AMER. ZOOL., 29:557-567 (1989)
Catalytic RNA and RNA Splicing1
CHRISTINE GUTHRIE
Department of Biochemistry and Biophysics, University of California,
San Francisco, San Francisco, California 94143
SYNOPSIS. The capacity of Watson-Crick base-pair complementarity to direct informational transactions basic to gene expression has long been appreciated. Among RNA
molecules, it mediates mRNA-tRNA codon-anticodon pairing and the 16S rRNA-mRNA
Shine-Dalgarno interaction. More recently, we have come to realize that the role of RNA
may transcend that of intermolecular recognition, per se, to include catalysis. Following
the tour-de-force studies of the self-splicing Tetrahymena rRNA precursor, the stage is
now set for the primary role of RNA to be revealed in nuclear pre-RNA splicing, which
is catalyzed by a large ribonucleoprotein (RNP) complex in the cell nucleus, called the
spliceosome. The removal of introns from nuclear pre-messenger RNA (pre-mRNA) shares
fundamental properties with certain RNA self-splicing reactions. It therefore seems likely
that the major catalytic strategies in nuclear pre-mRNA splicing are carried out by the
small nuclear RNAs (snRNAs), which are major constituents of the spliceosome.
INTRODUCTION
Most of us come to the study of molecular biology with certain articles of faith,
foremost among which has been the belief
that all enzymes are comprised of polypeptides. Indeed, it was an appreciation of the
structural sophistication afforded by the
twenty different amino acid side-chains that
prompted the Cambridge school (see Crick,
1958) to argue that only proteins, among
all naturally occurring macromolecules,
could fulfill the awesome requirements for
a biological catalyst and thus occupy the
"bottom line" of the Central Dogma (DNA
-> RNA -» protein). For the next two
decades, the attention of molecular biologists was focused on the synthesis of messenger RNA and its translation into protein.
In the last few years, however, a number
of discoveries in the field of RNA structure
and function have forced a revolution in
our conceptual thinking in two important
areas of biology. At the forefront of these
is the tour-de-force study by Cech and his
colleagues, which has revealed the existence of RNA-based enzymes (Cech and
Bass, 1986). The so-called ribozymes
exhibit the same fundamental characteristics as their conventional protein coun' From the Symposium of Science as a Way of Knowing—Cell and Molecular Biology presented at the Annual
Meeting of the American Society of Zoologists, 2730 December 1988, at San Francisco, California.
terparts and have now been found in a surprisingly broad biological niche (see, e.g.,
Greider and Blackburn, 1987). In addition
to the obvious impact on our mechanistic
understanding of the basic principles of
catalysis, an appreciation of the catalytic
potential of RNA has forced a sweeping
re-evaluation of theories on the origins of
life itself. Specifically, it now seems inarguable that biology began in an RNA world
(see Watson et al., 1987; Weiner and Maizels, 1987).
Finally, the study of catalytic RNA can
inform us about the likely solution to a
major problem of current biological interest, the mechanism by which introns are
removed from messenger RNA precursors
(pre-mRNA). Pre-mRNA splicing, as this
process is called, is known to require the
participation of a large, structurally
dynamic machine, dubbed the spliceosome. Major components include complexes of small RNAs (the small nuclear
RNAs Ul, U2, U4, U5 and U6) and a set
of 7-10 tightly associated proteins (Guthrie and Patterson, 1988); these small nuclear
ribonucleoprotein particles (snRNPs) now
seem likely to mediate the catalytic events
of splicing, by mechanistic strategies we will
consider below.
CATALYTIC RNA: LESSONS FROM THE
TETRAHYMENA RIBOSOMAL RNA
The premier example of an RNA catalyst is the intron within the large ribosomal
557
558
CHRISTINE GUTHRIE
it has been estimated that four to five
hydrogen bonds hold the guanosine in
position for nucleophilic attack (Bass and
Cech, 1984). The second strategy provides
a binding site for the intron/exon junction. This sequence, which is rich in pyrimidines, is held in place by complementary
base-pairing with a purine-rich stretch in
the intron. The use of Watson-Crick complementarity allows an elegantly simple
mechanism for programming the specificity of the cleavage site. Indeed, it has been
shown that the substrate specificity of the
Tetrahymena ribozyme can be altered in
a predictable manner by changing the
sequence of the so-called Internal Guide
FIG. 1. Transesterification mechanisms for RNA Sequence (IGS) within the intron (see, e.g.,
splicing, (a) Self-splicing of the Tetrahymena pre- ZzugetaL, 1986).
