In vitro selection of nucleic acids and proteins: what are we learning?

521

In vitro selection of nucleic acids and proteins: we learning? Richard W Roberts* and William W Ja For almost used

a decade,

to isolate

according

novel

to their

our perception appear

selection acids,

experiments peptides

function.

Selection

experiments

mimicry

and catalysis,

with

facile

than

desired

rational

properties.

design New

have

the

been

have

structural molecules

Information hlghlights

ratlonal

approaches.

testin::

altered

each

that

,\ dcr~lc

ago.

LYHII~ hc t’o~~nd that

:rnd/or

fllnction

I:ltion

of KiY,\

that

Addresses

in Structural

hind

Callfornta

lnstltute

Science

Ltd

ISSN

1999,

Introduction

and

aptan’crs

;;:encral

.Idrlitic)n

KY\

Kc\,), no

3s lrcll

nkitiiral \,ascul:ir

in3~

‘Ilie

ahilit)-

35 small

\ystcms.

I:ition.

‘l‘hcrc.

IX

ilii;igc,

molcciilca

hind

transcrip anJ

proteins :ictivit, [L~l<(;l:]

factor

tcstcd

allow for

function

all the ;It

the

tinic. 13> coniI)2risoii. sci-ccllinfi cupuriments th:lt t-:lch 1~)ol mcmhcr bc qllericd indi\iidllail\. selection cupcrinients ,~ wncrilll\ ,illou milch Iarxcito I)r sirr\,cvcd than \crccriiiig eYpcrimcnts. On

th;it

:tcici\

of the

it’ the mirror with naturkrl \t:lhlc

binds

to mimic

in

oli,qmu-

the

and

best

the

5triictiirc .i

‘l’hc

natural

rheory

crsir!

strtictiircs

121.

this

the

transition

seen1

of

the

that

i5 doniinatt4 howc\,cr,

an

‘KN/I

by proteins ha\,c

its hounds.

t:,ictol

ril~osonie

cnlr\

m;inv

bv folded in

ha\c the

cii!i;lr\otic

inlportaiit

3 p-cat

esperiniental

in tt‘r-

and

to in

internal

mimicry from

in\.ol\ed

.Siniil:irl~,

!)\ ;kn tlcncc.

pn~\~itlcs

th;it

dot<,

t:,ictors

min~icked of

kno\\ from exists in tr;lns-

1<1~-(~4;‘1’1’)

Kk-.34;‘1’1’)

portion

in sclcction

no\\

examples Ed

he mimicked (IKICS) 11.3,1-l]. di\

wc

protcin

I Il.

initi;ttion. can

proteins

eY:liiiplu~

One of the KKX

general

both

protein KNA strllc-

natur:il

and

support

for

c\orld

to a bio-

is possible. It follo~~s

thut.

if

any small molecrile target, c;italysts COIIIJ IK discovered tq sclccting lihrar): mcmhers that hind transition-statc andogs (‘I’SAsL ‘I’his approach has hccn iiscd to ,great adv3ntage in the immiinc sclcction of catal\,tic antiboclit3 1 IS]: howc~cr. success has been Iiliiitcd llsin:: KN.1 ;~ncl I)NA apt:lnlcr\. ‘l’he first ewmple was .ii)taincrs

euperirnents

sllhstitution

I~iologic:ill~

~onstrii~~tcd

(t~l~-‘li~~~~‘l‘l’~rK~~~ (KIY-1. KIT-2 :lnd

\liinicrk

libraric5

the

th;lt

rc\wsc

or hv the

is cchoetl

chcmistrv

i~lmc rcqllirt: ‘l’hiia.

isolated

HIV

(e.g. 191). .iltcrn:lti\ei~, protciil tar~:ct is iised

of nucleic

n:ltural

the

he

miniic

1101.

non-n:irirr;il

to

[7]) clone

hetcropol~-

to

iliiclci~-acid-t)iii~lili~ cndothelial ,qrowth

:I mirror

or I)k;\.

tal’!$?

of art

yclection

has

one

iiscd

been

(c.g.

of pho~~,horothioarcs

I leotide

tklrc\.

l~)oI

that

iso-

fit’

1.

have

\U~IIC~CC’ nmtifs can bc ftlnction:lli\

;,ifftv

‘I‘hc

proteins

hc

;I pi-ottzin

;Ipt;tniers

1’-01 I t\ith 2’-Ntl, irnak:c of the 0ri:in;rl

ti.;liisl;ition

the

:lt best.

2nd 1 asoprcwin) IX]. ‘l‘hcsc krptamcrs can Lx ackiptcd foi IISC i/l c,izw if thcv ;Irc protcctccl from degrad:ltion b! the

cll4‘4(;

i/l

:tnd

or :I I>NA strll<‘turc

the

c//)TN.F, ‘to

and

notion

protrins

thrombin,

ximc

in

KNA

(from

miiv

I1iX.A) (c.~.

I)N:\

,ind tlli’ t11:1t h:l\,e

clon,qtion min;ltion

nlolccllle

;I

in

cloning

either an mimicked

molcc~~lc~

the or

of :lnc)thcr

c\pcrinwnts

(kncra1ly.

rnelnbers

withollt

W;LS hrretical

I>NX

sm;ill

KN,A

t;isc

9:521-529

0959-440X

and terms

that

of ;1 protein

to advance

I<%\

quantum coherence polymerase chain reactlon single-chain vanable domaln fragment transitlon-state analog vascular endothellal growth factor

SCFV TSA VEGF

individwi

of

Abbreviations EF elongation factor FMN flavln mononucleotide HSQC heteronuclear stngle PCR

propn\:ll

:tncl

both

(e.g.

(c.K. c Elsevler

the

iiir~lci~-acicl-l,indinfi

Biology

of

known

and mimicry

fllnction

Opinion

it\

is not

m()IccIIIc

lnu(.h

Englneenng, USA

:Icti\

RNA and DNA

Aptamers

have

on a number of selected and future challenges for design via

Dlvislon of Chemistry and Chemical Technology. Pasadena, CA 91125, *e-mail: [email protected]

the

inciiviciual.

Functional

at generating methods

iller

Current

hand, experiment

and they

been developed improve the power of functional strategies In ways that nature has already dwovered - by expanding library size and facilitating the recombination of positive mutations. Recent evolved

