J. Mol. Hiol. (1992) 226, 651-660

Functional

Structures of the RecA Protein Found by Chimera Analysis

Tomoko Department

Ogawat,

Akira

Shinohara,

Hideyuki

Ogawa

of Biology, Faculty of Science, Osaka University Toyonaka, Osaka 560, Ja,pan

and Jun-ichi National

Institute Shizuoka-ken

(Received 8 November

Tomizawa

of Genetics,

Mishima,

4 11, Japan

1991: accepted 17 March

1992)

We developed a novel genetic method for finding functional regions of a protein by t,hr analysis of chimeras formed between homologous proteins. Sets of chimeric genes were made by intramolecular homologous recombination in a linearized plasmid DNA carrying bot,h recA genes of Escherichia coli and Pseudomonas aeruginosa. A recBCsbcA strain of E. coli was used for isolation of plasmids carrying recombinants between these genes. Examination of properties of E. coli st’rains deleting the recA gene and carrying a plasmid with a chimeric gene shows that chimera formation at certain positions inactivates a RecA function. Frequently, all chimeras with a junction in a certain region of the protein inactivate a function. Rather than a direct effect of the presence of the junction at a particular position. mismatching of the regions both sides of the junction that are derived from the different. species is responsible for the inactivation. For a chimeric protein to be functional, certain pairs of sequences in different regions of the protein must derive from the same parent. Four pairs of such sequences were found: two are involved in activities for genetic recombination and for resistance to ultraviolet light irradiation and the others in formation of active oligomers. Regions defined by these sequences are located in the looped regions of the protein. A pair of regions may (Lo-operate t,o form a functional folded structure.

Keywords: RecA protein; construction

Pseudomonas RecA protein:

of chimeric

genes: oligomerization

1. Introduction A protein must generally assume a specific folded structure to be functional. Functional structures could be produced both by a continuous stretch of amino acid residues and by the folded structures formed by co-operation of separate regions. Significant insight into functional structures of a protein may be obtained by studying properties of genetic variants. For example, mutations that inactivate protein function have been used to identify a functional region of a protein and the position of a subsequent suppressor mutation that restores the t Author to whom all correspondence should be addrrssrd.

chimeric protein; of RecA protein

fun&ion has served to reveal the location of regions that act tQgether for a function (Hecsht & Sker, 1985). However. it, is, of course, not always possible to suppress one mutation by another; and, even if a suppressor mutation is obtained. the suppression need not result, from alteration in a region separated from the region where the original mutation is present. These difficulties severely limit mutational analysis as a systematic method for identification of separate regions of a protein acting together for a particular function. We describe here a novel genetic method to find regions that participate in a specific function of a protein. The method depends on analysis of chimeric proteins produced as a result of legitimate rectombination between two genes that express

Line 1 Chimera

junction

Activity

t

tYt

t

t

+

+-+

+

++++

t

t

t

t +

The K,ets.4 protein Line 2 Chimera Activity

func~tiorral junction t

t‘itttt'tt

+

+--

t _

_

-++

+

Figure 1. Xctivit,~ of various chimeric proteins. Eitcil linr indicates a port,mn of a protein. The presence (+) OI ;tt)srnt.r (-) of an activity of each t~hirnrra with a thimrra-junction shown by an arrow is indit~ated. SW thr trxt for further explanation.

rtGdut’h

(Horii

t*arrying

out

tion

and

A’IY’.

SOS

proteins

(for

reviews.

oligorrieric~ t~ontlitions

19X8;

Stasiak

6ut.h

a t.ornplex

vontain

s;(v

Ii\,

ht~rnoltqq~ns il,blP

tionally

to

diffrrrlit

ional

R (*o\.

Prrbparatitm

t ht.

(A’1 ‘t)

: Fig.

2).

regions of

to

tlifftbrt-rtt

hrtu-tv~~r

I?. voli

(/‘.-I

SC’l)uriItf’

vf /I/.. I!W).

is likt~l>~ fi)r

proteins

BI’P tlrscrihrtl

2. Materials

physic)

pr,ot,eirr

of’

PA0

;ISSIIII~~Y ISgut~tri

ROW

t.himerits

svvrral

1979;

rvsponsihk~

prot.t>ins

nrrc~yinostr

analysis

I!N!:

I !W:

protein

making

co-operatv. their

Moi~~it.

‘I’htx

I9XX:

01 hr,r

(‘o\.

under

rr~gitmw

map

&

nl..