rRNA and other precursor RNAs containing Group
As a consequence of having the enzyme's
I introns. (b) Splicing of nuclear pre-mRNA and selfsplicing of precursor RNAs containing Group II active site structured solely from intron
introns. (Wavy line = intron.) Taken from Cech and sequences, the released intron is still catBass, 1986.
alytically active. As shown in Figure 3, the
linear intron (called here the L-IVS, for
RNA (rRNA) subunit from the ciliated Intervening Sequence) can re-fold to align
protozoan Tetrahymena. Pioneering stud- an appropriate cleavage site against the
ies by Cech's group revealed that this intron IGS, releasing a circular intron and a 15can be precisely and efficiently excised in nucleotide piece cleaved from the 5' end
vitro in the complete absence of protein of the intron. After a second round of re(Kruger et al., 1982). Under the appropri- folding, the still shorter intron (called here
ate conditions of ionic strength and diva- the CIVS, missing 4 additional nucleolent cation concentration, the only essen- tides) stops its catalysis because it has run
tial co-factor is guanosine. As outlined out of intramolecular substrate. Cech realin Figure la, the reaction proceeds by ized, however, that this RNA is still catatwo successive transesterification events; lytically active if an appropriate substrate
because there is no net change in the total is added in trans, rendering this RNA an
number of phosphodiester bonds, the reac- enzyme in the truest sense of the word, i.e.,
tion requires no energy. In the first step, it can re-cycle (see Zaug and Cech, 1986).
the 3'-OH of the guanosine cofactor ini- Perhaps as importantly, the demonstration
tiates a nucleophilic attack on the 5' splice that the shortened, excised intron could
junction; guanosine is covalently added to act in trans to elongate an RNA chain in a
the 5' end of the cleaved intron, leaving a sequence-specific manner provided an
free 3'-OH group at the end of the 5' exon. experimental paradigm for progenitor
In the second step, this hydroxyl group RNA replicases (Cech, 1986a). Thus the
initiates a nucleophilic attack on the 3' unique capabilities of RNA—to serve the
splice site; the products of this second phos- dual functions of genetic template and
photransfer reaction are the ligated exons enzymatic replicase—presumably allowed
life to evolve in an RNA-only world (Weiand the free, linear intron.
ner and Maizels, 1987).
Genetic and biochemical studies have
revealed two catalytic strategies used by
A major challenge for the future is the
the Tetrahymena intron. There is a bind- determination of the three-dimensional
ing pocket for guanosine (see Fig. 2), which structure of the Tetrahymena ribozyme, a
is somehow constructed from the three- non-trivial problem given the large size of
dimensional geometry of the folded intron; this intron (~400 nucleotides). A. signifi-
CATALYTIC RNA AND RNA SPLICING
559
FIG. 2. Model for the guanosine binding site in the Tetrahymena intron RNA. The proposed hydrogen
bonds (dotted lines) could involve one or two adjacent nucleotides of the RNA or, at the other extreme, five
different bases, sugars, or phosphates that are brought into close proximity by the RNA secondary and tertiary
structure. Also shown is the nucleophilic attack by the 3' oxygen of guanosine at the phosphorus atom of
the phosphate at the 5' splice site. The resulting transesterification has been proposed to be the first step of
RNA self-splicing. Taken from Cech and Bass, 1986.
cant boost to this endeavor has come from
the surprising realization that introns of
this type are remarkably widespread in
nature. In addition to their presence in
nuclear-encoded rRNA of ciliates, they are
found in mitochondrial and chloroplast
rRNA, in tRNA genes, and in three genes
from the bacteriophage T4 (Shub et ai,
1987)! As shown in Figure 4, the hallmarks
of these so-called Group I introns are a set
of short, conserved primary sequences and
a number of short- and long-range pairings
(Cech, 1989); the diagnostic characteristic
of the latter is the conserved ability to form
Watson-Crick base-pairs despite variation
at the primary sequence level. Extensive
phylogenetic comparisons of rRNAs have
revealed the important principle that
sequence conservation is a reliable indicator of functionally important residues,
while maintenance of helices that play a
more structural role is relatively flexible at
the sequence level (Noller and Woese,
1981).
