CA1339842C - Methods of generating desired amino-terminal residues in proteins - Google Patents

Abstract

Methods of designing or modifying protein structure at the protein or genetic level to produce specified amino-termini in vivo or in vitro are described. The methods can be used to alter the metabolic stability and other properties of the protein or, alternatively, to artificially generate authentic amino-termini in proteins produced through artificial means, The methods are based upon the introduction of the use of artificial ubiquitin-protein fusions, and the discovery that the in vivo half-life of a protein is a function of the amino-terminal amino acid of the protein.

Description

~ :~ 3 3 ~

M~THODS OF GENERATING DESIRED AMINO-TERMINA~
RESIDUES IN P~OTEIN~S

B ck~round of the Invention In bot~ bacterial snd eukaryotic cells, relatively lon~-lived proteinx, who~e half~ es are C108e to or eXoeet the cell generation ~ime, coexi~t with protelns whoc~ half-live~ c~n be less than one perc-nt o~ th~ cell generat~on time. R~te~ of intracellulAr protein degrad~tlon are a function of the cell's physiological st~te, and ~ppear to b~
con~rolled di~ferentially for individ~al proteins.
~n p~rticular, dama~ed and otherw~se abnormal proteins are metabolically unatable in vivo.
Although the specifie functions of selective p~otein degradat~on ~re in most cases still unknown, it is clear tha~ many re~ulatory protein-c are ext~emely .~hort-lived in vivo. Me~abolic inst~bility of such pro~cin~ allow~ for r~pid 4dJustmont o~ ~h~iX
intracellul~r concentrations thro~h regulated change.~ in ra~s of their ~yn~he~ls or de~rad~tion.
The few instances in whlch the met&bolic instablli~y o~ an intracellular protein has been shown C4 be ecsential for its fun~tion include the cII protein of ba~teriophage lambda and the HO endonuclease of the yeast Saccharomyces cer~visiae.
Mos~ of the selective turnover of intracellular protein~ under normal metabo~ic conditions ls ATP-de~endent and (~n ~k~ryotes) no~lyoo~om~l.

1~ 5~984~

~ecent biochemical an~ ~enetic e~idenc~ indicates that, ln eukaryotes, covalbnt conjugation of u~i-q~i~in to short-liv~d in~racellulfir proteinS is ~6~ential for their ~elect1~o degradation. The rules which determin~ whether a gi~en protein is ~e~abolically stable or unstable in ~ivo were pre~iously ~nknown.

SummAry of the Invcntion . f This ~-n~ention pertains to methots of eng~ne-r~
in~ the amino-terminus of pro~eins thereb~ co~- ;
trolling ~hs meta4011c stability and other proper-tle~ of ~ protein. Furth~r, this invention provides a method for either in vivo or in vitro production of pro~eins wi~ an~ of t~e tw~nty ~mino ~cid residue~ tor analogs thereof) at the pro~einls amlno-t~rminus. The invention l~ based in part upon the striking discovery ~hat the i_ vivo half-life of an intracellular prote~n is ~ function of its amino-to~m~nal ~m~no acid re~due ~nd upon a nov~l ~and more generAlly appllcable) technique th~t allows one to gener~cc protein~ wlth specii~ied amino-termini i_ vi~o or in vitro, Thc in~en~ion ~l~o pertains to ~ newly identified protease, ubiquitln-speciflc processln~ proteace, which hAs propertie~ th~t allow one to expose, ei~her in vitro or in ~ivo, any desired amino ~cid re6idue, o~her than p~ollne, ~ ~he amino-termin~s of a protein of intere~t, 1~398~2 The nature of the amino acid exposed at the amino-terminus of an intracellular protein was shown to be one crucial determinant that specifies whether a protein will be long- or short-lived i_ _i_o.
05 Individual amino acids can be categorized as either stabilizing or destabilizing amino acids with respect to the half-life that they confer upon a protein when exposed at the protein's amino-terminus. Destabilizing amino acid residues confer short half-lives, down to a few minutes for some of the destabilizing amino acids. Stabilizing amino acid residues confer long half-lives of many hours.
This striking and newly discovered dependency of a protein's half-life on its amino-terminal residue is referred to herein as the N-end rule.
For some proteins, the presence of a destabili-zing amino acid at the amino terminus is necessary but not sufficient for destabilization. This is so because the complete amino-terminal degradation signal in a short-lived protein comprises two distinct determinants, each of which is necessary, but each of which, by itself, is insufficient for efficient destabilization of a protein. One determinant, described above, is the amino-terminal residue of the protein. The second determinant, described below, is a specific internal lysine residue. The ability of this critical lysine residue to serve as the second determinant is to a significant extent independent of the amino acid I-.U~

~3~3.~'12 ~equence~ ~urrounding the reQi~ue, Instead, an es~ential fe~ture of this critic~l lysin~ resld~e include~ its ~patlal proximity to the procein's amino- terminus, Ba~ed upon the N-end rule, the amino-te~minu5 of 8 prote~n can thus be de~l~ned or altered to c~ange the in~racellular half-life of the protein and in thi~ way the lifetime A~d/or activity of the protein ~n vlvo can be regulated. This c~pability c~n bH exploited for rational p~otein deQi~n in meny different con~exts. Natural proteins can be modified to render them more or le~s resistAnt to degradstlon in vivo T~q desi~n or alteration of the protein can be done at ~he pro~eln level or se ~he 6enetic ~DNA) level. For exAmple, protein~ can be modlfled by che~ically ~lterlng or engineering the amlno-terminus to pro~lde for exposure ~t ~he amlno-terminus of ~n sminO acld residue of the 6tabiliz~ng or destabilizing class At the ~enetic level, 8enes encoding proteins can be m~de to encode an ~mino Acid of the desir~d class at the ~mino-terminu~ 80 that the expressed protein ex~ibits a prqdeter~ined amino-terminal ctructure which render~
it elthe~ metabolically stable or un.~table with r-~pect to the N.~nd r~l~ pathway o~ proteolytic te~ra~ation. Amino~terminal regions cAn be engine~red to provide approprlacely located lyslne residues in the context of a s~fflc~ently seg-~ent~lly mo~ile al~ino ~ernl~nU~ to produce P.06 3 ~.3~2 te~taoilized protein. Furthermore, pro~lns c~n be exp~o~s~d fused to ~ ~'masking" p~o~in sequen~e which m~sks th~ ~ngineered amino-terminu~ so that when unmasked the protein will exhiblt the deslred metabolic ~tabil~ty or other properties th~t drpend on the natu~o o~ ~h~ prateln's ~mino te~min~l recidue, In such const~uc~, for example, t~e junction betw~en the two protein s4quenceg can be de~igned to be ~leavet *pecifically, for in.~tance, by zn endop~o~ease Endoproteolytic cleava~e o~ the fu~e~ sequence unmasks the speoifi~lly engin~ered ~mino-terminus of the protein of int~re-~t and subjects the protRin to ~egr~dation ~ove~ned by the N~end rule. One specifio and new w~y to engine~
the protein' ~ a~ino-ter~inus is p~ovided ~n this inventlon by the identiflca~ion of ubiquiein-~peclf~c proce~Jing proteasQ ~nd d~termin~t~on of its 6ubstrate specificlty. UGing this p~o~eas~, fusions of ubiqui~in ~ith other proteins can be specific~lly proce6scd either in vitro or in vivo to gen~r~te protein~ with desi~ed amino-ter~inal residue~.
A different, and al~o new way to specii'ically engineer short-lived proteins 1~ provi~ed in this inv~ntion by the discove~y ~hat ubiqui~in-protein fusions, 6uch fis ubiqui~in-~ro-~galactosida6e, that cannot b~ iclently deublquitinat~d, er-metabolically unstable. Thus, by at~aching the ~mino-terminal ubiqultln moi~ty ~o ~ protein in a ~ P.07 1~39,~/~2 wsy that ma~es its removal either i~possiblq or in~fficien~, one can destsbili~e protein~ by a di6tinct techniqu~ that ls not directly based on the N-end rule.
In addition, v~riant cells can be developed which ~o~tain pu~ative mutation~ ln the 'IN-end"
de~rading protease~s~ which either condltlonally or nonconditionally stop d~r~ding shorc-livod p~o-teins. ~hese cells can be used to overproduce proteins that ordinarily would be short-llved within the cell, f DQscriptiDn of the Fl~ure~
__________ . . ____ Fig~e 1 show~ tho conctruction of ubi~ult~n-lacZ gene fusions.
Figure 2 shows ~xp0~iments in which the half-llve~ of engineered ~-gal proteins are directly m e ~ ~ u r e ~ , Fi~ure 3 ahow~ the changlng of alnino aelt re6id~es at the ubiqultin-~-g~l Junction (A) u-~lng the newly discovered p~ope~ties ~f ubi~uitin-specific proce~xlng pro~ea~e and the amlno acid sequence in the vicinlty of the ~unction ~B~.
~ i~ure 4 sho~s ~he presence of multiple ubi-qui~ln moieties in metabolic~lly uns~ab~e ~-g~l protelnQ, Figure 5 shows a series of ~ gal cpecie~
con~aining ublquitin in ~etabolically unstable ~-gal prot-ins.

Figure 6 shows that both prokaryotic and eukaryotic long-lived intracellular proteins have stabilizing amino acid residues at their amino-termini whereas secreted proteins exhibit a com-05 plementary bias.
Figures 7 and 8 show the construction ofubiquitin fusions with mouse dihydrofolate reductase.

Detailed Description of the Invention The elucidation of the N-end rule is described in detail below. Briefly, this rule governing protein degradation was revealed by examining the i_ _l_o half-lives of the enzyme ~-galactosidase having various amino acid residues at its amino-terminus and produced as a fusion protein with ubiquitin.
When a chimeric gene encoding a ubiquitin-~-galac-tosidase fusion protein is expressed in the yeast S.
cerevisiae, ubiquitin is cleaved off the nascent fusion protein, yielding a deubiquitinated-~-galac-tosidase (~gal). With one exception, this cleavagetakes place efficiently regardless of the nature of the amino acid residue of ~gal at the ubiquitin-~gal junction, thereby making it possible to expose selectively different residues at the amino-termini of otherwise identical ~gal proteins. The ~gal pro-teins so designed exhibited strikingly different half-lives i_ i_o, ranging from more than 20 hours to less than 3 minutes, depending upon the nature of ~' p~9 '3 ~ ~

the ~mino ~cid at the ~mino-terminus of ~gal. Amino acids can be thu~ o~der~d acçordlng eO the the half-lives they confer on ~gal when prcsent ~t its ~mino-terminus, For example, ~he amlno aclds me~hionine, ~erine, ~lanine, threonlne, valine, ~lycins and cysteine confer 8 hslf-life o~ mor~ than 20 hourx, Phenylslanlne, leucine, ~para~ine, and lyslne yleld half-llvec of sbout three minutes.
Arginine, the mos~ destabili7ing amino scid, confers a half-life of about two m~n~tec. (S~- ~able 1 below for complete 11st of smino ~cids ~nd the correspond-ing half-lives~
A simila~ resul~ is observed when 3$S lfibelled protein6 are synthesized in E. coli, lsol~te~, an~
added to a ~smmali~n cell ly~ate, specif~ c~lly, the well characterized rabbit reticulocyte ~ysate system, In such a system, for example the ~ollowing amino-te~minal residues can be character~zed as destabil~zing: ~rginine, lysine, histidins, phenylalanlne, leucine, ~ryptophan, tyrosine, alanine, serine, threonine, s.~p~rtlc &cid, ~lut~mic acid, glutsmine, cy~teine and asparagine. Whe~her a particular amino scid ls de~esbllizing in ~ny eukaryotic ~ystem can be determined Through the eourse o~ these studies, it has been determined thst the ~-end ~ule has hierarchical structure. 9pecifically, amino~ minal Glu and Acp ~and al80 Cy8 in reticulocyte~) a~e ~econdAry destabilizlng residues ~3~ 12 in that they are destabilizing through their ability to be conjugated to primary destabilizing residues such as Arg. Amino-terminal Gln and Asn are tertiary destabilizing residues in that they are 05 destabilizing through their ability to be converted, via selective deamidation, into secondary destabilizing residues Glu and Asp.
Currently known amino-terminal residues in long-lived, noncompartmentalized intracellular proteins from both prokaryotes and eukaryotes belong virtually exclusively to the stabilizing class of amino acids, exactly as predicted by the N-end rule.
This result stongly implicates the N-end rule in the selective degradation of intracellular proteins in general.
The appropriate amino-terminal amino acid appears to be an essential (though not necessarily a sufficient) requirement for the metabolic stability of a noncompartmentalized, intracellular protein.
Thus, in order for a protein to be relatively stable intracellularly, a stabilizing amino acid should be present at the amino-terminus. The presence of a destabilizing residue at the amino-terminus of a protein is often, though not always, sufficient for its metabolic destabilization i_ _i_o. When such destabilization occurs to a relatively small extent, further analysis shows either an insufficient accessibility of the amino-terminus or a lack of the second determinant of the complete amin~-terminal A
,,~L

~3~

degradation signal. In these instances, this second determinant, which by itself does not metabolically destabilize a protein, must be present in order for the half-life of a protein i_ _i_o to be strongly 05 dependent on the nature of its amino-terminal residue. The second determinant of the amino-terminal degradation signal was found to be a specific internal lysine residue. The ability of this critical lysine residue to serve as the second determinant was shown to be largely independent of unique amino acid sequences surrounding the residue.
Instead, an essential features of the critical lysine residue includes its spatial proximity to the protein's amino-terminus.
The presence of a stabilizing amino acid at the amino-terminus at least in some cases (for instance, as observed for ~-gal) will confer stability upon the protein. However, a stabilizing amino acid at the amino-terminus may not always confer a long half-life because other degradative pathways may be involved in determining the ultimate fate of the protein. For example, endoproteolytic cleavages (cleavages outside of terminal regions of the protein) may result in exposure of a destabilizing amino acid at the amino-terminus of a resulting product of the cleavage which is then rapidly degraded via the N-end rule pathway. Appropriate circumstances for use of a stabilizing amino acid can be ascertained empirically.