III \\.it h

antI

forms

multi-fur1t.t

Rjec.4

I’srztdottlontrs

assot+ic~s

1991~).

K- E:gelman.

I!#()).

l)NAs

I,ittIt~

4

xt.itl

rtv~c)rnl~iri;r-

‘t

(Ogawa

sprGfit*

timt%ons.

[)rottain

Ih~il

Kowaltzvkowski.

xrd

tht3

~1 ~1..

I

rnult

irmirrri

irit~lntlr

dou1~le-st~~~rr~tlrtl

I!4XX:

various

3.52

Snnt~rr whit~h

irrdut~tiori.

co/i i- a

of

19X0:

filnctions.

arid

Rndding.

i)FLPlI

of’ Eschcrichirr t~onsistinp

P/ ~1.. its

single

logit~al

homologous proteins. If a chimeritz protein is inactivt> for a particular function, it implies that t,htb parts of t’he chimerit: protein on either side of t,he tahirnera junction are incompatible in the formation of a functional structure. Formation of a junction t~uld abolish activity either by disrupting a lt~al region, or by preventing proper folding due Tao an improper combinat)ion of separate regions. Suppose onct isolates many chimeras with a junction at various positions and found chimeras that are not acativr. If the location of a junction is closer to an end of the protein, the structure of the chimera is ever closer to that of one of the parental prot,eins. both of which arp functional. Therefore, we are likely to encounter a situation in which the presence of a ,junction between point A and point I3 inactivates the function, but t,hat of a junction occsurring out’side of these points does not (Fig. I). If points A atld K are quite close to each other (Line 1). t)htl inat>tivaGng junct,ion can be considered as a mutat’ional alteration of a local structure. If the interval between points A and R is large (Line 2), either t.hc region between them is involved in the function or two separate regions located at, e&her end of the int,erval co-operate to form a functional structure.

protrirr

\vv

that

t hrst,

antI hi~vt, t’utrt~

t*hirlltlr;ts

br4tm

and Methods

AKI I.57 (Houartl-Flanders & Koyt~. IWX) with MI Hf’r strain that was construc+rtl by mating of KM1104 t’arryI!4771 with Ht’r KT,lfj (I,oM.. ing the rwAA gene (JlcEntetJ. 1965). Tt was ustd as a host for ii plasmid t~ontaining ii c.himrrica rrrrl gents. ,1(‘865!lrr(Hf’,sttc~ (Cillrn VI c/l.. IUXI) was used for isolation of plasmids cxrrying a t,himrric,

gene. HfrH5 (Ogawa K: Tonrizawa. the chromosomt~ abilit,y.

kb.

7 Abbreviations IO3 basr-pair(s);

used

for

t~sarnining

used: Ef’. /C coli: u.v.. ultraviolrt.

1967) was t)hr tlortt~r of thtb

/‘.,I

r.t~c~ornl,in~lt~iorr

I’. r/rr~~qinoscr:

Figure 2. Comparison of amino acid sequences of RecA proteins of E. coli and I’. neruyinosa PAO. The amino avid sequences of the EC and PA RecA proteins are aligned to give maximum homology. Thr regions where amino avid sequences are identical are shown in boxes. The positions of amino acid residues. shown by the numbers of residues f’rom t,he N terminus of the EC’ RecA protein, are used as the common co-ordinates. Brackets below the amino at%1 seyurnc~ indicate the critical residues and sequences described in the tcaxt.

Chimera

Analysis

of Structures

(a)

of RecA

Protein

653

(cl

(b)

Figure

3. Construction of plasmids carrying ehimerit: genes. The structures of plasmids used for const.ruction of the EC recA gexle or its chimeric genes are schematically shown at the top of the Figure. A filled arrow indicates fragment. An open arrow shows the PA recA gene or its fragment. An arrow indicates the direction of transcription. An open arrowhead shows the lac operator-promoter sequence. The short linker sequence is present between the two recA genes. The sites of cleavage by restriction enzymes are indicated. The product of cleavage in (a) the linker sequence or (b) Plasmids carrying various types of chimeras that are and (c) in both genes of the plasmid shown above is presented. made by homologous recombination at different positions are established by transformation of the r~cBCshc.4 bacteria. See Materials and Methods for details.

