Authors

John F. Atkins, Eccles Institute of Human Harry F. Noller, University of California—Genetics, University of Utah Santa Cruz

Andrew Ellington, Indiana University Stephen M. Pasquale, BSCSB. Ellen Friedman, Consultant Richard D. Storey, The Colorado CollegeRaymond F. Gesteland, Eccles Institute of Olke C. Uhlenbeck, University of

Human Genetics, University of Utah Colorado, BoulderAlan M. Weiner, Yale University

Biological Sciences Curriculum Study (BSCS)Pikes Peak Research Park5415 Mark Dabling Blvd.

Colorado Springs, Colorado 80918-3842

Copyright © 2000 by the BSCS. All rights reserved. No part of this work may be reproduced or transmitted inany form or by any means, electronic or mechanical, including photocopying and recording, or by any infor-mation storage or retrieval system, without permission in writing. For permissions and other rights under thiscopyright, please contact the BSCS, Pikes Peak Research Park, 5415 Mark Dabling Blvd., Colorado Springs,CO 80918-3842.

This material is based on work supported by the National Science Foundation under grant No. 9652921. Anyopinions, findings, conclusions, or recommendations expressed in this publication are those of the authorsand do not necessarily reflect the views of the granting agency.

BSCS Development TeamProject LeadershipJoseph D. McInerney, Principal InvestigatorStephen M. Pasquale, Project Director

Project CoordinationRose Johnson, Production AssistantDee Miller, Production AssistantBarbara C. Resch, Editor

Project SupportLinda Ward, Senior Executive AssistantKevin Andrews, Graphic ArtistLisa M. Chilberg, Graphic Designer

BSCS Administrative StaffTimothy H. Goldsmith, Chair, Board of DirectorsRodger W. Bybee, Executive DirectorSusan Loucks-Horsley, Associate Executive DirectorMarcia Mitchell, Financial Officer

Field-Test TeachersDavid Allard, Texarkana College & Texas

A&M University-Texarkana, TexasGail Carmack, The Science Academy of

Austin at LBJ High School, TexasAlix Darden, The Citadel, South CarolinaDenise M. Foley, Chapman University, CaliforniaDavid A. McCullough, Wartburg College, IowaJohn Ohlsson, University of Colorado, BoulderLeLeng P. To, Goucher College, Maryland

Design and layout by Angela Greenwalt

Illustration CreditsFigure 1.4a: © A. Barrington Brown, ScienceSource/Photo Researchers; Figure 1.4b: Courtesyof Harry F. Noller; Figure 1.4c: Genes and Genomesby Maxine Singer and Paul Berg, published byUniversity Science Books, Sausolito, CA; Figure2.3: John T. Ohlsson, University of Colorado, Boulder; Figure 3.6: stopwatch © 1999–2000www.barrysclipart.com; Figures 3.9, T3.1, andT3.2: Gerald F. Joyce and Kathleen E. McGinness,The Scripps Research Institute; Figure 4.1b:Noller, H.F. (1993). On the origin of the ribosome.In Gesteland, R.F., & Atkins, J.F. (Eds.), The RNAworld (p. 141). Cold Spring Harbor, NY: ColdSpring Harbor Laboratory Press.

Figures 8 and 1.6 are adapted from Gesteland,R.F., & Atkins, J.F. (1993). The RNA world. ColdSpring Harbor, NY: Cold Spring Harbor LaboratoryPress.

Evaluation Form for Bringing RNA into View: RNA and Its Roles in Biology

Your feedback is important. After you have used the module, please take a few minutes and return this formto BSCS, Attn: RNA, 5415 Mark Dabling Blvd., Colorado Springs, CO 80918-3842.

1. Please evaluate the Faculty Background by marking this form and providing written comments or sugges-tions on a separate sheet.

Sections Used not helpful very helpful

Overview 1 2 3 4 5

Background on RNA 1 2 3 4 5

2 . P l ease evaluate the activities by marking this form and providing written com ments or su g g e s t ions on aseparate sheet. Rate the activities for their eff e c t i ve ness at teaching concepts of RNA and its role in biol o g y.

Activity 1: RNA Structure: Tapes to Shapes 1 2 3 4 5

Activity 2: RNA Catalysis 1 2 3 4 5

Activity 3: RNA and Evolution 1 2 3 4 5

Activity 4: RNA Evolution in Health 1 2 3 4 5and Disease

3. What are the major strengths of this module?

4. What are the major weaknesses of this module?

5. Please rate the overall effectiveness of this module: not effective very effective

1 2 3 4 5

6. Please provide a description of the classes in which you used this module: (circle response)

College: 2 year 4 year High school: grade 9 10 11 12

freshman sophomore junior senior Level of class: basic honors 2nd year

How many students used the module? _______ How many students per class? _____ _ _ _ _

Ethnicity (approximate % of minorities): ____________________________________ _ _ _ _ _ _ _ _ _

Description of school: ____________________________________________________ _ _ _ _ _ _ _ _ _

College: liberal arts science High school: urban suburban rural

7. Have you used BSCS materials before? ❒ yes ❒ no

8. Please provide your name and contact information below:

Name__________________________________________________________________________

School_________________________________________________________ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _

Mailing address__________________________________________ ❒ home ❒ work

___________________________________________________________________ _ _ _ _ _ _ _ _ _ _ _ _

______________________________________________________________________ _ _ _ _ _ _ _ _ _

Phone_________________________________________________ ❒ home ❒ work

FAX__________________________________________________ ❒ home ❒ work

E-mail address______________________________________________________ _ _ _ _ _ _ _ _ _ _ _ _ _

Was the address on your mailing label correct? ❒ yes ❒ no

Faculty BackgroundOverview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Background on RNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5References and Related Literature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Annotated Faculty PagesAbout the Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Activity 1 RNA Structure: Tapes to Shapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Activity 2 RNA Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45Activity 3 RNA and Evolution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Activity 4 RNA Evolution in Health and Disease. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

Copymasters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

Templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

Student PagesActivity 1 RNA Structure: Tapes to Shapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133Activity 2 RNA Catalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145Activity 3 RNA and Evolution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161Activity 4 RNA Evolution in Health and Disease. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

The impetus for developing Bringing RNA into Viewfor college biology classes is the recent and rapidgrowth in knowledge of the structures and diversefunctions of RNA molecules. First described in1947 as a cellular constituent involved in proteinsynthesis, RNA has since been shown to play sever-al other essential roles in gene expression, includinggenome maintenance, processing and editing of pri-mary transcripts, and localization of proteins with-in the cell. New RNAs continue to be discoveredperforming unexpected tasks in the cell.

Perhaps more than any other single discovery aboutnucleic acids since Watson and Crick’s elucidationof the DNA double helix, the finding that RNA (andDNA) has catalytic ability has expanded our view ofthese molecules’ potential, as evolutionary progeni-tors and contemporary biochemical players. Thenotion that an RNA world phase occurred early inprebiotic evolution (deriving both its informationencoding and catalytic functions from RNA mole-cules) has inspired a burst of research into themolecular origins of life and the biochemical poten-tial of nucleic acids.

P a rt i c u larly re vealing of the impre s s i ve bioc he m i c a lp otential of these molecules are recent studiese x p anding on the 30-yea r- old observa t ion of Dar-w i n i an sel e c t ion and evol u t ion of nucleic acids int he test tube. Te c h n ological advances of the last 15y ears have greatly increa s ed the pow er and ver s a t i l-ity of such in vitro e x per i ments to create and sel e c tt a l e n t ed nucleic acid molecules, some of which havebegun to re veal RNA’s potential for true sel f - re pl i c a-t ion, for synthesizing its own mon omer building

b l ocks, and for practical use in combating viral an dm i c robial infections as well as genetic disorders.

BSCS sel e c t ed the topic of this module, RNA and itsrole in biol o g y, as a key area for synthesizing the s ei m p ort ant new re s ea rch findings into the fundame n-tal concepts of col l e g e - l e vel biol o g y. In addition, thistopic off ers a useful op p ortunity to shift stude n t sf rom focusing on isola t ed facts to approaching biol o-gy conceptually; in short, the module helps stude n t sthink about biological proc e s s e s .

To de vel op this module, biologists at the Biol o g i c a lSciences Curriculum Study in Col or a do Springs an dt he Eccles Institute of Human Genetics at the Uni-versity of Utah in Salt Lake City wor k ed with ane x t ernal advisory committee of scientists and ed u c a-tors, plus a variety of college faculty who con d u c t edf i eld tests of the module. This process ide n t i f i ed thefol l owing major concepts for the module:

• Nucleic acids (DNA and particularly RNA)have two major functions: as informationalmolecules and as biochemical catalysts.