A final point to be made about Group I
introns is that they are unlikely to be self-
splicing in vivo (and, in fact, only some are
autocatalytic in vitro). It is of course difficult to prove that proteins are required in
an organism like Tetrahymena, where
genetic manipulations are difficult to perform. However, in more tractable systems
such as Neurospora, mutants have been
isolated which are defective in splicing due
to alterations of proteins. A likely role for
such factors is the stabilization of the specific tertiary structures of the intron
required for biological activity. Recent data
suggest that some of the proteins that initially evolved to recognize tRNA may have
been exploited for this purpose (Akins and
Lambowitz, 1987).
OTHER SELF-SPLICING INTRONS
USE A NOVEL INTERMEDIATE
Autocatalytic splicing is not restricted to
Group I introns. Other organellar mRNAs
are interrupted by intervening sequences
with quite distinctive structural characteristics; as shown in Figure 5, these Group
II introns contain six domains which radiate
from a central "core" which is quite con-
560
CHRISTINE GUTHRIE
COM
Pre-rRNA
c
G
uga cu c u c uAAAUAGCAA,,
I • I.I.I.!,! I I
*J
A A A G G G1AG G!U U U C C A U U U
-uga
*1
^
Cucuc
1 Splicing
Ligoted exons
L IVS
c
,
GAAAUAGCAA
i i i
y
A A A G G G : A GG:U u u e
C A uu
u*
Conformotionol change
(translocation)
GAAAUAGCAAUAUUUAC,.
C A
G A
A-U
U-A
A-U
A-U
G A'UA CCu'uU.-,
.' .L-..-I-/G.
, A A A:G G G;A ^
Minor cyclization
CA
G A
A-U
15-mer u-A
G
C A
A
A-U
19-mer u-A
+
A-U
A-U
A-U
A-U
. A-U
G
ucu 0OM-J
rA
A
"UA C C U
oiigonucleotide binding
A AG GGIA G G! U U
Reverse cyclization
UCU-L 1
FIG. 3. Model for self-splicing and other self-processing reactions of the Tetrahymena intron RNA. Lower
case letters, exons; capital letters, intron. The pre-rRNA is shown with the 5' exon paired with the Internal
Guide Sequence. The critical oligopyrimidine binding site within the internal guide sequence is boxed; it is
part of the active site for transesterification. Following splicing, the oligopyrimidine binding site is unoccupied,
so a local conformational change can bring another tripyrimidine sequence into the binding site. The 3'
terminal guanosine residue of the IVS RNA can occupy the guanosine binding site and undergo transesterification to produce a circular IVS RNA (CIVS or CIVS). Following cyclization, the oligopyrimidine binding
site is again unoccupied and available for binding tripyrimidines, such as UCU, which can attack the cyclization
junction and linearize the IVS RNA. Taken from Cech and Bass, 1986.
561
CATALYTIC RNA AND RNA SPLICING
luEQSSnSfc.
k
A
«
UUAUAUAUAUAUAA
I iII I I IiI I I > U
AAUAUAUAUAUAgA
c
.a''
FIG. 4. Secondary structure models of Group I introns according to Michel and Dujon, 1983. Large arrows
designate splice sites. Heavily boxed residues are invariant in Saccharomyces cerevisiae mitochondrial Group I
introns. The arrows above sequences 9L and 2 indicate an established base-pairing interaction between these
two 5-nucleotide long sequences, (a) The fourth IVS in the yeast mitochondrial cytochrome oxidase premRNA; solid triangles point to the sites of splicing-defective mutations in this IVS or the closely related fourth
IVS of the cytochrome b gene, (b) T. thermophda (upper case letters) and T. pigmentosa pre-rRNA IVSs: solid
triangles indicate splicing-defective single-base or double-base mutations. Taken from Cech and Bass, 1986
(see references therein).
served at the primary sequence level
(Michel et ai, 1989). Several years ago it
was shown that at least certain of these
introns can be self-spliced in vitro (see, e.g.,
Peebles et al., 1986). Like Group I intron
removal, splicing of Group II introns is a
two-step transesterification reaction, as
shown in Figure lb. In contrast to Group
I, however, Group II splicing is initiated
by the 2'-OH of an intramolecular nucleotide, an adenosine residue a short distance
upstream of the 3' splice site. As a conse-
562
CHRISTINE GUTHRIE
S6-2
...U*UU»UU*U»
S6-3
«.
E1
l..GUUAUUGUUGUGUUUAUGGACAc
I?