(h)

Cnnstruction,

(f plasmids used for isolation recA genes

of chimeric

The process of construction of a plasmid carrying a chimeric gene began by isolation of a plasmid carrying the EC or PA rec.4 gene. A plasmid containing the E. coli recA gene was clonstructed as follows. First, the DNA fragment that carries the coding region of the EC’ recA gene from position +4 to + 1142 (according to the coordinates of Horii rt al., 1980) was obtained by digestion with BamHI and XbaI of pTH453 (Horii et al.. 1992). The fragment was inserted between XbaI and lYamH1 sites in the polylinker sequence of pUC19 to construct pEC19. in which the recA gene is transcribed from the lac promoter. Similarly. for construction of a plasmid containing the PA reeA gene. the whole coding region of the PA recA gene from position 145 to 1378 (according to the co-ordinates of Kano & Kageyama. 1987) was obtained by partial digestion of plasmid pMC2101 (Sano 6 Kageyama, 1987) with NruI followed by a complete digestion with HindIII. The fragment was inserted between SmaI and Hind111 sites in the polylinker sequence of pUC18 to make plasmid pAK610. To construct a plasmid carrying both EC and PA recA genes, pECl9 and pAK610 were cleaved at the ScaI site in the /I-lactamase gene and at the 8~1 site in the linker sequence. Then the fragments that contained the recA gene from pEC19 (2.8 kb) and those from pAK610 (2.1 kb) were ligated. Plasmid pAK650 (Fig. 3(a)), thus made, contained both EC and PA recA genes downstream from the Zac promoter--operator in direct repeat in this order. Similarly, t,he 2.9 kb fragments from pAK610 and the 2.0 kb fragments from pEC19 were obtained after digestion with AcaI and Hind111 and ligated. Plasmid pAK651 (Fig. 3(a) and (b)), thus made, carries the two lac operator-promoter regions, PA recA gene and EC’ recA gene. in t,his order. Plasmid pTO853 (Fig. 3), which carries both a chimeric gene and thr native EC recA gene, was construrted as

follows. First. the 3.7 kb fragments were obtained by cleavage of pAK651 at’ the ScaI site in the /?-lactamase gene and the RgZTT site in the PA reed genr. Then the fragments were ligated with the 1.9 kb fragment.s similarly obtained from the plasmid caarrying thr gene for chimrra no. 1 (Fig. 4).

((.) Isolation of plasmids carrying a chimrric gene and determination of the sites of chiwwra jun.ctions ,4 total of 3 plasmids, pAK650, pAK651 and pT0853, were cleaved with a restriction endonuclease at a unique site in the linker sequence or appropriate sites in the two recA genes to make defined overlapping regions in the resulting fragment. Then recRCsbcA bacteria were transformed for ampicillin resistance by the linear DNA. Most plasmids thus obt,ained were found to be in oligomeric forms. In addition, some oligomers contained 2 or 3 different kinds of chimeric recA genes. To isolate a plasmid carrying a single monomeric chimeric gene, the alkaline mini-preparation of plasmid DSA was linearized by cleaving at a single site with a restriction enzyme. After extensive dilution, the cleaved D?JA was self-ligated and used to transform AR1 157 rrcA” for ampicillin resistance. To locate the approximate sit’e of a chimera junction, plasmid DKA was cleaved by each of several restriction enzymes and the lengths of fragments were determined by agarose gel electrophoresis. Then the exact site was located by the det,ermination of nucleotide sequences.

(d) ICxamination

of thP

abilities for recombination a.~). resistancr

and for

The bacterial ability for recombination was judged from the efficiency of formation of recombinant,s after mating with an Hfr strain. AB1157 recAA barteria carry ing a plasmid with a chimeric gene grown to 5 x lo8

The rxknt of sensitivity to u.v. irratliatiorr \viis rsanillrtl fi)r rac~h strain of transformants yt’own to .5X IO* c~rlls/ml in Lbroth. After irradiation with U.V. light. t ht, t’rac+iorr of surviving tvlls was tiet~rrminvcl. ‘I’hr ability of N c~himrric~ protrin to form misrtl oligomers fi it)h the, wilt1 t v1w Rrc~A protein \vas t~xaminrcl by nirasurillg tht. ahilitit3 for u.v. rrsistancr and for rt,t.orul)irlatiorl of thtl wiltltypv bacteria (AI.

‘I’h

amount

measuring srprated

of

Rw.4

the density

protjrin

of

hy elrctrophoresis

was

rstimattd

I))

the hand of’ ItrcA prokin in SL)S/pol!‘“‘.r.Vlamnitlr grls.

l’rotrin from a lysate of 10’ bac%rrial c~rlls was used for rat-h det,ermination. The density of the protein band wits

tlrterrnined

by a Laser

Scanner

(MIWlO=\.