other

selection

and proteins

of molecular

to be more

biopolymers

in vitro nucleic

what are

:I Irll>o/.wnc

could

th:lt

hind

enhanced

the

isomerization

ofs~~l~rtitutetl

522

Engineering

and

design

biphenyls by M-fold over background [ 161. Subsequently, RNA and DNA aptamers that bind /V-methylmesoporphyrin (NMMP) were found to stimulate porphyrin metalation [I 7,181. ‘I’he scarcity of enzy-mes compared with catalytic antibodies may be a result of the different mechanisms used by nucleic acids to recognize small molecules compared with those used by proteins (see discussion below). On the other hand. the relative paucity of ‘I’SA-binding protein catalysts that have been isolated via phage display may imply that in vitro selection techniques need to be improved in order to match the ability of the immune system. Amide- and peptide-bond-forming catalysts One central tenet of the ‘R&A world’ hypothesis is that the mother of all enzymes, the ribosome, contains RNA in its catalytic heart. In addition to in\,estigations that point to rRNA as being essential for ribosome activity [ 19-21 1. :I number of laboratories ha\,c set out to test whether or not ribozyncs could be found that arc capable of catalyzing amide-bond formation. ‘I’he first C~ICCCSS involved an ac\l transferasc ribozyme that was capable of forming an amide bond between a .i’-NI-I,-R’VA and .Y-biotiny-l methionine attached to the 3’ end of ;I hexanlrcleotidc (k, - 0.6 min-1) [27]. Recently, E;aton and co-norkcrs (2.31 reported the isoIation of amide-bond-forming rihozymcs that do not require an oligonuclcotidc tether ro the substrate. a .j’Ahll’ actixrated ester of biotin. ‘l’hcse riboqmes do rcquirc .j’-imidazole uridine, which ws incorporated uniformI\ in place of uridine in the RNA pool. Although neither of these ribozymcs was reported to contain sec~wzr~cc: homology to natural RNA sequences, me of the peptidyl transferasc riboqmcs isolated by %hang and Tech did [24,25-l. The clone 25 ribozyme contains ml internal loop with regions of’ sequence identity to the centrdl loop of domain c’ of 2% rRNA, a region that is central to the peptidyl transferasc function of the ribosomc. Surprisingly, aptamcrs that bind the pcptidyl transferasc inhibitor (:cdAp-I’urotn)lcin share identity with the peptidy1 transferasc ribozyme and with 23s rRNA 1261. I>cspite the recent retraction of bvork arguing that domain \’ RNA transcripts ha\,e peptidyl transferase activity [27/> the selection data represent proof that RhA has the potcntiul to carry out a \,ariety of amide-bon~l-forming reactions. New ribozyme reactions A number of rihozymes with new activities have been isoIatcd recentI>, including molecules that are capable of catalyzing the s\.nthcsis of ;I IlUc~e~Jtidc [28”]. cnzylcs selected to mim;c reactions in the spliccosome [29’] and riboaynes capable of performing a I)iels-Alder cycloaddition reaction [.iO]. In the first example, I Jnrau and I3artel [28”] uncovered ribozymcs that are capable of joining a tethered I-thio-uridine nucleotide to pRpp (5’ phosphoriI,os)jl-a-l’~rol’hosphatc), the activated ribosc (ised in nucleotide biosvnthesis (I~,..,, = 0.1 .i minmI). ‘I’his reaction is almost identical to the natural reaction carried out I)\ uracil

phosphoribosyltransferase pRpp, making uridine

(IIPK’I’) 5’ phosphate,

to couple

uracil

and

In the second example, a selection was designed to test whether modern spliceosomal RNA ll2/1J6 could be used to resurrect RNA-based spliecosomal activity. Each of the four ribozyme classes found performs some kind of cleavage reaction, followed by a ligation; however, none of the selected molecules produces an activity that is similar to that of the spliceosome [29’]. Interestingly, the templates generated by the ribozymes uncovered a previously unknown activity for reverse transcriptase. nontemplated jumps in sequence. Adding functional group diversity - ‘If I only had a histidine’ A number of laboratories have recently worked to expand the chemical di\.ersity of RtiA and 1)NA. either through covalent modification or through the addition of cofactors and ions to the selection mixture. Examples of the first category include ribo/.ymes that catalyze carbon-carbon bond formation though 3 IXels-Alder reaction (.iO]. a5 well as the previously mentioned smi,ie-bond-forming rihJ~);lll~S [Z.3] that contain 5-pyridy~meth~lclirhox3mid uridinc and ii-imidazole uridine, respcctivcly. Kxamplcs of the second case include the routine addition of low (micromolar to millimolar) concentrations of’ divalcnt cations, including hip:+‘, (:a+‘. %n+?, Rin+L. ( :u*l and eI’en l’b”, or cofactors, \uch as histidinc 1.31‘I. ‘I’hc itnportancc of these prosthetics can bc xc11 in that their omission abrogates ftlnction.

DNA catalysts Although DNA tion storage. it

plays a centrdl role in biological informaa functionail>impoverished bi(Jplylller comparc~ with cithcr proteins or RNA. Despite its limitations, the list of DNA catalysts has grown significantly in the past few years (see ‘Iliblc 1 in [32]; [U]). It is not surprising that many of the DLL4 catalysts isolated require ;I cofactor of some kind. It is surprising that the highest catalytic cfficietq, 10” IL-~*s-’ (although not the highesr k,;,,), is held by the 10-33 RNA-cleaving DKA enzyme [.34,35j. ‘Ii) date, the scope of I>KA-based catalysts has significant di\,ersity. including ligation ( .VI]. RNA phosphoester cleavage [.3-!,.37-391, DNA clea\,age (40j. porphyrin metalation [1X], peroxidase acti\,ity [41’], fl~~orophorc oxidation [3.2] and 5’-IIN, phosphor)ilation [4.3*]. It is interesting to note that C;-quartets seem to play a ccntral role in the folded structure of many DNA catalysts, ;IS \vell as of DNA aptamers. is

Engineering allosteric enzymes and templates One potentially powerful application of nucleic acid aptamers and catalysts is the engineering of allostericall) controlled enzymes and templates. Breaker and co-workcrs [-M.lrS*] have shown that, through combinatorial enainecring, chimaeras between either the A’I’I’ or I:hlN aptamer with the hammcrhcad ribozyme may be allostcricaily regulated to function onI> when they bind their

In vitro

selection

of nucleic

acids

and proteins

Roberts

and Ja

523

Table 1 Peptides Initial

evolved

pool*

MVSKGEEX,,DAQAPKA

Pool:

X&X&X, x,cx,cx, X&X&X, x,cx,cxs x,cx,cx, x,cxgcx,

from

randomized

sequences. Diversity+ 2x10’3

5 x 10s

Target c-myc

(each)

References

Result

mAb

Epitope X(Q/E)XLISEXX(L/M) Mimic

VEGF

3.5 x 1 Og (total)

(a) [69-l

(i) RGWVEICVADDNGMCVTEAQ (ii) GWDECDVARMWEWECFAGV

x4cx,ocx4

X,CX,GPX,CX,

X, ,GGGS

SRX,

,SR

SADGAX,,GAAGA

ADGAXsGAAG

X 10

5x10s

VEGF

1.9 x 109

1 xl09

Paclitaxel

Tropomn

2.5 x 108

C

CTLA4

4x107

CSF

2 x 1 Og (phage)

Mimic GERWCFDGPLTWVCGEES

[69-l

Epitope HTPH

11071 P

Epitope (VIL)(DIE)XLKXXLXXLA

mAb

IgG

Ii peptide

[751

Unknown (i) GFVCSGIFAVGVGRC (ii) APGVRLGCAVLGRYC

11081

Epitope RRPFFX

I1091

Epitope WFSWGFPCIWW

I1101

LEX,TS

2.8 x 10s

(clones)

PAI-

Unknown GCIFWHLTSMSGSYFLEPFDLIST SQQRNIPLEIRDADT

11111

LEX , ,TS

1.4 x 1 O7 (clones)

PAI-

Mimic PVSQFVFLCGHQPCFTSEHAHDVP DPAPPHHPLELITGRCIATPISVGMS

11111

*Flanking sequences shown where reported. ?Where further differ, values given refer to only the initial round of selection. represent diversity, except where Indicated (clones = number transformants; phage = number of input phage). Sequences type represent ammo acids encoded by DNA sequences

respective small molccr~le. Wcrstuck demonstrated that allosteric mRN.4 erated by inserting 3 drug-binding S’-untranslated region (S’-IU‘R) of tion allow both it/ s&u and it/ repressed in the presence of a small

rounds Values of in bold

and (;recn [Ih’] have templates may be genaptamer into the an mRN.4. ‘I‘his inserGsv translation to bc molecule.