E2
AUUOAOGUAAUAUAAAUAUCGCC..>'
-ccallaiillla'lagcgg
V>
O
«*
S6-4S
*£,
* "V
S6-48-I
V^
S6-4S-2
FIG. 5. Schematic presentation of the core secondary structure of a Group II intron, showing m vitro generated
mutations of Stem 6. Left panel: wild-type intron; arrows mark the 5' and 3' splice sites. Right panel: mutations
affecting the branchpoint or neighboring sequences; altered sequences are indicated by arrowheads. The
branchpoint adenosine is marked with an asterisk. Each of these mutations blocks splicing. Taken from
Schmelzer and Schweyen, 1986.
quence, the bond formed in the first step
of splicing is a 2',5'-phosphodiester bond
that links the 5'-phosphate from the 5' end
of the intron to the 2'-OH of the adenosine, which is itself joined by normal 3',5'phosphodiester linkages to its adjacent
nucleotides. This adenosine residue is thus
said to be in a "branched" structure, and
the resultant product is referred to as a
"lariat," due to its characteristic shape as
a tailed circle (Fig. lb).
T h e adenosine that will form the
branchpoint is always found embedded
within a helical region as an unpaired—or
"bulged"—nucleotide. The importance of
this novel structure has been demonstrated
genetically; for example, as shown in Figure 5, insertion of a complementary residue, which restores base-pairing, strongly
inhibits splicing (Schmelzer and Schweyen,
1986). Apparently this structural motif
achieves the specialized geometry needed
for nucleophilic attack on the 5' splice site.
The second requirement of autocatalytic
splicing is the ability to retain the free 5'
exon that is liberated in the first step so
that it is available for nucleophilic attack
on the 3' splice site in the second step. As
with Group I splicing, the Group II 5' exon
is held in place by intramolecular basepairing to complementary sequences within
the intron (Jacquier and Michel, 1987).
NUCLEAR M R N A SPLICING ALSO
USES "LARIAT" INTERMEDIATES
The particular significance of Group II
autocatalytic splicing derives from the discovery that nuclear pre-mRNA splicing also
proceeds via a lariat intermediate in a twostep reaction (Ruskin et ai, 1984). In contrast to the highly conserved structural elements that reside within Group II introns,
however, the only conserved features of
nuclear pre-mRNA introns are restricted
to short regions at or near the splice junctions. As alluded to in the Introduction,
splicing takes place in association with a
large trans-acting machine called the
spliceosome, which contains the U-snRXPs.
The low information content within premRNA introns is compensated by the participation of the snRNPs, which impart
563
CATALYTIC RNA AND RNA SPLICING
the appropriate three-dimensional architecture to the intron. The short sequences
required in and adjacent to the intron
appear to serve as binding sites for the
snRNPs (Guthrie and Patterson, 1988);
presumably, the binding of the snRNPs
allows the intron to fold into a catalytically
active structure.
In principle, both RNA- and proteinbased interactions can contribute to communication between the various snRNPs
and between them and the intron. In
light of self-splicing reactions, however, it
seems evident that nuclear pre-mRNA
splicing evolved from an RNA machine.
Thus the mechanism can be most directly
approached by seeking to establish the
RNA-mediated aspects of the reaction
pathway. The archetype of the recognition
events mediated by snRNAs is the interaction between the Ul snRNP and the 5'
splice site. This complementary base-pairing interaction has been verified genetically in both mammals (Zhuang and Weiner, 1986) and yeast (Siliciano and Guthrie,
1988; see Fig. 6), using the strategy of
introducing compensatory changes in the
snRNA and the intron; that is, a splicing
defect caused by mutation of a conserved
nucleotide in the 5' splice site can be suppressed by restoring base-pairing potential
via a complementary nucleotide change in
Ul.
As shown in Figure 7, a similar approach
has been used to identify base-pairing
interactions between the U2 snRNP and
nucleotides surrounding the branchpoint
adenosine (Parker et ai, 1987): in yeast
introns this sequence is highly conserved.
Interestingly, it is proposed that this intermolecular interaction results in the bulging of the branchpoint adenosine, creating a structure similar to that found intramolecularly in domain VI of Group II selfsplicing introns (see above). An attractive
notion is thus that the RNA components
of the snRNPs represent "escaped domains" from these autocatalytic RNAs
(Cech, 19866).
If this theory is correct, how can we provide experimental evidence for a direct
function of the snRNAs in catalysis per se?
That is, the foregoing genetic analyses of
A.
B.