• The sequence of monomers in RNA dictates itsthree-dimensional structure and, consequently,its function.

• Molecules can be subject to natural selection.• E v ol u t ion requ i res an iter a t i ve process of

molecular replication, mutation, and selection.• Modern roles for RNA suggest its major role in

the origin of life.• Studies of the origin of life and its evol u t ion wor k

with su b s t antial data, such as modern observa t ionof naturally oc c urring ribozymes (RNA catalysts)and in vitro m ol e c u lar sel e c t ion exper i me n t s .

• M o dern roles for RNA include serving as ag e n ome to a variety of viruses and smaller viroi d s .

• Most RNA-based viruses and viroids are signif-icant pathogens of humans, animals, and agri-culturally important crop plants, and as suchthey have a major social and economic impact.

The module provides background materials for fac-ulty and a set of four educational activities for stu-dents. Background on RNA (in this section) providesyou with an update on RNA research in a form thatis accessible to busy faculty. This material is for yourown use. It may extend your knowledge and thus behelpful for teaching the activities, but it is notessential to teaching the activities. The activities areinquiry-based explorations that offer you an alterna-tive to lecture and stimulate student interest andresponsibility for learning.

Figure 1 shows the layout of the materials. Themodule features four core classroom and laboratoryactivities that explore RNA’s structure and function,RNA catalysis, RNA replication and evolution, andRNA’s role in health and disease.

Notice that each activity appears in two sections,Annotated Faculty Pages and Student Pages. The Stu-dent Pages consist of introductory text and For YourInformation essays that provide context and elabo-ration for the activity. Detailed protocols are provid-ed for students, and Challenge Questions stimulatestudent thought and synthesis of ideas. The Student

Pages and the Copymasters and Templates may bephotocopied for classroom use.

The Annotated Faculty Pages contain the studenttext (bold type) plus annotations for the faculty(regular type). The annotations contain suggestionsand hints for teaching each activity, answers toChallenge Questions, optional extension exercises,and reagent preparation instructions.

A summary of the activities is provided in Figure 2.Time and resources may not permit you to teach allfour of the activities. However, we recommend thatyou teach Activity 1 (RNA Stru c t u re: Tapes toShapes) before the other activities, to ensure thatstudents understand the structural basis for theRNA functions explored in the later activities.

As your students proceed through the module, werecommend that you encourage them to ask ques-tions, seek outside resources, and be aware of theway in which science attempts to understand natu-ral processes. For example, call students’ attentionto citations in essays so that they begin to appreci-ate the significance of discussions that are based onprimary scientific data rather than hearsay.

F i g u re 1 T he module at a glance. Bringing RNA into Vi e w con-sists of three main com p onents (Faculty Backgro u n d, A n n o t a t e dFaculty Pages, and Student Pages) as well as su p p ort mater i a l s .

Faculty Background • Overview• Background on RNA• References

Annotated Faculty Pages• Instructions and background for Activities 1–4

Copymasters and Templates• Handouts and masters for Activities 1–4

Student Pages• Student materials for Activities 1–4

Activity 1 RNA Structure: Tapes to ShapesStudents apply rules of base pairing and folding toconstruct physical models of RNAsequences. Theyuse their models to explore structure-function rela-tionships and the effects of mutation.

Activity 2 RNA CatalysisStudents explore catalytic RNA in the laboratoryusing a self-splicing group I intron. Students applythe techniques of in vitro transcription, RNA isola-tion, and acrylamide electrophoresis to study thekinetics of the splicing reaction.

Activity 3 RNA and EvolutionStudents explore the replication of a catalytic RNAin the laboratory using a continuous in vitro system.

Activity 4 RNA Evolution in Health and DiseaseStudents explore the continuing evolution of RNAinthe context of the emerging resistance of bacteriaand viruses to therapeutic agents.

Figure 2 Summary of the student activities.

The classic formulation of the flow of genetic infor-mation during gene expression holds that DNA iscopied both to itself and to RNA, and RNA is thendecoded to synthesize protein. This view was for-malized by Francis Crick in 1968 as the centraldogma of molecular biology (Figure 3).

Several assumptions are inherent in this traditionalview: Inform a t ion flow is unidire c t ional, asexpressed by the one-way arrows; only DNA is atemplate that can be replicated; nucleic acid codinginformation must ultimately be translated to pro-tein form if working catalysts are to result; andDNA, as the ultimate molecule that encodes genet-ic information, likely preceded RNA during the for-mation and subsequent evolution of biomolecules.

This straightforward view of gene expression hasgiven way to a more detailed understanding ofi n form a t ion flow in biol o g y. Mol e c u lar gene t i cresearch, particularly during the last 15 years, pro-vides exciting new insights that reveal the originalformulation of the central dogma to be incomplete.Some of the assumptions listed above have beenshown to be limiting. For example, the study ofviruses having an RNA genome revealed that RNA,like DNA, can serve as a primary information-encoding molecule. Also, the discovery in 1970 ofreverse transcriptase, an RNA virus-encoded pro-

tein catalyst that copies RNA-based informationinto DNA form, revealed that information flow inthe biological world is in fact a two-way streetbetween DNA and RNA.

Watson and Crick appreciated DNA’s potential to bea template for its own replication as soon as theysolved the double helical structure of the molecule.We now realize that RNA molecules likewise canplay template roles for their own replication. In theprocess known as t e m p l a t e - d i rected nucleic acidre p l i c a t i o n, an RNA sequence can serve as ani n form a t i ve scaffold on to which com pl e me n t a rybases align to produce a complementary strand(Figure 4a). Many viruses infecting plants and ani-mals, such as plant viroids and polio virus, employa two-step replication strategy in which their single-stranded RNA genome, acting as a template, is ini-tially transcribed by a replicase enzyme to yield acomplementary RNA strand. This molecule in turnserves as a template whose transcription regeneratescopies of the original genome strand, thereby effect-ing replication (Figure 4b).

The discovery of RNA’s informational and templatefunctions expanded our ideas of the early evolutionof information storage and expression mechanisms.However, the most dramatic and unexpected dis-covery to influence these ideas came early in the1980s, when scientists in two different laboratoriesindependently discovered that RNA has catalyticability. Thomas Cech and his group at the Universi-ty of Colorado were studying the splicing of largeribosomal RNA (rRNA) precursors in the protozoonTe t r a h y m e n a. These scientists sere n d i p i tou s l yobserved that the RNA precursor spontaneouslyFigure 3 Central dogma of molecular biology.

changed size, becoming smaller after incubating ina protein-free buffer solution containing only Mg2+

ions. Realizing how unusual this result was, theypursued it in earnest. In a series of brilliant experi-ments, Cech and his coworkers proved that thisRNA, the “group I intron,” has the inherent abilityto catalyze its own excision from the RNA precur-sor. This RNA intron catalyzes its own self-splicingwithout the aid of any protein.

Meanwhile, in the lab of Sidney Altman at Yale Uni-versity, researchers were continuing a long series ofbiochemical experiments to characterize an enzymeactivity in Escherichia coli, called RNase P, thattrimmed the 5’-end of a transfer RNA (tRNA) pre-c ur s or. After exhaustive pur i f i c a t ion, the activeenzyme was found to contain both a protein com-ponent and an RNA molecule. The conventionalprejudice was that the protein must act as the“enzyme.” But any effort to remove the RNA com-ponent eliminated the catalytic ability. Again, pur-suit of an une x pe c t ed finding and diligentexperimentation showed that the RNA moleculeitself was sufficient to catalyze the trimming reac-tion. The RNA, in this case, was a catalyst actingnot on itself but on another RNA.

Here were two different RNA molecules that cat-alyzed biochemical reactions, just like proteins.Nobel Prizes followed for both Cech and Altman—

a new era had begun. Since this discovery, sevennaturally occurring classes of ribozyme have beenrecognized, and hundreds of specific examples havebeen identified in a wide variety of organisms (Fig-ure 5). The true prevalence of ribozymes in con-temporary cells is unclear, but it is quite likely thatnew examples will be discovered.

Interestingly, the independent discovery of RNAcatalysis by Cech and Altman occurred in researchfields that were relative backwaters at the time. Inthe early 1980s, the hot research fever was in thefirst thrust of exploiting the new molecular biologyt e c h n ologies (for example, re s t r i c t ion enzyme s ,cloning, and sequencing) to explore emerging top-ics such as split genes, tumor viruses, and onco-genes. The unlikely discovery of catalytic RNAs inthis setting is testimony to the importance in sci-ence of observing and attending to unexpected andunusual results—and being willing to pursue themdespite prevailing fashions.