The thirst for structural information I,ogic:~lly, structural analysis providch ;I clear path to understanding the mechanisms of recognition jnd catalysis pcrformcd by functional nucleic acids. Indeed. Jacks [-I71 argued that, for protein enqmcs, the most striking generalization WLS that many of the expectations bawd on chemical studies were simple cwnfirmrd once crvstal strllcturns were in hand. In rhis l.cin, the striicnirc of alptamers in S(JlLltiOn solved using NhIR spectroscopy has generally pro\idcd significant insight inro the recognition process, indic:lting the presence of compact binding pockets for slnall inolcclilcs (for rwiws, set [-C--SO]). Analysis of

corresponding RW Roberts, observations. co-stimulatory MHC class

to restrictlon sites. (a) P Burgstaller, S Hale, M Wright, R Liu, JW Szostak, RW Wagner, unpublished CSF, cerebrospinal fluid; CTLA4, anti-T lymphocyte molecule; Ii, Invariant chain peptide from cytosollc tail of II; PAI-1, plasminogen activator Inhibitor 1.

three pockets (present in A’l‘l’, FRIN and @nine aptamers) indicates that binding appears to provide somehvhat less of a tight fit than protein sites and relies heavily on stacking, 3s opposed to hydrogen bonding [SO]. ‘l’he path has been somewhat bumpier when examining nucleic acid catalysts. ‘I’he central difficulty is crystallizing either RNA or DNA enzymes in their active form. In the crystal structure of the 10-23 DNAqme, the molecule has rearranged to form a dimer that is not likely to be the active conformation [Sl’]. ‘I’he sn-ucturc of the lead-dependent rihozyme crystallizes in two forms. one of uhich is ‘precat3tytic’ and rationalizes the cleavage chemistry (S?]. IXfficulties are seen in structural analyses of natural ribozvmcs as well. hluch has been made of the conundrum tlut the hammerhead ribozyme structures do not rationalize the cleavage chcmistr?; [S.i,%]. whereas larger RN.4 catalysts. such as the group I intron catalytic core [SS] and the hepatitis delta virus rihozyne (after cleavage) I.561,

524

Engineering

Table

2

Proteins

and design

displayed

Protein

on phage. Clones*

modlfled

Heregulin (nine mutagenized Knottins

1 .O-6.4

sections separately)

(CBD)

Llpocalln bindlng

(16 residues site)

in

hGH (site 2: four sectlons mutagenlzed separately)

*Values

represent

the

ErbE3

x 1 O*

receptor

References

>50-fold

higher

afflnlty

Cellulose tr-Amylase Alkaline phosphatase [&Glucuronldase

Epitope Unsuccessful K,=lOpM Unsuccessful

3.7 x 10s

FluoresceIn

K, = 35.2

nM

(>102-told

hGH

K, = 0.16

nM

(38.fold

higher

afflnlty)

domaln;

hGH,

human

growth

x 107

of transformants

in the first

round.

receptor

CBD.

C-terminal

cellulose-bIndIng

therein).

[87”]

H>- comparison.

the

hxttisc

the

imcnt that

cm

IX

ribosome Ability

and proteins

Selections ‘i‘hc tnosT

using

prorcitls

and

Tc’iii

dcvelopcd

!n

:ttf‘init!,

toh,

2nd

c;iI

panning

of ~tly~ro~itmttcl~

iising

;I \3ri:iriori

infi-ctive

phtig:c

wlccrioti int’cctt\,it!, libr-;ir)-

of or

‘S11”

iii

\\ ith

:I tot31

Iligh

in

binding I I I f(tsions

libr:tr\

rccovcred

dtthbcd

(rc\ic\+cd

to

\I hcreb\

uliarion.

h:ts also. t)ccn

1,

0.0001 ‘,C IX cl cn selccli\c-

ISX],.

111

i\ rcclttired containinK

constraints

iniposcd

translated

prowins

ril)osontc

iticrcasing ribosomc

one

IIillion

the

lihrar\ prouzina

six

has

to otticr

mainI\ In addirion, ph;tx:c pwt2rftil techniqites.

4s

t-ccoliil)inacioii

in tuctcria.

;ttld

nlolcculcs.

(WC

~tscd

for

been

the

i~/c.it/~~

limircd

frOmi

prescricc \vhich

*I

31ltl

the

or 101.i pcplidc4 rc\

prowin\

ic\vS).

‘4

hid

G;I c’ompruiicn-

IW].

cl

i-cfcrcncc,

In this

itscd cl.oltition

scl~~\~s

cull.

In ribosotn~

remain

of Ioueritig

display.

complexed

the

with

tcnipwttttre

and

[O.i]. Selections dctnonstratcd the

(single-chain

is Ihat

approach. .Afwt-

pcptid)l u ith the

lnatel\

host

mKN4

trrtn\lation or other

\ 3riat)lc

libr;iricS

fb~‘].

libraries

mitsr

ttsing aKinit\’

dotnain (hlc

frq-

Of

~illlir~i~iOi~

he screened

in rhe

rhc entireI~i~JiJSOtllc iindcicondition:, in tcr-itar\ mKN~i-ril,osolne-pepri~i~l-tliN ~1 is stahlc ;liigh hlgJ+ arid 1014~ remperatttre). A nicthod for producing co~.alcntl\ linked fitsion~ ha hccn rcccntl\ introdrtc.cd

.i’-puromycin.

abottr

cell-free in I itro :t11\ wnsfortn;lCon

of

(4,5”].

cliie C(J the trarisfcction liinit disl>la\, is nc,r easily poj-tzl>lc includln~ i/l E+//.o rnttt3gcI1C. I.SY.f~O”*.f,l

nietfiods,

pro\.idc

the

coii~l~lcx IlO\

by chc and

iilliiiilnized

displa);

fortltc

selection

to onl!

01‘

tiieil~s)

\.arietv

and

expcr-

transl:ttcd fOr

is littkccl

the hl$’ concetitracion displa), rcccntl! ha\,c

rlii\

target.

rccentl~~.

pcpridcs

10”

[~?,h]

reporred

34 3 resttlt

in;itiir:lfioti

linkage I ltitil

infortn;triott

:\II of these stratcgics ttTili/c \) siccms. therch\, a\ aiding

tilKNA-protcitl

New methods of

than

the

of tiiolccttle:,

fusions.

(XX

sclccin

is impotunt

dc\~ulolxx1

kcepin::

IJ!

inKk/j

approach.

ril~osornc

niulriplict-

01‘ pliage

displ:t!.

rcccp-

In ;i typlihr;tr\ is

the

i ii

O\vIl

acid sreps

combinatorial

rccentl)

of more’

rcspccti\,cly.

hormone.

tlutnlwr

niK\/\-prowin

lil)rarics

11121

nttclcic

ofa

to the

‘I’\~-0

and

ICI dlcir

the

bacteria

hctuccti 0.01 % :ind ‘I‘his pcrcctltag:’ can

proccdttre. ligaticlLtar,p3 itsing hipartatc g:enc ~iid

uittt

mcml)crs.

ph;ig:c

or

selccrcd 01

IrottnJ~.

fr:tctiott

lo\\ (rvl~ic:tll~ popiilation).

\‘cI-\

31~

mixctl

(>lOOO).

tltc

(- l-7700 1’111

substraws

ticrlcc

IO-10”

Ixxaiisc

iripttt

lower

scq

\)s-

in the K tcrniiiitis hactcriophagc

\trh\ccl”cnr

phagc

inany

i\ csscntial,

~cncr;ill\ of the

f‘or

prorcins.

with

b38’1

;irc

gciic

hlolcc~tles arc

phagc

t5CIl

I,\

I0

imnrot~ili~cd

pliaqc

cXpcrinierir.

\i7c

I>

rcsitltirifi

pro~ctl)

represented r\

on

display

formar

fiirtons

[7..57’]).

of

Scqitencrs

tntilticopv

sw

selection

phqc

(O-S col~ies/pha,gc) sitrf2cc of filametitotts

rebic\v.

tlic

prmliicc

it/ Ct/u

Snlich.

Y-terminal

‘single’ copy fortii;rt 01‘ ;gctic II I on the 3 rccwlt

the rhc

(;corgc

in

cithct 3s

ccJl~ics/phafic)

for

screened.

ro scrccii

linked

is currently IJ!

cxprcsscd

(for

display

techniclttc

common

pcptides

or

phage

proportional

display

affinity)

bccausc all the ‘l’his diftbrencc

pwvcr

~Otll~ltl~~~~Oll~i~

is dirccrly

higher

litnit

tions is 1 ()t5-1()t(~ seqttrnccs, process arc pcrformcd ;?I s.itm.

Peptides

[ 104”]

5.5 r 108

1.3-7.3

number

Result

Target(s)

transfuasr protein allous

that

a mKNA

Cr3ns13tion.