D.
pre-mRNA
NITON
MUTATION
SUPPRESSED
none
none
IVS-A5
none
IVS-C1
IVS-A1
—
VlARH.n*Y
+
+
(2X)
(1.5X)
(2.3X)
^ — 3' END YEAST Ul
FIG. 6. Base-pairing between yeast Ul snRNA and
the yeast 5' splice site consensus sequence. (A) The
proposed interaction is shown, with the pre-mRNA
in lightfaced type running from the 5' end (below) to
the 3' end (above). The heavy slash indicates the site
of 5' cleavage. Positions +1 to +6 are conserved in
>85% of yeast introns; positions — 1, —2, and —3 are
drawn in lowercase letters because they are conserved
<60%. The 11 nucleotides at the 5' end of Ul are
shown in boldface type, running from the cap (above)
toward the 3' end (below). (B) The U1 point mutations
are indicated. (C) The intron mutations predicted to
be suppressed by each Ul mutation are listed. (D)
The viability of the U1 point mutations in the absence
of wild-type Ul is indicated, with the increase in doubling time caused by the mutation in parentheses.
Taken from Siliciano and Guthrie, 1988.
Ul and U2 (together with substantial biochemical evidence; see Green, 1986; Padgetts al., 1986) establish their role in intron
recognition, but do not address their possible involvement in bond cleavage/formation. While such evidence is difficult to
come by in the absence of a protein-free
in vitro system, recent studies of two other
snRNAs, U4 and U6, reveal properties
consistent with such a role. Specifically, it
is suggested that U6 comprises the "catalytic core" of the spliceosome and that U4,
which is generally tightly associated with
U6, functions as a negative regulator that
is specifically released only when certain
pre-conditions have been met.
The major evidence supporting this
hypothesis is of two types. First, the identification of the spliceosomal snRNAs from
the yeast Sacckaromyces cerevisiae has
revealed a 1:1 analogy between these five
RNAs and U l , U2, U4, U5 and U6 from
mammals; surprisingly, however, there are
564
CHRISTINE GUTHRIE
a
UACUAAC- snR?O PAIRING
pre-mRNA
7Gwp
I2H>
11*11
-UACUA^A-
I
IGUAUGU-
UACUA
IIII in IIII
IAUGAU
CA
in II
AG|
|QH
-AUGAU
I J9i
v i l d type intron
G U - v i m type mR20
(331
GUGAACU..
u
I
MUTATE UACtMAC 6 O x \
-UA A UA A CA- A256 mlron
- A U Q A U GU- Vila type snR2O
A- A257 nitron
-AUG A U GU- viio type snX20
COMPENSATORY CHANGE M «
-UAAUA A CA-
A256 imron
-AUSIAU GU- U37 suppressor
-UACAA*CA- A257 uilron
-AUGUU GU- U36 suppressor
FIG. 7. Base pairing between the yeast U2 snRNA (snR20) and the sequences surrounding the branch site
(the so-called "UACUAAC box"), (a) shows a model for the base-pairing interaction. The pre-mRNA is shown
as the upper molecule; exon sequences are represented as open boxes and the intron is shown as a line with
the 5' and 3'junction sequences as indicated. The yeast U2 analogue is the lower molecule; the sequence of
a portion is shown, (b) shows base-pairing between the UACUAAC box and snR20 for wild-type as well as
mutant combinations. Intron sequences are shown in normal type, with the UACUAAC mutations produced
in a pre-mRNA shown as bold letters. snR20 sequences are shown as bold type, with the compensatory changes
produced in mutant snR20 molecules shown as outlined letters. The numbers in parentheses represent the
nucleotide positions of the residues shown; for snR20, position 1 is the 5' end of the RNA; for the intron,
the numbers are based on the actin intron, with position 1 being the G at the 5' splice junction. Taken from
Parker etal, 1987.
significant differences in size and structure
in all but U4 and U6 (see Guthrie and Patterson, 1988). While U4 is similar in size,
there is little homology at the primary
sequence level. In striking contrast, U6 is
virtually identical in size and shares 60%
identity over its full length with the mammalian snRNA (Brow and Guthrie, 1988;
see Fig. 8); indeed, with the possible exception of 5S rRNA, U6 snRNA appears to
be the most highly conserved RNA in biology! This suggests that U6 is under severe
size and sequence constraints, consistent
with its presence at the physical and functional center of the splicing machine.