In addition to destroying the orthodox assumptionthat only proteins function as biological catalysts,the discovery of RNA’s chemical versatility led to adramatic change in how scientists view the likelysequence of molecular events during early evolu-tion. Taken together, RNA’s informational, template,and catalytic abilities led to the hypothesis that

Figure 4 Viral RNA replication. The replication cycle of viruses typically consists of two steps. In (a), the positive-sense RNAgenome is transcribed into many copies of a complementary negative-sense strand by RNA polymerase. In (b), the newly formedcomplementary strands become templates for synthesizing many copies of the original positive-sense genome. The new genomicstrands may be translated as they are being synthesized.

RNA evolved, before the appearance of DNA or pro-tein, in an RNA world. During this proposed phaseof evolution, RNA is assumed to have provided boththe coding and the catalytic abilities necessary andsu fficient to initiate biological evol u t ion. Spe c i f i c a lly,if RNA’s catalytic abilities during this time extendedto its own self-replication, then molecular evolutiona u tomatically would have start ed as ran domly va r iantRNAs were naturally selected on the basis of ever-more efficient replication ability. This property ofu n a i ded sel f - re pl i c a t ion, although an essentialassumption of the RNA world h y p ot hesis, remains tobe de m on s t r a t ed. No nucleic acid sequence possess-ing RNA re plicase catalytic activity has yet beenfound in nature, but re s ea rc hers have taken the firststeps tow a rd creating such a sel f - re pl i c a tor RNA mol-ecule in vitro. In a la t er section, we examine pow er-ful new in vitro t e c h n ologies for directing thee v ol u t ion of nucleic acid molecules toward this andmany other functions, and we discuss their impl i c a-t ions for re s ea rch in both evol u t ion a ry biology an db iomedical science.

The coding, template, and catalytic abilities of RNAalso led to the current, expanded formulation of thecentral dogma, one that incorporates the RNA worldhypothesis and places information flow in an evolu-tionary, historical context (Figure 6).

The discovery of ribozymes clearly was the catalystfor the current intense interest in the RNA worldhypothesis. But the idea that RNA perhaps was thefirst genetic molecule is not new: Francis Crick andLeslie Orgel first proposed the possibility in 1968.Indeed, today’s RNA world hypothesis builds upona long history of research findings related to the ori-gin of life and molecular evolution (Figure 7).

Next, we discuss RNA shapes and the growingappreciation of their complexity and contribution tothe diverse functions of RNA.

A key principle of molecular structure is that shapedetermines function. For biological polymers com-posed of multiple subunits, the fundamental deter-m i n ant of shape is the linear inform a t ionrepresented in the sequence of individual subunits;the molecule is in effect an information tape. A spe-cific three-dimensional shape emerges in the mole-cule as thermodynamically favored physical

Figure 5 Naturally occurring ribozymes.

Figure 6 Modified central dogma.

Class

group I intron

group II intron

RNase P

hammerhead

hairpin

hepatitis Delta virus(HDV)

Neurospora VS RNA

Size

large: 413 NT in Tetrahy -mena thermophilia

large: 887 NT in yeastmitochondria

large: 350–410 NT

small: 31–42 NT (enzymestrand can be 16 NT)

small: 50 NT (minimumsequence)

84 NT (required)

881 NT (164 sufficient)

Reaction

intron excision

intron excision

hydrolytic endoribonuclease

RNA cleavage

RNA cleavage

RNA cleavage

RNA cleavage

Source

eukaryotes, eubacteria, andviruses

eukaryotic organelles andeubacteria

RNA subunit of eubacterialRNase P

viral satellite RNA in plants,viroids, and newt satelliteDNA

(-) strand satellite RNA oftobacco ringspot virus

HDV

Neurospora mitochondria

1800s The idea of spontaneous generation, belief in the ongoing creation of living organisms from nonlivingmaterials, persists.

1828 Friedrich Wöhler synthesizes urea in the laboratory, eliminating “vital-force” as an agent in the synthe-sis of organic chemicals.

1859 Charles Darwin publishes On the Origin of Species, in which he proposes a theory of biological evolu-tion based on the mechanism of natural selection.

1864 Louis Pasteur experimentally disproves ongoing spontaneous generation by showing that when liquidsare boiled in order to kill any microorganisms present, and are subsequently kept sterile, no organismsappear in them.

1924 Alexander Oparin employs geological evidence in proposing that the early earth had a reducing atmos-phere lacking oxygen and that the first single-celled organisms might have arisen from simple organicmolecules present in this early atmosphere: in effect a restricted version of spontaneous generation.

1929 Biochemist J.B.S. Haldane proposes that life might have arisen on earth when Oparin’s early atmos-phere was subjected to energy in the form ultraviolet radiation and heat from the cooling earth.

1953 Graduate student Stanley Miller provides experimental support for the Oparin-Haldane hypothesis bymixing gases of the “primitive atmosphere” in a glass reaction vessel and subjecting them to electriccurrent for one week; amino acids are formed de novo.

Biologist James Watson and physicist Francis Crick publish their findings on the structure of DNA.

1961 Marshall Nirenberg and his colleagues begin their five-year project of cracking the genetic code by dis-covering that a messenger RNA made up entirely of the base uracil can be translated into a peptidemade up entirely of the amino acid phenylalanine.

1962 Watson and Crick share a Nobel Prize for their work on DNA structure.

1967 Sol Spiegelman demonstrates the replication and evolution of RNAmolecules in the test tube.

1968 Francis Crick and Leslie Orgel propose that the first information molecule was RNA; Crick advances thecentral dogma of molecular biology.

1970 David Baltimore and Howard Temin independently discover reverse transcription of viral RNAgenomesinto DNA.

1972 Harry Noller proposes a role for ribosomal RNAin the translation of messenger RNA into protein.

1982–83 Thomas Cech and Sidney Altman independently discover the first examples of catalytic RNAmolecules:ribozymes.

1986 Walter Gilbert coins the term “RNAworld” to describe the hypothesized time during which RNAwas theprimary informational and catalytic molecule.

Kary Mullis develops the polymerase chain reaction (PCR) technology that allows rapid copying of DNAand RNA sequences in vitro and enables large-scale laboratory studies of molecular evolution.

1989 Cech and Altman share a Nobel Prize for their discovery of catalytic RNA.

Gerald Joyce develops the technique of in vitro amplification and selection of RNA(that is, directed evo-lution) using the PCR technique.

1992 Noller presents evidence for the catalytic involvement of the 23S rRNA in peptide bond formation.

1993 Mullis receives a Nobel Prize for his development of the polymerase chain reaction.

Joyce further develops in vitro RNA amplification and evolution experimental procedures.

1 9 9 5 Jack Szostak’s laboratory takes the first steps toward the in vitro selection of a self-replicating RNA m o l e c u l e .

1998 David Bartel and Peter Unrau use in vitro selection to demonstrate that RNAcan catalyze the formationof individual nucleotides.

Figure 7 Milestones in the evolution of the RNA world hypothesis.

interactions, typically noncovalent, occur betweencompatible subunits located at a distance from oneanother. Indeed, this important structure-functionprinciple provides the entire rationale for evolution’sinvention of a genetic information-coding strategyalmost 4 billion years ago.

T he import ance of linear inform a t ion for mol e c u la rs h a pe and function was first appre c i a t ed for prot e i n s ,t he first biop ol y mers to be sequ e n c ed. Their co va-lently linked amino acid subunits are now known toi n t eract furt her through a variety of wea k er non co va-lent assoc i a t ions; these include van der Waals forc e s ,h y d rogen bonds, ionic bonds, and hydrophilic inter-a c t ions. These non co valent inter a c t ions of aminoacids, both locally and with more re m ote ne i g h b or st h rough folding of the molecule, give rise to higher-order protein shapes. More basic, local shape el e-ments, such as hydrogen bon d - s t a b i l i zed alpha-helices and beta-sheets, can interact in a variety ofways within the fol ded protein. These higher- order,t h re e - d i me n s ional inter a c t ions are typically stabi-l i zed by hydrophobic and ionic forces to create a va s ta rray of specific protein shapes. We re co g n i ze a va r i-ety of functional sites in proteins that re sult fromt heir shapes: enzymatic active sites, binding poc k e t s ,re g u la tory sites, and domains for prot e i n - prot e i ni n t er a c t ion .

The science community’s recognition of diversity ofshape among nucleic acids developed more slowly,however. DNA’s extended double helix, the firstnucleic acid structure to be revealed, gave no hint ofmore complicated shapes. Only later, when the basesequence and three-dimensional structure of trans-

fer RNA (tRNA) was worked out, was the ability ofnucleic acids to adopt complex shapes confirmed.