2

selccrions

ttr bc Kccenrl!.

arid

scrccii

x 1 ()I.“ diffcrcnr to the anti-c-tn\ic,

mcnT tltcror5 (1’ Hurgstallcr. .I\L’ S/.oartik,

is tagged

puromycirl

site and bccotnes it encwdcs. ‘l’his

of conditions. to generate

librq rlic

of more S Hale, 1<\2~ \Vagrtcr,

co\~alently chctnically

pcrformcd rhic

3 lihrar\ peptidcs antihod!-

rh:rn 700 hl \Vright. ttnl>ttt)lishcd

lvith

enters

linked rohttsr

i~t~clct-

techniqrtc conraininx for YElO.

:t the

;I

has

wide

IXCII

approxiIjinding and wirh enriclt-

per rotlticl (‘li~hle 1) RM’ Roberts. K I,itt, ot,aer\,ations).

In vitro

Se~cral other technical ad\,ances arc also expected to techniques, cnhancc the power of ill r:ifro selection including improved in tvtro recombination and mutagrnsis. and interaction mapping by XMK spectroscopy. New methods for in vitrn recombination may bc technically simpler and/or more eff*ecti\,e than the original IINA shuffling technique devised by Stemmer [hS]. Among these is family shuffling [60”], whereby di\,crsity is created b> recombining two or more homologous gents. ‘I’his has the ad\,antage of creating di\.ersity that ha\ alrcady been prescrecned 1,~ nature to weed out mutatil)nb thar are deleterious to :I particular fold. generating libraric\ that are much richer in functional proteins. Alternatives to the original I)Yase-I-based shuffling pietow1 ha\,e also been dc\,tzlopcJ. including random priminK recombination (KI’K) [M] and the staggered extension process (61’1. Kandom clon,gation rnirtagcnesis (Kk:kl). the addition of a partially randomized sequence to the (: terminus of’ the protein. i\ rrnlikel\~ to change catal) tic propcrtics. hiit 111;1\ bc ;I siinplc apl)roac‘h for enhancing enzyme \tabitit! (073. III an increasing nurnbcr 01‘ casts, thcb conforn~ation~ 01 sclcctcd n~01cc11Ies (peptidc\ and pmtcins) bound to thcirtarget4 h;I\.c’ been sol\Tcd 1~) crystallographic (see discussion below). NhlK spcctrowlp~ has pro\ed to bc increasingly uaefui ti)r char-actchri/.ing l)oth dynamical sy+ tcnis and those that arc rccalcirrant to cr\srallization. ‘li\o rqJrcscntati\ c examples arc given. ‘I’hc Schultz and \Ycmmcr laboratories [6X’] ha\c rcccntl\ iised transt‘crrcti NOI< (nuclear O\whauscr cnhancsmcnt) cxpct-inicrits t(~ dctcr-niirrc the qzonrctr\~ of‘ t\I 0 \rrbstratca in tlic ;rcri\c site ofarr a(,\.1 tr-snsferasc antiljod!. I~;rirlJrothcr- /I/ N/. [OO**/ irscd t\\.o-~iiiiiciisioii;iI I I I-l’K I I.S(>C 1 (hctcronrrcIc3r billfilC (1uantirm cohcr-crrc,e) cxlwr-inicnt5 to dcl‘inc the interactiori bct\vccn bclectcd 1Jcptidc.s 2nd tlic rl l,ri%labclcd target VE(;IC I’cptidc intct-action5 with \~l
selection

of nucleic

acids

and proteins

Roberts

and Ja

525

.An increasing number of laboratories have reported peptides or small protein tigands that can functionally mimic the three-dimensional surface that is normally bound by a full-sized protein. ‘I’hese examples are particularly interesting because they emphasize both the plasticity of protein surf&es and the range of structures (both in size and topology) that can functionally substitute for one another. ‘I’he functional substitution of small peptides for much larger proteins runs counter to the notion that protein interactions arc driven in proportion to the burial of h!rdrophobic surf&e (see [7h] and references therein). One of the most compellink cr e.uamples of this mimicry arc the erythropoetin mimetic peptides (ti;2ll’). short (-13-20 residues) p hairpins that functionally mimic erythropoetin, a .M kl)a glycoprotein. but bear no sequence homology 177.7X.79’]. Similarly, prptides or minimized proteins have bc*en found that mimic the function of thrornbopoetin IXO], an antagonk VkL(;I; lhY”.70’], or bind to constant tcgions in antibody structures (XI 1. ‘l‘lrc peptides isolated by both types of selection generatI! bind to the same regions utilized by the natural counterparts they arc meant to mimic. I’or linear epitopes, this is perhap” not surprising. I’Or’ the peptide mimics. it borders on scandalous, as, often. most of the target protein surface i\ available for interaction during the aelcction. ‘l’trc clear implication is that sonic regions of protcin srrri’ac~~ arc more ‘stickv’ than others and, thus. act as hot spots for interaction iX.!I. Antibody selection: the importance of large libraries Although complete co~eragc ofpha~c antibody selection is tjcvond the scope of this war-h. a recent re\.ic\r highlights the importance of high-colnlllc~it~ starting librdrics [.57’]. \z’lrci~ lo\\~-con~~~lcsit~ libraric~ containing 3 x 10’ clones \\crc rised. single-chain antibodies with affinitic\ ranging from 10 to 0.1 X 10~” pZl IX.%.XAJ \\cre isolated. With lrbi-arics containing 10”’ clonc~, affinities of 10 to O..‘, nhl \+‘cr’c tijund (X.51. ‘I’hc commensurate increase in binding constant with increased library \i/c is in line with thcorctical pi-edictions dc\,clopcd to cstirnatc the binding ~onst;~nts from IitJrar~~ complcxit~~ (XOI. Alternative scaffolds II‘ Iqtidcs can bc found that mimic proteins. one might ekpcct that snia11. strrrctured pcptidcs could be used as scaffold\ to display function. Kather than choose ;I single motif’. I~‘airbrothcr otcr/. [W”] used 3 variety of small, disulfi~lc-coiitaininC pcptide\ to isolate binders to \‘K(;f; c’I~IIJIc I). Similarly, knottiw. small, disulfide-rich proteins h.r\,c been irsctl to generatr librark that recognize alkalint phosphatasc, a ~H’Oteill Lrith no wild-type knoctin binding 1X7”]. j’r.otcin scaffolds that cordd IX imbued with antibody di\.crsit\: h:r\,e also been sought. particularly beca~~se ofrhe rcfractorq natrtrc of some scl*‘\~. Skcrrd and co-workers (HX’I ha\ c rcccntlv reported that lipocalin HI3I’ variants

526

Engineering

and design

(a 174 amino acid p barrel) could be isolated that bound fluorescein with a dissociation constant (K,) of 35 nlvl. Direct selections

for catalysis

There is great interest in using selection both to refine known catalysts and to isolate new ones. TSAs have been used to isolate a great diversity of antibody catalysts via ill v&o immune selection (151. In a few cases,irt vitro selection has been used for affinity maturation of the catalysts using ‘1‘SAs [89,90’,91,923 or via a mechanism-based approach (discussed below). The potential limitations of the ‘TSA approach are highlighted by work from the Wells’ laboratory. Baca rt ul. [89] used affinity panning of an esterolytic antibody library derived from immune selection to isolate variants with two- to eightfold improvements in their binding affinity for a phosphonate ‘EGA. Notably, a correlation between binding afftnity and catalytic ability was not observed. A weaker binding variant was identified with twofold greater catalytic activity than the parent antibody, emphasizing the need for an approach that selectsfor catalytic activity [WY]. In the direct strategy. enzymatic activity is required to either covalently attach or releasethe library membersduring selection. This approach hasbeen successfully applied to isolate RNA and DNA catalysts that are capable of performing chemistry with rates ranging from 0.1 to more than 100 min-1. significantly faster than many catalytic antihodies [33]. Lerner and co-workers [93,94] ha1.eapplied this strategv to immune selection with great success.In one interesting combination of reacti\,e and ‘I‘SA approaches, phage display was used to improve catalytic antibodics designed to cleave a glycosidic bond by 700-fold (k,,,/k,,,,,, = 7 x l(P versus 102 from immune selection) [OSj. Substrate-bond cleavage generated a quinone methide that reacted with rhe phage to afford immobilization and selection. Reactive approaches can also be incorporated in phage selections to afford specific release. The Schultz laboratory [96’] has recently demonstrated that phage harboring active staphylococcal nuclease can be enriched IOO-fold in a single round, based on cleavage of an oligonucleotide tether in C/S.A similar strategy involving immobilization demonstrated enrichments of more than SO-fold for two model proteins using phage-displayed enzymes and the calciunl-depelident binding of substrate to calmodulin [97]. Structural proteins

studies

of selected

and evolved

peptides

and

‘I’here are now a significant number of selected prptides and proteins whose structures arc known. \Ve include catalytic antibodies and i/f d/a screened proteins in this discussionbecausethe structural conclusions seem gcnerally applicable to all functional approaches. One \‘crv striking feature is the way in which protein structure can be modulated by seemingly subtle changes in sequence.