The second source of evidence comes
from studies of the spliceosome assembly
pathway. Based on a variety of experimental approaches it is now generally believed
that the snRNPs assemble onto the mRNA
precursor in a highly ordered, stepwise
process. As summarized in Figure 9,
reversible binding of U1 is followed by the
ATP-dependent addition of U2. U4/U6
and U5 bind next, again in a reaction that
requires ATP. Curiously, this supposedly
"mature" (i.e., completely assembled) spliceosome does not yet undergo the first
nucleolytic step of splicing (5' splice site
cleavage and lariat formation). Rather, U4
must first be released, leaving the other
snRNPs in stable association with the
spliceosome (see, e.g., Cheng and Abelson,
1987). Since the dissociation of U4 from
U6 appears to require the breakage of some
two dozen base-pairs, as shown in Figure
8, this process must be an active event, conceivably requiring the participation of an
ATP-dependent helicase.
Thus we can imagine that a primary role
of U4 is to sequester a catalytically active
domain of U6 until other conditions of
spliceosome assembly have been met;
according to this view, the high stability of
the U4/U6 complex is required to allow
what may be the final proofreading step in
splicing. Verification of this idea may come
by exploiting yeast genetics, which makes
it possible to modulate the stability of the
helical interaction domain. One predicts
that a class of mutants can be designed that
might have an impaired splicing fidelity, in
which the release of U4 occurs prematurely.
WHY SPLICEOSOMES?
Perhaps the most perplexing question to
be answered in this field is why nuclear
565
CATALYTIC RNA AND RNA SPLICING
stem I
60
GUUCGCGAACUAACCCUUCCU--GGACAU UUGGUCAAUUUCAAACAAUACAGAGAU CAUCACCA GUUCCCCUGCAUAAGGAU
GAACCGUUUUACAAA---CAGAUUUAUUUCG--UUUU
CUUCCG—GGACAU CCGAUAAAAUUGGAACGACACAGAGAA GAUUAGCA UGGCCCCUGCGCAAGGAU GACACG
GUUCUUG
mammal GUGCOCG-
CACAAA-UCGAGAAAUGGUCCAAAUOUU
COUCGGCAGAACAU AOACUAAAAUUGCAACCAUACAGAGAA GAEJOAGCA DGGCCCCUGCGCAAGGAO GACACG
CAAAAUCCUCAAGCCUUCCACAUOUU
•CUUCGGCAGCACAU AUACUAAAAUUGGAACCAPACAGAGAA GAUUAGCA UGGCCCCtJGCCCAAGGAO GACACG
S' terminal domain
CAAAI
3' terminal domain
U4:U6 interaction domain
d
,
"uc
U G
C-G
u c
< » u: A
U-A
'
G-C .
A
••
u)
(J
U-A
G-C."
S.
Ay
mammals (fly)
-wc
FIG. 8. (a) Evolutionary conservation of U6 snRNA primary structure. The nucleotide sequence of the yeast
U6 snRNA (snR6) is shown aligned with that of U6 RNA from bean, fruit fly, and mammal. Four domains
are boxed, and sequences contained in U4:U6 intermolecular stems I and II are labelled. Dashes represent
gaps introduced to improve the alignment and dots indicate identical nucleotides. (b) Proposed structure of
the yeast U4/U6 complex; snR6 is shown above yeast U4 (snR14), and adjacent parallel lines represent base
paired stems. The regions elaborated in c and d are boxed, (c) Detailed structures at the 5' end of U6. (d)
Conservation of the U4/U6 interaction. Nucleotides 45-82 of yeast U6 and 1-68 of yeast U4 are shown
arranged in a "Y" structure, reminiscent of 5S rRNA. Substitutions found in mammalian U4 and U6 are
indicated with arrows. Taken from Brow and Guthrie, 1988 (see references therein).
mRNA splicing requires such a complicated machinery, when Group I and Group
II introns attest so eloquently to the potential for RNA self-splicing. A plausible
explanation is provided by the observation
that alternative splicing appears to be a
fundamentally important mechanism for
achieving genetic diversity in eucaryotes.
That is, we are becoming increasingly
aware of instances in which cellular genes
are differentially spliced to produce a family of related gene products. While in some
cases these variations in splicing pattern
are stochastic (i.e., fixed), in others they are
regulated—in a tissue and/or temporally
specified manner (see Breitbart^ al., 1987;
Binghame/a/., 1988). Because Group I and
Group II introns carry the structural
requirements for catalysis internally, they
are poor candidates for regulation. In contrast, by moving splicing information to
the external machinery, the intron is free
to diverge, in sequence, and intron recognition can be differentially regulated.
