signal RNAs. After waxming
to 37C, MgCl2 was added to 5 mM to initiate the splicing
reaction. Incubation periods ranged from 10 to 120 min-
utes at 37C, at which times samples were removed andmixed with an equal volume of formamide/dye. Samples were
analyzed on denaturing gels as described above.
If self-splicing were reactivated, more ligated exon
pro~ucts would be expected to be produced in the presence
of these signal RNAs than in their absence. Results of
experiments demonstrating reactivation of the splicing
rea~tion are given for pTET1419 RNA in Fig. 5 and for
pTET21-24 RNA in Fig. 6.
As seen previously in Fig. 5, incubation of pTET1419
RNA under splicing condi~ions in the absence of any signal
RNA generated a small amount of ligated exon product.
With gel-purified transcript, this was again the case
(Fig. 5). It may be that the P(~ tem in pTET1419 RNA
is not stable enough to completely inhibit the formation
of P(l), so a small amount of splicing still occurred.
The amount of ligated exons produced increased, however
when either of two specific signal RN~s was present in
incubation. Even with an extremely low signal-to-
transcript ratio (0.1:1), a slight elevation in the amount
of ligated exons was seen. As the signal-to-transcript
ratio was increased (up to 1000:1), the production of
ligated exons also increased. These experiments showed
8lJB8TlTlJTE SHEET
wog~/~3o7~ PCT/~'S92/0060~
2 1 ~
36
that the ability of pTET1419 RNA to correctly self-splice
and produce ligated exons responds directly to the pres-
ence of a specific signal RNA, and that a significant
level of self-splicing is recovered.
S A similar response to signal RNAs was seen with gel-
purified pTET21-24 RNA (Fig. 6). As noted before, with
pTET21-24 RNA, no ligated exons were visible when the
transcript was incubated alone (see also Fig. 4). This
indicates that the P(-1) stem in pTET21-24 RNA is suffi-
ciently stable to completely inhibit the formation of
P(1). Upon addition of either of two signal RNAs (8S4 or
12S) specific for this transcript, however, ligated exons
are produced. That the 32P-labelled RNA products axe
l~gated exons can be seen by comparing their length to
that of ligated exons produced from pTETBLU`RNA. Splicing
of transcripts produced from EcoRI-digested templates pro-
duced ligated exons of 135 nt in length. Transcripts from
templates linearized with ~HI produced ligated exons
that were correspondingly shorter (114 nt). Thus, even
though the splicing reaction was turned completely "off"
in the pTET21-24 RNA itself, it was still possible to
reactive the splicing reaction with a specific signal RNA.
For the EcoRI runoff transcripts shown on the left of
Fig. 6, there was a second major product tindicated with
a dot) that also seemed to respond to the presence of the
~- signal RNAs. This RNA is shorter than the correctly
ligated exons, and at this time its origin in unknown.
Splicing at an alternative site or a specific breakdown of
the RNA are possibilities.
Example 7: Colony_Color Assay
When grown on LB or B agar plates containing S-bromo-
4-chloro-3-indoyl-~-D-galactoside (X-gal), a chromogenic
substrate of ~-galactosidase, pTETBLU-sontaining colonies
are dark ~lue as expected for a colony producing ~-galac-
tosidase. Since the coding region of the ~-complementa-
tion fragment of ~-galactosidase on pTETBLU is interrupted
8UB8TITIJTE SHE~ET
WO~/13071) PCT/~S92/00603
~1 ~' iJ~ jl
by the Tetrahymena IVS, this RNA must be correctly self-
splicing in order to produce an active ~-fragment. If
self-splicing is not occurring/ stop codons present in all
three reading frames in the IVS would not allow transla-
tion into the downstream portion of the gene. For compar-
ison, a control plasmid (pTETULB) in which the intron-
containin~ HindIII fragment from pTETBLU is inserted into
pTZ19R in the reverse orientation was constructed. For
this control, where no splicing can occur due to the wrong
orientation, the resulting colonies are white.
Theoretically, then, cells containing mutants which
are deficient in splicing should produce lighter blue
colonies, while colonies o~ non-splicing mutants would be
white. Under standard growth condition~, cells containing
pTET1419 and pTET21-24 mutants grew as colonies that were
considerably lighter in color than cells containing the
parent plasmid pTET8LU, but not white. This appears to
indicate that even the strongest non-splicing mutant,
pTET21-24 (as judged by its inability to form ligated
exons ln vitro) is still capable of for~ing the minimal
amount of spliced message necessary to support translation
of a level of an ~-fragment of ~-galactosidase that could
confer blue color to the colonies. Other scientists have
noted ~-galactosidase activity (blue colony color) with
IVS-containing constructs in which self-splicing should
have left the ~-galactosidase message in an untranslatable
~rame ~Been and Cech, Cell 47:207, 1986; Price and Cech,
Science 228:719, 1985). It may be that alternative splice
sites exist.