In general, proteins that have been evolved either in viva or in vitro contain positive mutations that are distal to the active site ([98-l 001;B Spiller, A Gershenson, FH Arnold, RC Stevens, personal communication). These mutations often have the effect of remodeling the shapeof the active site and may act by biasing the conformation towards lockand-key-type recognition [99,101]. In addition, the selected mutations often make significant changes to the local structure (B Spiller, A Gershenson, FH Arnold, RC Stevens, personal communication) and can even alter topology [1OZ”]. As interesting as where mutations are is where they are not-the active site or critical residueson the interface of the proteins. For example, residues conserved in phage selections of heregulin variants (binding to the ErbB3 receptor) correlate strongly with positions that were found to be important using alanine-scanning mutagenesis [103’,104”]. Selection and evolution thus seem to avoid changing residuesthat are in direct contact with the target. Rather, beneficial mutations enhance affinity by reconfiguring essential binding residues or by forming new contacts. Indeed, these complicating secondary effects prohibit the delineation of a recognition code, even for highly modular protein contacts, such aszinc-finger-DNA interactions [lOS**]. The importance of long-range interactions and plastic local structure are likely reasons why rational design has yet to produce the results seen with functional approaches, such as in vitlz, selection [lob]. Computational approachesmay significantly enhance pool design. however, allowing us to ‘stack the deck’ of initial libraries, making selections more fruitful or reducing the number of cycles required.

Conclusions IN eli~o selection has proved itself to be a powerful technique. For the foreseeable future, functional approaches provide the most facile path for engineering nucleic acids and proteins alike. The new methods developed will allow larger, better libraries to bc generated and more powerful \vays to screen them to be elaborated - further expanding our ability to generate interesting molecules for studv. A fundamental question is how the information from selection experiments can be used to further insight into biopolymer folding and function. l;or the present, the results of selection and screening cxperimcnts provide existence proofs and hard data on some very interesting questions, including the number of nucleotides needed to specify a catalyst, the number of amino acid changesneeded to improve protein thermostability and the free energy consequences(or lack thereof!) of introducing or removing a small functional group (e.g. a methyl group) from either the macromolecule or its target. Future experiments are likely to addressissuessuch asthe generation of new function. the minimum size of protein needed to perform catalysis and the likelihood that such a protein could be found in random sequence libraries.

In vitro

Acknowledgements 1%‘~ thank Frances H Arnold and rhcir manu&pt prior IO publication.

References

Kdymond

Stcvcns

and recommended

Papers of particular have been highlighted

interest, as:

published

for providing

a copy

the annual

period

Szostak JW, Ellington RF: ln vitro selection of functional RNA sequences. In The RNA World. Edited by Gesteland RF, Atkins Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 1993:51 I-533.

2.

Smith GP, Petrenko 97:391-410.

3.

Jermutus L, Ryabova L, Pliickthun A: Recent and selecting functional proteins by using Curr Opin Biotechnol 1998. 9:534-548.

4.

Roberts selection 94:12297-l

VA: Phage

RW. Szostak oipeptides 2302.

Roberts fusions 273.

RW: Totally and ribosome

display.

JW: RNA-oeotide and protein;.

LIU R, Barrick J, Szostak RNA-protein fusions Enzym0/2000, 317:ln This paper details sigmflcant fusion synthesis. In addition, tern can be generated as specificity as naked fusions. 6.

Chem

JF.

23.

Wlegand synthesis.

24.

Zhang B, Cech TR: Peptide bond ribozymes. Nature 1997, 390:96-l

proteins

Roberts

JW: Ribozyme-catalyzed 1996, 381:442-444.

and

amino-acid

TW, Janssen RC, Eaton BE: Selection Chem &o/1997, 4:675-683.

527

transfer

of RNA

formation 00.

Ja

by in vitro

amide selected

advances cell-free

fusions Proc Nafl

for the in vitro Acad SC; USA 1997,

in vitro protein selection using mRNA-protein display. Curr Open Chem Bioi 1999, 3:‘268-

Ellington A, Szostak JW: ln vitro selection of RNA bind specific ligands. Nature 1990, 346:818-822.

8.

Gold L, Polisky oligonucleotide

molecules

that

9.

Wills MC, CoIlIns B, Zhang T, Green LS, Sebesta DP, Bell C, Kellogg E. Gill SC, Magallanez A, Knauer S ef al.: Liposomeanchored vascular endothelial growth factor aptamers. Bfoconjug Chem 1998, 9:573-502.

10

Willlams KP, LIU X-H, Schumacher ‘TNM. Lln HY, Auslello DA, Kim PS, Bartel DP: Bioactive and nuclease-resistant L-DNA ligand of vasopressin. Proc Nat/ Acad So USA 1997, 94:l 1285-l 1290.

0, Yarus M: Diversity of Annu Rev &o&em 1995, 64:763-797.

11

Nakamura translation

Y, Ito K, lsaksson LA: Emerging understanding termination. Cell 1996, 87:147-l 50.

12

Wilson engine.

13.

Pestova TV, Shatsky IN, Hellen CU: Functional dissection of eukaryotic initiation factor 4F: the 4A subunit and the central domain of the 4G subunit are sufficient to mediate internal entry of 43s preinitiation complexes. MO/ Cell Viol 1996, 16:6870-6878.

14.

Hentze MW: elF4G: 1997, 275:500-501.

15

Schultz lessons

PG, Lerner from the

16.

Prudent catalysis.

JR, Uno T, Schultz Soence 1994.

17.

Conn MM, by a small

10.

LI Y. Sen D: A catalytic Bioi 1996. 3:743-747.

19.

Noller HF, Hoffarth V, Zlmnlak L: Unusual resistance of peptidyl transferase to protein extraction procedures. Soence 1992, 256:1416-1419.

20.

Samaha RR, Green R, Noller HF: A base pair 23s rRNA in the peptidyl transferase centre Nature 1995, 377:309-314.

KS, Noller HF: Molecular Cell 1998, 92:337-349.

movement

a multipurpose

inside

ribosome

of the

adaptor?

translational

the

Prodent JR, Schultz PG: Porphyrin RNA molecule. J Am Chem Sot DNA

for porphyrin

Green R, Swttrer C, Noller HF: Ribosome-catalyzed formation with an A-site substrate covalently ribosomal RNA. Soence 1998, 280:286-289.

Welch M, Majerfeld for peptidyl-transferase 36:6614-6623.

27.

Nitta I, Kamada Y, Noda H, Ueda T, Watanabe peptide bond formation with Escherichia domains. Science 1998, 281:666-669.

scope

842.

Nat Struct

between tRNA of the ribosome.

linked

peptlde-bond to the 23s

M: 23s rRNA similarity from mimicry. Biochemistry 1997,

RNA-catalyzed

selection

K: Reconstitution co/i 23s ribosomal

nucleotide

of RNA

synthesis.

Nature

known ribozyme reactions to include nucleotide pRpp (the activated nbose used In nucleotlde 4-thlouracil (k,,, = 0.13 mln-‘1.