Toyobo

(‘0.)

after staining with (‘oomassk brilliant blur It al., 1987). However, we added the criticall) improvement using recB( ‘&cd important of bacteria for transformation under t)hr conditions that) ensure recombination to be legitimate as described below. One group of’ chimeric recil genes was obtained from plasmid pAK650. which contains both fiF:(’ and /‘A rec.4 genes downstream from the operatorpromoter region of t,he Zac operon. The rrcA genes are located in this order wit.h a short linker sequence between them (Fig. 3(a)). The presence of the Zac operator-promoter region makes the expression of thtl rrcA genes independent, of SOS indut6on. Thtl plasmid DNA was t,leaved in the linker sequent’e and used for transformat,ion of rrcN’.stxA-l bac%eria. All transformants werp found to t*ontain a plasm mid(s) with a chimeric rrcA gene(s) (Materials and Methods). The approximate locat,ion of t,he chimera ,junction in eatzh plasmid was determined I),I.

\Vr iLlSO UliltlP chinit~r~;ih with t \I 0 jrrlrc.1 iottb I0 t~saniint~ thtl tirrrt3iorrs of’ ir p;rrtit~ulitr~ rty$olr rot’ t trfb prot,ein, =\ yronp of t~hinit~rns was rriatlc~ t’rcrnr ;I plasmitl pTOX.i:I that t*olrtains t htl g~rrt~ for ant’ 01’ the 1 c~trinirrit* [)rott~ins (t~hirrrerir 1111. I ilr Fig. l(a)) atid the Iiativtl EC’ rrr.-1 gent’ (Alat~~rials plasniitl L\.;IS c.lrnvrtl hot tr iri t tic> and Mrt,trotls). I’,1 rrgiori of thts t~tiirrrt~rit~ gt~rrt~ alit1 in the EC’ gtbIr(a and ust4 for tr~nrrsforlrri~tiorr (Fig. :$(c’)). ‘I’htk fjosit,iorr of the rita\vly formed ,junt*t ioil ot wc*tl of’K3 ~~tlinlt’r’iis was tl~~tt~rnrirrt~tl. Thus. f)lasrnitls that sptsc.if:v I (i different kinds of t*hilrrt>r;rs. c~allt~d ‘I’yprs I II tahimeras. wert’ obtainetl (Fig. -C(t,)). The N-tt~rrnirr;\l 53 residur~s of t tresr c4iimt~ras art’ of t hr, E(’ tJ.f,tl. Anot hf~r grol1p of’ c.h~rnt~rws wit tr rnultiplt~ junt~tions t.onsisls of c*tiirnerws nob 53 and 5-C (Fig. 4((a)). prtspart’ a plasniitl c,;rrr.Tirig t*himtlrw no. 53, thtb 2.1 kb /+‘.spI fragment c)tlt;trn~~tl by frart ial digestion of (~trirut~ra no. I I iLlltl t hr, I .X klr Fspl f’rngrrlr~rlt K1ittlt~ I)?- c~ornplf~tt~ tligSc\st ion 0t’ chimera IIO. 1X \vt’rt’ ,jc)intttl. (‘hinrtkr;r Iro. .?-1 LV:IS t’r;rgnrc~rr( prepartltl 1,). joining t trfL I,.? kl) f‘rc ,111 t*himcra Ii,,. I7 ;r1itl thtl 2.5 klr f’rag,rrrit~lrt fro 111 chimrra no. 53. I)oth obtained by cYrnrpl4~1f~ tliythstion with /~JcoRl ant1 Scrrl. Examination of t)he nut*lrotitie srqupntat’h arountl the tahilrrtxra ,junctions of thr tot,al of Ii:! c*himttras revealed that’ rt~t~ombinatiorr had always occurrt~tl at tsorrrsponding poxit.ions in idfantical regions. \vit haul any dt~letion or irisert.icjn. \l’c found t,hal thtl t+K t&ivy of rfwrmlrination in the ovtdappirrg rrgit~rr was rotrptrty proportional to tht. Icngth 01’ t hcs I75 to I200 fQlsf~-l)iiirs. region _ in 1 llfs range Iwtkvrf~ll All t~hinrt~rah c~s;rnri~ic~tl \\tlrt’ rnatitb Itygitilrrattl [)J I~f~t~trnll)ina,tiorr. F’or rrtwmhinitt ioii in twR(‘.u/~~~~l bacteria to Iw efCt+nt. t tit, ovrrlapping rrgion must t~oritai~r an tIntI of t htl rnolct~iitt~. \YhrLn t tr(s Itxrrgth 01’

rryfjt,

nr

(pt)cs I \.I

nj

1 T I234-

Type I chimera Positionsof junctions 100 200 300 I : I ; I I ““R?C 58 ‘+ + 76 : I+ + 85

1 113

1

19

29

:

:

:

:

21r

1 I 3733-

76:

:

.'+'

22'

,++:

t?-

*++

150

10. 11 B 12.