P.lZ

Al~hough ~he N-end rule msy be only one com-ponent ~albelt a centr~l one) of a more complex "half-life rule" whic~ embraces oth~r ~spect.~ of selective prote~n deg~ad~tion in vivo, the N-end rulc provides a ratlonal, practical ~ppro~ch for de-signing or changing protein structure in ord~r to produce proteins which are mor~ or les~ resistan~ to degradation by the N-end rule pathway than nficural, unmodified protein. Proteln~ cfin be de~igned or modlfied ~t the protein or gen~ le~el to provide a desired amlno acld of either ~h~ st~b~lizing or destabilizing cla~s at thelr amino-terminus, Where necessAry for destabilization, ~ddition~l modific~-tlon~ csn be made to th~ a~ino - terminal region to providR appropriately loca~ed lysine r~sidue~. The ability to ~egulate the h~lf-life of R protein will allow one ~o modula~e the lntracellular activity of the protein, A straightforward approach to modifying a protein ln order to increase or decre~se i~s meta-bolic stability or to modul~te other properties of the protein i~ to directly englneer the amino terminus of the protein at the protein level. To provide a de~ired amino-te~minal amino Acid, ~he amino-~e~min~s of the protein o~ interest csn be chemic~lly sltered, for exfimple, by addi~ n ~mlno acid of the ~tabillzln~ or d-~t4bilizLng cla~ eo the amino-terminu~ of a protein or polypep~ite, em-ployin~ ~n appropri~te chemistry. Thu~, for P.01 1~3~i2 example, an unstable protein can be rendered more ~table by adding a s~abillz~n~ ~mino acid ~esldue ~e,g. methionine gerine, alaninq, ~hreonine, v. line glycine or cysteine) to the amino- terminus of the pro~sln, Con~cr~ely, a stable protein can be testabilized by adding a d~stabillzlng g~in4 acld to ehe amlno - terminu6. One di~tln~t way to modi~y the amlno-terminus of a protein wo~ld be eo employ sp~-cific enzy~e~, amino ac~d-protein li~ases, which c~talyze po~ttranslational ~dltion of a single amino acid to the protein'~ amino-terminu6. Other methods for nongene~ic alte~ations of the same type can readily be accertained by thos~ ~killed in the art.
In some pro~ins, ehe amino-~erminal qnd is obQcured a~ a re~ult of the pro~ein's conform~tion ~l.s., it~ terti~ry or ~uatern~ry struc~x~). In the~e ca6e6, more ~xtensive ~lteratlon of the amino-eermin~s may be nece~sary to make the protein subject to the N-end rule pathway. For example, where ~imple addi~ion or replacement of the .~ingle a~lno-termlnal residue is insuffioient because of an inaccessible amino-termlnus, ~ever~ nino acids (including lysine, the site of ubiquitln ~olning to substrate prot~ins) may be added to the original amin4-terminus to increas~ the acce~sibility and/or soemencal mob~l~ty o~ the en~ineer-d amino t~rminus.
Mod~flca~io~ or design of the amino- terminus of a protein can also be accomplished at the genetic p~z 3 ~ 2 level, Con~entiOnal techniques of site-direoted mut~g~nesis for addition or ~ubstit~tion of ap-propriate codons ~o the 5' en~ of an Isol~ted or 6yntheslzed gene can be employed to provide a de~lred amino~terminal structure for the encoded prot~in. ~or ex~mple, ~o th~t the pro~eln expre~se~
hs~ the desired amino acid at its amino-ter~inus the appropriate codon fo~ ~ sc~b~lizing ~mino ~ci~ can bo ln~-rted or b~llt ~nto the amino-termlnu~ of the prote~n-encodin~ sequenc~. Uher~ nec~ssary, the DNA
sequence encoding the amino- terminal region of a proteln can be modified to introduce a lysine resLd~e in an appropri~te context. Thi6 c~n be schie~ed mo~t conveniently by e~ploying D~A
construct~ encodin6 "universsl dest~bilizing segments~'. A univer~l destabilizin~ se~nent ço~prises ~ DNA conx~ruct which encodes a polypeptide ~tr~cture, preferably segment~lly mobll~, containin~ one or moro lysino resLd~, th~
codons for lysine re~ldue~ bein~ positioned within the con~truc~ ~uch that when the construçt is inserted Into the structural gene, the ly~Lne re~Ld~les ~re sufficiencly spa~Lally proximate to the a~Lno-t4rminus of the encod~d protein to serve AS
the second det~rmin~nt of the complete amino-terminal degr~d~tlon s~nal, An example of a deht bilizing segment i~ 6hown in the exempli~ic~tion be~ow (~ee Flgures 7 and 8). The ~nsertion o~ Juch con~tructs into the S' portion oi RP ~ p,~3 1~3~8~2 ~ s~ructural ~ene would provide ~he encoded pro~eln with ~ lysine re~idue (or resldue~) ~n an approp~ate context for destabiliz~tion.
At ~he s~me tl~e, expres~ed proteln~ ~re often natura~ly modifled w~ thin a cel} af~er translation.
The~o modifications can include changes at the protein'- ~mlno-te~mlnu-. F~r ex~mpl~, ~he amlno-terminus c~n be Mcted on by a~ aminop~ptidase which cleaves one o~ seve~al amino acids from the amino-terminus. Amino ~c~d~ msy also be added to the amino-termlnu.~ by p~s t-tr~nslational proe~s s ing.
This Inven~lon provide~ a way to "by-pass" still und~ined rules of amino-terminal proteln proce~ing to expo~e exactl~ ~nd specific~lly the desired amino ~ci~ residues at the amino-~e~minus of ~ mature pro cessed protein species. ~o mi~imize the impact of such postt~anslaelonal eventq on ~he ultimate structure of the amino-terminus of A protein of in-~eres~, ~pecific fuslon protein.~ can be designed wherein the amino-ter~inus of a protein of interest (desi~nod to have tho desired stabilizin~ or de-stabilizing st~ucture) is preceded by a ~masking"
protein sequence fused to the amino-terminu~. The fusion proteinc Qre de~gned ~o that the m~sklng protein sequence fused to the amino-terminus of the pro~ein of interes~ is susçeptlbl~ to specific cleava~e at tho junc~ion bçtwe~n the two. Remov~l of the protein sequenç~ thus unmas~s the amlno-t~rminus of the protein o~ in~erest and thç half-r.~

~i3 39~

life of the relea~ed proteln ~ thus governed ~y theprede~l~ned amino-~rmlnu~. Th~ fu~lon protein can be de~lgned fo~ gpeci~ic cl~avage ~n vLvo, for sxample, by a host ~ell endoproteas~ or fo~ speciflc cleavagc in a in vitro sys~em where i~ can ~e clefived afeer l~olatlon from a producer cell (which lacks ~hc capability ~o oleav~ th~ fusion proteln).
Ubiquitin is a broadly use~l fusion partne~
for con~truction of a fused prote~n wlth a protein of lnteres~: t~a di~covery t~at a~tificial ubi~uitin-p~otein fusions can be aleavcd precisely by a cytoplasmic eukaryocic protea~e wlth little or no dependence on ~he protein to whioh ubi~uitin ls fused can be Applied both 'n vivo nd in vitro in protcin enginee~ing s~ra~egiec, and is a major aspect oi ~his invention. For exa~ple, ~he ublqult~n-protein fusion method c~n be used to artlflclally ~enerate ~uthentic amino-termlni in p~oteins produced through artifici~l ~eans. Thus, ~mlno-termlnl çharacte~istie of natural euka~yotio or p~okaryotic prote~n~ can be.generated by in vi~ro clea~age of ubi~uitin-protain f~sions produced in 3 p~oka~yotic ho.~t, A specific m~hodology for praducing ubiquitin-~-galacto~ldas~ fusion proteins is de~cribed in det~ll below, Genes encoding ~ny other proceins can be ~ubs~ituted for LacZ (the ~ ene) in th~s msthodology.

133~8~

In general, ubiquitin fusion proteins are expressed by a chimeric gene construct comprising, in S' to 3' orientation, a ubiquitin gene linked to a gene encoding the protein of interest. The codon 05 for the amino-terminal amino acid of the protein of interest is located immediately adjacent the 3' end of the ubiquitin gene. The fused gene product is cleaved endoproteolytically either i_ _i_o or i _itro (using either pure or partially purified ubiquitin-specific protease identified in the present invention) at the junction between ubiquitin and the protein of interest to generate the protein of interest having the desired amino acid at its amino-terminus.
There are a number of specific uses for the described ability to specifically engineer the protein's amino-terminus. One such use is es-tablished by the fact that the intracellular half-life of the released protein is governed by the principles of the N-end rule. Other applications of the specific method for engineering the protein's amino-terminus described herein range from adjusting the desired functional properties of a protein of interest, to modulating its antigenicity, and again, to other uses that can readily be ascertained by those skilled in the art.
This method of generating the desired amino acid residue at the amino-terminus of a protein of interest involves two novel components: one, the , ~

-17- 133~12 use of ubiquitin-protein fusions, and the other, the use of ubiquitin-specific processing protease that has been identified, and whose striking substrate requirements were discovered, in this work. Although the initial identification of the ubiquitin-specific protease has been made in vivo, the enzyme is also relatively stable and active in vitro (in extracts), and can readily be purified to homogeneity by techniques known to those skilled in the art.
Furthermore, the substrate specificity of the ubiquitin-specific processing protease is highly conserved in evolution, being the same in yeast and mammals. The enzyme can be purified chromatographically from a crude extract by sequential chromatography on phosphocellulose, DEAE cellulose, and SH-SEPHAROSE* among other methods known to those skilled in the art. Alternatively, the gene for this protease can be cloned by those skilled in the art.
The cloned protease gene can be used either in vivo, or, alternatively, the gene can be over-expressed in a suitable host, the overexpressed ubiquitin-specific protease purified and used for the same or similar purposes in vitro. The discovery of this enzymatic activity, and detailed characterization of its substrate specificity herein provides for the in vitro and in vivo use of this enzyme.

*Trade Mark ~, P.07 1339~'12 The use of ~biqui~in-pro~ein fusions to allow the gene~ation of ~ de~ired amino acid residue ~t the amino-termln~ of a protein of interest can be ex~en~e~ to facili~ate the purification of ~uch protelns ~rom producer cells. A gene can be re~dily constructed that encodes A convenlent msrker pro-tein, ~h as streptavidin, linked to a ublquitin-protein fusion construct de~crlbed above, The resulting ~merker prote~n)-ubiquitin-protein ~usion can be ~imply igolate~ fro~ produc~r cells by usin~
the preselected prope~y of the marker protein, for instance ~ the known ability of strepe&vidin ~o be lQolatable by affinity chromato~raphy on a biotin colu~n. Thus, purified (~rker protein) -~biquitin protein fu~lon can then be speçifically cleaved by the ublqultin-~pecifiç protease described in ~hi~
~nventlon to genera~e the final pro~uct, ~ protein of intere~t with ~he deslred ~mino acld resldue at ~t~ amino-terminus.
The codon for th~ amino-terminal amino acid o~
the protein of interest c~n be m~de to encode ehe desired ~ino acid by, for example, site-directed mut~gene-~ls techniques currently -~andard in the ~ield. If the gene encod~ng the protein of interest iG a synchetlc go~e the approprlate 5~ codon can be bullt in du~inR the ~ynthet1c proc-~, Alter~
natively, nucleotlde~ fo~ a ~p-c$fic codon can be added to the 5' end of ~n lsolated or syn~hes17ed gene by lig~tion af an approp~iate D~A ~equence to ~ 33~8~2 the 5' (amino-terminus-encoding) end of the gene.
DNA inserts encoding appropriately located lysine residues (such as the "universal destabilizing segments" described above) can be inserted into the 05 5' region to provide for the second determinant of the complete amino-terminal degradation.
Ubiquitin-like fusion partners capable of being cleaved by the ubiquitin-specific protease can also be used. In addition, fusion partners other than ubiquitin for masking the amino-terminus of a protein of interest can be used. For example, functional homologues of ubiquitin from eukaryotes or prokaryotes may be used. In appropriate cases, the fusion proteins can be designed to contain a proteolytic cleavage site for a restriction endo-protease which has sufficiently narrow specificity so that only one target site is cleaved in a fusion protein. A crucial property of such a protease must be a sufficiently relaxed requirement for the nature of the amino acid residue(s) abutting the carboxy-terminal side of the cleavage site. The target site for cleavage is the junction between the fusion partner and the amino-terminus of the protein of interest and thus the recognition site for the endoprotease is located to provide for cleavage at this location. The commerically available protease, complement factor Xa, exhibits these properties and thus can be used to directly generate proteins with predetermined amino acid residues in the ultimate ~' p~9 133~8~2 .

position of their ~mino-termini (see, ~. No~ai ~nd H.C. Thoger~en Nfiture 30g;B~0 (lg84)), The recog-nition site for the endopxotçase can be en~ineered into the ~unction between the mssking protein se~u~nce and the 3' re~lon encodin~ the ~ino-terminu~ oof th~ pro~ein of in~erest.
A diffe~en~ and distinct method f or engineering .chort-lived proteins ls provided in thi~ invention by th~ discovery that ublqui~in-protein fusions, 6uch as ubiquitin-Pro~ alactosid~s~ fu~ion ~Tab~e 1), that cannot be ef~ici~ntly d~ubiquitina~ed are metabolically unstable. Thus, by attaching ~he amino-ter~inal ubiqultin molety t~ a p~o~ein in a way th~t ~akes i~s re~oval either impos.~ible or ineifi~ient, one can destabil1ze ~ protein by a dis~inct technique which is qualitativ~ly differe~t i~rom the ~thod o~ g~noratin8 the de~red a~lno-ter~inus oi2 A protein according to the requi~m~nts of ~he N-end rule. Prevention of the efficient deubiquitlnatlon of ~ ubiquitin-protein fusion can be ach~eved ~n s~veral w~y4, for instance, by using a prolinH residuc ~t ~h~ ubiquitin-protein ~unction as shown in Table 1, or by changing the ~mlno acid 4equence of ublqui~in near its c~rboxyl-~erminus in such n~ way that the ublquitin moi~ty is no longer recognized by the ubiqui~in-speciiic proces~lng prote~ss but q~n ~till b~ r~cognizod b~ the ~est of the degr~dAtive pathway. The~e and other way~ ~o reduce the rat~ of deubiquitination of a ubiquitln-13398g~

protein fusion can be readily ascertained by thoseskilled in the art.
The methods of this invention can be employed, i_ter ali_, for regulating the half-life of a 05 protein intracellularly. There are many instances where this capability is useful. For example, when a gene is introduced into a cell for expression therein, the expressed product can be designed for a long or short half-life depending upon the particu-lar need.
In general, destabilized proteins which haveshort half-lives are more amenable to regulation of intracellular levels of the protein. The ability to finely regulate the intracellular levels and ac-tivity of a protein can be useful in therapy or inthe work with i_ _itro cell cultures. In gene therapy, for example, a gene may be introduced into a cell to compensate for a genetic deficiency or abnormality. The gene can be inserted under control of an inducible promoter. Induction results in enhanced expression of the gene product and con-sequently, higher levels of the product within the cell. If the gene is designed to encode an unstable protein, the intracellular concentration of the expressed protein will be more quickly responsive to a later reduction in the rate of its synthesis because it does not persist within the cell. In this way, the intracellular level and/or activity of ~, r~ P 01 1 3 ~3 9 ~3 the protein enc~ded by the inserted ~ène can be more flnely regulat~d.
The metho~ of this in~enti~n can ~lso be used to expand the uses of selectabl~ ~arkers by shorten-ing ~he time nec~ssary for a phenotype rel~ted to the marker to become manifest. Toward thl~ end, a p~oduct ~ncoded by ~ marke~ gene c~n be destabllized by altering its amino- terminus ac40rding to the N-end ruls. ln this w~y selection ~or the negAtive phenotype can be ~acilitated because the prod~ct of the marker ~ene will ~e more quickly ex~inguished after the ~unction of the ~ene oncoding the m~ker is abolishod, An example is the thymidine kin~se (tk) gen-. The tk gone can be en8intered to eneode a le.~ ~ta~le enzym~ by introducLng an appropri~te de.~tablllz~ng a~ino acid at the a~ino-ter~nus.
~ene mutation re6ulting in tk pheno~ype wlll ~e moxe qulckly manifested by o~lls because re~iduhl tk will bH more qulckly de~raded. This ean be especiAlly us~~ul in 510w growing cell~ where more time is required to "dilute out'~ tk synthesized prior to transformfition to the tk type.
Th~ principles of protein modification based upon she ~end rule m~y ~l~o be omployed ~n the desi~n of Cytotoxln~, Pro~einac~ous cy~otoxins c~n be designed as unstable prote~ns de~rad~ble by the ~-en~ rule pathway go that they d~ not persiS~ ~fter their toxic ~otion has been exert~d on a target ~3~9842 cell. Reduci~g the lifetime of thH toxin ~educes the likelihood of ~illing nontarge~ed cells.
Dl~covery of the N-end ~ule pa~hway of degra-dation allows development of m~tant cell6 ha~1ng muta~i~n~ in genes encoding es~ential componen~ of the N-end rule pa~hway. For example, cell-~ can 4e produ~d that either permanently or con~itionally are unable to efficiently de~r~de o~he~wise short-lived protein~ These cells can be used to prod~ce desired proteln~ tha~ ordinarily would be unstable with~n a cell.
The invention ls illust~ted furthe~ by the followlng detailed description of the elucidation o~
the ~-end rule, Mçthod_ Protein_Seque_cin~
S. cerevLsLae cell~ c~rrying p~B~3 ~Fl~. 1), which encodes ub~Me~-~gal ~Fi~. ~A), were labeled with [355~ methionine, followed by exeract pre-paraeion, immunoplecipitation of ,~E;al and elec~-~ophore~is ~8 de~aribed below. The wet poly~cryla-mide gel w~q su~ ected to ~utoradio~rAphy, the band of ~gal was excised, ~nd th~ electroelu~ed ~al was ~ub~ected to ~ix cycl~s of radio~hemlc~l sequencin~
by Edman de&radat~on. The ~eq~encin~ carried out by W. Lane at the Micro~hem Faclli~y of Harv~rd Unlver 5 i ~y .