The most important determinant of folding andshape in single-stranded nucleic acids, both RNAand DNA, is complementary base pairing via hydro-gen bonding, according to the base-pair rules firstestablished by Watson and Crick. For RNA, pairingof A with U and of G with C is the primary basis forfolding. Even though each RNA molecule in a cellnormally consists of a single continuous strand,RNA molecules frequently contain linear runs ofbases that are complementary to other runs locatedelsewhere in the molecule. This allows the moleculeto fold back and form double-stranded regions with-in itself.

T hese dou b l e - s t r an ded re g ions of the mol e c u l eadopt a helical configuration similar to that found indouble-stranded DNA. (A subtle but characteristicdifference, however, is that helical regions in RNAadopt the A-form geometry, whereas those in DNAare most often B-form.) The helical regions alternatewith more flexible single-stranded regions. Activity1 demonstrates that even an RNA molecule of mod-est length can fold in several possible ways bybringing together different, more or less comple-mentary regions. The degree of match and resultingthermodynamic stability of one structure over analternative determines which form predominates inthe cell.

Until re c e n t l y, it was difficult to de t erm i ne the fol deds t ru c t ures of RNAs, and only a few were know n .Te x t b ooks typically show tRNA as a fol ded stru c t ure ,

Figure 8 Some examples of RNA shapes and structural elements. Molecules consisting of one (a), two (b), or three (c) stem-loopelements. (d) A pseudoknot configuration.

w hereas me s s e n g er RNA (mRNA) is shown as a lin-ear thread lining up to be de co ded. In fact, mRNA isk n own to fold into com plex stru c t ures, and it is on l y“ i roned out” by the passing ribosome assembly dur-ing tran s la t ion. Recent advances in X-ray cry s t a l l o g-raphy and nuclear magnetic re s on ance spe c t ro s cop yh a ve ope ned up the inve s t i g a t ion of RNA stru c t ure ,and the number of re s ol ved stru c t ures is increa s i n gr a p i d l y. Stru c t ural and sequence databases are begin-ning to re veal com m on motifs in RNA. Most su c hm otifs are formed by con ve n t ional A-U, G-C basepairings, although oc c a s ional non - Wa t s on - C r i c kpairings, such as G-G and G-A, can form by usinga l t ernate hydro g e n - b on d - forming sites. (For Activi-ties 1 and 2, pairings ot her than A-U and G-C are an ds h ould be ignored . )

RNA’s ability to fold in complex ways causes it toresemble proteins by having secondary and tertiarylevels of structure. The determinants for foldingthese two molecules are quite different, however.Most proteins inside the cell are stabilized in a fol dedstate with hydrophobic amino acid re s i d u e ssequestered on the inside of the molecule, awayfrom water, and by hydrophilic residues on the out-side surface, exposed to the aqueous environment.Formal ionic interactions between amino acid side-chains of opposite charge also contribute to proteinstability. Like RNA, proteins have many hydrogenbond interactions that help determine their shape.However, amino acids are not complementary in themanner of nucleotide bases, and thus there are noone-to-one pairing rules for amino acids. This factmakes it difficult to predict protein shapes fromamino acid sequence alone. In the case of RNA, theaccelerating accumulation of new sequence andc rystallographic data makes the pro s pect loo kb r i g h t er for eventually predicting RNA thre e -dimensional shape from its sequence alone.

The folded structure of RNA is stabilized primarilyby helical regions that form within the moleculebased on Watson-Crick base pairing. Such helicalregions differ in their degree of thermodynamic sta-bility, depending on the number and nature of basepairs engaged in hydrogen bonding. Base stacking(the interaction between the electron clouds of theplanar, cyclic bases when positioned on top of eachother like a stack of dinner plates) also contributesto the stability of folded nucleic acids. Other morerecently recognized contributors to folding and

shape stability include the so-called ribose zipper, ajuxtaposition of the minor grooves of two helicalregions that is stabilized by hydrogen bonding oft he 2’ OH group of ribose; the use of single-stran dedloop regions that pair with “receptor” sequenceselsewhere in the molecule; and the use of ions suchas Mg++ to shield the uniform negative charge of thephosphate backbone. We next briefly describe someof the important structural motifs currently knownto occur in folded RNAs, with emphasis on theirknown or proposed functions.

DOUBLE-STRANDED HELICESWith more than 50 percent of its bases in double-stranded form, a typical RNA contains a great dealof secondary structure. As mentioned previously,RNA helices are in the A-form whereas those inDNA are B-form. This difference in helix geometrycreates significant differences in surface geographybetween RNA and DNA. Specifically, the majorgroove in RNA is quite deep and narrow, and theminor groove is shallow and wide, just the reverseof the B-form DNA helix. This difference in surfacetopography, along with the specific base sequence ofthe helix, determines the recognition and binding ofother molecules to helical nucleic acids. For exam-ple, the shallow, wide, minor groove of RNA appearsto be more accessible to protein side-chains and topresent more hydrogen bonding opportunities thanthe major groove. Despite the different shapes of thetwo grooves, examples of protein binding to RNAappear to involve both to different extent. Becausethe binding of proteins to helical RNA and DNA caninduce local bending and “melting” of base pairs,the limitations of helix groove size can be overcometo some extent. Indeed, there is some evidence thatthe single-stranded regions of the RNA, instead ofthe helices, may be more important as actual con-tact and recognition sites; the helical regions in thiscase serve to properly orient the single-strandedre g ions for pre s e n t a t ion. Tr an s f er RNA’s single-stranded anticodon loop, which recognizes bothmRNA and the synthetase enzyme that aminoacy-lates the tRNA, is a notable example of single-strand recognition ability.

HAIRPIN LOOPSLike the anticodon loop, many important single-stranded recognition regions in RNA arise as part ofa structure called a hairpin loop. The loop is createdwhen a single strand of RNA bends back on itself to

form a double-stranded region. This creates a dou-ble-stranded stem and a single-stranded loop thatcaps the helix (see Figure 8). The number and sizeof hairpin loops vary among different RNA types;for example, the three loops in tRNA have 7–8 baseseach, whereas the Tetrahymena self-splicing group Iintron has six larger loops. So-called tetraloops,which have four unpaired bases atop their helicalstem, are common in ribosomal RNAs (rRNAs); onesuch tetraloop in the 23S rRNA molecule appears tobe the ribosome binding site of the toxic proteinsricin and sarin. In the Tetrahymena self-splicingintron, a tetraloop, along with its conserved recep-tor site elsewhere in the molecule, facilitates self-folding of the intron into the proper thre e -dimensional shape (Cate et al. 1996). Stem-loopstructures are also found within the catalytic site ofmost ribozymes; for example, plant viruslike agentsk n own as viroids contain sel f - c l eaving RNAgenomes whose catalytic site adopts the so-calledhammerhead structure made up of three stem-loops. Smaller loops, known as bulges, are formedwithin a helical stem rather than at its end; theyresult when opposed bases are mispaired, causingthem to pucker out from the helix.

T here are many examples in which diff ere n ta s pects of gene expre s s ion, from mRNA tran s c r i p-t ion and tran s la t ion to mRNA de g r a d a t ion, empl o yhairpin loops as con t rol el e ments. Single-stran dedl oops may in general be pre f erred sites for thei n t er a c t ion of RNAs with re g u la tory molecules. Fore x a m ple, RNA is generally less susceptible tode g r a d a t ion by RNases when it contains a highprop ort ion of hairpin secon d a ry stru c t ure, and thisc an affect RNA’s half-life in the cell. A wel l - s t u d i ede x a m ple of stru c t ure con t rolling RNA de g r a d a t ionis the tran s f errin re c e p tor mRNA. This mRNAe n co des the cell surface re c e p tor re s p onsible forbinding the plasma iron tran s p ort molecule, tran s-f errin. A feedback me c h anism re s p on s i ve to pla s m ai ron level increases the density of these tran s f err i nre c e p tors on the cell surface when iron levels arel ow, thereby enabling the cell to more eff i c i e n t l ys c a venge tran s f err i n - i ron com plexes from pla s m a .O per a t ion of this system at the RNA level invol ve sa re g u la tory protein, pro d u c ed when iron levels arel ow, that binds to the tran s f errin re c e p tor mRNA ata stem-loop stru c t ure near its 3’-end. The bou n dre g u la tory protein stabilizes the stem-loop, makingt he RNA less susceptible to de g r a d a t ion and thus

able to be re u s ed to make more copies of the re c e p-tor prot e i n .