Tuschl T, Sharp PA, Bartel DP: Selection in vitro of novel ribozymes from a partially randomized U2 and U6 snRNA library. EM60 J 1998, 17:2637-2650. The attempted resurrection of RNA molecules with spllceosomal activity. Four separate catalysts were Isolated, none of which corresponds to the spllceosomal mechanism. Interestingly. one has a complex three-step mechanism. 30.

Tarasow TM, Tarasow SL. Eaton BE: RNA-catalysed bond formation. Nature 1997, 389:54-57.

carbon-carbon

31

Roth A, Breaker RR: An amino acid as a cofactor for a catalytic polynucleotide. Proc Nat/ Acad Sci USA 1998, 95:6027-6031. ;hs first demonstration of free hlstidine as a necessary cofactor for a DNAcatalyzed (or RNA-catalyzed) reaction. The paper highllghts the utility of enhancing the functional group diversity of nucleic acids. 32

Joyce GF: Nucleic acid enzymes: playing Nafl Acad Sci USA 1998. 95:5045-5047.

with

33.

Breaker RR: ln vitro selection Rev 1997. 97:37 I-390.

polynucleotides.

34.

Santoro enzyme.

35

Santoro SW, DNA enzyme.

36.

Cuenoud activity.

37.

Breaker RR, Joyce GF: A DNA Biol 1995, 1:223-229.

38.

Faulhammer dependence DNA/RNA 202.

39.

Geyer CR, Sen D: Evidence of an RNA phosphodiester-cleaving 4:579-593.

40.

Carml DNAs.

41.

of catalytic

SW, Joyce GF: A general Proc Nat/ Acad So USA Joyce GF: Mechanism Bochemistry 1998,

a fuller

Wilson aptamer

deck.

and utility 37:13330-l

that

DNA

of an RNA-cleaving 3342.

B, Szostak JW: A DNA metalloenzyme Nature 1995, 375:61 I-614. enzyme

Proc Chem

purpose RNA-cleaving 1997, 94:4262-4266.

with cleaves

DNA RNA.

ligase Chem

D, Famulok M: Characterization and divalent metal-ion of in vitro selected deoxyribozymes which cleave chimeric oligonucleotides. J MO/ Bloll997, 269:188for the

metal-cofactor DNA. Chem

N, Schultz LA, Breaker RR: In vitro Chem Biol 1996, 3:1039-l 046.

independence Biol 1997,

selection

Travasclo P, LI Y, Sen D: DNA-enhanced peroxidase DNA aptamer-hemin complex. Chem Biol 1998, ;he authors demonstrate that the N-methylmesoporphyrin enhances the peroxidase activity of hemin by two orders tlve to free hemin. 42.

and

PJ, Bartel DP: 395:260-263. expansion of A 3’+nked IS joined to

I, Yarus

its surcen-

29. .

of RNA

metallation catalyzed 1996, 118:7012-7013. metallation.

28. Unrau .. 1998, The further biosynthesis. blosynthesis)

Soence

RA: From molecular diversity to catalysis: immune system. Science 1995, 269:1835-l PG: Expanding 264:1924-l 927.

26.

in producing translation.

JW, Roberts RW: Optimized synthesis of for in vitro protein selection. Methods press. improvements in the yield of mRNA-protein tt IS the first to demonstrate that a full-sized proa fusion and that RNA-bindlng peptides retain

8, Uhlenbeck functions.

Zhang B, Cech TR: Peptidyl-transferase ribozymes: trws reactions, structural characterization and ribosomal RNA-like features. Chem Viol 1998,5:539-553. structure determination of clone 25 from 1241 reveals The secondary pnslng homology to three distinct sequences in the peptidyl transferase ter of the ribosome.

Rev 1997,

7.

21.

and

25. .

of special interest **of outstandlng Interest

5. ..

acids

Lohse PA, Szostak reactions. Nature

of revnew,

l

1.

of nucleic

22. of

reading

within

selection

of self-cleaving activity of a 5505-517. DNA aptamer of magnitude rela-

C, Szostak JW: Isolation of a fluorophore-specific with weak redox activity. Chem Viol 1998,

DNA 5:609-617

43. Li Y, Breaker RR: Phosphorylating DNA with DNA. Proc Nat/ Acad . SC; USA 1999,96:2746-2751. The authors describe the isolation of more than 50 different DNA enzymes that are capable of autophosphorylation. The enzymes In the pool demonstrate very high specificity for ATP as a substrate versus the other three nucleotides. or specificity between dNTPs versus NTPs.

528

44.

Engineering

and

design

Tang J, Breaker RR: Mechanism for allosteric sensitive ribozyme. Nut/e/c Acids Res 1998,

inhibition of an ATP26:4214-4221.

45. Soukup GA, Breaker RR: Engineering precision RNA . switches. Proc Nat/ Acad SC; USA 1999, 963584-3589. The combmatron of aptamers with the hammerhead ribozyme ate artrficral allosteric ribozymes 8s described. 46. .

Werstuck G, Green MR: Controlling gene expression through small molecule-RNA interactions. Science 282:296-298. The use of aptamer structures rn the 5’.UTR (untranslated allows the down-regulation of translatron rn the presence cule target. Binding energy, specificity, Adv fnzymol Relaf Areas

molecular to cre-

in living 1998,

cells

65.

Stemmer WPC: DNA shuffling by random fragmentation and reassembly: in vitro recombination for molecular evolution. Proc Nat/ Acad Sci USA 1994, 91 :10747-i 0751.

region) of mRNAs of the small-mole-

66.

Shao 2, Zhao H, Giver L, Arnold FH: Random-priming recombination: an effective tool for directed evolution. Acids Res 1998, 26:681-683.

67.

Matsuura T, Miyar K, Trakulnaleamsar Yamamoto K, Urabe I: Evolutionary random elongation mutagenesis.

Jencks WP: Circe effect.

48.

Fergon Chem

49.

Pate1 DJ: Structural Chem Biol 1997,

50.

Marshall KA, Robertson MP, Ellrngton AD: A biopolymer by any other name would bind as well: a comparison of the ligandbinding pockets of nucleic acids and proteins. Strucfure 1997, 5:729-734.

T, Smrth 3:61 l-61 analysis 1:32-46.

FW: 7.

and enzymic MO/ Biol 1975,

Aptamer

of nucleic

structures acid

Hanes J. Jermutus L, Weber-Bomhauser S, Bosshard HR, Pltickthun A: Ribosome display efficiently selects and evolves high-affinity antibodies in vitro from immune libraries. Proc Nat/ Acad So USA 1998, 95:14130-14135. The frrst report of the construction and Implementation of an rmmunrzed munne s&v ltbrary using ribosome display. The best clone appears to have 65.fold higher affinity than the sequence it was derived from.

rn order

47.

J, Terkmann B/o/ 1996,

64. .

catalysis: 43:219-410. from

aptamers.

the

A to 5.

Curr

Driggers EM, LIU CW, Wemmer DE, Schultz PG: Structure of the Michaelis complex of an efficient antibody acyl transferase determined by transferred nuclear Overhauser enhancement spectroscopy. J Am Chem Sot 1998.120:7395-7396. ..-Transterred Nut experiments were used to determlne the structure 01 two bound substrates rn the presence of an acyl transferase antrbody. Farrbrother WJ. Chnstrnger HW, Cochran AG, Fuh G, Keenan CJ. Quan C, Shrrver SK, Tom JYK, Wells JA, Cunningham BC: Novel peptides selected to bind vascular endothelial growth factor target the receptor-binding site. B,ochemisfry 1998, 37:17754-17764. Lrbrarres of drsulfide-constrarned peptrdes drsplayed on phage ytelded peptrdes that btnd to the receptor-bindrng domain of vascular endothelial growth factor (VEGF). HSGC spectra of ‘H. ‘5N-labeled protein mdrcate the sites of peptrde interaction, suggestrng a bindmg ‘hot spot’ that can be readily targeted. l

51. .

Nowakowskr J. Shim PJ, Prasad GS, Stout CD, Joyce GF: Crystal structure of an 82nucleotide RNA-DNA complex formed by the 1 O-23 DNA enzyme. Nat Sfruct Biol 1999, 6:151-l 56. An tnterestrng nuclerc acrd structure IS described that probably does not conform to the actual geometry ot the to-23 DNA enzyme. 52. ..