For a more quantitative determination, ~-galactosi-
dase assays were carried out on plasmid-containing cells
growing in culture. (Miller, Experiments in Molecular
Genetics, Cold Spring ~arbor Labora~ory, Cold Spring
Harbor, N.Y. (1972). For this assay, o-nitrophenyl-~-D-
galactoside (ONPG~ was used as the chromogenic substrate
because its product after cleavage with ~-galactosidase
can be~measured spectrophotometrically. A control plasmid
..
8UE~8TITUTE SHEET
W~92/l307() PCT/~'S92/00603
38
(pTETULB) was constructed in which the intron-containing
~clIII fragment from pTETBLU was inserted into pTZ19R in
the reverse orientation and was used to determine back~
ground levels of spontaneous breakdown of ONPG. In these
experiments, cells containing either the parent plasmid or
the splicing mutants were grown under inducing conditions
(i.e., in the presence of IPTG, a lactose analog). Pro-
duction of active ~-galactosidase in cells containing the
pTET1419 and pTET21-24 splicing mutants was reduced to a
few percent of the parental values, thus indicating that
the changes in the RNA were reflected, not only by a
decrease in the amount of n vitro self-spliclng, but by
a concomitant decrease in the amount of active protein
produced in the E. coli cell.
Use
The responsive RNA molecules of the invention are
useful for producing cells that respond to the presence of
a given virus. In particular, in the case of large-scale
in vitro cell culture for commercial production of bio-
logical compounds, viral contamination is a serious prob-
lem which can lead to destruction of the entire culture.
In many instances there is no way to prevent viral infec-
tion of such cells. The molecules of the invention will
allow creation of cell lines that are resistant to any
given virus in that any cells whiHP LaserJet Series
IIHPLASEII.PRSffect the physiological state or viability
of a particular cell type. In the case of responsive RNA
molecules tha~ are regulated by the ~ormation of a base-
paired domain within a protein-coding region the method
requires construction of a responsive RNA which encodes a
protein which will affect the physiology or viability of
a cell; and identification of an signal RN~ which is
specific to the cell type, i.e., an RNA molecule which
carries a nuclectide sequence that is only present or
accessible in the RNA popula~ion of the cell type which is
to be affected. For responsive RNA molecules regulated by
self-splicing introns the method requires construction of
~3VB8TITIITE ~3HEET
W092/l3n7l~ PCT/~'592/00603
2 ~
a responsive RNA which encodes a protein which will affect
the physiology or viability of a cell. The active protein
must be translated from the spliced message and not the
unspliced message. It also requires identification of a
S signal RNA which is specific to the cell type, i.e., an
RNA molecule which carries a nucleotide sequence that is
only present or accessible in the RNA population of the
cell type which is to be affected.
For example, a responsive RNA can be designed to
specifically kill: virus-infected cells containing viral
RNA and not uninfectPd cells; cells containing mutant RNA
and not cells containing wild type RNA; cells in a parti-
cular tissue or organ and not other kinds of cell in the
organism; and neoplastic or cancer causing cells contain-
ing abnormal RNA within a population of cells and notcells containing normal cellular RNA.
The efficacy of such a responsive RNA in altering the
physiological state of a cell will depend upon the respon-
sive RNA being delivered to the iocation in the cell where
the signal nucleic acid resides; the responsive RNA having
all of the nucleoside sequences required for all the pro-
cesses leading to production of the encoded protein
including splicing, poly-A addition, capping, transport
across the nuclear membrane, and translation initiation;
and the responsive RNA must carry sequence elements which
confer stability to RNA in the nl~cleus as well as the
cytoplasm.
A responsive ~NA molecule can be delivered into a
cell in the form of RNA or in the form of a gene made of
DNA or RNA. Delivery of RNA into a cell can be accom-
plished by needle injection, or by the use of liposomes
including those made of cationic lipids. Delivery of the
responsive RNA in the form of a gene can be accomplished
by the use of a nonvirulent virus. This would require the
insertion of the responsive RNA-encoding gene along with
the transcriptional or replicative signal elements into
the genome of the virus. Retroviruses, polyom~ viruses,
8~1B8TITUTE SHEE~T
~VO 92fl307~ PCT/~S92/00~03
2 i ~92~ 1
and vaccinia virus have been engineered which are capable
of delivering and expressing genes, and other viruses
could be developed and used for this purpose.
Another general method of using a responsive RNA to
control the physiology of a particular cell type involves
a responsive RNA gene integrated into the genome of a
cell. The activation of splicing of the responsive RNA
could be caused by exogenously added polynucleotides.
Other embodiments are within the following claims.