BASE TRIPLESUnlike the two-way interaction within a standardpair of bases, a base triple is an interaction betweenthree bases. Base triples are formed when a single-stranded region of an RNA nestles into the major orminor groove of a double-helical segment of themolecule; hydrogen-bonded triplets such as G-C-Acan result. These triple-stranded regions help stabi-lize tertiary, three-dimensional structure and maybe essential for certain RNA functions. For example,they have been found at the proposed catalyticregions in ribozymes such as the group I intron ofTetrahymena.

PSEUDOKNOTSPseudoknots represent a higher order, tertiary levelof structure found in RNA. A pseudoknot resultswhen some of the bases in an otherwise single-stranded loop pair with bases located outside thatloop. This kind of interaction can potentially form avariety of distinct topologies, but all pseudoknotshave two loops and two helical stems, usually withthe stems sharing a common axis (see Figure 8).The multiple stems and loops of this RNA confor-mation provide more complex sites for interactionwith proteins. In the 20 years since they were firstidentified, pseudoknots have been implicated inseveral examples of the regulation of gene expres-sion. For example, pseudoknot structures in certainmRNAs of retroviruses, bacteria, and yeast appearto stimulate a gene-regulatory phenomenon calledribosomal frame shifting. In this process, pseudo-knots in the mRNA, along with other specific basesequences, cause the ribosome to stall and slip dur-ing translation. The result is a change of readingframe that allows more than one protein sequenceto be synthesized from a given mRNA. This is aclear example of an RNA structure that increasesthe compactness and efficiency of genetic informa-tion storage. Another gene regulation example, inbacteriophage T4, involves the binding of a tran-scription regulating protein to a pseudoknot in thepromotor region of the gene 32 mRNA.

Pseudoknot structures also are found at the 3’-endof genomic RNAs of certain bacterial and plantv i ruses. Most intriguingly, these pseudo k n ot sresemble the stem-and-loop configuration of tRNAs.

The exact function of these tRNA-like ends is unset-tled, but they appear to be required for replicatingof the virus RNA genome. Similar structures arefound in a variety of other RNAs, such as the shortmolecules used to prime the reverse transcription ofretroviral RNA genomes to cDNA copies, and theRNA transcripts made from certain fungal plasmids.Also, the chromosome-capping telomeres at theends of eukaryotic chromosomes contain a TGG-rich sequence that is potentially able to base pairwith the CCA sequence at the ends of these tRNA-like molecules. Weiner and Maizels (1987) haveproposed the so-called genomic tag hypothesis,which posits that the tRNA-like sequence evolvedearly and functioned as a recognition tag that iden-tified certain RNAs as genomes and somehow facil-itated their replication. According to this view, theRNAs of similar shape found in today’s viruses andcells can be considered molecular fossils—vestigesof a much earlier, RNA-dominated world.

These examples of RNA structure illustrate the gen-eral stru c t ure - f u n c t ion pr i n c i ple stated ea r l i er :Polymer shape is determined by the linear order ofsubunits and their interactions within the molecule;three-dimensional shapes essential for the mole-cule’s biological function emerge as a result. Thestructural motifs discussed in this section give onlya hint of RNA’s potential for shape diversity. Ongo-ing structural research employing a range of mod-ern mol e c u lar techniques, such as RNA basesequence determination, site-directed mutagenesis,comparative analysis of sequences from differentorganisms, and advanced methods of X-ray crystal-lography, will no doubt reveal even more variety.That prospect is made all the more likely by the newRNAs that are being discovered in unexpected loca-tions—and carrying out surprising functions—inthe cell. In the next section, we discuss some ofthese newly recognized RNA functions.

Until recently, most biologists were aware of threet y pes of cel l u lar RNA: the stan d a rd trinity ofmRNA, tRNA, and rRNA. Research largely withinthe last 20 years makes it clear that many moreRNAs exist. They are found in diverse cellular loca-tions, from the nucleus outward, carrying out avariety of key functions related to gene expressionand metabolism. Some of the newly recognizedRNAs are highly conserved between species as dif-

ferent as yeast and humans. And some, such asthose bearing tRNA-like regions, appear to be relat-ed to familiar RNAs. The question of how thesediverse and widely distributed RNAs arose andbecame so thoroughly integrated into cellular econ-omy is central to the RNA world hypothesis.

Everything we know about evolution suggests thatit is a conservative process that builds upon existinginformation to create novel structures and func-tions. That being so, if the RNA world hypothesis iscorrect and RNA was indeed the first information-encoding and catalytic entity, then we can predictthat vestiges of those early RNA structures andfunctions, molecular fossils, should be present incontemporary organisms. Indeed, the diversity ofRNAs and ribonucleotides in contemporary cellsand their widespread involvement in key cellularfunctions provides much support for this hypothe-sis (Figure 9).

Most of the newly recognized RNAs share a com-mon functional theme: gene expression. Unexpect-edly, the roles of these molecules extend to aspectsof gene regulation beyond simple coding and tem-plate functions. They play key roles at several levels,from participation in DNA replication and mainte-nance to regulation of transcription, RNA process-ing and editing, translation of mRNAs, proteinlocalization within the cell, and modification of pro-tein function. Because some of these RNAs areknown to be catalysts, we next highlight someribozymes and their functions.

RIBOZYMESSeven categories of naturally occurring ribozymeshave been found in the 15 years since their discov-ery (see Figure 5). They occur in a wide range oforganisms, including viruses, bacteria, fungi, andplants; however, the prevalence of ribozymes intoday’s biological world is presently unknown. Therelatively few catalytic roles apparently still left forRNAs may represent evolutionary vestiges, formerroles having been taken over through natural selec-tion by the later appearing but structurally moreversatile proteins. Indeed, if RNAs were the firstbiological catalysts, as proposed by the RNA worldhypothesis, they were far from the most efficient:Proteins catalyze thousands to millions of timesfaster. The kinetic data of ribozymes reflects thissluggishness; ribozymes have quite low Km values,

indicating very high affinity for their substrates andeasy saturation. These kinetics seem well suited tor i b o z y me function, how e ver; unlike prot e i nenzymes, they typically catalyze only one reactioncycle (for example, their own removal from a largermolecule). Only two catalytic reaction mechanismsare known for naturally occurring ribozymes, trans-esterification and hydrolysis, and both employ OHgroups as nucleophiles for cleavage of the RNAphospohodiester backbone. Notable examples arethe coupled endonuclease-ligase reactions involvedin splicing, and the coupled endonuclease-phos-phatase phosphotransfer reactions that remove asubstrate 3’ P and transfer it to the ribozyme.

Among the best-studied catalytic RNAs are thegroup I and group II introns, which autocatalyzetheir own excision from a larger precursor RNA andligate the flanking exons. These introns are enco dedin the genomes of a wide assortment of organisms,including bacteria, fungi, and plants. In eukaryotes,t hey are more com m only found in org anel l egenomes, such as fungal and plant mitochondria,and plant chloroplasts.

Group I and group II introns are rich in stem-loopsecondary structure, particularly in their catalyticregions, and their higher order tertiary shape (atleast in the case of the group I introns) appears tobe stabilized by divalent cations like Mg++. Most

group I introns do not appear to require the aid(catalytic or otherwise) of proteins to self-splice. Incontrast, the splicing of group II introns appears tobenefit, both in vitro and in vivo, from the aid ofmaturase proteins that presumably help stabilizethe correct tertiary structure of the intron requiredfor self-splicing. Interestingly, at least some of thesematurase proteins are encoded within the sequenceof the intron itself; the intron carries coding “soft-ware” needed to make “hardware” that facilitates itscatalytic role. Likewise, certain mobile intron scapable of transposing to new genomic locationsencode proteins, such as endonuclease and reversetranscriptase, that facilitate their movement. Enzy-matic and coding abilities such as these are a far cryfrom the view held until quite recently that intronsare nonfunctional genetic baggage or “junk DNA.”

Unlike the autocatalytic processing of group I andgroup II introns, the processing of most eukaryoticmRNA precursors does not involve self-splicingintrons. These introns instead require for theirremoval the actions of complex RNA-protein assem-blies in the nucleus, known as spl i c e o s ome s .Spliceosomes are in effect macromolecular “splicingmachines” and are reminiscent of ribosomes inbeing RNA-protein assemblies. The relative roles ofthe RNAs and proteins of spliceosomes, that is,which are catalysts and which are the structural ele-ments, are not yet known. However, the catalytic

• RNA is informational and catalytic in vivo; no other biomolecule has both properties.• The nucleotide sequences of RNAs common to all organisms (for example, rRNAs) are highly conserved (similar)

among the many different species studied, suggesting that RNA was a key molecule present early in evolution.• RNA or ribonucleotides are involved in most critical cellular functions in all three domains of life:

- Adenosine triphosphate (ATP) is a universal energy carrier.- Universal metabolic pathways employ adenine nucleotide coenzymes (NADH, NADPH, FAD, CoA).- Protein synthesis employs mRNAs, rRNAs, and tRNAs.- rRNA by itself can catalyze peptide bond formation.- DNA synthesis requires the prior conversion of ribonucleotides to their deoxy form.- The ribonucleotide uracil, found only in RNA, is the precursor for DNA’s thymine.- RNA is the primer for DNA replication.- Ribonucleotide derivatives function as key signaling molecules in the cell (for example, cAMP, ATP).