25"'

a+

a9-

I ..

:..2"

136 7-

159

+

26"'

I++

167 '170

:

:

:178

:

28' 20 ' L3P

228 13, 235

14.

j 256

:

18.

7 ‘1

:

:I16

:

:129

:

:

159

:

162

:

167

266 294

:+* ..... .. .

+' + ___........

:

:

:

:

:

116

4ov

..

424344 ,/“-45”“~ 4647-

I

; 129: , : 150 ,--: 159 , : 167 I

i170

:

I , I ,

:

_ _ . ... ... ... ... . + + 226 -+ 235 -+

+ +

..... 31'

: .

:

,113

33,

. . . . +..+.,

16. 17-

:

325

251

15-

110

271

'++ ... ... ....~

:

-

31

4,129;

Type IU chimera Positionsof junctions 100 200 300 I’ ; I 1 I “Y F&c 56 Id2 ; ++ 110

++

20

'+ + . . . . . _I *++:

:

Type II chimera Positionsof junctions 100 1 200 300 I 1’: : I I 7 “Y Ret : : 25 ++

:

34,

+

+

*,., 35

r

8+

+

36

I

:

:

:

:

:

:

189 -++

:

226 -++ 235 -++: 251 -+ 266 -+

/

505,-

: : : : :

52: + : ___________, +

: , : , : . :

264 -+ 294 -+ 306 -+

+ + +

(cl

Type IY chimera Positionsof lundions

(b’

170 53170

251 -++ 251256 t+ +

54-

idi

Figure 4. Stru(%ures and properties of’ various chimeric genes. Series of chirntaras are each arranged a,cc.ortiing to the positions of’ their t,himera junctions. Filled and open bars show portions of E(~’ and P4 proteins. rrsprc.tively. The position of the junction is assigned as described in the t,ext and shown for each chimera. For chimera no. 5,4. the junction made by the in &TO techniyue is located at, amino acid residue 266. hut non-identicaal amino acids around t,hr region are only at posikms 25 1 to 253 and 256. The abilities of each chimrara for ~1.17.resistance and for rc,c~omhinatiorl that are tlesrrihed in the text and present,ed in Figs ii and 6 are summarized in this Figure. 4 + indicates the presence of activity similar to thr h’:(’ or PA recA gene and - shows a practical ahsrnc*ta of thr ability. Properties of c,himrras Lvith the signs + * and -- * art’ t~xplained in the text. the overlapping region became shorter. the eficdiency of recombination decreased greatly and the proportion of illegitimat,e recombination became significant. For example. when a region 34 nucleotides long was overlapping, about 70To of recombinational events were illegitimate (T. Ttoh. personal c.ommllnication). A certain minimum length appears to be required for efficient legitimate recombination. When wild-tvpe bact’eria instead of r~cRCshcA bacteria were used as the recipients of transformation. the efficicncv of transformation was very much lower. and in addition. a large fraction of t,he plasrnids obtained n-as formed by illegitimate recomillegitimak bination. The prominence of recombination in wild-type bacteria, between two homologous genes in a linearized plasmid I>KA has been reported (Mizuno ef rrl., 1987). An advantage of using rrcH(‘Cs6c.-l bacteria for formation of monornrricb molecules from linearized dimeric molecules hv intramolecular recombination has been report)ed (Syminyton it 01.. 10X5).

The rrcA gent’s of the EC and PA strains are 70 (& identicaal in their nucleotide sequences and the proteins they specaif), are 74’1,, ident’ical in their amino acid residue sequences. To examine functional similarity between the tC:C’ and Pil RecA proteins, the