P.03 9~i2 Site-direc~ed ~utagenesis pUB~3 (flgure 1) wa~ treatod sequHnti~lly wi~h Acc I, the Klenow frsgment of pol I, and Bam HI. A
fragment containing the Xho I .~ite waq purified ~nd inserted between ~ filled-in Hind III s~e and A B~M
HI sit~ of ~h~ M13mpg pha~e DNA. (J. Messing and J.
V~eirs, Gene 1~, 263 ~1~8~)). Si~e-direceet m~ta-~eneci~ ~M, Smlth, Annu. Re~. Ge_et. 19, 423 ~1985)) wa~ carrled ou~ a~ deccribet by ~r~er, W. ~t al.
Nucl. Acids Res. 12, ~441 (l~a4) using a synchetie 25-resldue oll~odeoxyrlbonucleotid~, contalning ten bascs o~ the 5~ side ~n~ twel~e bases on tho 3' side of the Met codon of g~l. All fo~r ba~e~ were ~llowet to occur ~e the originAl ~et codon po.~itions durln~ ~ynthe~is. Primary pha~e plaques were sc~ocn~d by ~ybri~iz~tion (Wood, N . I . et al, PNAS
82, 1~8S ~19~)), with thn use oi a 12-re~idue oligonucleot~de probe spanning the ~egion o~ codon changes ~nd hy~ridizin~ to the original sequence~
Nonhybridizing plaques containing inserts of ~he expected slze were sequenced by the e~ain termina -tlon ~ethod. (Sanger, ~. et al., P~AS 71 5463 (1977)). To trAnsfer the te~i~ed congtructs ln~o th~ pUB~3 backgro~nd, replica~ive for~n DNA of m~ltant ph~geh wa~ dig~s~ed wi~h Xho I ~nd ~am HI, and added to the sa~e digegt of the plasmid pLGSD5-ATC ~see Fig. 1 4nd L. G~-~nte, ~ethod~ ~n~ymol-, 10l lal ~1983) ) . The ligated mixture was u~ed to transform the E, coli strain MC1û61. ~M,J. C~sadabarl And S.N.

P.04 ~33~

-~5-Cohon, J. ~ol. Biol., 13~ 179 (1~80)~. ~oloniescontaining plasmld5 of lnterest (~n which the open re~dln~ fr~me of ~al had been ~es~ored) were recognized by their light blue color on X-~gal plates.

Pulse-Cha~e Experlments S, cerevisl~e Cell8 of che s~rsin BWG-~s-l ~MA~
his4 ur~3 ~de6), tr~nsfor~ed ~F, Sherman et fil .
Methods in Ye~t Genetlc~ Cold Sprin~ Harbor L~bo-ratory, N.Y,, lgBl)) w~th plasmid~ of interest were grown ~t 30 C to A600 of approximately 5 in ~ medium of 2 percent ~alao~se, 0.67 per~ent Yeast Nitro~en ~ase without amino flcid~ ~DIFC0), ~denine ~10 ~g/ml~
and ~mino acids ~nclud~ng meth~onlne ~ShermAn, F. et al., su~ra). Typically, cells from ~ 5 ~1 culture were h~rvested by filtr~tion th~ough the well of a M~llipore mlcrotlter ~lltr~tion pl~te, w~hed se~eral ti~es on ~he ~ilter with the same medium lacking methionine and resuspen~ed in 0,3 ml of 1 percent ~alactose, 50 mM potassium phosphate buffe~
(pH 7.4). t35S~me~hionine (S0 to 100 ~Ci~ w~s then added for 5 minutes at 30~C; the cells we~e col-lect~d by filt~ation and ~suspended on 0.4 ml of the ~rowth medium çontaining cycloheximide at 0.5 ~g/ml. ~mples tO.l ml) were withdrawn at indicated eime~, and added to 0.75 ml of cold buffer A (see below ~or buf~e~ compo~ltion~ containl~ leup~pt~n, pep~tatln A, ~ntlpain, ap~o~inin and chymos~atin ., , 13 ~. ~ 8 ~2 (Sigma), (each at 20 ~g/ml) in addition to 0.4 ml of glass beads. Immediately thereafter, th~e cells were disrupted by vortexing for approximately 3 minutes at 4 C; the extracts were centrifuged at 12,000g for 05 3 minutes and the radioactivity of acid-insoluble S in the supernatants was determined. Aliquots of the supernatants containing equal amounts of the total acid-insoluble S were processed for immuno-precipitation with a monoclonal antibody to ~gal.
Ascitic fluid containing a molar excess of the antibody (at least tenfold) was added to each aliquot, with subsequent incubation at 4 C for 2 hours; protein A-SEpXAROSE (Pharmacia) was then added, the suspension was incubated with rocking at 4 C for 30 minutes and centrifuged at 12,000g for 1 minute. The protein A-~EPHAROSE*pellets were washed three times in buffer A (see below) containing 0.1 percent sodium dodecyl sulfate (SDS), resuspended in an SDS, dithiotreitol (DTT)-containing electro-phoretic sample buffer (~.K. Laemmli, Nature 227 680(1970)), heated at 100 C for 3 minutes, and cen-trifuged at 12,000g for 1 minute. Equal aliquots of the supernatants were subjected to electrophoresis in a 7 percent discontinuous polyacrylamide-SDS gel (15 by 15 by 0.15 cm) with subsequent flourography.
In some experiments, the above protocol was not used, but the extracts were prepared by boiling cells directly in the presence of SDS, with es-sentially the same results.

* Trade mark ~ ' .

~33~2 Analysis of ub-~gal proteins produced in E. Coli Plasmid pUB23 (Figs. 1 and 3) was introduced into DS410, a minicell-p,-oducing E. coli strain. (N.
Stoker, et al, in Transcription and Translation: A
practical Approach B. D. Harnes and S. J. Higgins, Eds., IRL press, Oxford, 1984, p. 153). Minicells were prepared and labeled for 60 minutes at 36~C with [35S]methionine ( 600 Ci/mmole, Amersham) as described by N. Stoker et al, supra.
Labeled minicells were centrifuged, resuspended in 2 percent SDS, 10 mM DTT, 10 mM Na-HEPES (pH 7.5) and heated at 100~C for 3 minutes. After centrifugation at 12,000g for 1 minute the supernatant was diluted 20-fold with buffer A (1 percent TRITON* X-100, 0.15 M
NaCl, 5 mM Na-EDTA, 50 mM Na-HEPES, pH 7.5), followed by the addition of phenylmethylsulfonyl fluoride (PMSF) and N-ethylmaleimide to 0.5 mM and 10mM, respectively.
After 4 hours at 4~C, the sample was dialyzed against buffer A containing 0.5 mM PMSF overnight at 4 C, and processed for im~unoprecipitation (as described above).

Analysis of ub-~gal proteins produced in yeast S. cerevisiae cells carrying plasmids of interest were grown in 800 ml of a uracil-deficient medium, then harvested and disrupted with glass beads in buffer A containing leupeptin, pepstatin A, antipain, aprotinin and chymostatin (each at 3 ~g/ml).
The extract was centrifuged at 12,000g for 3 minutes.

*Trade Mark ,., ~:. i 1~3~2 Saturated ammonium sulfate was added to the supernatant to a final concentration of 57 percent. After overnight incubation at 4 C, the precipitated protein was collected by centrifugation at 23,000g for 30 minutes. The pellet was redissolved in buffer A
containing protease inhibitors. After clarification at 12,000g for 3 minutes, the sample was passed through an affinity column which had been prepared by crosslinking at IgG fraction from an ascitic fluid (containing a monoclonal antibody to gal to AFFI-GEL* 10 (Bio-Rad).
The IgG fraction used for crosslinking had been purified from the ascitic fluid by affinity chromatography on protein A-SEPHAROSE. After washing with buffer A lacking TRITON X-100, the antibody-bound proteins were eluted with 0.25 M glycine-HCl (pH 2.6).
The eluate was immediately adjusted to pH 7.5 with 1 M
Na-HEPES (pH 8.5), and thereafter made 0.1 percent in SDS. The sample was concentrated by ultrafiltration in Centricon 30 (Amicon), and subjected to electrophoresis in a 7 percent discontinuous polyacrylamide-SDS gel (U.K. Laemmli, Nature (London) 227, 680 (1970)).
Electroblotting of proteins to nitrocellulose, and immunoblot analysis with a peptide-mediated antibody to ubiquitin were performed as described by P.S. Swerdlow, D. Flnley and A. Varshavsky, Analyt. Biochem. 156, 147 (1986). The same results were obtained with a different *Trade Mark t ~, P.03 ~339~2 -2~-~ntibody to ublq~itin obtained from A. Haas (Univ of Milwaukee Med. School).
Con~truction of E. coli Expression_~ectors_E_codin~
Twenty Ub-X~ l Fusion Proteins _____ ______ __________________ Four of the pKKUb-X-~al vectors (chose encod~n~ Ub-MET.~g~l, Ub-Gln-~g~l, Ub~Arg-~al, and Ub-Pro-~g~l) were constructed ~s followg. Site-directed mu~agenex~s ~M. S~ith, Annu. ReV.-~en~t~
19, 423 (1985); T, M~nlatls, et ~ Molecular Cloning", (Cold Spring ~a~bor Laboratory, N.Y., 1~82); "Current Prococols in ~olec-ular Biology", F.M. Aus~bel, et al., ~Wiley-Interscience, ~.Y., 1987~ w~s used to in~ert the sequence GTAC between the first and second codons of the ubiquitin re~dLng frsme ln the yeast expression vector pUB23 (which encodes Ub-Met-~gAl) and in i~s deri~a~ives, ~A, Bachm~ir, et al., Science 234, 17g ~lg86~, encoding Ub-Ar~ al, Ub-Gln-~gal, and Ub-Pro-~gal. The insertion created a Kpn I ~itk posi~ioned such th~t when the vector is cut wi~h Kpn I and the ends bl~nted by mung be~n nuclease. the second co~on of the ubiquitin reading frame s~arts precisely at one of the fragment'~ ends. Thu~, dl~estLon of e~ch of the above fou~ voo~o~s with Kpn I and Tth lllI, followed by tr~at~ont with ~ung bea~ nuclea~e, yielded four fr~gments whlch conta~ned the corresponding Ub-X ~gal-coding sequences but lacked the flr~t (ATG~ codon of the ubiquitln readlng fra~e. ThesH f~gm~nts we~e ~ubcloncd into an E.

P.01 ~ 3 3 ~

coli ~xpression v~cto~ pKK233-2 [E. Amann and J.
Brosius, Gene 40, 183 (1985)] whioh h~d been prepared by dl~es~in~ it wi~h Nco I ~nd filling in ~taggered ends u~ing Klenow fra~ment of Pol I [M.
Smlth, Annu. Rev. Genet. 19, 4~3 ~1985); T.
Maniatis, et al,, ~Molecular Clonlng", (Cold Sp~ing Harbor Labora~ory, ~.Y., 1~82); "Current Protocols in Molecul~r Biology", F.M. Ausubel, et al_, ~Wiley-Interseience, N.Y., 1987)~. This seep yielded the complctc Ub-X-~al sequence ~in whi~h the ATG codon was supplied by the pKK233-3 ~eceor), optimally posltioned downstre~m of tho regulat-~ble Ptrc promotor of ~he vector. To construct the remainlng ~lxteen pKKUb-X-~g~l expres~ion vectors, pKKUb-Arg-~gal wa~ digested with S~lI and B~mHI.
One of the ewo BAmH~ si~os in pKKUb-Arg-~gal ls located ~t the junction be~ween th~ ubiquitln- and ~gal-codin~ sequences; the other Ba~H~ si~e, presen~
in ~he initial pKK233-2 vector ~E. Amfinn ~nd J.
~ro-~iu~, ~eno 40, 183 (1485)], w~s removed in a prellminary con~r~q~ion s~ep. The ~mall SalI/Ba~n~I
~ragment (containing the Ptrc pro~o~e~, the complete ubiqu~tin-codlng ~equence, and the ~rg codon at the ~b-~gal ~unc~lon~ wa.~ subcloned into ~ M13~p~ vec~or [M~ Smith, Annu. Rev. ~enet. 19, 423 ~19~j); T.
Mani~tis, et al,, ~Molecular ~loning~, tCold Sprin~
~arbor L~boratory, N.Y., 1982); "Current Pro~oco~s ln Molecular Biology", F.M. Au~bel, et al,, (W~ley-Inter~cience, N.Y., (1~7); J. Messing ~nd J.

P.01 ~ 3 ~

Vie~ra, Gene l9, 263 (198~)~, A B~tXI/BamHI
fr~ent oi thi~ con~truct that contained a portion of the ubiq~ltin.coding sequence ~nd ~he Arg codon at the Ub-~gal ~unction, was th~n oxchanged for the sixteen otherw~se Identical Bs~XI/BamHI fra~ents ~from the previou~ly ~de, M13mp~-b~sed ~ons~ructc A. Bachmair e~ al.l S4~4nc~ 234l 17g (lg8~)], whieh differed ~xclu~ively In ~ codon at the ~b ~ga~
~unction. The resulting sixteen M13-mp~-baPed con~truct~ werc treated wLth SalI an~ BamHI, and th~
~mall SAlI/BamHI fra~ments aontaining the ubiquitin-coding sequence and different 6inglc codons at the Ub-~al ~unction were cloned bac~ into p~KUb-Arg-~gal, replaaing t~c origlnal SalI/Bam~I
fr~ment, and yielding the remainin~ slxteen pKKUb X-~gal expression ~e~ors. ~n all c~es, the Ldentity of the a~lno acid encoded ~t the Ub flgal junction of a final pKKUb-X-~const~uct was verified by subcloning lnto M13 and nucleotide sequençi~g by the ch~i~ term~nstion m~thod ~M, Smlth, Annu, Re~, ~enet. 19, 423 (19B5); T. Maniatis, et al., "Molec~lar Clonin~ Cold Spring Harbor L~boratory, .Y., lg~2); "Curren~ Protocols in Molecular Biology", F.M. Au~ubol, ot al., ~Wiley-Interscience, N.Y., ~19B7)].

PurlficAtion o~ 35S L6beled Ub X ~al Proteins_Fro_ E, coli ~3~3~12 An ov~rni~he culture (lml) of ~ coli JM101 c~lls bo~ring one of the twenty pKKUb-X-~gal exp~e~sion v~ctor~ was dilu~ed into 50 ml o~ Luria b~oth x~ppl~m~nted with ampicillin at 40 ~g/ml, and the cells we~e ~own wlth ~h~king fo~ ~pproxl~ately 2 h~ur~ at 37~C. The cells were h~rvested ~y cenerlfu~ation At 4, OOOg $or 10 minu~es, w~shed twi~a wlth M9 buf~er, ~nd resusp~nded ~n 2~ ml of M9 minimal medium supplem~nt~d wlth glucose (0.22%, w/v), thiamine ~18 ~g/ml), ampiqillln (40 ~g/ml~, O.Sm~ isopropylth~ogAlactoside (IPTG), and 0.15 ml o~ 10,5~ (w/~) Methionine Ass~y Medium (Di~co).
A~er inc~b~ion.with shaking for one ho~r at 37~C, 0.5 to l.O MCi of 35S-TRANSLABEL ~ICN: 85~ [3$S~
methionlne, ~15~ [35S] cys~ine) W~8 added a~d shaking was contlnued for 5 minutes. Unlabeled L-methionins wa~ than adde~ to lmM ~nd sh~kin~ was eont~nued for anothor 1~ min~t~s. C~ wer~
ha~vested, wa~hed twice w~h M9 buffer, and resuspended ~n 0.5 ml of 25~ ~/v) s~cro~e, 501nM
Trl~-HCl (p~8.0), Thereafter, 0.1 ml o~ lysozy~e (lOmg/ml, Sigma) in 0.25 M Trl~-HCl (pH 8.0) was added, and the mixture was ~ncubated at 0~ C for fl~e minutes, ~ollowed by ~he addltion of 0.1 ml of 0,5 M Na-EDTA (pH 8.0) ~nd fur~he~ incu~ation at O~C

fcr ~ive minute~. The cell su~pension w~s t~en added to a lysla solu~ion ~O.B ~1 H~0, 50 ~1 of 1 M
TriY~Hcl ~pH 8.0), 1~5 ~1 of 0.5 M ~4~EDTA (p~ ~.~), 10 ~1 of 10~ (w/v) TRIToN~X-100), and ~ently mixed.