• RNAs function as primers in DNAreplication and in reverse transcription of retroviral genomes.• tRNA-like molecules are involved in nontranslational (nonprogrammed) polymerizations (for example, cell wall syn-

thesis, polypeptide antibiotic synthesis).• A tRNA-like molecule may have given rise to the RNA component of telomerase, the enzyme that maintains the

ends of chromosomes.• Enzymatic processing of mRNAs involves other small RNAs (snRNPs, RNase P).• Protein sorting into the endoplasmic reticulum of all eukaryotes involves RNA(SRP-RNA).• Ribonucleotides are used to activate and carry sugars during polysaccharide synthesis.

Figure 9 Lines of evidence supporting the RNA world hypothesis.

mechanism of splicing by spliceosomes resemblesthat of a group II intron, leading to speculation thatthe several spliceosomal RNAs may be an “intron-in-pieces”: fragmented descendants of a once self-splicing intron.

RNA EDITINGAnother level of gene expression regulation thatinvolves the functional RNAs is the process of RNAediting. During editing, particular bases in the pre-cursor RNA are added, deleted, or changed follow-ing transcription. The resulting edited RNA has abase sequence different from that encoded in theDNA and transcribed RNA precursor. Most exam-ples of RNA editing involve eukaryotic mRNA pre-cursors, and when these are edited the maturemRNA specifies a different protein product thanthat enco ded in the gene. RNA editing thus increa sesthe number and diversity of products from a givengene. As with gene regulatory processes like alter-n a t i ve splicing and t r a n s- s plicing, RNA ed i t i n ge ff e c t i vely increases the inform a t ion coding capacityof a genome. Although the detailed chemical andcatalytic mechanisms of editing are not yet known,one version of the process employs small RNAsknown as guide RNAs. The base sequences of theguide RNAs enable them to pair with the pre-editedt a rget RNA and specify the bases to be added, del e ted,or altered by complementary pairing.

Two additional examples illustrate the variety ofRNA effects on gene expre s s ion. During the tran s-la t ion process of protein synthesis, a pe p t i d y lt r an s f erase catalytic activity co valently linksi n coming amino acids to the growing pol y pe p t i de .This catalytic activity is known to re s i de in thela rge ribosomal subunit, which in E. coli con s i s t sof some 31 proteins and 2 distinct RNA mol e c u l e s .Which of these ribosome com p onents catalyze spe p t i de bond form a t ion has long been a mystery.T he con ve n t ional assu m p t ion was that it must beone of the proteins. Recent re sults, how e ver,s t rongly suggest that the 23S rRNA com p onent ist he catalyst. Nol l er et al. (1992) extracted moret h an 95 percent of the proteins from the la rge su b-unit, leaving the 23S rRNA, and the peptidyl tran s-f erase activity, intact; trea t ments that damage theRNA eliminate the activity. More re c e n t l y, Nitta etal. (1998) de m on s t r a t ed that cloned segments oft he 23S rRNA ne ver exposed to ribosomal prot e i n scould be re con s t i t u t ed in the test tube and cou l d

c a t a l y ze pe p t i de bond form a t ion. The emerg i n gp i c t ure is that of an RNA catalyst in the lead rol eof a rea c t ion essential for life, with ribosomal pro-teins playing stru c t ural su p p orting roles. This viewis consistent with the not ion that com plex ribo-s omes evol ved when functions once carr i ed ou ts ol ely by RNAs were impro ved upon by the addi-t ion of prot e i n s .

RNAs also play a role in the cellular localization ofnewly made proteins. In eukaryotes, a complextranslocation machinery enables nascent polypep-tides bearing an amino-terminal “signal sequence”to cross into the lumen of the endoplasmic reticu-lum (ER). A key element of this translocationmachinery is the so-called signal recognition parti-cle (SRP), a cytoplasmic RNA-protein assemblyconsisting of a core 7S RNA, to which six differentprotein components bind. The signal recognitionparticle binds to the free-floating ribosome-nascentpeptide complex. It then carries the complex to thesurface of the ER membrane, where the SRP binds toa re c e p tor. The SRP in effect acts as a bridge to tetherthe ribosome-peptide to the surface of the ER, andthus facilitates move ment of the newly madepolypeptide into the lumen. The 7S RNA is neces-sary for SRP function and may be the componentthat binds directly to the ER membrane receptorprotein.

A basic assumption of the RNA world hypothesis isthat early in the evolution of life, one or a few mol-ecules came to dominate the pool of RNAs that hadbeen ran domly gener a t ed by non b iological proc e s ses.A s suming that RNA’s appea r ance pre c eded that ofcatalytic proteins, the RNA must have been able toreplicate itself in order for evolution at the molecu-lar level to get under way. The demonstrated cat-alytic ability of RNA makes the possibility ofself-replication much more plausible. Even thoughno RNA capable of catalyzing its own replicationhas yet been found in nature, recent experiments indirected evolution (molecular evolution in a testtube) have demonstrated that RNA does indeedhave this potential.

S ol Spiegel m an and his group first de m on s t r a t ede v ol u t ion and sel e c t ion of RNA molecules in a testtube in the 1960s. Spiegel m an ’s in vitro e x per ime n t s

s t a rt ed with a mixed pop u la t ion of RNA genome sf rom the bacter iophage Qβ, which were tran s f erreds erially to a series of tubes su p plying only ribon u-c l e ot i des and re plicase (the RNA pol y merase thatcopies the viruses’ RNA genome inside host cel l s ).By limiting the time ava i lable for re pl i c a t ion ineach tube, the exper i ments imposed spe ed of RNAre pl i c a t ion as the sel e c t ion criter ion. Several cyclesof in vitro re pl i c a t ion and tran s f er led to va r i an tRNAs that could be cop i ed at greatly increa s eds pe ed. By altering the physical or chemical sel e c-t ion con d i t ions, pop u la t ions of RNA adapted to thei m p o s ed con d i t ions came to predominate in them i x t ure .

The modern era of in vitro selection (also referred toas directed evolution) began in 1989–90 and wasmade possible by technical advances in RNA andDNA synthesis (for example, automated oligonu-cleotide synthesis), nucleic acid amplification (forexample, polymerase chain reaction), and selectionmethods (for example, affinity chromatography).Starting with a large synthetic population (pool) ofrandomly varying RNA or DNA molecules, the goalof in vitro selection experiments is to amplify thosevariants that are able to meet some experimentallyimposed selection criterion, such as the ability tobind to a particular target molecule or catalyze aparticular chemical reaction. After 10 to12 cycles ofselection and amplification with mutation, mole-cules well adapted to the selection criterion pre-dominate in the pool (Figure 10).

A parallel can be drawn between the variation,selection, and amplification aspects of these in vitroexperiments and natural, biological evolution. How-ever, the action of selection in these experiments ismore direct than in biological evolution: In organ-isms, mol e c u lar stru c t ure - f u n c t ion is sel e c t edsomewhat indirectly through a complex organismalphenotype; with in vitro methods the molecule’sbase sequence, which directly determines its func-tion, is the selected phenotype. It might be said thatthe generation of highly functional molecules froma random assemblage of sequences is akin to a tor-nado assembling a 747 from random parts, cre-ationist views notwithstanding.

The RNA and DNA molecules found by this power-ful experimental approach tell us a lot about thepotential range of nucleic acid functions and theirevolutionary potential. In fewer than 10 years, themethod has unco vered many new synthe t i cribozymes, supplementing the seven known naturalclasses. Significantly, one of these synthetic RNAscan copy an RNA template, forming short comple-mentary strands. Further laboratory refinements ofthis molecule may yield a bona fide RNA-dependentRNA polymerase, a key ribozyme activity requiredfor self-replication and assumed in RNA world ideasabout the origin of life.