plasmid carrying the EC’ or 1’A WC=I gene. whose transcription starts from t)he Inc-promoter was introduced into AR1 157rrcA* lncf +. \I:e found that u.v. sensitivit,B- of the bacteria carrying the EC’ protein was about &S-fold more resistant than that carrying the I’.4 protein down to a t&&on of surviving cells of about 10°;I (est~imat~rtl from t)he slopes of’ survival curves), but when the surviving frac.t ions were less than 1 “%Cjtheir srnsit ivit’ies were almost identical. Tn the case of recombination ability, estimated from the efhicienq. of the Hfr bacteria in yielding lru+ rrc~ombinants. the rpc.JA bacteria carqing the plasmitl-specitied 1’L1 protein were about twice as active as those carrying the plastnid-specified EC’ protein. The properties of the rrc:l* skain ca,rrying the plasmid with the Et’ rw,-l gene were similar to those of the wild-t vpr E. co/i cells without a plasmid. ln the c~lassi‘fka~tiorr of chimeras, we ignore a few-fold dif%rences in the extent of u.v. resist.ance and in the effic~ienc~- of rr~c~ombinat~iorr. In spite of t)he functional similarities between bac.trria carrying the plasmid-specified ICC’ or I’.4 rrr,-1 gene, the cbellular concentration of the plasmidspecsified EF:(’ RecA prot,ein was itt)out 1.3) times and F’,1 RrcA protein about 20 times as high as that of t,he Et’ RecA protein specified I)?; the E. roii chromosome (data not shown). The concentrations of all Type I Type TIT and Type IV chinreric pro-

concentration of’ t hc, t’lssmid-sl)rc~itirc~ 1’. I I:~~(~.\ protein is due to greater instability 01 lb01Ii lli(~ protein anti thfx mRSA (data not shower). Ttrt~ results descritxxl above shop. that t he c.c)nc*rllt rat ion of a plasmid-specified I&CA prot& in h;tc+t~ria IS apparently In CX(Y:SSfor rx~xrssion of the wil&typr~ phenotypes. Therefore, t I-IV properties of’ l)i\(*tc‘rii\ probahl,v reflrct the charact~c~ristic~s rather t hwn t I)(. c*oncentrat,ioll of’ the Ret;\ protrin I hat t.hc\ (~itrr~..

IO-

I I

I 2

I I ,23,26,39;40, I 3

recA I 4

u v dose (J/m’)

Figure 5. Fract,ions of surviving bacterial cells after U.V. irradiation. Bacteria (rrcA”) with a plasmid carrying one of the representative rec4 genes were irradiated. Fractions of surviving cells are plotted against the II.V. dose. The number for each curve refers to the chimera number assigned in Fig. 4. Control values for bacteria clarrying plasmids with either the nat~ive EC or PA WCA gene are shown by broken lines.

ta:ins were similar to EC’ RecA protein and proteins were similar fkd PA RecA protein

that of thr ptasmid-specitird those of all Type 11 chimeric to that, of t,he ptasmid-speciin E. coli. The relatively tow

The linear structures of series of’ (*himeras arty presented in Figure 4. The position of a chimera junction is defined as that of thtl first difkrent amino acid residue after t.hr identical region whertb the recombination occ*urrrd. tGac*h c~tiitTlrrit was examined for the cl?ct,ent, of’ U.V. rt~sistanct~ of thfb rrc.4 A hnc%eria c-arrying a plasmid with a c*hirnrric, gent’. Tht> r~ults ohtainr>d with st~vt~rat strains carrying represent,ati\-cl c+imeras arc showrr in Ti‘igurc~ 5. in \vhich the fractions of t)ac,trriat c.rlts surviving afier irradiation with various doses of’ 11.v. tight art’ prrsrnted. S~~vt~rat cbhimeric. t)rottBins ga\-cl slopes of the survival (‘urves that wcrt’ not VW\ different, from the EC’ or I’d RecA protein: t ht,se arc* classified as having activity for U.V. rrlsistanc*e. White czhimera no. 24 is f’airty acntivcx (dt+nated + *). chimeras 110s 2% and 3X art’ almost ttt~fcyt,ivc~ with some remaining activity (designated *) ant1 several ot’hers arc (YHnpletety drft>ctivch. ‘T’htl remaining cahimrras werc~ Icsted for their ability to mediak 11.~. resistanc~c~ aftrr irradiat.ion wit I-I a single dosc~ of 11.v. light (Fig. 6). Each was found to he either almost as act ivc> as the nat ivr, f)rot tbill or c*omptrt,ely defectirc.

Types Ill and El

Type I

0

50

100

150 200 250

Positrons of chlmero 10)

300 350

junctions

0

50

100 150 200 250 300350

PosItIons

of chlmero ib)

junctions

0

50

100

Positions

150 200 250 300 350 of chlmero

]uncrlons

Cc)

Figure 6. U.V. resistance and recombination efficiency of various chimeras. For each strain, (0) the fraction of surviving cells after irradiation with 3.4 *J/m2 of U.V. light and (0) the recombination frequency relative, to that supported by the wild-type EC recA gene were determined as described in Materials and Methods. The numhrrs in thr Figure are those assigned (Fig. 4) for several chimeras of particular interest.