* Trad2 Mark . .

nrr~ o~ o no~~~ s~ ~

~3~9~2 The lysz~e ~s centrl~uged at 40,000g for one hour and Ub-X-~g~l W~5 purified ~rom ~h~ superna~nt by affinity chrom~ogr~phy o~ fin ~m~noph~nylthio-pyr~nogal~cto~ide~ag~rose ~APTG-~g~ros~) colu~n as described (~. Ullm~n, ~ene 2~, 27 (1~4~.
Ubiqu~tin~X~ l was el~ted ~rcm APTG-a~arose wi~h 10 mM 2~mercap~oe~hanol, ~ .1 M N~-borat~ (p~ lQ.0), dialyz~d overnigh~ ~ 4~ aga~ n3t 50~ (V~V) glycerol, 0.1 mM EDTA, l mM di~hiothrei~ol ~DTT~, 40 mM Tris-H~l (p~ 7.5~, ~nd ~tored at -~O~C in the ~ame bu~fer. C~ntrol experiment~ showed ~h~t the trsnsient exposure of Ub-X-~al purifie~ by th~
abo~e proceud~e were 0.5 - 1 mg, wl~h enzym~tic a~tiv~ty of 4 - 6 x 104 uniesfmg Pnd specific r~dio~e~vity of 1 - 2 X 105 cpm~g. Unlabeled Ub X ~g~l wes prepar~d essentia~ly ~s describod a~ove except tha~ after two hours of growth in L~rla bro~h with ampicillin, IPT~ was sdded to 4.5 ~M, and ~he c~lls were grown fo~ one more ho~r ~efore har~estin~ and lysis, Pre~aration o~ RetlcuLocyte L~s~te and Assay for De~rad_tion of Test Proteln~
Uashed re~icu~ocy~s ~rom phe~ylhydr~zine~
t~ated r~bb~ts wers pur~hAsed from Green Hectares (ore~on, Wls~onsin), ~nd shipped overnight at O~C.
Ths r~iculocytes were w~shed thr~e ~im~ with 3-4 voL~mes o~ sts~dard phosphat~-buffered saline (P~S) (centr~fugations ~t ~OOOg for 10 minu~eS ~t 4~C~

P.~4 ~3~2 To deplete intr~çellular ATP ~J. Etlln~er and A.
Goldber~, Proc Natl. Ac~d. Sci. USA, 74, 54 ~1977~;
A. Her~hko, et ~1 ., Proa. N~tl. Ac~d._Sci, USA_77:
1783 (1980); Hershko et al., ~. Biol. Ch~m. ~ 258, 8206 (1982~], the cell~ were incuba~ed for 90 m~n~te~ ~t 37~ in Kr~b~-Rlnger phosphate buffer cont~in~ng 0.2 mM 2,~-dlnltrophenol ~nd 20 mM
2-d~oxy~lucose, and then wa~hed three timeS ln PBS.
PelletHd ret~culocytes wqre ehen lysed at 0~C by re~u~p~ndlng the pelle~ in 1.5 volumes o~ 1 mM DTT.
Aft~r ~10 mlnutes ~t 0~C, the A~mple was eentr~f~ges At 80,000g fo~ gO m~nutes at 4-~. The ~upernat~nt was removed, divided into aliquot~, and ~tored under llquid nitrogen. Only once-frozen aliquoe~ were u6ed ln ~11 çxperlments. Unle~s st~ted other~ise, the ATP- depleted reti¢ulocyte extract wa9 used directly a~ter thawing, without further processin~.
In co~e exper~ments~ the thawed extr~et was at first dialy~ed overnight at 4~C ~gain6t 1 mM DTT, 10 mM
~ri~-HCl ~pH 7.5) in d~aly~i~ tubing with a ~.w.
cueoff of -3 kD, Fr~ction II w~ prepared by DEAE
chromatogrAphy of ATP-depleted retlculocyte e~tract ~s previously de~cribed ~D, Finley and A, Var~h~vsky, ~rend~ Biochem. Sci. 1~. 343 (1~85); A.
Her~chko and A, Ci~chanover, P~o~r. Nucl__Ac__R~s.
Mol. Biol. 33, 19 (1~86); $. Pontremoli and E.
Mslonl, Annu. Rev. Biochem. 55, 455 (198~
Rechsteine~, Annu. R~v. Cell, Biol. 3, 1 ~1987);
J.S. Bond and P.~. Butle~, Annu. Rev._Biochem. 56, P. 1~31 398'~2 333~1987) J.F. Dice, FASE~_J. l, 3~g (1987); ~.
Etlinger and A. ~oldberg, Proc. Natl._Acad Scl USA, 74, 54 ~19~7); A. Hershko, et ~l . Proc._Natl.
Acad, Sc~, USA 77 17~3 ~ 0); Her~hko e~ ~l., J.
Biol. Chem., 258, 8206 (lg82~] and stor~d under liqu~d nitro~en, R~action mlxtures for gssAyin~ the de~ada~ion of tost prot~ins in either ehe tot~l reticulocyte exeract oP Fraction II cont~ined ~fin~l aoncentrat~ons) 5~ (v/v) glycerol, l mM DTT, 5 mM
MgC12, 50 mM Trl~-HCI (pH 7.5), 70~ v) reticuloeyte ex~ra~ (or Fxactlon II ~t 6 mg/ml of ~he ~o~al protoin), [35S~Ub-X-~g~l fusion pro~ein a~
20 ~g/ml, ~nd when present, 0.5 mM ATP A~d an ATP-r~generating ~ystem ~lO mM c~eatin~ phosphate, 0.1 mgJml creatlne phosphokinsse). Reaction mlxturoa wcre p~epar~d ~ follow~: a mixture ~o~pletc except ~or ATP and ATp re~ener~ting .~ys~em was ~ncubatod for ten ~nuces at 37~C, to allo~ for the deubiquitination of a ~b-X-~g~l ~usion protein ATP And ATP-regenera~in~ cystem were then sdded to ~rt the ATP-depend~nt rHaation~ in ~he extrace ~nd the 37~C inc~b&tlon continuet. Control reaction~
with the ATP-depletet extr~ct ~ere performed identic~lly except that ATP and ATP-~egener~tlng syotem ~ere om~tted, The ATP-tependen~ degradation o~ l25I~labelod bovin~ serum albumin, hon 1Y90ZYmQ~
and cytochrom~ c from S. cerevl~ae (purch~sed from Sigma, S~. ~ouis, Mi~souri, and lab~led using the chloramine T me~hod (A, Ciech~no~er, et al., Proc P.02 ~ ~ 39$~ 2 Nstl. Acad. Sci. USA 77, 13~5 (198~) w~s ~ssayed as d~ribod above except th~t the 10-minut~
preincubAtion o-f the ~st protein at 37'C in the AT~-depleted reticulocyte extract w~s omltted. To ~ollow the de~radation of test proteins, allq~ot~
we~e tak~n from the reaction mixturo a~ the indicated times, and ~ieher a.cs~yed for the ~moun~
of 5~ TCA-soluble radioactiv~ty present, or analyzed by SDS-PAGE ~U.K. Laemuli, N~tur~ 227: ~B0 ~1~70~]
(8~ polyscryls~ide, 0,05 blsacrylamide, 15x15xO.lS
cm gels), with subsequent fluo~ography.

Detalled Descrlpt~on of the Fi~ure~
Figure 1 shows const~uction of a ublquitin-la~Z
gene fucion. pUB2, ~ p~R322-based ~eno~ic p~A clone ~E, Ozkaynak, et al. Nhture 31~, 663 (1~84) contains ~ix repest-C of the yeast ublqultin-codlng seq~enc~
(open box~s) ~og~he~ with the flan~lng regions (jagged lin~s). pUB2 was modifi~d as shown ln the d1 a~ram by plac~ng a ~am HI site six b~ses down-~roam from the fir~t ~biquitin repeat. Thi~
sllowed the c4n~truc~on of an ln-fr~e ~usion tconfirmed by nucleotide sequencing) betw~cn ~
slngl~ ubiqui~in repaat and the lacZ gene of the expre6sion ~ector pLGSD5-ATG tcalled G2 ln L.
Guarente, ~ethods Enzymol . 101 181 (~983) ) . Th~
term ~ m~ denote* a region of tho pLGSD-ATC that conta~ns the ~plioa~ion origin and flanking se-quenccs of the yesst plasmid c~lled 2~m ci~cle tSee =r. Ll la rvl 17 I'IH

133~2 L. Guaren~e~ ~u~ra). Fi~,ure 3B shows the amino ~cid sequence of the ~usion protein in the v~cinity of the ubiquitin-~gal ~unceion~
Figu~e 2 shows th~t the ln ~i~o ~slf-life of ~g~l iB a functlon of it~ ~mino-~ermin~l re~idue.
(lAne ~ ~inicell isoLated ~rom ~n E. coli strain carry~ng pUB23, the initial ub-l~o~ ion ~Figs. l and 3B), were 12beled with ~35S]~th~onino for 60 minutes et 36~C, with ~ubsequent analysis of ~gal as described. The ~am~ result was obtained when the lab~lad minioell SDS extr~ct was combinad with an unlabeled yeas~ SDS extract b~fore i~munoprecipita-tlon of ~g~l. tl~ne b) S. c~revisiae oells carrying pUB~3 (Fig, l), which encodRs ub-~et-~gal (~lg. 3~), were labeled with ~ S~ethionine for 5 mlnu~es 30~~, with subsaquent analysls oi ~gal. The ~ame rssult wa~ obtained wit~ ehe lcn~hs of ~he ~3 S~met~ionin~ labeling period~ ~rom 1 to 30 ~inutes, And wit~ yO~e extr~c~ produce~ elth~ by mechanlchl disruption of cslls in the presence of protease inh~bLtors or by ~oilin~, the calls directly in sn S~g~containing b~ffer. (lane c) Same G~ l~ne ~ but with E. coli cells c~rryLng thç con~rol pla~id pLGS~5 ~called Gl in L. Guarente, su~r_.
wh~ch enoode~ ~gal. ~lane~ d ~o g) S. cerevisiae cells carryLng pUB23 ~Fig. 1), which encodes ub-Mee~
~g21 (Fi~, 3A), were labeled wi~h [ 5S]me~hionina for 5 minute~ ae 30~C ~lane d) f~llowed by a ch~s~
in ~he presence of cyclohexlmlde for lO, 30, and ~0 , . . .

~ ~ 3 ~ 1 2 -~3-minutH~ ne-~ e to g~, extraction, immunopreaiplt~-tion, and an~ly~is of ~gal. ~lane~ h to J) Sam~ as lanes d to f, but wlth ub-Ile-~gal (see ~ig. 3A).
(lanes k to m~ S~me as lAne~ h ~o ~, b~t wlth ub-Gln-~gal, ~lanex n to q~ S~me as lan~s d to g, but wlth ~b-Leu-~gal. (l~nes r eo u~ game as lanes d to g, but wlth ub-Arg-~al. D~si~na~ions: ori;
o~igin of the separ~ting gel; ub, ubiqui~in; ~gal, an electrophoretiç band of the ~gal protein con-taining a specified amino-te~minal residue; in thi~
t~rminolo~y, the Met-~al portion o~ ub-Met-~gal i6 designa~ed as ~gal. Arrowh~ds denote a ~et~bolic~lly .ctablel abou~ 90kD de~radaeLon produçt of ~gal whlch i~ formed ~pparently as the result of an ln vl~o endoproteolytic cle~vage of a proportion __ ____ of short liv~d gal pro~eins such as Leu-~gal And Arg-~gal (lanes n to u).
Figure 3 ~hows the changing amino acld residues of gal ~ the ubiquitin-~gal junction, ~A~ The initial plasmid, pUB23 t~ig.l). which encode~
ub-M~t-~gal, was mut~g~nized as descri~ed above to convert the original Met codon ATG at the ub- gal ncelon into codons hpec~fying 19 amino acid~ other than Met, ~The original roun~ of m~tagenesi6 shown in Figure 3, produaed 15 out of 19 posslble ~ub-stitutlons. The remaining four subs~itutlon~ were prod~ced later ~see Table 1)). ~he ar~owhe~d ln-dicates the s~te of the deubi~uitinating in vivo cleav~ge In the nascent fu~lon protein ~hat occu~s ... . _ , . .. . . . . ..

F.~5 ~33~2 -3~- -wlth all of the fusion proteins eX~epe ub-Pro ~gal ~see text). All of th~ constructions shown encode His a8 the second gal re~idue. Tn addition, in some of the construction~ (ub-Met-His~Gly-~gal, ~b-Met~Gln-Gly-~gal, ~nd ub-K~t-Gln-His-~ly-~g~l, the last one produced by a~ in~Hrtion mut~tion, see Table 3), eith~r ~ or Gln were following Met a~
the ubiqultin-~al j~netion, wlth indist~n~ul~h~ble consequences io~ the meta40Llc stsbilities of the oorrespondln~ ~gal proteins. ~B) The amino acid sequence ~ln ~ingle-let~er ab~rov~at~on~) o~ ub-Met-~gal, the initial fu~ion pro~ein ~Fi~, 1), in the ~iclnity of the ub-~gal Junction. Sin~le-letcer ~mino acld abbre~iation~: A, Ala; C, Cys; D, A~p; E, ~lu; F, Phe; G, Gly; H, Hls; I, Ile K, Ly~; L, Leu ~, Me~; ~, Asn; P, Pro; Ql ~ln; ~, A~g; S, Ser; T, Thr; V, V~l; W, Trp; Y, Tyr.
Fig~e 4 6~w~ that ubiquitin-~gal is sho t-l~ved ~f not deubi~u$~1nated. (lanes a to g) S.
cerevlslae cells carryin~ plas~lds encoding ub-X-~gcl fu~lon proteins ~n which X is the resi~ue indlcated a~ the top of each lane, were la~eled for 5 minute-~ at 30~C with [35S~methlonine, followed by extraction, immunopreclpitation and analysi~ of ~gal. Fluorographic exposures for these lanes were several tlme~ lon~er than those for similar psttern.~
in Figu~e 2 to reveal ~he ~ultipl~ ubiqu~inAtion of ~hort-lived ~gal proteins. (lanes h, 1) Fluoro-~raphic o~erexpo~ure of lanss n, o In F~g. 2 to P. IZIl tj 3 39 ~ ~

rev~al the "la~der~ of multiply ubiquitina~ed Leu-~gal proteins in a p~ls~-~hase experimen~ ~zero .
~nd 10 minutes cha~e, respectively) . ~lane ~ Same a~ l~nex a to ~, but wlth ub-Pra ~gal. ~lane k) Same aE l~ne ~I bu~ ~lth ub-Gln-~al ~lane 1) S~me as lane j. (lanes m to p~ S. cerevi.~l~e cells carrying a plasmid encoding ub- Pro-~gal were lAbeled for 5 minutes at 30~C wi~h t35$~methionine (lane m) followed by a ch~se in the presence of cyclohexLmide for 10, 30, ~nd 60 minutes (lanes n to p), The upper sm~ rrow to the ~lght of lane p deno~es ub-Pro-~al, a ~m~ll proportion of which is stlll prc~ent af~er l ho~r chase. The lower small arrow indioates an apparHntly d~ubiqultinated Pro-~ga~
that slowly accumulates during cha~e And Is ~eta bolically stable, The dot to the left of l~ne m deno~es an endogeneo~ yeast protein that is p~e-cipi~ated in some experiments by the ~ntibody ~s~d.
Square bra~ets denote the m~ltiply ubiquitin~eed ~-gal speeies ~see Fig. 5~. Other design~tion aro as in Figure 2.
~ igure 5 Qhows the "ladder~ ~gAl species containlng ubiquitin, (lane a) S~ cerevisi~e eells carrying a plasmid which ~ncod~ ub.Gln-~gal, were grown and di~rupted, and the oxtracts proce~ed for isola~ion of ~g~L proteins by affinity chromato-graphy on a column with lmmobilized antibody to ~gal. The ~gal proteins thus ob~ain~d wcrc ~lectro-phorosed in a polya~ryla~ide-SDS gel, transferred to P.02 ~33~i2 nitro-cellulose, ~nd probed with an antibody to ubi~itin. (lane ~) Same as lane a, b~t wlth ub-P~o-~g~l. (lane c~ S~me ~ b but a longer autoradiographic expo~ur~. (lane d) S, cerevisl e cells carrying a pl~smid whio~ encodes ub-Le~-~gal ~ere labeled with ~3$S~methionine for 5 mi~u~es, with ~ub~quent extraction, immunopreclp~tation and electrophore~is of ~gal ~the sa~e ~ample as in F~gure 4, l&n4 f). Square bracket~ denote the multiply ublquitinated Gln-~gal speole~ de~ectsd with ~ntibody ~o ubiquitin. The arrow indicates the band of ub-Pro~al, the initial ~usi~n pro~ein seen in l~ne~ ~ and c. The arrowheGds indicate the posltion of the band of deu~iquitinated ~1 (d6-tectable by either ~oomassi~ ~tainlng or met~bolic l~beling, but not wlth antibody to ubiquitin) derived from the ub-~ln-~gal fucion prot~in.
Figure 6 ~hows both prokaryotic and ~uk~ry~tlc long-liv~d intracellular protein~ h~ve st~bilizing amino acid residues at their amino-ter~inil whereas ecreted protein~ exhlbit a complemen~ary bi~.
~ A) ~8 long-lived, dlrectly sequenced, iner~
cellula~ ~noncompsrtmentalized) proteins with unblocke~ ~mino-termlni from bo~h prok~ryotes (77 protelns) and eukaryotes ~131 proteins) we~e dl-qtri-buted into three groups aocordlng to the nature of their amlno-term~al residues as defined by the ~-end rule (Table 1). All of the long-lived intra-cellular protein~ examined ~ear exclusively p.0~