There is, of course, a price to be paid
for such variation within introns; as constraints for conserved sequences/structures are relaxed, how is the fidelity of
splicing maintained? Great precision is
demanded of each splicing event, since an
error of even a single nucleotide will generally have the same consequence as a
genetic frameshift mutation, destroying the
protein encoded by the RNA molecules.
The splicing pathway must have evolved
multiple strategies to achieve the required
accuracy. The discovery of such strategies
poses one of the most exciting challenges
to those who would study splicing. We have
already suggested one example of how an
ATP-dependent process could be used for
such an end, in achieving the timed release
of U4 from the spliceosome; indeed, NTP
hydrolysis appears to be a hallmark of
proofreading mechanisms (Hopfield,
1974), which appear to be used in translation (Thompson, 1988) and in other suspected biological clocks (Selick el al, 1987).
In the meantime, at least one general solution seems apparent: the employment of a
566
CHRISTINE GUTHRIE
} RNA 3, 5, 7, I,
and John Moore for providing this valuable forum. I am particularly grateful to
Tom Cech, whose work provided many of
the figures. I owe many thanks to Judy Piccini for her help in the preparation of this
manuscript. Cited work from this laboratory was supported by grants from the NIH
(GM21119) and the NSF (DMB-869 3926).
REFERENCES
\ A2-2
Fie. 9. Proposed spliceosome assembly pathway. PremRNA is indicated as boxes (exons) and a line (intron).
The snRNPs are indicated as ovals labelled with the
snRNA contained therein. The "?" indicates that one
or more unknown factors are required in addition to
Ul for early assembly steps. The designations B, A21, etc. refer to distinct spliceosomal complexes that
can be resolved by gel electrophoresis (see Cheng and
Abelson, 1987). RNA3, etc. are genes with temperature-sensitive defects in splicing. Taken from Vijayraghavan and Abelson, 1989.
complex, multi-component machine—
which must be assembled in an ordered,
highly cooperative pathway—appears to be
a common strategy in macromolecular processes requiring high biological fidelity
(Echols, 1986).
ACKNOWLEDGMENTS
I thank myy colleagues
in "the RNA
g
ki
h
World" for making these remarks possible
Akins, R. and A. Lambowitz. 1987. A protein
required for splicing group I introns in Neurospora mitochondria is mitochondrial tyrosyl tRNA
synthetase or a derivative thereof. Cell 50:331 —
345.
Bass B. and T. Cech. 1984. Specific interaction
between the self-splicing RNA of Tetrahymena
and its guanosine substrate: Implications for biological catalysis by RNA. Nature 308:820-826.
Bingham, P., T.-B. Chou, I. Mims, and Z. Zachar.
1988. On/off regulation of gene expression at
the level of splicing. Trends Genet. 4:134-138.
Brow, D. and C. Guthrie. 1988. Spliceosomal RNA
U6 is remarkably conserved from yeast to mammals. Nature 344:213-218.
Breitbart, R., A. Andreadis, and B. Nadal-Ginard.
1987. Alternative splicing: A ubiquitous mechanism for the generation of multiple protein isoforms from single genes. Ann. Rev. Biochem. 56:
467-495.
Cech, T. 1986a. A model for the RNA-catalyzed
replication of RNA. Proc. Natl. Acad. Sci. U.S.A.
83:4360-4363.
Cech.T. 1986*. The generality of self-splicing RNA:
Relationship to nuclear mRNA splicing. Cell 44:
207-210.
Cech, T. 1989. Conserved sequences and structures
of group I introns: Building an active site for
RNA catalysis—a review. Gene. 73:259-271.
Cech, T. R. and B. L. Bass. 1986. Biological catalysis
by RNA. Ann. Rev. Biochem. 55:599-629.
Cheng, S.-C. and J. Abelson. 1987. Spliceosome
assembly in yeast. Genes Dev. 1:1014-1027.
Crick, F. H. C. 1958. On protein synthesis. Symp.
Soc. Exp. Biol. 12:138-163.
Echols, H. 1986. Multiple DNA-protein interactions
governing high-precision DNA transactions. Science 233:1050-1056.
Green, M. R. 1986. Pre-mRNA splicing. Ann. Rev.
Genet. 20:671-708.
Greider, C. and E. Blackburn. 1987. The telomere
terminal transferase of Tetrahymena is a ribonucleoprotein enzyme with two kinds of primer
specificity. Cell 51:887-898.