Wedeklnd JE, McKay DB: Crystal structure of a lead-dependent ribozyme revealing metal binding sites relevant to catalysis. Nat Strucf B/o/ 1999. 6:261-268. The frrst crystal structure determrnatron of a laboratory-Isolated RNA enzyme. The structure contarns two molecules, one of which is in a precatalyttc. geometry that can be used to ratronalrze the cleavage chemrstry. McKay DB: an unfinished

54.

Verma S, Valsh NK, Ecksteln F: Structure-function studies of the hammerhead ribozyme. Curr 0p111 Chem B/o/ 1997, 1:532-536.

55.

Golden BL, Goodlng AR. Podell ER, Cech active site in the crystal structure of the Soence 1998, 282:259-264.

56.

Structure story.

Ferre-D’Amare hepatitis delta

and RNA

function of the 1996, 2:395-403.

hammerhead

57. Grrfftths AD, Duncan AR: Strategies . phage display. Curr Opz Bfofechnol An excellent review of antrbodres drsplayed

ribozyme:

TR: A preorganized Tetrahymena ribozyme.

AR, Zhou K. Doudna JA, Crystal virus ribozyme. Nature 1998,

structure 395:567-574.

58.

Spada S. Pluckthun A: Selectively infective phage technology: a novel method for in viva selection protein-ligand pairs. Nat Med 1997, 3.694-696.

59.

Tsang J, Joyce GF: In vitro evolution of randomized Methods Enzymol 1996, 267:41 O-426.

69. *

70 .

Wiesmann C, Chrrstrnger HW, Cochran AG, Cunningham Facrbrother WJ, Keenan CJ, Meng G. de Vos AM: Crystal the complex between VEGF and a receptor-blocking BYochem/stry 1998, 37:17765-i 7772. One of the selected peptrdes rn [69”] was crystallized bound peptrde adopts a largely extended conformatron resemblrng Interactron more closely than the VEGF-receptor interactron.

(SIP) of interacting ribozymes.

60. ..

Cramerr A, Rarllard SA, Bermudez E, Stemmer WPC: DNA shuffling of a family of genes from diverse species accelerates directed evolution. Nature 1998, 391:288-291. In nfro shufflrng of four related cephalosponnase genes produced a recombinant with greatly enhanced actrvrty in fewer cycles than through mutagenesis and recombtnatron of the tndrvtdual genes. Zhao HM, Grver L. Shao ZX, Affhoiter JA, Arnold FH: Molecular evolution by staggered extension process (STEP) in-vitro. Naf Biofechnol 1998, 16:258-261, Short cycles of annealrng and extension rn a PCR-lrke reaction produce prematurely terminated products that can randomly anneal to parent strands rn the next cycle, creating recombrnant sequences. Thts novel method was used to evolve a subtrlisrn E variant that IS equivalent to Its thermophrlrc homolog.

63.

Hanes J, Pluckthun proteins by using 94:4937-4942.

A: In vitro selection ribosome display.

and evolution Proc Nail Acad

of functional SC; USA 1997,

to VEGF. The Fab fragment

72

Sparks AB, Hoffman NG, McConnell SJ, Fowlkes DM, Kay BK: Cloning of ligand targets: systematic isolation of SH3 domaincontaining proteins. Nat Biofechnol1996, 14:741-744.

73

Sparks AB, Rider JE, Hoffman NG, Fowlkes DM, Qurllam Distinct ligand preferences of Src homology 3 domains Yes, Abl, cortactin, p53bp2, PLCy, Crk, and Grb2. Proc SC/ USA 1996, 93:1540-l 544.

74.

Pasqualinl R, Koivunen E, Ruoslahti E: uv lntegrins as receptors tumor targeting by circulating ligands. Nat Biofechnoi 1997, 15:542-546.

75

Pierce HH, Schachat F. Brandt PW, Lombard0 CR, Kay BK: Identification of troponin C antagonists from a phage-displayed random peptide library. J Biol Chem 1998, 273:23448-23453.

76.

Spolar RS. Ha J-H, Record MTJ: Hydrophobic folding and other noncovalent processes Nat/ Acad So USA 1989, 86:8382-8385.

effect involving

77.

Wrtghton NC, Farrell FX, Chang R, Kashyap Mulcahy LS, Johnson DL, Barrett RW, Joilrffe peptides as potent mimetics of the protein erythropoietin. Soence 1996, 273:458-463.

AK, Barbone LK, Dower hormone

78.

Lrvnah 0, Stura EA. Johnson DL, Mrddleton SA, Mulcahy LS, Wrrghton NC, Dower WJ, Jolliffe LK, Wilson IA: Functional mimicry of a protein hormone by a peptide agonist: the epo receptor complex at 2.8 A. Science 1996. 273:464-471.

61. .

62. Tawfrk DS, Gnffrths AD: Man-made cell-like compartments for . molecular evolution. Nat Biofechnoi 1998, 16:652-656. In v/fro transcrrption and translatron were lrnked vra compartmentakzatron rn reverse mrcelles. A selectron for targeted DNA methylation using Haell methyltransferase was demonstrated.

of

Kay BK, Kurakin AV, Hyde-DeRuyscher via phage display. Drug D,scovery

by

R: From Today 1998,

BC, structure peptide.

71

of a

for selection of antibodies 1998, 9:102-l 08. on phage.

S, Yomo T, Shima Y, Miki S, molecular engineering by A!af Biotecbnol1999, 1758-61.

68. .

Op/n

53.

in vitro Nucleic

peptides 3:370-378.

to drugs

LA, Kay BK: from Src, Nat/ Acad

in protein proteins.

for

Proc

FP, WJ: Small

79. .

Johnson DL, Farrell FX, Barbone FP, McMahon FJ. Taullat J, Hoey K, Ltvnah 0, Wrrghton NC, Mrddleton SA, Loughney DA et al.: Identification of a 13 amino acid peptide mimetic of erythropoietin and description of amino acids critical for the mimetic activity of EMPl. Biochem,sfry 1998, 37:3699-3710. ,.. Alanrne scanning anu truncatton were applred to the erythroporetrn mrmetrc peptides (EMPs) isolated prevrously, resulting rn a 13 ammo acrd sequence contarntng the mrnimal agonrst eprtope. 80.

Cwrrla Singer

SE. Balasubramanran SC, Davis AM. Tansik

P, Dufftn DJ, Wagstrom CR. Gates CM, RL. Mattheakts LC. Boytos CM ef al.:

Peptide natural

agonist cytokine.

of the thrombopoietin receptor Soence 1997, 276:1696-l 699.

81.

Starovasnlk MA, Bratsted AC, Wells JA: Structural native protein by a minimized binding domain. USA 1997, 94:i 0080-i 0085.

82

Clackson receptor

T, Wells interface.

JA: A hot spot of binding energy Soence 1995, 267:383-386.

as the

mimicry of a Proc Nat/ Acad SC{

selection

of nucleic

protein model

of phage. enrichment

acids

and

Intramolecular catalysis using staphylococcal

proteins

provides nuclease

Roberts

and

Ja

529

the basis for selection. is demonstrated.

97.

Demartls S, Huber A, Vitt F, Loui L, Glovannoni Neri D: A strategy for the isolation of catalytic repertoires of enzymes displayed on phage. 286:617-633.

L, Neri P, Winter activities from J MO/ Biol 1999,

98

Patten PA, Gray NS, Yang PL, Marks CB, Wedemayer Bonlface JJ, Stevens RC, Schultz PG: The immunological of catalysis. Soence 1996, 271 :I 086-l 091.

A G,

in a hormoneevolution

Marks JD, Hoogenboorn Wmter G: By-passing gene libraries displayed

99

84

Gnfftths AD, Malmqvist M, Marks JD, Bye JM, Embleton MJ. McCafferty J, Maler M. Holliger KP. Gonck BD, Hughs-Jones NC ef al.: Human anti-self antibodies with high specificity from phage display libraries. EMBO J 1993, 12:725-734.