~ ~3~

9tabilizin~ r~sidues At the~r amino-termini. In p~nels B to D, An~logo~s di~gr~m~ are presen~ed for 2~3 secreted eukaryotlc protelns ~B), for 37 light and hea~y immunoglobulin chains ~C), and fo~ 94 secreted eukAryotiç toxins ~. Entries in ~ a~d are ~ubsets of entrie~ in ~. For proteins in B to ~, ~he amino-ter~inl compiled cor~espond, whenever the a~signment i5 po~sible, to the mo.~e proce-~ed form of n protein that ~s ~till located within a secreting cell, Th~ da~a in A to D were manually co~piled from the entire ~et of complete protein cequences avsilable before 19~1, The same con-clusion~ heve been recently reaçhod ~ter a more detailed and exten6ive, compu~er-a~sisted tabul~tion of protein amino.termini usLng the current Natlon~l 8iomedlc~1 Resea~çh Foundatlon d~&b~se. The amino-ter~lnal residues o~ Asn, Cys, His, ~nd Trp were exclud~d from ehe compllatio~ because in vivo half-li~es of the corre~ponding ~al pro~ins are ~till unknown (ses, howe~, the legend to T&hle 1).
Inclusion of the residues (T~ble l) ~nto ~ recently compil~tion of the ~me type d~d not change the o~i~inal concluJion. Althou~h the amino-terminal Pro w~ also excluded from the compila~ion, Pro ~ppear.~ to b~ a stabillzing rosidue for ~g~l (T~ble l), consisten~ wi~h the frequçn~ pre.~ence of Pro at th~ amino-~ermini of long-lived noncompartmentaliz~d prote~n~, 1~3~2 Figure 7 shows the construction of ubiquitin fusions with mouse dihydrofolate reductase.
Figure 7 describes the use of ubiquitin-protein fusions to generate X-~gal and X-DHFR test proteins bearing different amino-terminal residues. Figure 7A
is a diagram which represents changing amino acid residues of ~gal at the ubiquitin-~gal-junction into codons specifying 19 amino acids other than Met. The black rectangle denotes the sequence between residues 2 and 45 of an X-~-gal protein that is absent from the wild-type E. coli ~-gal. The bulk of this sequence is identical to an internal sequence of the lac repressor which was joined to the amino-terminal region of the E. coli ~gal as the result of a construction route employed in making the yeast expression vector. The arrowheads indicate the site of deubiquitinating in vivo cleavage. Figure 7B is the same as 7A but with mouse DHFR as a test protein. Figure 7C is the same as B but with the original Ub-X-DHFR fusion proteins modified by an insertion of the 38 residue amino-terminal region of the X-~-gal test protein shown in A.
Figure 8 shows a greater range of specific dihydrofolate reductase (DHFR)-based protein con-structs which have in common the DHFR moiety, anddiffer exclusively in the specific amino-terminal extensions attached to DHFR. Structure I is an initial DHFR construct. Structure II contains the 40-residue amino-terminal extension derived from the amino-terminus of ~gal (see Figure 3). Structures III-V are variants of Structure II in which either one or both of the lysine residues (denoted as K in the single-letter code, see the legend to Figure 3) were replaced by the arginine residues (denoted as R).
Structures V-X are variants of the Structure II with , ~ 43a- ~ ~3~2 increasing deletions in the carboxyl-terminal half of the ~gal-derived extension. Structures XI-XIII are variants of Structure II with increasing deletions in the amino-terminal half of the ~gal-derived extension.
Single-letter amino acid designations of the amino-termini of Structure I-XIII indicate variants of these protein constructs that differ exclusively in their amino-terminal residues. These variants were obtained through the use of the ubiquitin-protein fusion approach described herein (see Figure 3). Each of the Structures I-XIII was constructed at the DNA level using site-directed mutagenesis and other methods of recombinant DNA technology currently standard in the field. These -~ .......................................... P.05 1~3~3 1;~

~NA eonstructs were intxoduced into ~he ye~st S.
cerevisiae, ~nt the half-llv~ (left col~mn b~
Fl~ure 8) o~ th~ corre~ponding protelns I XIII were dlrectly determined usin~ the methods described abov~ for ~al ~nd a monospec~flc ~ntibody to DHFR.

~esul~ and Di~cussion ______ . .

~apid in vivo de~biqul~ination o~ a nascent_ub~-quitin-~al ~u6 i O n p r o t o In Branched ubiquitin con~u~ates in which the ~rboxyl-terminal glycine of ubiquieln moie~ies lq ~oined via an isopeptid~ bond to ehe ~-amino groups of internal ly6ine residues in proteins apparently c4mpri~e the bulk of ublqultin con~ugates ln eu-karyo~i~ cell~. Joining of ubiqultin to ~he amlno-terminal ~-a~ino ~roup~ of ~Arget proteLns, to yield lin~ar ubiq-~itin con~u~te~. may also be che~nically feasible. See A. Hershko, ~t ~1., PEAS USA 81: 7021 (1984). Whether or not linear ubiquitin-pro~ein fusions are act~ally synthesized in vlvo through pos~ransl~tlonal enzymatic conj~ation of ubi~uitin to proteln amino-termini, s~ch proteins can also be pro~uced by con6tructing apprOpriaee chimerio ~enes and expr~ssin~ ~hem in vivo. Construction o~ one such ~ene, which encodes yeas~ ubiqui~in llnked to ~gal of Esc_erichia coll, is ~hown in F~ ~u~e 1.
When this ~ene is expressed in E. çoli, the resulting ~ containlng protein h~s an apparent .. ~ . .

i 2 molecul~r m~ss which is ~pproximately ~ kD gre~ter t~ t that of ~h~ control ~al, ~ value consi~tent with the pro~onc~ of ubiquitin ln the protein ~naoded by the ~himeric gene. In contrast, when the same gene i8 expresced in ye~t, ~he cor~e~ponding ~gg~l protcln is ffloçtrophoretically in~is~in,~uish-able from the control ~gal. This resul~ is in depentent of the len~th of the [35Slmethionine l~b~ling period ~be~ween 1 ~n~ 30 minu~es).
Fur~her~or~, de~erminatlon of the amino-terminal ro~iduo in tho putative Met-~gal ~half-life, tl/~ 20 hours) by Ed~an degrad~tion of the in vivo-l~bel~d, gel-purlfled ~al (Fi~ure 2, l~ne d) directly conflrmed tho pres~noe of th~ expected Me~ residue (~igure 3A ~nd Table 1) ~t its amino-terminus.
Independent ovidence tha~ ubiq~ltin oleava~e of the fu~ion protein occurs immediately after the last Gly residue of ublquitln 1~ presented below. We con-clud~ t~at in yea6t, ubiquitln is efficlently cleaved off ~he nAsCant ubiqultln-~g~l fusion protein, yleldlng a deubiquitln~ted ~gal The ~b~ence of the deubiquitination reaotion in E. coll eonsi~tent wlth Oth-r lines of evidence in-~icating th~t proka~yote~ l~ck both the eukaryo~lc ubiqul~ln and ubiquitin~specific enzymes, At the samo timo, the po~sibility remains that a function.~l counterpart of ubiquitin eXi5~9 ln bActeri~ but is different ln its amino acid ~equenee fro~ that of eukaryotic ubiquitin. Tho pre~ent lnvention cle~rly 13338~2 ~pplies not only to the extrem~ly çlose ~mlnb ac~d homologs of ub~quitin such as those found in eukaryo~es but also to f~nc~ional homologs of ubiquitin such as those ~hat may exi~t in b çteria.
The ubiqui~in~ l junction en~oded by the chimeric ~ene, Gly-Met (F~ g~res 1 and 3B), ls identlcal to the ~un~tions be~w~en ad~cent repeats ln the polyub~uitin precursor proeein, whioh is efficiently processed into m~ture ubiq~itin. Thu-~it ~5 likely that thc same protease, as yet un-characterized b~ochemic~lly, is responslble both for the conversion of polyublquitin into m~tu~e ubi qultin ant for the deubiquitin~tlon of t~e nsscent ublquitin-~al prote~n. If ~o, one potential way to inhibit the in vivo deubiquitination of the ubi~
quitln ~g~l ~and thereby to allow ~nalysis of metabPlic consequences of ~ s~able ubiquitin aet~ch-~ent to ~al) would be to con~e~t ~he Met ~esidue of ~g~l ~t the ublquitin~ l junction (Figure 3B) into othe~ amino ~cid residuos (Flg~r~ 3A~. The unex-pected results of ~uch an approaoh ~re d~scribed below.

~ho_in_vivo half-life of_@~al_i~ funotion_of_~t_ a~ino-~e~in~l residue_ The ATG codon which speci-~ie~ the origin~l Met ro6~du~ of ~1 ae the ubi-quLtin ~unc~ion (Figure 3B~ was ~onverted by ~ite~
directed m~t~genesis into codons ~pecifyin~ lg other ~mino aoids (S~o ~igure 3A ~nd Table 1~. These ~3~9842 construc~ions di~fe~ exclusively in the first codon of ,Bgal a~c the ubiquitir~-,Bgal Junction ~Figure 3A).
After each of the 16 plasmids thus desi~ned was int~odueed into yeast, ~naly~is of the ~orrespon~i~g ~gal proteins pul~e-labeled in vivo led to the following results (Fl~ure~ 2, 4, and Ta~
1) With one exception ~ee below), the effi-clent deubiqultination of the na~cent ubi~uitin- gal occur~ irro~pect~ve of the nature of ehe amino ~cid resldue of ~gal at the ubiquit~n-~gal ~unction.
Thus, th~ apparen~ly ubiquitln-specific protease that ele~ve.~ the orlglnal ubiqultin-~gal pro~ein At the Gly-Met junction is ~enerally insensitiv~ to the nature of the fir6t residue of ~1 at the junction ~Figure 3A and ~Able 1). This result, in e~fec~, makes lt po~ible to expose di~feren~ amino acid ~esidues at the amino-termlnl of the otherwise identical ~gal proteins produced In ~ivo, 2) The ln vlvo half-lives of the ~gal proteins thu6 designed vary from more ~ha~ 20 ho~rs to less than 3 mlnutes, depend~ng on ~he nature of the amino acld re~idu~ expos~d a~ the amino-terminus of ,~gal (Figures 2, 4, and Table 1). Speclflcally, de-ubiquitinated ~gal proteins with eith~ M~t, Scr, Al~, Thr, Val, Cys or ~ly ~t the amino - te~inu-~ have relatively long in vivo half-lives of 20 houxs or more (figure 2, lanes d ~ ~, an~ Table 1), similar ~o the half-life o~ a con~rol ~gal whose gen~ had not been i'u~ed t~ thnt o~ ublq~ltin. In strlking ~ 3 3 9 ~

çontrast, the ~gal protelns wlt~ either Arg, Lys, Phe, Leu, A~p or Trp ~t the a~ino-ter~inus have very short half-li~es, ~oetween ~pproxi~a~ely 2 minutes for Arg-~gal ~nd approximately ~ minutes for Lys-~g~l,.Phe-~gal, Leu ~gal, A~p-~gal, Asn-~g~l and Trp-~g~l (Figure 2, lanes n to u, and T~ble 1~. The ~alf-ll~e of ~al pro~eins with amino-t~rmin~l ~esidues of eithe~ ¢ln, His or Tyr is approx~mately 10 mlnute.c (Fig~re 2, lane~ k to m, and Table 1 ), whlle an a~ino-terml~al ~le o~ Glu confers on ~g~l a h-l~-life of ~pproximately 30 min~tes ~Figure 2, lanes h ~o ~, and Table 1). Both pulse-ch~e and continuous labelLng cec~niques were used ln these experiments and yielded similar re~ults The set of individu~l amino acids can be orde~ed with respect to the h~lf-live~ th~t they con~er on ~g~l when expoged At its ~mi~o-te~minu~.
The resultin~ rule ~Table 1) ls refer~ed to ~9 the ~N-end rule".

~,~ 3 3 9 $1j ~49 -Tab le 1 _ T}le N - end rul e .. . .. .. . .... .. . . ..
In vivo Radius o~ dqubiquitina~ion Regidue X ingyraeion of n~scent u~-X-~al of X(A)ub-X-~l tl/2 of X~gal ~ . . . . . .. . .
Met 1. ~0 +
Ser 1. 08 +
Ala 0,77 +
~hr 1. 24 + ~20 l~ours V~l 1. 2g ~ly . O +
Cy5 Ile 1. 56 +
Glu 1. 77 ~ -30 minu~es Tyr 2 .1 3 Gln 1. 75 + -10 minutes H 1 ~

~he 1.~0 +
Leu 1. ~4 + ~ 3 ~ninutes Trp Asp 1. 43 +
Aon Ly~ 2 . 08 +

_.. , ,.. ,.. , ~...... . .

r . 1 1 Arg, 2 . 38 ~ - 2 ntinute5 ~ . ~ ....... ... ~ .. ...... ... .. .. .. . .. .. .. . .
Pro 1.2S -* ~ 7 minu~es *The rnt~ of in ~ivo d~ubiquitinatlon o~ ub-Pr~ ~gal is extremely low, ~he tl 2 ~ho~n ig ~h~t of ~he i~itial ub-Pro-~al fusl~n pro~ein (see Fig. 4, lanes ~ eo p~.