Guthrie, C. and B. Patterson. 1988. Spliceosomal
snRNAs. Ann. Rev. Genet. 22:387-419.
Hopfield, J. 1974. Kinetic proofreading: A new
mechanism for reducing errors in biosynthetic
processes requiring high specificity. Proc. Natl.
Acad. Sci. U.S.A. 71:4135-4139.
Jacquier, A. and F. Michel. 1987. Multiple exon-
CATALYTIC RNA AND RNA SPLICING
binding sites in class II self-splicing introns. Cell
50:17-29.
Kruger, K., P. Grabowski, A. Zaug, J. Sands, D.
Gottschling, and T. Cech. 1982. Self-splicing
RNA: Autoexcision and autocyclization of the
ribosomal RNA intervening sequence of Tetrahymena. Cell 31:147-157.
Michel,F. andB. Dujon. 1983. Conservation of RNA
secondary structures in two intron families
including mitochondrial-, chloroplast- and
nuclear-encoded members. EMBO J. 2:33-38.
Michel, R., R. Umesoho, and H. Ozeki. 1989. Comparative and functional anatomy of group II catalytic introns—a review. Gene. (In press)
Noller, H. and C. Woese. 1981. Secondary structure
of 16S ribosomal RNA. Science 212:401-411.
Padgett, R. A., P. J. Gradowski, M. M. Konarska, S.
Seiler, and P.A. Sharp. 1986. Splicing of messenger RNA precursors. Ann. Rev. Biochem. 55:
1119-1150.
Parker, R., P. Siliciano, and C. Guthrie. 1987. Recognition of the TACTAAC box during mRNA
splicing in yeast involves base-pairing with the
U2-like snRNA. Cell 49:229-239.
Peebles, C , P. Perlman, K. Mecklenburg, M. Petrillo,
J. Tabor, K. Jarrell, and H.-L. Cheng. 1986. A
self-splicing RNA excises an intron lariat. Cell
44:213-223.
Ruskin, B., A. Krainer, T. Maniatis, and M. Green.
1984. Excision of an intact intron as a novel lariat
structure during pre-mRNA splicing in vitro. Cell
38:317-331.
Schmelzer, C. and R. Schweyen. 1986. Self-splicing
of group II introns in vitro: Mapping of the branch
point and mutational inhibition of lariat formation. Cell 46:557-565.
Selick, H. E.,J. Barry, T. A. Cha, M. Munn, M. Nakanishi, M.-L. Wong, and B. M. Alberts. 1987.
567
Studies on the T4 bacteriophage DNA replication system. In R. McMacken and T. J. Kelly (eds.),
DNA replication and recombination. UCLA Sym-
posium on Molecular and Cellular Biology, New
Series, pp. 183-214. Alan R. Liss, New York.
Shub, D., M.-X. Xu, J. Gott, A. Zeeh, and L. Wilson.
1987. A family of autocatalytic group I introns
in bacteriophage T4. Cold Spring Harbor Symp.
Quant. Biol. 52:193-200.
Siliciano, P. and C. Guthrie. 1988. 5'splice site selection in yeast: Genetic alterations in base pairing
with Ul reveal additional requirements. Genes
Dev. 2:1258-1267.
Thompson, R. 1988. EFTu provides an internal
kinetic standard for translational accuracy. Trends
Biochem. Sci. 13:91-93.
Vijayraghavan, U. andj. Abelson. 1989. Pre-mRNA
splicing in yeast. Nucleic Acids and Molecular
Biology. (In press).
Watson, J. D., N. H. Hopkins, J. W. Roberts, J. A.
Steitz, and A. M. Weiner. 1987. The origins of
life. In J. Watson et at. (eds.), Molecular biology of
the gene, pp. 1098-1124. Benjamin/Cummings,
Menlo Park, CA.
Weiner, A. and N. Maizels. 1987. tRNA-like structures tag the 3' ends of genomic RNA molecules
for replication: Implications for the origin of protein synthesis. Proc. Natl. Acad. Sci. U.S.A. 84:
7383-7387.
Zaug, A., M. Been, and T. Cech. 1986. The Tetrahymena ribozyme acts like an RNA restriction
endonuclease. Nature 324:429-433.
Zaug, A. and T. Cech. 1986. The intervening
sequence RNA of Tetrahymena is an enzyme.
Science 231:470-475.
Zhuang, Y. and A. M. Weiner. 1986. A compensatory base change in Ul snRNA suppresses a 5'
splice site mutation. Cell 46:827-835.