Chimua

Analysis

of Structures

The ability to carry out. recombination in mating with Hfr bacteria was examined for recipient. bacteria carrying each chimeric protein. The results presented in Figure 6 show that most chimeric proteins are either as active as the native protein or cbompletely defective. A small number of chimeric proteins have abilities intermediate between these exkemes. We cla.ssify chimeras nos 22 and 38 as partially active ( + *) and chimeras nos 23 and 39 as partially defe&ve ( - *). (d)

t’attPrn.r

of’ loss of artivities

among

oJ’ Krcil

Protein

657

Chimeras NO.

251

15 -

16(b) 24’ 25’

“Y Ret ,+ + - _

167 170

co) 1o11~

110

u 167

256 u 2512f.z

.

_

+ +

chimeras

The phenotypes with respect to the U.V. resistance cc) 2,, 3’ + + and recombinat,ion ability of various chimeric proteins are summarized in Figure 4 to facilitate cornparison of their properties. All Type T chimeras that contain a junction between residues 165 and 251 (both exclusive) were found t’o be defective in both (d) Z.~ 58 102 37 -+ + abilities (Fig. 4(a)). In progressing from chimera 1 L 1-Y Ollgomer no. 10 to chimera no. 11, we found loss of the RecA fomation 58 102 functions. Because these chimeras are different only -+ + + (e) 3758 110 at residue I67 (see Fig. 2), the residue must be responsible for the difference in their entirely properGs. On the other hand, we found the gain of the RecA functions in progressing from chimera no. 16 to chimera no. 16. The differencta in the sequence at residues 251. 252 and 253 is responsible for t,he gain of the functions. A sequence that is Figure 7. C’ritical residues and sequences and bases of responsible for transitSion in a property t,hat is found thtbir assignments. Properties of srlect,rcl cahimrras of by analysis of R series of chimera is named a critical particular interest are shown. (‘ritical residues and sequence. In Figure 7. we list) the positions and the sec~rr~~~z: are shown by brackets. The structures and bases of assignments of these a.nd other critkal propvrtirs of chimeras shown are srkvtivv rrproduct,ion of t,hoscsshown in Fig. 4. The capacity for oligomer formasequences found by the present study. tion hy tiach chimtira is also shown (e). ,111chimeras with The results described above show that the pret,hr weld-type R,e(,A activities are assumed to be proficient sence of a junction betwcsen residues 167 and 251 in oligomer formation. The data shown in the last 1 of Type I chimeras. inac+ivat)es the functions columns (LM and 1X19)are on t,he deletion proteins (Horii However, because chimclra no. 53 is functional rt (I/.. 1992) described in the trxt,. A bracaket.. Z-59. is a (Fig. l(d)): the region between these residues need rrgiolt involved in oligomer fnrmation. not to be the EC t’ype sequence. Furthermore. because (-himera no. 54 is func%ional. the region that is necessar? to be the li:(1 type does not extend is of’ the EC’ type, the sequence 31 to 5X has to be of beyond residue 257. Tn other words, when the region the same type. Therefore, we assign the sequence 31 up t,o residue 1tii is the EC’ type, t’hr EC’ type to 58 and t.he sequence 102 to 105 as critical sequence between residues 251 and 2% is sufficient for the activit,ies. sequences (Fig. i(c)). In addition. the result) that the The results obtained for Type II chimeras are TJ-pe TT chimera (no. 23). with a junction at the 102nd residue, is u.v. sensitive while the corresummarized in Figure 4(b). Tn contrast to Type 1 sponding Type III chimera (no. 37) is U.V. resistant, chimeras. the presence of a junct,ion between shows that the first crit,ical sequence mentioned residues 167 and 2.50 had no effect. Instead. the above ma.y be narrowed down to t.he sequence 31 to presence of a junction in the region between residues 63 (Fig. i(d)). 1%~the analysis of Type 111 chimeras 113 and 167 inactivated funct.ions for both U.V. (Fig. 4(c)), the critical sequence 167 to 170 found by resistance and rcc~ombination. (Critical sequences the above analysis can be defined more precisely as found in this region are sequences 110 to 111 and residue 167 (Fig. 7(b)). 167 to 170 (Figs 4(b) and 7(b)). Tn contrast, the presence of a junction in the region between these carit,ical sequencers does not inactivate Type I (r) I ~Lhihition qf activities of ~wild-typp l&c.4 protein (*himeras. hy chimeric protein The pattern of’ toss of U.V. resistance by formation Because the RecA protein probably carries out, of T,vpe IT chimeras shows t,hat when the sequence most) of’ its functions in oligomeric forms. loss of 31 to 58 is of the I’d t,ype. t,he sequence 102 to 105 some functions in chimeras could be due to the has to be of the same type for an active chimera failure to form oligomers. Previously. we found that (Fig. -C(h)). Similarly when t,he sequence 102 to 105