~ 3~.3~8~2 LeRend to ~able 1 The N-end rule. In vlvo half-lives of ~
proteins in ~he yeast S. corevi~l~e were deeermined either by the pulee-ahs~e ~echni~ue ~for short-lived ~l's see below~ or by measuring the enz~atie ac~ivlty of ~gal in ~rude extrac~s. For ~he me~surements o~ ~81 actlvity, cells growing In a ~lactose-qontaining mediu~ were tr~n~ferred to an oehe~wise id~neic~l medlum lsaking g~lactose and containing 10 percen~ gluco0e. After ~urth~r ~rowth for at least 5 hours ~t 30~C, the raeio o~ ~g~l ac~ivitiec per cell before and after shift to glucose was determined for e~h o~ the ~gal pro-teins. ~GAL promoter~dri~en expression of the fusion ~enes tFigs, 1 ~nd ~) is repressed in glucose ~edium]. For shorter-ll~ed fl~al prote~ns ( 1 2 1 hour), the pulse-chase te~hniq~ was used as ~ell (Fig.~, 2 ~nd 4). El~trophorotic bands af ~gal protein5 labeled wi~h [ Slme~hlonine ~n pulse-cha~e experiments were cut out fro~ scintillant-impre~na~ed3~ried gels ~imilar to ~hose of Figs. 2 an~ 4, and S in the bands wa~ de~erminod. The ~n vivo decay of ~hor~-live~ ~gal pro~elns devia~cd fro~ first-order kin~tic~ in that ~h- r-t~ o~
de~radatlon WA~ lower wh~n ~eflsured at lator ~1 hour~ tl~e point~ of the chase, the lower rat~
re~loctin~ ei~her a tlme dependen~ ~oxic effect of cyclohex~ide or intrinsic char~cteristlcs of the in vivo d~gr~dat~on proc~s. [Arrcst of ~ransla~ion i~
requir~d ~or an sffic~en~ short- ~erm chase in S .
cerevisiae because of t~e amino acid pool equili~
bratlon proble~s rclaced to the presence of vacuoles ln this organi~mJ. Tho hal~-lifc ~alues listet 4elow were detormined for ehe ~i~st 1~ minutes of chase. Sevsr~l linos of svidoncc (see descrlption of F~gs. 4 and 6) s~g~es~ th~t Pro i8 a s~ab~lizing residue . The llsted rsdii of gyr~tion of amlno ~cids ~re from. M. Levitt, J. Mol~ Biol. lQ4,59 (197~).

'1 2 Deublq~itinatlon_o~_Ub X @~al_f_6ion ~roteins in ATP-d~leted re~laulocyte extr~ct~
Each of twenty 35S-labelled Ub-X-~gal proteins prepared in E. co~ descri~ed ~bove, w~ added to an extrac~ prepared from ATP-deple~ed rabbit reticuloçy~es (E~lin~er et al., Proc, ~atl._Acad.
ScI. USA 74: 54 ~1977~; Hershko et__L,, Proc~-Nat Aoad. Scl. USA 77: 1783 (1980) H~r~hko et_al ., J_ B~ol. Chem. 258: ~206 (lg82)), and the fa~es of the ~dded proteins were followed by SDS~polyacrylfi,nide gel electrophore~ SDS-PAGE). As h~d been o~ser~ed in_vi~o wi~h the ~ama ubiquitln fusions in yes~t, an app~ren~ly ubiqui~in-specific prote~se ln reticulocyte extrac~ deubiqui~in~ted ~he added Ub-X-~gal fusion protein~ to yield ~he corresponding X ~gal te-~t protein~. The de~biquitinatlon of 19 out of the 20 Ub-X-~al proteins in th~ ATP-depleted extract was more than gO~ complete in 5 ~inu~8 At 37- (T~ble 2~. The sin~rle exception, both in yeAst and In ret~culocytes, is Ub-Pro-~g~l, which w~s deubiquitina~ed ~pproximately 20 tl~es more slowly than were the other ~b-X-~ral proteins.
Amino acid ~equencing (by Edman degradation) of deubiquitinated ~g~l pro~ein~ rei~olated ~ro~ either the r~ticulocyte ex~ac~ or yeast cells showe~ thatr in every case tested, the proteolytic clea~age oooured precisely Bt the Ub-~gal ~unction, Althougrh seq~encing re~aled that the amino termini o~ ~omo X-~gal pro~eins underwent ~peeific ~3 .. , ~ ~3~'12 modificiations (see Table 2), in no case did these modifications involve proteolytic cleavages beyond the amino-terminal residue X.
All of the deubiquitinated X-~gal proteins were 05 metabolically stable in the ATP-depleted reticulocyte extract as judged from SDS-PAGE
analysis and from the negligible production of acid-soluble radioactivity in the extract. Thus, preincubation of Ub-X-~gal fusion proteins in the 10 ATP-depleted reticulocyte extract makes it possible to generate twenty X-~gal test proteins which differ exclusively at the amino-terminal residue X.

., ~,".~

9~2 Half_life oi a ~al proteln in ATP-sup~leme_ted reticulocyte extr~ct lg functio~_o~_~he_~gal's ami_ o terminal residue.
While all of the twenty X-~al proteins were met~bolically stable i~ the ATP-deple~ed reticulocyte extract, most o~ them became-~ho~t-lived upon addition of ATP ~o ~he extr~ct We refer to an amino terminal resi~ue ~g stAbillz~ng if th~ oorresponding X-~g~
relatively long-livsd in the ATP-~pplemen~ed ex~rAct (less than 10~ ~e~r~dation in 2 hours at 37~C~, ~nd as dest2bilizing if the degradation of the corre6ponding X-~l in ~he extrac~ exceeds 15 under the ~am~ conditions.
The t~ co~rse6 of degradat1 on for several X-~g~l protein~ 6howed reproduaible initial lags.
However, ~emilogarithmic plots of the ~ime cOurse~
showet ~hat, after the ini~ial ~ags, the degradation of X-~al in the ATP - supplemented rcticulocyte extract obeyed first-order klnetics for At lea~t the fi~t ~wo hours, making it possible to compfire the degrada~ion of different X-~gal pr4~eins by comparlng their h~lf-llv~s in the extract.
Th~ r~nge of ~gal half~l~ves ln the reticulocyte extract encompas~es ~ore ~han two orders of ma~ni~d~, from approximately 50 minutes for ~ln ~g~l to approxl~tely 100 hours f~r Va}-~gal, Ihe hal~-lives of X~ l protein~

- "

173~

-55.

bearing s~ab~ ng am~no-terminal res~dues r~nge from approxl~ately 20 hours ~or Ile-~gal to approximately lO0 hours for V~ al. Half-lives of the metabolically uns~able X-~g~l protein~ in the r~ticulocye ex~ract wq~e comp~rable to the half-lives of othcr proteoly~le ~ubs~r~te~
(iod~nated serum albumin, lysozyme, and cytochrome c) ~n the same extract. These lat~er test proteins have been used in ~arlier ~tudie~ o~
ublqu~tLn-dependen~ prot~in tegrada~lon in retlaulocye extract (Flnley et al., Tren_s Biochem.
Sci 10: 343 tl~85); Etlin~er e~ 1., Proc._Natl.
Acad. Sci. UsA 74: 54 ~l977)~. Recentl~, at le~st some of these pr~teins have been shown to be t~rge~ed for de~rad2t10n via their destabilizing amino-terminal res1dues [Reiss et al., J._B~ol.
chem. 263: 2~93 (1~88)] a~ defin~d by the N-end rul e, - 56 - 1~39~4~
Table 2 The N-end rule in yeast and in mammalian reticulocytes Residue X Half-life of X-~gal Amino terminus of r~ .oldlG~d X-~gal as d~te~ led by in protein sequencing Ub-X-~galYeast ''-nllll-' -n Yeast Retic~' ~ cytes (S. cerevisiae)r~,tic~' yte~s Jn vivo In vitro /n vivo In vitro Val >20 hours 100 hours - Val-~gald~e Met >20 hours 30 hours Met-~gala Met~galdle Gly >20 hours 30 hours - Gly-~9aldle Pro>20 hoursl >20 hoursl i i J
Ala >20 hours 4.4 hours Ala-~galb Ala-l~gald~f Ser >20 hours 1.9 hours h Ser-~gald~f Thr >20 hours 7.2 hours Thr-~galb Thr-~gald~f Cys >20 hours 1.2 hours - [?]-¦~gal9 lle30 minutes 20 hours lle-l~galb~Clle-~gald~e Glu30 minutes 1.0 hours Arg-Glu-~galCGlu-13gal+Arg-Glu-~gald Arg-Glu-~galf His10 minutes 3.5 hours - His-~gald Tyr10 minutes 2.8 hours Tyr-l~galb~C Tyr-~gald Gln10 minutes 0.8 hours [?]-Glu-~gall[?]-Glu-~gal + Glu-~gald~kArg-Glu-~galf Asp 3 minutes 1.1 hours Arg-Asp-~galCAsp-~gal +Arg-Asp-~galdArg-Asp-~galf Asn 3 minutes 1.4 hours Arg-Asp-~galCrAsn-~gal + Asp -~gald Asn-~gal +
LArg-Asp-~galf Phe 3 minutes 1.1 hours - Phe-~gald Leu 3 minutes 5.5 hours - Leu-~gald Trp 3 minutes 2.8 hours - Trp-~gald Lys 3minutes 1.3hours - Lys-~gald Arg 2 minutes 1.0 hours - Arg-~gald a Determined by ~I.Ji~h ~. ' ' sequencing (Bachmair et al., 1986).
b The S. oerevisiae strain used for e~.. ~ , of this X-~gal was BWG-9a-1 (MATa, his4, ade6, ura3).
c The S. oerevisiae strain used for a, . ~ , of this X-~gal protein was a mutant (obtained in the background of the BWG-9a-1 strain) in which all of the otherwise short-lived ( 1PI Ihiql '- I ' ~ X-~gal test proteins are ~ ' " 'Iy stable, whereas Ub-Pro-~gal is still short-lived (I.
Wl~nning, A. Bachmair, and A. Va.J,~ , unpublished data). This mutant (whose use allowed the isolation of the otherwise short-lived X-~gal proteins in quantities sufficient for sequencing) retains both the intact "d ,_~. dGyl ' " , pathway and the Ub-X-~gal Ihjql ~ ' I ~ - 19 activity but is impaired in the amino-terminal, ~. ,' ' , of at least the X-~gal proteins.
d This X-~gal protein was incubated in ATP-depleted reticulocyte extract for 20 min at 37OC before reisolation and sequencing.
e This X-~gal test protein was incubated in ATP-su~",L...~ 'ie ~'- ,' extract for 1 h at 37~C before reisolation and sequencing.
f This ~gal protein was incubated for 2 h at 37OC in ATP-su~ h Fraction ll before reisolation and sequencing.
9 Cys-~gal was incubated in ATP-depleted reticulocyte extract for 30 min at 37~C before reisolation and sequencing. The amino-temminal Cys, unmodified by alkylation before sequencing, could not be idenb'fied by the c.l .. . ' _ _. hi-, prooedures used; howGver, the second and subsequent sequencing steps unambiguously identified the protein as ~gal.
h No signal v,~as seen upon sequencing of Ser-~gal reisolated from yeast, strongly c--__ '' ~ that the protein's amino terminus was blocked. Note that Ser-~gal was not blocked when reisolated from ATP-sL~",I~,..._.,' ' ~- '' " h extract.
In both yeast oells and n " ~' , '.e extract, Ub-Pro-~gal is d~l ' ' l ''', ' ' _,, . ' ' '~ 20 times more slowly than are the rest of the Ub-X-~gal fusion proteins (see main text). Pro~gal, the product of slow ~lol Ihlql I " I '' I of Ub-Pro-~gal, is a long-lived protein in both yeast oells and ~G'' ~ ' extract.
J The amino-temminal residue of this sequenoe could not be identified unambiguously with the amount of ~gal used (~1~ pmol), but, from the data obtained, was most likely Arg. The data clearly identified Glu as the second residue.
k The frame-shifted sequenoe (?)-Glu-~gal was the more abundant (~90%) of the two ssquenoes present. With the amount of ~gal used (~15 pmol), the amino-temminàl residue of this sequenoe could not be identified unal ~ '_ mll~ly but, from the data obtained, was most likely Arg.

~ 33~k2 Amino-termin~l location of an amino ~cid is essenti~l for its effect on @~al h lf-llfe_a~_te~ted in ~oast Site-direct~d mutagenesis was employed to insert n codon speçifying a "stablll~ln~ amino acld (in t~s experiment, the Met resid~e~ before the ~irst codon of ~al at the ubi~uitin-~gal ~unction (Table 3). Insertion of a st~blllzln~ resid~e (Met) befo~e either ~nother ~eabilizing residue (Thr) or va~ie~y of de~t~bilizing residue~ ~Gln, Lys, ~nd hr~) at the ublquitin-~al ~unction lnvariably reoult~ i~ a long-lived deubiqultinated ~al ~Table 3). Furthermore, in contrast to ubiquitin-Pro-~g~l which is noc only short-li~ed but ~l~o resistant to deubiquitination (Figure 4, lanes J to p, and T&ble 1), ubiqui~in~Met-prb~ s;al ls efficien~ly deub~-quitinated in vi~o to yicld a lon~lived Met-~ro-~gal (Table 3). These rexults show th~t both the identi~y of ~mlno aoid ~ssidue And lt* amino-te~minsl loeation ~presumably ~he presence of a free ~-~mino group) are essential for its effec~ on ~gal hslf-life, In addition, these results (Table 3) f~rth~r support ~he expectation th~t ubiquitin specifio cleav~ge of ths fusion protein oocurs lmmediat~ly after the last Gly resldue of ubiquitin (FiRure 3A~.

~ ~.3~2 s~

T~le 3. ~ er;~inal location ~f an arnino acid is esset~tial ~qr its effec~ o~,~gal half-l~f~

t3 o~ deubiqui tinatet Fusion ~rot~in fusi~n pr~tein ~ llb ! Thr - ~g~l 720 h~rs ub - ~e~ - Thr - ~gal ~20 hours ~ ub - Gln - ~gal ~tO minutes ub - ~et; Gln - ~g~l ~2~ ho~Jrs ub - ~ys - ~gal ~ 3 minutes ub - ~lee - Lys ~,~gal ~20 hours ~ u~ ! A~g - ,Bgat ~2 min~tes ub - ;let - A~g -,~s~t ~2~ hours ~ ub - Pro - ~gal ~7 minutes ub - ~let ~ Pro -,~gal ~20 hours .

Amillo termln~l location of an ~mino acid i~ essen-ti~l for its effec~ on ,~gal l alf~l~ ~e The in~er-tlon mu~an~s wore obt~ined essentially as doscri~ec~
for the lnitial ~t of mu~nts exeept t~a~ a 32-rssidue oli~;onucleotid0, S' CccGGGATccGTGc:
T~) (G~ CA~ACCACCTCTTAC~ w~s usod, con~slinitt~
b~sao on the S' ~ic.le ancl L5 basas on the 3' side o~
~he am~iguo~s cod~n inserted behind the Mot oodon, ~es ln pllrontheses ~eno~ Ambiguitie$ at the posi~lons 16 And 17 in the sequence. Half~ es o~
the corr~spondin~ ,~g~l proteins ~ere deterrolned ~s ~çs~r~ bed ln the legend tP Table 1.

~ ~3~2 .s~ .

A_lon~-llved_cleava~_product of ~a~ form~d durlng tecsy of short-lived ~al protei_s_ The electrophoretic p~ttern~ o~ short-lived (but not oi long-llved~ ~al proteins Invarlably contain a ~pecific, ~bout 90 kD clea~ge product of ~gal (Figure 2, l~ne~ n to u) which, unlik~ the p~rent~l ~g~l specle~, accumul~tes durin~ the postlabeling (cha~e~ perlot ~Figure ~, lane~ m-p).
The 90 KD ~gal fr~gm~nt conctitute.~ a relatively small proportion of the init~al amount of the pulJe-labeled ~gal, Nonet~ele~s, its ~xise~nce implies that An in v~o endoproteolytlc cleavag~ ean rescue a protoin fragment from the metaboliG f~te of its short-live~ parental proeeln. It remainS ~o be seen whether the resulting po~sibility of multiple h~lf-lives with~n a single protein speeies $s exploited in the de~ign of naturally short-lived proteins.