In vitro e v ol u t ion exper i ments selecting for D N A m ol-ecules have shown that single-stran ded DNA, likeRNA, has catalytic potential. Perhaps the distinction

Figure 10 The strategy of directed evolution in vitro. Starting with a large, randomly synthesized population of nucleic acid,repeated cycles of selection followed by amplification can yield a particular se quence well suited to the selection criterion.

between RNA and DNA is not as great as we havecome to bel i e ve. Recall that the only chemical diff er-ence is one extra oxygen atom in RNA (in the 2’ O Hg roup of ribose). Also, both RNA and single-stran dedDNA can fold into com plex shapes. If there really islittle diff erence between these two closely rela t edm olecules, why is DNA today the storage form ofmost genetic inform a t ion, and why is RNA the activeform performing a variety of cel l u lar tasks? The cur-rent thinking is that DNA appea red la t er in the RNAworld, most likely as an RNA der i va t i ve, and DNA’ssu c c e s s ion to the role of inform a t ion re p o s i tory maybe rela t ed to its grea t er chemical stability (due to itsfully dou b l e - s t r an ded con f i g ur a t ion and lack of a 2’OH nucleop h i l e ). Or perhaps DNA pro ved to havetoo small a range of catalytic ability to ou t - com pe t eRNA in a functional role. It remains an intriguingqu e s t ion whe t her life could have evol ved equ i va l e n tl yif the earliest catalytic events had been DNA-basedand RNA had appea red la t er as the storage form .

Many viruses have RNA genomes. Some are impor-tant human pathogens: polio, HIV, flu, and measles,among others. Numerous examples are also foundamong plant viruses, several of which have majoreconomic impacts in agriculture and forestry. Acommon feature of all RNA viruses is a high level ofvariation among their genome sequence, the resultof a high mutation rate (10-3–10-4 per base pair perreplication). The frequency with which mutationsarise in a population is determined in part by howoften the organism reproduces. Viruses replicatemillions of times each day, so random mutations areconstantly arising. Another factor affecting the fre-quency of mutation is the fidelity of the replicationprocess itself. The replicase enzyme that copiesRNA genomes occasionally makes random errors,inserting the incorrect monomer (for example, Aopposite G, or U opposite C). Indeed, the poly-merases that copy DNA also make errors, but cellshave evolved a molecular quality-control, proof-reading mechanism that can correct mistakes mostof the time, keeping the DNA mutation rate low(less than 10-7). RNA lacks such a correction mech-anism, a condition that may be related to its morelimited role as a genome in nature. Although popu-lations of RNA viruses have greater variation ingenome sequence than DNA viruses, variation isessential to the adaptation and evolution of allviruses, indeed, all life forms.

Because of this high mutation rate, the many pro g en yviruses made in an infected cell constitute a pop-u lat ion of va r i ants, with typically one to thre ec h anges in each viral genome. The genome sequ e n c eof a pop u la t ion of RNA viruses thus is not uniqu e ,but rather is a pop u la t ion of va r i ant “qu a s i s pe c i e s .”A n ot her source of genetic va r i a t ion in viruses is the i rability to exc h ange whole blocks of genes amon gprogeny by the process of re com b i n a t ion. Most of theg e n ome alter a t ions are del e t er ious, and those viru s e swill not surv i ve. Other genome changes, how e ver,c an con f er advantages for diff erent aspects of viralb e h a v ior: faster re pl i c a t ion, increa s ed virulence, orde c rea s ed sensitivity to antiviral dru g s .

In the case of HIV, it has been estimated that eachof its 10,000 RNA bases is mutated more than10,000 times each day in an infected person (Cof-fin, 1995). HIV’s particularly rapid mutation rateand the chronic nature of the infection make it pos-sible to detect the emergence within an individualpatient of new mutant strains during the course ofthe disease. The virus thus can evolve within themicrocosm of a single human host. In similar ways,new strains of cold and flu viruses continue toemerge and plague humankind.

With both RNA and DNA viruses, the war betweenthe hosts and the viruses that infect them is sophis-ticated, subtle, and evolving: The host inactivatesthe virus or limits its propagation, the virus subvertsthe cell to make more virus. The battle shifts backand forth with move and countermove. A criticalmove by the host is to mount an immune responseagainst the virus. Unfortunately, the RNA virus’sability to rapidly generate genetic diversity pro-duces an ever-moving target for the immune sys-tem, as well as for developers of vaccines andantiviral drugs. In the case of HIV, the virus hasevolved an additional powerful strategy to ensure itspersistence in the host, that of attacking and dis-abling the immune system itself.

Some viruses subvert cellular defenses by employ-ing RNA molecules as weapons. The cellular inter-feron response is one example. When infected withvirus, cells typically react by secreting the potentsignaling molecule interferon. This protein protectsneighboring cells from becoming victims by tem-p orarily shutting down their prot e i n - s y n t he t i cmachinery. This clever cellular defense strategy

denies the virus access to the one source of new pro-teins needed to complete its life cycle. Adenovirus,however, has evolved a more clever way to under-mine this defense: Specialized viral RNA molecules(for example, VAI RNA) effectively block interfer-on’s action and prevent the shutdown of cellularprotein synthesis. The cell has no option but to con-tribute slavishly to the production of viral progeny.

Vi ruses can likewise de f eat medical ther a p i e s .Although vaccines have been dramatically success-ful in curtailing some viral diseases, such as polio,they are less effective at combating the rapidlyevolving RNA viruses, such as rhinovirus (commoncold), influenza, and HIV. For these, the currentapproach is to develop antiviral drugs, chemicalagents that either block entry of the virus into hostcells or, once inside, block key steps in the viralreproduction cycle. Current treatment of HIV, forexample, employs drugs such as Acyclovir, AZT,and prot ease inhibitors. By targeting diff ere n taspects of viral reproduction, these drugs are prov-ing effective at slowing the course of HIV infection.

E v ol u t ion a ry con s i der a t ions, how e ver, dictate thatt he w a y in which these drugs are administered is akey factor in their lon g - t erm eff e c t i ve ness. Adminis-t ered singly, any given drug soon loses eff e c t i ve ne s sas the rapidly mutating virus generates a pop u la t ionthat is re s i s t ant. If the drug is re pla c ed by an ot her inan attempt to hit the remaining re s i s t ant pop u la t ion ,that pop u la t ion will in turn give rise to a new pop u-la t ion that is re s i s t ant to both drugs . . . and so itgoes. The unintended but predictable ou t come of thist y pe of serial drug trea t ment is the eventual crea t ionof a viral pop u la t ion that is re s i s t ant to several dru g s .Such a multiply re s i s t ant virus could spread qu i c k l yin the host pop u la t ion. A much better strategy, onethat makes rational use of evol u t ion a ry pr i n c i ples, isto use all three drugs in com b i n a t ion. The key to thisa p proach is that each drug i n d e p e n d e n t l y and s i m u l-t a n e o u s l y hits a diff erent step in the viru s ’s re pro d u c-t ion cycle. For a triply re s i s t ant viral pop u la t ion toe merge in the presence of all three drugs, three inde-pe n dent re s i s t ance mutations would have to oc c urto g e t her in a single fou n der v i ru s, a much less pro b-able oc c urre n c e .

The alternative to devising separate drug treatmentsfor each different type of virus is to design a broad-spectrum antiviral agent that is effective against

many viruses. Unfortunately, because viruses aredependent on host cell machinery (such as that forprotein synthesis), any drug that interferes with akey synthetic step common to all viruses likely alsowould inhibit cellular function. The situation is dif-ferent for bacteria, which have evolved enough dif-ferences at the molecular level from eukaryoticcells, including RNAs, that it is possible to developdrugs that specifically target bacteria. Agents suchas penicillin, which impairs synthesis of the uniquemolecular outer layers of bacterial cells, were dra-matically effective when first introduced. Also, theprotein synthetic apparatus of bacteria is su ff i c i e n tlydifferent in molecular detail from that of eukaryotesthat it is a good target for antibiotics. Some of theseantibiotics inhibit bacterial protein synthesis bybinding directly to RNA com p onents of the bacter ialribosome. Structural differences between bacterialand nonbacterial rRNAs account for the specificityof the antibiotic in this case. Mutations that changethe structure of the bacterial rRNAs, such that theantibiotics no longer bind, confer resistance on thepathogen. Once again, we see that the ongoingprocess of evolution, fueled by mutations such asthese in the bacterial and viral populations, ulti-mately undermine the effectiveness of even the mostpowerful therapeutic agents in our arsenal. Activity4 of the module explores RNA’s role as a target forcertain antibiotics and antiviral agents, as well asthe acquisition of resistance.