Wedemayer GJ, Patten PA, Wang LH. Schultz PG, Steven RC: Structural insights into the evolution of an antibody combining site. Science 1997. 276:i 665-l 669.

100.

Vaughan TJ. Willlams AJ, Pritchard K, Osbourn JK, Pope AR Earnshaw JC, McCafferty J, Hodlts RA, Wllton J, Johnson KS: Human antibodies with sub-nanomolar affinities isolated from a large non-immunized phage display library. Nat B;ofechno: 1996, 14:309-314.

Oue S, Okamoto A, Yano T, Kagamiyama H: Redesigning the substrate specificity of an enzyme by cumulative effects of the mutations of non-active site residues. J B/o/ Chem 1999, 274:2344-2349.

101.

Romesberg FE, Spiller B, Schultz origins of binding and catalysis Science 1998, 279:1929-l 933.

86.

87. .*

Lancet D. Sadovskv molecular recog&ion significance to the 1993, 90:3715-3719

Smith GP, Pate1 SIJ, Gnfflths AD: Small repertoire of knottins 277.317-332. Tht: knotttns represent sequences maintain their fide bonds (typically 2-3 folds to display chemical knottln library and derived

TP, McCafferty J, Gnffiths AD, human antibodies from VI MO/ 8101 1991, 222:581-597.

GJ,

83.

85.

HR, Bonnert immunization: on phages.

as potent

In vitro

E. Seidemann E: Probabilitv model for in biological receptor repertoires: olfactory system. Proc Nat/ Acad So USA Wlndass JD. Thornton JM, Winter G. binding proteins selected from a combinatorial displayed on phage. I iflo/ B/o/ 1998,

103. .

some of the smallest proteins known. These folded conformations due to the presence of dlsulout of 30 amino ar:lds! and have potentcal as scat diversity The authors screened a cellulose~blndlng binders for alkaline phosphatase (see Table 2).

Beste G, Schmidt FS, Stlbora T, Skerra A: Small antibody-like proteins with prescribed ligand specificities derived from the lipocalin fold. Proc Nafi Acad So USA 1999, 96:1898-l 903. The authors demonstrated that I~pocal~ns may represent an alternate franxwork to antIbodIes for developing general bIndIng proteins vfa ,,I vitro selcctlon Baca M, Scanlan TS, Stephenson RC, Wells JA, Phage display of a catalytic antibody to optimize affinity for transition-state analog binding. Proc Nat/ Acad So USA 1997, 94.10063-l 0068

90. .

ArkIn MR, Wells JA: Probing the importance of second sphere residues in an esterolytic antibody by phage display. I MO/ B/o/ 1998, 284:1083-l 094. Phage display was used to improve and probe estrrolytlc antbody functton The report provides another example in which residues remote from the active site are Important for blnding and catalysis A correlation IS seen between transitton state analog bindlng and catalytic function for single mutations. 91.

92.

FUJII I, Fukuyama S, lwabuchi antibodies in a phage-displayed Rmfechnol 1998, 16:463-467.

Y. Tanlmura R. Evolving catalytic combinatorial library. Nat

Gao C, Lavey BJ, Lo C-HL, Datta for catalysis from combinatorial acid probe: primary amide bond 120:2211-2217.

A, Wentworth PJ: Direct selection antibody libraries using a boronic hydrolysis. / Am Chem Sot 1998,

93.

Wirsching P, Ashley JA, Lo C-HL. Janda immunization. Sc/ence 1995, 270.1775-l

94.

Barbas CFI, Helne A, Zhong G, Hoffmann T, Gramatlkova S, Blornestedt R, List B, Anderson J, Stura EA. Wilson IA, Lerner RA Immune versus natural selection: antibody aldolases with enzymatic rates but broader scope. Science 1997, 278:2085-2092.

95.

96. .

KD, Lerner 782

Jones JT, Balllnger MD, Plsacane PI, Lofgren JA, Fitzpatrick VD. Fairbrother WJ, Wells JA, Sllwkowski MX: Binding interaction of the heregulinR egf domain with ErbB3 and ErbB4 receptors assessed by alanine scanning mutagenesis. J Biol Chem 1998, 273:l 1667-l 1674. A thorough study of protein speclficlty through exhaustive alanlne scanning mutagenesis.

Balllnger MD, Jones JT, Lofgren JA. Fairbrother WJ, Akita RW. Sliwkowskt MX, Wells JA: Selection of heregulin variants having higher affinity for the ErbB3 receptor by monovalent phage display. J BKI/ Chem 1998, 273:i 1675-l 1684. Variants of a growth factor domaln with higher affinities for Its receptor revealed that beneflclal mutations generally avotded critlcal blndtng residues, Evolved proteins had enhanced affinity either by reconfjgunng essential blnalng reslduos or by formlng new contarts

l

104. *

105 ..

Elrod-Erlkson M, Benson TE. Pabo CO: High-resolution structures of variant Zif268-DNA complexes: implications for understanding zinc finger-DNA recognition. Struc~re 1998, 6:451-464. A thorough structural analysis of zinc-finger-DNA variants revealed the dlfflcultles in determInIng an exhaustive recognition code defimng zlnc-finger DNA-bindIng spectftclty. The results cast serious doubt on developing a code for sequence~spectflc protein-DNA recognltlon. 106

Street 1999,

107.

Rod) DJ. Janes RW, Sanganee HJ, Holton Screening of a library of phage-displayed human Bcl-2 as a taxol-binding protein. 285:197-203.

108.

Fukumoto T, Tongoe N. Kawabata S, Murakaml M. Uede T, Nlshl T, Ito Y, Sugimura K: Peptide mimics of the CTLA4-binding domain stimulate T-cell proliferation. Nat Bmtechnol 1998. 16:267-270.

109.

Rand KH, Houck H. Denslow to find target(s) for oligoclonal / Neural Neurosurg Psychiatry

110

Bremnes displayed peptide

RA: Reactive

Janda KD, Lo L-C. Lo C-HL, Sm M-M, Wang R. Wong C-H. Chemical selection for catalysis in combinatorial antibody libraries. Science 1997. 275:945-948

Lerner

Immunological antibody.

102. Cordes MHJ, Walsh NP, McKnight CJ, Sauer RT: Evolution of a .. protein fold in vitro. Science 1999, 284:325-327. The substltutlon oi two residues Into the Arc repressor changes the local secondary structure from a b sheet to a short u helix, with the preservation of core packing. The results Imply that modest changes in amino acid sequence may allow signlflcant changes in local secondary structure

88 .

89

PG, Stevens RC: in a Diels-Alderase

RA:

Pedersen H. HBlder S. Sutherlln DP. Schwltter U. Klna DS. Schultz PG: A method for directed evolution and f&ctional cloning of enzymes. Proc Nat/ Acad SC/ USA 1998, 95:i 0523-l 0528. A novel strategy tar the direct selection of catalysts by phage display IS demonstrated by attaching both the enzyme and substrate to the coat

AG, Mayo 7:R105-R109.

SL: Computational

protein

design.

RA, Wallace peptides J MO/ B/o/

Sfructure BA, Makowskl identifies 1999,

ND,

Hellman KM: Molecular approach bands in multiple sclerosis. 1998. 65:48-55.

T, Lauvrak V. Llndqvlst B. Bakke 0: Selection peptides from a random IO-mer library target. lmmunotechnoi 1998. 4:21-28.

of phage recognising

11 1, GBrdsvoll H, van Zonneveld A-J, Helm A, Eldering E, van Meijer M, Dano K, Pannekoek H: Selection of peptides that bind to plasminogen activator inhibitor 1 (PAI-1) using random peptide phage-display libraries. FEBS Letl 1998. 431 :I 70-l 74. 112.

Pearce KHJ. Cunmnaham hormone binding aifinity requirements for cellular

L:

BC. Fuh G. Teerl T. Wells JA: Growth for its receptor surpasses the activity. &ochemfstry 1999, 38:81-89.

a