Ubiquit~n @~l is short-lived when not__eubiqui tin~ted.
Ubiquitin-Pro-~gal, che only ubiquitin-~al fusion that Is not ~eubiqultina~ed ln vivo ~Fig. 4, lanes ~ ~o p), h~s a half-life of ~pproximately 7 minutes ~Table 1) which is les~ than 1 percent o~
the half~ e of met~bolically ~table ~al proteins (~able 1). One interpretation of this re~ult is th~t a ~etabolically ~table ublquit~n attachment to prote~n a~ino-terminl i8 s~ lc~ent to ~i~n~l de~rad~tion of acceptor proteins. Thls inter-pre~ation i~ ~on~stent wi~h earlie~ bioohemical and gene~ic evidenoe that ubiquitination of s~ort-llved pr~te~ns in 8 mam~alian oell is essen~ial for thelr degradation. At the same eim~, all ublquiein-~al fusion proteins other th~n ubiquitin-Pro-~al ar~
rapidly deubiquitinated in vlvo ~Tabl~ 1). T~us, the po~ttran~lational amino-termfil ubiquitinatlon o~
proteins may not be ln~olved in an ini_ial recognl-tion or co~mitment step that dssi~nates pro~e~ns for degradation in ~vo. Whether posttranslational ~mlno-ter~inal ubiquitin~ion (if lt ~ctu~lly occurs ln vlvo) is essentlal for lat~r sta~es of the degradAtion pathway remains to be dete~mlned.
Earlier in vltro experimen~s indicated that prei'er-ential chemical modifioation of amino-termini of proteolytic ~ubstrates inhibits t~eir degr~dation in an in vierO ubiqultin-depend~nt proceolytic system.
B~sed ~n these data, it wa~ proposed that amino-ter-minal ~biqultlnation of proteins is e.~sential ~or their tegr~dation. An alternatlve int~rpretation of the ~ame result~ is that chemic~l bLocking o~
protelna' amino ecrmini provenes the rcco~nition of t~eir amino-~erm~n~l rosidue6 b~ the "N end rule"
pathw~y whoso initi~l st~es aro not necessarily ~biquitin dep~ndent.
., ~33~2 Short-llved-B~al pro~eins are multiply ubiguitina~ed in vivo, Ove~expo~ures of the puls~-chase fluorograms (Fl~. 2~ reve~l that the ~a~or band af a deubi-~uitin~ed, short~ ed ~g~l proee~n coexists with a "ladder" of larger molecul~r mas~, ~g~l-concainin~
band~ irregularly sp~ced ~t 4 to 7 kD intervals (Fi~. 4, lanes c to ~ o such lar~er specles appe~r when the fluorograms af long-lived ~gal proteins sre ~lmilarly overexposed (~ig. 4, l~nes ~
and b). Immunological an~lysi.~ with both antibodies to ~gal ~nd ~ntibodies to ublquitin demonsC~ate~
that the "ladder~ ~g~l species ~ontain ubiquitin ~Fig. S).

A model for the sel~cti~e de~r~_~tion_~athw_~.
Wlth th~ exception of n~tur~l or eng1neered ubiquitin fusion prote~ns (Fig. 1 and Table 1), n~scent proteins apparently l~ck ubiquitin moietles.
Th~ in vivo ~mino-t~rmin~l proce~slng of nascen~
noncompartment~lized protelns genera~e4 the~r mature amino-terminl via the action of amino~terminal peptidases ~hose .c~bstrate specificities have been partially characterized. ~See Tsunasawa, S. et ~1.
J . BiO1 . Ckem- j~60 5382 ~19~5~; BO5.~J~1 ~ J ~ P~ ~t 4ll~
PN~S USA 82, ~448 (1985) ) . We sugges~ that the ~mino-term~ni thus generated ~re recognized by an "N-end-re~ding~l enzyme. One ~pecific model is tha~
a c~mmitment ~o degrade ~ prote~n molecule is made ~ ~ 3 ~ ~ L~ ,7J

as a re~ult of the re~ognltio~ o~ its amlno-terminal re~ldue by a ~tochastic~lly oper~ting enzyme whoce probabLli~y of ~clamping" at the target's a~ino-terminu~ $s determined by ~he N-~nd rule ~Table 1~.
Once the co~mi~ment is msde, it i9 followed by hlghly processive ~biquitination of the target protein whi~h in the c~se o~ ~gal lg conjugated to more than 15 ubiquiein moietie~ per molecule o~
tF~g- 4, lsne~ c to g, and F~g. 5~. The ~ul~iply ubiqultinated target protein ls the~ ~egrad~d by 8 "down ~tream" enzyme (1) ~or whlch the ubiquitin moietle~ of ehe targ~t serve ~s oither ~ecognition signals or d~naturstion ~unfoldin~) devices, o~
both The ubiq~ltin-containing "ladder~ ~gal ~pecies (Fig. 4, lane~ ~ to 1, and Fig. 5) con-~ist o~
app~rently br~nohed ubiquitin moeitie~ ~olned ~ the ~-~mino ~roups of Internsl lysine resldues in ~g~l.
Surprisin~ly, thè "ladder" ~gal species ~erived from ubiqultin-Pro-~gal are electrophoretically indisti~-~ui~hable from the ~nalogous ~pecie~ of ~gal who~e ~mino-t~xminal ublquitin i~ ~lea~ed off ehe n~-ccent fuslon protein (Fig. 4, lanes ~ to 1, and Fig. 5).
If ~he electrophore~ically in~istingu~shable ubiqui.
tinated ~gal species are lndeed struo~urally homo logous, thHse re~ults wGuld be compatible with e~o alt-rn~ive mo~ls in whlch, lmmodlacely a~ter th~
flr~t ublquitins are branch-con~ug~ted ~o ~gal, either a branch-ub~quitinated ubiquitin-Pro~,~gal t~ 3 ~ 2 -~3- -undergo~s ~mino-termlnal d~ubiquitlnation or, alterna~ively, ~n an~logous ~gal spocie~ lacking ~he amino-termlnal ubiqui~in moiety reaoqulres it.
Experlmental resolution of thi~ ambigui~y m~y estebli~h whe~her the po~ttr~nsl~tion~l amino-termlnal ubiq~itlnation of proteins tlf i~ ocours in vivo) plays a rolq~ in selective protein turnover, Although both prokaryocic and hukaryo~ic proteins ~ppear ~o follow the N-end rule (see below), bacteri~ apparencly lack the ubiqu~tin system. Thus it is possible that t~e hypothe~c~l N-end-recognlz~ng pro~eln i8 more strongly conserved between prok ryotss and eukaryot~s than is the rest of the selective degradation pathway. Intere~t-in~ly, the properties of ~ m~m~ n prot~in E3 whose presence is req~ired for u~iquitination of proteolytlc sub~tratos by ubi~uiti~-conjugating enzyme~ ln vitro are con~is~ent ~ith it being a component of the N-end-reoogn~zin~ protein.

The_N-end rule ~nd the known amino ~ermini_o~
~ntracellular protein~.
The unblockad amino-termlnal res~dues in metabolic~lly st~ble, no~compartmcntal~zed proteins ~rom both prokaryo~es and eukaryot~s ~re excl-usi~ely (Fig. 6A) of the s~abilizing class (Met, Ser, Ala, Gly, Thr~ , tha~ is, the cla6s that con~ers lon~
In ~ivo hal~ os on ~al ~T~ble 1), The one ~hort~ d intr~celluln~ protein ~or which the 3 3 ~

~64-mature ~mlno-eermlnus i~ known ls the cII protein of phAge lambd~, ~he contral componqnt of a ~r~er that dete~mines whether lambd~ ~rows lytically or lyso~enize~ ~n ~nfected csll . (Y, S, Ho, ~, Wulff, . Ro~enbergl in Re~ulation of ~ene_Ex~ression, I, Booth and C. Higgins, Eds. ~Cambridge ~niv. Press, ~ondon, 1~86), p. 79; F, ~nuett, M,A. Hoyt, L.
McFarlane, H. Echol.~ Her~kowitz, J._~ol ,_Biol, 187, 213 ~1986); M.A. Hoyt, ~.M. Knight, A. Da~, H,I. Miller, H. Echols, ~ell 31, 5~5 ~1~82); K.
Na~myth, Nature (london) 320, 670 (lg~3~), The h~lf-life of cII $n lambda-infected E, eoli is less than 3 m$nuteS. Strikingly, the mature amino-terminus of cI~ ~tarts wlth Ar~ ~HD, Y,W, et al , J
Biol. Chem. 257, gl~8 ~ 2)~, the most desta~lizing residue $n the N-end rule (Table 1).
While the de-~t~bllizing amlno acids can be eithsr hydrophob$c, uncharged hydrophilio or charg-ed, they 6hAre th~ property of having larger radii of gy~ation than any of the s~bilizing ~mino acids except Met ~Table 1). -Amlno-term~_ 1 residue~ in com~artmental~zed_pro~
tein~ ar~ la~elx of_ ho de6~abillzin~ ClA~.
Figur- 6 illu~trato6 a 6trikin~ difference betweon the choice of ~mino-terminal re~idue~ in lon~-lived, noncompartment~l~zed intr~cellular pro~ln~ ~A) and in compArtmental~zed pro~eins, such a~ ~ecreted protei~ (B), ~any o~ whlch ~ru al~o long-lived i~ their respective extr~cellular com-partmene~. One lmplic~tion of thi~ flnding is that a sin~le intracellul~r degradatlon pathw~y operating Accordin~ to the N-end rule could be recponsible both for the dlv~rsity of in vivo half-lives of intra~ellular protein~ ~nd for the selective de-~tr~ction of compart~ntali~ed p~oteins that a~e aberrantly introduced into the intracollul~r space.
Some miscompartmen~llzed protein~ ~ay be more harmful to the cell than othera, It 1J tharefore of interest that secroted euk~ryotic toxlns aon~aln Q~rongly destabilizing re~idues tAr~, Lys, Leu, Phe, Asp) at their a~ino-termini more often than the general pop~lation of secreeed ptoteins (Fig. 6, pan~ls B to D).
The above consider~tiDn also ~uggest that, ii the topological outside o~ a cell, such as lumens of the endoplas~ic reticulum and golgl, ~nd the extr~-cellular Qpaae, wero to h~ve degradation pa~hw~ys analo~u6 to the N~nd rule p~thway, th~y could be b~d on "inverted" ~ersions ~f the N-end r~le in which ~he a~ino-termi~al resid~es tha~ are d~-stabiliz~n~ inside the cell ~re now th~ s~abilizing one6 and vlce ver3a. Thus, th~ methods of the present lnvention should ~130 be useful for manipu-lating the metsbolie stability nd oeher properties of compartmentalized protelns, including secreted ones.

~33~8 ~2 Possable role of ~he N end rule pathw~ in the turnover o~ lon~ ved protei~ns.
Long~lived intracellular proteins with d~abilizlng (Table 1) penultlmaee rea~dues generall~ re~ain th~ir inltial ~mino-termlnal methlon~ne re~idue.
~h~ sm~no-terminal residues In lon~.lived intracel-lul~r protein~ tha~ do undergo ~mino- terminal proces$ing ~re invari~bly of the ~tabillzinR class tTable 1), An interestlng possibillty that would in~olve the N-end ~ul~ pathway in thç ~urnov~r o~
long l~ed prot~ln~ 19 t~st t~e raee~limiting step in the in vivo degradstion of long-li~ed proeelns ~ay be ~ slow ~minopep~lda-~e cleavage ~ha~ expo~e~ a destabilizin~ re~ldue, followed by ra~id degrad~tion via the N-~nd rul~ pathway. Note tha~ fine-~uni~g of the ra~ of de~ra~ation may in this case b~ A
~unoeion o~ the rate of ~minopeptid~se cleav~ge expos~ng a destabili~ln~ resid~e rather th~n ~
funceion of t~e residue's de~tabilizing capacity acco~d~n~ to the ~-end rulc.

~he N-end r~le and_~lective de~radation of ~ort lived and_d~m ~ed proteins.
The reco~nition of polypeptide çhain foldlng pa~terns or of lo~l che~ical ~eatures ~hae target nn otherwise long~ ed but dsma~ed pro~ein for selective degradatlon in ~ivo i~ unl~kely to be medi~t~d direc~ly by the N-~nd r~le pathway.
~n~e~d, ~e sugges~ tha~ specific proteases ~ ~3~ ~2 -~7- -~8nalogou4 in function to nucle~-~es that reoognize sp~cific lesions in DN~ oleave a targe~ed protein so as to expose a d~s~abilizing residue Bt the amlno-ter~-nus of one of the two pro~uct~ of a cut.
On~ testsble prediction of this model is that ~he initial cleh~age produc~s of the deg~adation pathw~y should besr de~tabiliz~ng r~siduer at their N-ter-minl. The preferential exposure of destab~lizlng resldues a~ the a~ino-t~rmini of product~ of th~
inltial proteln clea~ages may be due ~ither to intrinsic ~pecificities of the proteases involved or simply to the fact that a m~Jority of the amino acid~ belong to the dostsb~lizing class (Tnble l).
Fur~hermore, initial cle~vages of a protein would be expected to de~tabilize aspect~ of its ori&i~al conformation, thu~ incr~asin~ th~ probability of further lnternal cuts. Whe~her the initial cle~vage product~ of ~ proteln would be de~raded excluslvely via the N-end rule pa~hw~y or would ha~e ~o be proces~ed fur~her by addit~onsl internal cleavages should depend on sev~ral façtor~, such as the exposure of d~stabiliz~n~ resldues ae the amino-tOrm~n~ of in~tlal clenv~g~ produc~s, And th~
rel~tivo rAte~ of i~e~oduccion of in~ernal cu~s. In thi~ model, ~he N-end rule pathway should be ~s sential for te~radation of most of ~he metabolically un~able proteinc, from ~emically d~aged, pre-maturely terminated, lmproperly ~olded and misco~-partm~talized one~ to ~hoco thst cannot a6se~ble into na~ive ~ultisubu~it ~g~egat~s, and fin~lly to o~herwise normal proteins that are short-lived in vivo. T~us, the ~etabolic inst~bllity of a protein may be medi~ted no~ only by the exposure of a destabili~in~ residue st it~ a~ino-terminUs, but al80 by local conformational and o~emical ~e~tu~es of Its polyp~ptid~ ch~in ~ha~ res~lt ln proteoly~ic eleavagea exposing destabilizin~ residues at the amino.termini of cleavage produo~s.
For any glven protein, a variety of ~actors in adtition to the N-end rule mby combine to modula~e it~ half-life in viva. Among such factors may be the flexib-lity and accessibility of the proteln's a~ino-t-rmin~ ho~n~on, J M. ~nd SibAnda, B.~., J.
Mol. B~o, 167 443 ~1983~), th~ presence o~ ch~ical-ly blocking amino-t~rm~nal groups ~uch as the acetyl gro~pl th~ distribution of ublquitinatable lysine residue-~ near ~he a~ino-terminus, a~d other vAriabl-es, such as the str~c~ure of the carboxy-ter~inus.
Since amino-t~rminal region~ of ~ultlsubu~i~ pro-teins are commonly involved ln t~e interface~
betwe~n sùbunit~ ~Thornton, ~.M and Sibanda, B.L,, 3. Mol. ~io. 1~7 443 (1983)), quarterna~y ~tr~cture of protein~ Ls y~t anothe~ parAmet-r eh~c i~ e~
pected to modulfite th~ impact of the N-end rul~
path~ay on protein half-live-~ in ~ivo. Finally a~
suggested above, the N-end rule path~ay ~Ay al~o be essential for the degradation of protein~ whose ini~ial recogni~lon A~ t~rget~ for de~r~dation Is 1~39~2 independent of the structures at their amino-termini.