Two novel strategies focusing on RNA with poten-tial appl i c a t ion against viral pathogens are curre n tlybeing explored in the laboratory: in vitro selectionand antisense technology. The former appro a c hemploys the in vitro selection strategy describedabove to select functional nucleic acid moleculesthat bind to and inactivate viral com p one n t s .Ribozymes able to cleave DNA have been devel-oped, as have deoxyribozymes able to cleave RNA.Looking ahead, it may one day be possible to engi-neer into the genome of affected cells a ribozymesequence that has been selected in vitro for its abil-ity to inactivate the RNA, DNA, or protein of a viralpathogen. The coding sequence for a ribozyme hav-ing, for example, anti-HIV activity might be incor-porated into a population of the patient’s immunecells. Subsequent expression of the ribozyme “bul-let” in these cells and their descendants conceivablycould have significant therapeutic effects against thevirus.

Another potential application of in vitro-selectednucleic acids is as ligand-binding molecules; suchmolecules could be useful as biochemical reagentsor as potential therapeutic agents. The term aptamerdescribes in vitro-selected RNAs or DNAs that havethe ability to bind with specificity to another mole-cule (aptamers, unlike ribozymes, are not catalytic).Aptamer sequences with unique shapes have beendeveloped to bind a wide range of ligands, includingsmall organic dyes, coenzymes, amino acids, vita-mins, and viral proteins. The binding specificity ofaptamer RNAs is high enough that they can distin-guish between ligands as similar as theophyllineand caffeine, which differ by only a single methylgroup. These designer nucleic acids promise to beuseful additions to the growing list of functionalbiomolecules, including protein enzymes and anti-bodies, that already have been generated by in vitrotechniques.

T he second RNA-rela t ed approach to ther a pe u t i c s ,antisense technology, attempts to inactivate and ne u-

t r a l i ze unwan t ed mRNAs that der i ve from a mutatedt a rget gene. This approach takes advantage of thet e n dency of single-stran ded RNAs to bind tos equences com pl e me n t a ry to the m s el ves. The strateg yis to engine er into aff e c t ed cells an antisense copy oft he de f e c t i ve gene, that is, one in which the 5’- 3’or ie n t a t ion of the gene has been re ver s ed rela t i ve toits prom otor. Insert ion of the invert ed gene copy cre-ates a situation in which RNA transcripts from thei n vert ed gene, known as antisense transcripts, arisef rom what is normally its non t r an s c r i b ed co d i n gs t r an d. T he antisense transcripts are thus com pl e-me n t a ry to the sense transcripts from the de f e c t i veg e ne, which arise from the opposite, templa t es t r and. Antisense and sense RNA molecules canbind and ne u t r a l i ze one an ot her, eff e c t i vely silencingt he expre s s ion of the de f e c t i ve target gene, perhaps apro v i rus or an on co g e ne. Successful ther a pe u t i ca p pl i c a t ions of both the in vitro s el e c t ion and an t i-sense appro a c hes await advances in gene ther a p yt e c h n ologies ne eded to incor p orate engine ereds equences into the genome in a functional state.

A b b ott, I., et al. (1977). Com p a r a t i ve ecology of Galá-pagos ground finches (Geospiza G ould): Eva l u a t ion oft he import ance of floristic diversity of inter s pe c i f i ccom pe t i t ion . Ecological Monographs, 4 7: 151–184.

Cairns-Smith, A.G. (1985). Seven clues to the originof life. New York: Cambridge University Press.

Cate, J.H., Gooding, A.R., Podell, E., Zhou, K.,Golden, B.L., Szewczak, A.A., Kundrot, C.E., Cech,T.R., & Doudna, J.A. (1996). RNA tertiary structuremediation by adenosine platforms. Science, 273:1696–1699.

Cech, T.R. (1998). Structure and mechanism of thelarge catalytic RNAs: Group I and group II intronsand ribonuclease P. In Gesteland, R.F., & Atkins, J.F.(Eds.), The RNA world (pp. 239–269). Cold SpringHarbor, NY: Cold Spring Harbor Laboratory Press.

Cleland, A., Watson, H.G., Robertson, P., Ludlam,C.A., & Brown, A.J.L. (1996). Evolution of zidovu-d i ne re s i s t an c e - a s s oc i a t ed genot y pes in humanimmunodeficiency virus type 1-infected patients.Journal of Acquired Immune Deficiency Syndromesand Human Retrovirology, 269: 6–18.

Coffin, J.M. (1995). HIV population dynamics invivo: Implications for genetic variation, pathogene-sis, and therapy. Science, 267: 483–489.

Ekland, E.H., Szotak, J.W., & Bartel, D.P. (1995).Structurally complex and highly active RNA ligasesderived from random RNA sequences. Science, 269:364–370.

Eldredge, N. (1995). Reinventing Darwin: The greatdebate of the high table of evolutionary theory. NewYork: John Wiley & Sons, Inc.

Ellington, A.D., & Joyce, G.F. (1990). In vitro selec-tion of RNA molecules that fold into specific ligand-binding ligands. Nature, 346: 818–822.

Gesteland, R.F., & Atkins, J.F. (Eds.). (1998). TheRNA world (2nd ed.). Cold Spring Harbor, NY: ColdSpring Harbor Laboratory Press.

Gilbert, W. (1986). The RNA world. Nature, 319:618–620.

Howard Hughes Medical Institute Holiday Lectureon Science. (1995). The Double Life of RNA [video-tape set]. (A four-tape series of lectures by NobelistT h omas Cech, discussing his disco very of ribozymesand other important functions of RNA. Summariesof the lectures can be found on l i ne at http://www.hhmi.org /.)

Jenison, R.D., Gill, S.C., Pardi, A., & Polisky, B.(1994). High resolution molecular discriminationby RNA. Science, 263: 1425–1429.

Joyce, G.F. (1992, December). Directed molecularevolution. Scientific American, 267(6): 90–97.

Joyce, G.F., & Orgel, L.E. (1993). Prospects forunderstanding the origin of the RNA world. InGesteland, R.F., & Atkins, J.F. (Eds.), The RNAworld (pp. 1–25). Cold Spring Harbor, NY: ColdSpring Harbor Laboratory Press.

L an d w e b er, L.F., Simon, P.J., & Wa g ner, T. A .(1998). Ribozyme engineering and early evolution.BioScience, 48(2): 94–103.

Mills, D.R., Pe t er s on, R.L., & Spiegel m an, S.( 1 9 6 7 ). An extracel l u lar Darw i n i an exper i me n twith a self-duplicating nucleic acid molecule. Pro-ceedings of the National Academy of Sciences, 58(1):217–224.

Nitta, I., Kamada, Y., Noda, H., Ueda, T., & Watan-abe, K. (1998). Reconstitution of peptide bond for-mation with Escherichia coli 23S ribosomal RNAdomains. Science, 281: 666–669.

Noller, H.F., Hoffarth, V., & Zimniak, L. (1992).Unusual resistance of peptidyl transferase to proteinextraction methods. Science, 256: 1416–1419.

Orgel, L.E. (1994, October). The origin of life onthe earth. Scientific American, 271(4): 77–83.

Simons, R.W., & Grunberg-Manago, M. (Eds.).(1998). RNA structure and function. Cold SpringHarbor, NY: Cold Spring Harbor Laboratory Press.

Smith, J.M., & Szathmary, E. (1995). The majortransitions in evolution. Oxford: W.H. Fre e m an(Spektrum).

Szostak, J.W., & Ellington, A.D. (1993). In vitros el e c t ion of functional RNA sequences. In Gesteland,R.F., & Atkins, J.F. (Eds.), The RNA world (pp.511–533). Cold Spring Harbor, NY: Cold SpringHarbor Laboratory Press.

Unrau, P.J., & Bartel, D.P. (1998). RNA-catalyzednucleotide synthesis. Nature, 395: 260–263.

Weiner, A.M., & Maizels, N. (1987). 3’ TerminaltRNA-like structures tag genomic RNA moleculesfor replication: Implications for the origin of proteinsynthesis. Proceedings of the National Academy ofSciences, 84: 7383–7387.

Wright, M.C., & Joyce, G.F. (1997). Continuous invitro evolution of catalytic function. Science, 276:614–617.

Wyatt, J.R., & Tinoco, I. (1993). RNA structuralelements and RNA function. In Gesteland, R.F., &Atkins, J.F. (Eds.), The RNA world (pp. 465–496).Cold Spring Harbor, NY: Cold Spring Harbor Labo-ratory Press.

Zaug, A.J., McEvoy M.M., & Cech, T.R. (1993).Self-splicing group I intron from Anabaena pre-tRNA: Requirements for base-pairing of the exonsin the anticodon stem. Biochemistry, 32: 7946–7953.