Functional si~nificance of posttranslational addi-tion of amino acids to amino-termini of proteins.
05 It has been known for many years that in both bacteria and eukaryotes there exists an unusual class of enzymes, aminoacyl-transfer RNA-protein trans~erases, which catalyze posttranslational conjugation of specific amino acids to the mature amino-termini of acceptor proteins in vitro (R.L.
Soffer, in Transfer RNA:Biolo~ical Aspects, D. Soll, J.N. Abelson, P.R. Schimmel, Eds. (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 1980), p493;C. Deutch, Methods Enzymol. 106, 198(1984): A.
Kaji, H. Kaji, G.D. Novelli, J. Biol. Chem. 240, 1185 (1965)). The posttranslational addition of amino acids to proteins in vivo dramatically ac-celerates in a stressed or regenerating tissue, for example, after physical injury to axons of nerve cells (S. Shyne-Athwal, R.V. Riccio, G. Chakraborty, N.A. Ingolia, Science 231, 603 (1986); N.A. Ingolia et al., J. Neurosci 3, 2463 (1983)). The N-end rule provides an explanation for this phenomenon. We suggest that selective changes in metabolic sta-bility of otherwise undamaged, longlived proteinsthat may be required by a changed physiological state of the cell are brought about by posttransla-tional addition o~ des~abilizing amino acids to the ,. ~

~mino-term-ni of target proteins in vivo. Striking-ly, the known reactions of po~ttransl~tiOnal ~d~i t~on of amino ~cids to proteins ~R.L. Soffer, ln Tr~ns~er RNA:Biolo~ic~l As~ects, D. Soll, J.N.
Abelson, P.R. Sohimmel, Eds. ~Cold Sprin~ Harbor ~bor~tory, Cold Spring Harbor, NY 1980), p493;~.
Deutch, ~ethods Enzymol. 106, 198(1984); A. Ka~i. H.
KaJi, G.D. Novelli, J._Biol. Chem, 240, 1185 ~196$~;
S, Shyne Athwal, R.V. Ricclo, G. Ch~k~aborty, N.A.
In~oli~, Science 231, 603 (1~86); ~.A. Ingolia et al,, J. Neuroscl 3, 2463 (1983)) involve lar~ely those amino acids (Ar~, Lys, ~eu, Phe, and Tyr) th~t Are teQtabilizing ~ceording to the ~-end rule ~Tabl~
1). ~hysiologic~l states in whlch addi~lon of dest~bilizlng amino ac~ds to protein~ could ~e expected to o~cur ~nclude entry to ~nd exit from the cell cycle, r~spon.~es to ch~mic~l ~r physlcal 4tre~s, and speciflc diff~rentla~ion events, such as ery~hroid differenti~ti~n and spermatogeneqls, in which ~ proportion of pre~xis~in~, o~herwise long-ll~Hd intraeellul~r proteins i~ selectlvely de-graded, The in vitro degr~d~tion of some proteolytic subxtrates in a ubiqu~in-dependent system f~om m~mmali~n ~etlculocytes has recently been s~own to depend on the p~e~ence o~ c~ ln ~inoaoyl cRNA~
~Ferber, S. and Ciechanover, A., J. Biol, ~hem. 261 3128 (1986)). We ~u~gest that ~his phcnomenon alQo ro~lect~ a require~en~ for pos~tr~n~lation~l 9~

~ddition o~ speclfio destabili~ing g~ino acids to th~ amino-term~ni of proteolytic substr~tes. The In1tifll prot~olytic sub~rate-~ in que~tion have Amino-t~r~inal re~idues of Asp or Glu, both of which are dest~billz~ng according to the N-end rule (T~ble l~. This rsi~q~ ~n interesting ~nd testable pos~ibiliey thst certain ~mino termin~l res1dues in proteins may not be directly destabili~ing as such bu~ only through their ability to ~e conJ-lg~ted to other destabilizing residues.

Ub~quitin ~usions wlth dihydrofol_te reduct~se In ~ 9et of Ub ~u~ons constructed with mouse dlhydrofolate reductase (DHF~ monomeric -2~-kd protein who~e structure is ~nown at a~omic resolu-tion, the m~ture am~no ter~lnu~ of the "natur~l"
~FR i~ extended by 7 re~ldue~ d~e ~o a oonJ~rUction rou~o taken (Fi~ure 7). After clea~a~e of ~b from the na~cent ubiquitin-DHFR fu~ion p~o~eins in vlvo, the d~ubiquitinated D~FR proteins differ excluslvely a~ their ~ino-t~rm~nal residues, These con~truc-tionq Aro analogous to ehe set o~ ~al ees~ proteins (Figure 3), As expeeted, the DHFR proteina be~ring those amino-terminal residues that ~r~ stabilizlng acoording to ths N-end rule (Table l) are lon~ ed in ye~st (Fig~re 7 and data not shown), Al~hough the presenee of a res~due that i~ de~t~bil~zln6 acoordln~ to tho N -nd rul- Bt the amlno-termlnl o~
an otherw~s~ identical DH~R pro~ein do~s de~tabilize it ~n vivo, ~e extent of dest~bllization is s~all ~Fi~ure 7A) in ~omparl~on to t~e results with ~gal of analogous de~ign ~Table 1). To sddress the m~c~anistic signific~n~e of these finding~, a 40-residue amlno ter~inal re~ion of ~al was positioned upstream of the original D~FR's amino-termfnu~ (Fi~ure 7~. The ~H~R pro~eins bearin~ a de~tabilizing residue ~ollow~t b~ che ~g61 ~erived extenslon ~re approximately as shor~ lived in v1~o as thelr unstable ~gal counterparts, in striking contrase to tho otherwi6e ldentical DH~R proteins that la~k the ~al-speqlfic amlno-terminal ex~ension ~Fig~re 7B and d~ta not shown; c~. ~lgure 7A), Furt~ermor~, ~he extension-bearing ~R proteins that have stabilizin~ re~idues ac their a~lno~
termini are long~ ed ~n vivo (~lgure ~B~. This latter result proves ~hat the ~gal-spccific ex-t ension ~ *uch t in th~ absence o~ a d e s ~ ab I 1 ~ z i n a~lno ter~inal residue, does no~ oonf~r a short hal~life on DHFR, These find~ngs also indicate ~h~t the reason for the striking ~ifference between ~alf-llves of t~e DHFR's t~at either lack or contain the ~g~l-6pecific extension (and bear id~ntical, dest~bilizln~ ~ino~erminal residues) ls d~e to differen~e~ in amino-eermlnal tar~eting elements in these protein~ and not to differences between t~e overall structure~ of ~HFR and ~gal, ~ 3 3 ~ 2 W~en DHFR is fitted with a 26-re~idue, ~gal-terlved amlno-terminal ex~enslon instead of the ori~nal 40-resldue extension, t~e dependenoe of ~he ln vi~o hal~ e of the resulting protein on the nature of i~9 amino-ter~inal residue iB intermedia~e botween thnt of the origlnal DH~R ~nd th~t of the D~R bearing a 36-resldue ~al-dorived exten~ion (F~ure 7C; cf. Figur~ 7B), Thu~, the sequences req~ired ~or the ef~ect of the origin~l ~gal-~pecific exten~ion ~re not ¢on~ine~ to a short stretch within ~he extension but are distribuced over t~e length of ~he extension. These insigh~s indlca~e that the complete amlno-termlnal ~egrada tion signal contains a di~tinct deter~inant addi e~onal to the determin~nt represented by ~he a~ino-termln~l amino a~id re~idue. To addres$ the n~t~re of the ~econd determinant in ~rea~er d~tail, a number of otherwi~e identical D~F~-based pro~eins bearing different vsriants of the ~gal-deri~ed extension ant elther ~ stabilizing or a de-st~bilizing amino-t~r~inal residue were expre~sed in th~ yeG~t S. core~isiae ~nd th~ir h~lf-llve~
determined (~igure 8~. The first conclusion from the data 6hown in Figure 8 i~ that the tWo ly~ ine tK) re8~ues present in the ~gPl extRnsion, although ~y themsolves they do not render the protein ~etabolically unstable, are absol~tely essen~ial for conferring sensitivi~y to the N-end r~le upon the t~9~ p~o~e~n. ~nt~ed, ~hile the con~er~ion of just ~ ~3~3~

,~ .

one Gf the two lysine residueR into ~ si~ilarly char~ed arginine (R) re~idue st~ll results ln ~
prot~in who~ hal~ life is ~ 6tron~ functlon of its amino-termlnal residue ~truatures II-IV in Figure 8), the conversion of both lysine residues in~o argi~ine residues resul~ in a long-lived eesS
pro~ein whose half-life is ecsentially insensisitive ~b the nature of its amino-term1n~1 residu~
(Structur~ V in Figure 8). At the same time, lysins residues are the only amino ~cid resid~-s in p~o-tein~ th~t c~n bo post~ranslationally Joined to the c~rboxy-t~rmlnu~ of ubiquitin, wi~h the formatlon of b~anched ubiquitin-prot~ln conju~te6 Strikingly, our direct determination of the positions of u41quitin moieties in multiply ubiquitin~ted, short- l$v~d proteins of the ~ype shown in Figure 8 h~s shown that all of the multiple ubiqultin ~oietie~ attached to a g ven molecule of the ~est protein r~ide in branched Ub-Ub structures attach~d to one of the two lysine re~idue that h~ve been 1 dentifled above by gen~t~c me-thods as being es-seneial compon~nt.~ of tho comple~e amino-terminsl degrad~tlon signal, Uhst thsn distlnguishes the sbove lysins re~idues (Structuro~ IV) f~om the numerous oth~r lysine re~idues in the rest of ~he D~F~ ~est p~ot~ln~ A al~e to the unlque role of the lysin~ residues ~ tho s~cond determin~n~ of tho amino-t~rminsl de~rsdation slgn~l is provlded by the ~ 3~'12 fact that, due ~o the de~ign of ~he origin~l ex pres~ion vec~or used in our work ~ee Fi~ure 1), Our ~al te~t proteins bear a 45-rH~idue amino-~erminal Hxten6ion deri~ed from an intern_l sequenee of the lac repre~or encoded by the lacI gene. Thus the "~al-derlved" amino-terminal extenslon di~cu~sed above ~Fig~e~ 7 and B) 1~ derlved not from th~
amlno-terminsl oequence of th~ wild- type ~gAl but fro~ an unrela~ed s~quence presen~ at the amino-ter~lni o~ our ~gal eest proteins, It is likely that tho lac repressor-specific extension at the amino-termini of ~hese ~gals i8 more dlsortere~
(~e~mcnt~lly mobile) th~n ~he ~mino-t-rminal r~ion of the wild-type ~al. If so, this exten~ion, while not metabollcally dest~billzln~ ~gal as such, could allow the obser~ed extr~me dependenee o~ the ~gal'~
hal~-life on the nature of its amlno-terminal reffid~ ~Table 1), and thereby, in hlnd~lght, could have greatly facili~ated the discovery of th~ ~ ~nd rule. Tho disordered (segm~ntally mobile) s~a~e o~
thH ~g~l exten~ion p~ovides an explAna~ion for the unique nature of lysine residue~ within the ~x-tenslon vorsus th- ly~ine r~sidue4 in the ~pscially orderod PHFR portion of the tcs~ protein ~Fig~re 8~.
Thus, th~ si~plest interpretation of this ant ~elated evidence 1~ thnt the co~plete amino~ter~nal degratation ~ignal co~prlscs not one b~t twc di~-tinct teter~ln~nts, each of whlch ic necessary but ~y itself not ~ufflcien~ to render the p~otein ~3~2 m~eaboliçally unstable, One dete~minant, described in the fi~t part of thi-c application, i~ the protein' 5 amino-~e~minal re~idue. The s~cond determinant, tescribed imm~dia~ely above, is a ~pe~ifi~ ineern~l ly~ine residue. A~ indicated by ~he da~- of Flgure 8 and the consideration~ above, th~ ability qf thls critlcal lysine residue to serve as the se40nd dete~minant 1~ to a signifi~nt degree independent of the unique amino acid sequenoqs sur.
rounting the lyeine resid~e. Inaeead, an essential feature o~ the crit~c~l lysine r~sid~e includes its spatlal proximlty to the protein'~ amino-~erminuS.

Eq~ivalHn~s Thoae ~killed in the ~rt will recogniz~, or be able to a~certai~ uaing no more than routine experi-mentation, many equi~lents to ~he specific embodi-m~nts of the invon~ion desoribed herein. Such equiv~lents are intended to be enoompassed by the followlng clalms,

Claims (10)

1. A method of regulating the stability of a protein of interest in a reticulocyte extract, comprising the steps of:
a) selecting an amino acid residue from either a destabilizing class or a stabilizing class of amino acid residues as defined in a reticulocyte extract, wherein the destabilizing class of amino acid residues consists of arginine, lysine, histidine, aspartic acid, glutamine, glutamic acid, cysteine and asparagine and the stabilizing residues consist of glycine, proline, isoleucine, valine and methionine, and b) engineering the N-terminus of the protein of interest such that the amino acid selected in step a) becomes the N-terminal amino acid thereof.

2. A method of claim 1, wherein the amino-terminus of the protein of interest is engineered by:
a) producing the protein of interest as a fusion protein containing a masking protein fused to the N-terminus of the protein of interest, and b) removing the masking protein by proteolytic cleavage to expose the selected amino acid residue as the N-terminal amino acid of the protein of interest.

3. A method of producing a protein of interest of predetermined stability in a reticulocyte lysate comprising the steps of:
a) selecting an amino acid residue from either a destabilizing class or a stabilizing class of amino acid residues as defined in a reticulocyte extract, wherein the destabilizing class of amino acid residues consists of arginine, lysine, histidine, phenylalanine, tyrosine leucine, tryptophan, alanine, serine, threonine, aspartic acid, glutamine, glutamic acid, cysteine and asparagine and the stabilizing class of amino acid residues consists of glycine, proline, isoleucine, valine and methionine;
b) providing a DNA construct comprising DNA
which encodes the protein of interest having the amino acid selected in step a) as the N-terminal amino acid thereof; and c) expressing the DNA construct to produce the protein of interest having the amino acid selected in step a) as the N-terminal amino acid thereof.

4. A method according to claim 3, wherein the DNA
construct encodes a fusion protein comprising:
a) a gene encoding the protein of interest, the gene having a sequence at its 5' end encoding the selected amino acid; and b) DNA encoding a masking protein linked to the

5' end of the gene encoding the protein of interest.

5. A method of regulating the stability of a protein of interest in a mammalian cell, comprising the steps of:
a) selecting an amino acid residue from either a destabilizing class or a stabilizing class of amino acid residues wherein the destabilizing class of amino acid residues consists of arginine, lysine, histidine, phenylalanine, leucine, tryptophan, aspartic acid, glutamine, glutamic acid, tyrosine and asparagine and the stabilizing residues consist of glycine, valine, proline and methionine, and b) engineering the N-terminus of the protein of interest such that the amino acid selected in step a) becomes the N-terminal amino acid thereof.

6. A method of claim 5, wherein the amino-terminus of the protein of interest is engineered by:
a) producing the protein of interest as a fusion protein containing a masking protein fused to the N-terminus of the protein of interest, and b) removing the masking protein by proteolytic cleavage to expose the selected amino acid residue as the N-terminal amino acid residue as the N-terminal amino acid of the protein of interest.

7. A method of producing a protein of interest of predetermined stability in a mammalian cell, comprising the steps of:
a) selecting an amino acid residue from either a destabilizing class or a stabilizing class of amino acid residues wherein the destabilizing class of amino acid residues consists of arginine, lysine, histidine, phenylalanine, leucine, tryptophan, aspartic acid, glutamine, glutamic acid, tyrosine and asparagine and the stabilizing residues consist of glycine, valine, proline and methionine, and b) providing a DNA construct comprising DNA
which encodes the protein of interest having the amino acid selected in step a) as the N-terminal amino acid thereof; and c) expressing the DNA construct in a mammalian cell or lysate to produce the protein of interest having the amino acid selected in step a) as the N-terminal amino acid thereof.

8. A method according to claim 7, wherein the DNA
construct encodes a fusion protein comprising:
a) a gene encoding the protein of interest, the gene having a sequence at its 5' end encoding the selected amino acid; and b) DNA encoding a masking protein linked to the 5' end of the gene encoding the protein of interest.

9. A method according to claim 2, 4, 6 or 8, wherein the masking protein consists of ubiquitin.

10. A method according to claim 2 or 6, wherein the masking protein is removed with a ubiquitin specific protease.