4. Discussion

39,41 0

I

2

3

dl

u v. dose (J/m’)

Figure 8. Suppression of U.V. resistance of wild-type bacteria by the presence of chimeric RWA protein. Thr plasmid c-arrying a chimeric recA gene producing thra protein of the designated number was introduced into the TWA + bacteria (continuous line) or recAA bacteria (broken line). They were irradiated at various dose levels of U.V. light.

t’he presence of the inactive RecAl protein reduces the extent of U.V. resistance of wild-type bacteria and demonstrated that, the reduction is due to formation of inact’ive mixed oligomers bet’ween the active and inactive RecA proteins (Ogawa et ol., 1979). We examined inhibition of the activity of the chromosome-specified wild-type RecA protein 1)) tto-existence of various Type III chimeric proteins that are defective in the function for u.v. resistancatb. The results presented in Figure 8 show that chimeras nos 38 (position 110) and 39 (position 113) inhibited the wild-type activit’y poorly while chimera no. 40 (position 116) did so very strongly. Other Type III chimeras with a junction between positions 116 and 167 were strong inhibitors (da,ta not shown). Inhibition by co-existence was observed also for the recombination activity. The presence of chimeras nos 38. 39 and 40 reduced the efficiency of recombination of wild-type bacteria to about 50 “5C1. These resu1t.s show 1O”/* and O+l ‘y;,. resJjectively. either t)hat, chimeras nos 38 and 39 form mixed oligomers poorly with the wild type protein or that, they form fairly a,ctive mixed oligomers. The ability to form oligomers was directly examined for chimeras nos 38 and 39 by a gel filtration analysis of the purified proteins in the presence of ATP. These proteins were found to form oligomers poorly (dat,a not shown).

The central region around an~ino acid rvsiducbs 100 t.o 2.50 can be charact,erized t)y t,hrec critical seyuencrs at positions 110 to 11 1. 167 and 251 to 253 (Fig. i(a) and (1))). (‘himera format~iou in the region br%wren posit,ions 1 13 and I67 of’ Type I I chimeras and that in t hv region brtwern positions 170 and 251 of Tyl~ J c*himcras inact~ivat~t~s t hr, functions for u.v. resist anc~ and for rt~(:on1t)iriati(,n. These rthsults could be explained most cbasily if each of t,hesr regions specify a fbc+ion alld t-himera format,ion inac-tivates it. However,. (*hinit-ra form:rtion in the same region dors not inac*tivatr tht> function when the orientation of the part~nt.al t,yprs is reversed (cornparc, Type 1 and 11 chimr~ras). In addition. a 1ar.p part of thr region t)rt.wcvn the pair of the c:rit)ical sequences (‘au be repluctd by a homologous secluencae without, inactivation of t h(J f’nllt,tion. as was found for chimeras nos 53 and 54. Tht~se rrsults t3nnot br explainrd simply 1)~ tlirtv.1 inat+va,tioll of’ the furlotion and strongly i;~vor 1h(a alternative interpretation that rt,gions;. rac~ll including one of’ a pair 01 critical seque~~cw. Iltvbtl 10 match in formation of the functional st ruc+urcL. 111 of’ the cbomparison of’ 1hfs proprrtit~h particular. (-himeras nos 11 and 54 shows that t hv prt’scan, itstllf. t hts cause of inactivation of t.he funcstion. Rather. formation of the mismatched P.-l t.ypr squenct’ Mwet:n positions 2.51 t)o 2% t)y t hr cahirnrra f’ormation iit position 170 is the c~ausc of’ the irlitc~ti\at ion. ‘I’hc>stt result,s show that c*hirnera forrriat~ion (‘All inacativatr a function t)$- creating mismat,ching bet wern ii pair of regiorls. Thus, this genet,ic* analysis provides iniport,ant information on qions of’ it I)roteiri Lhilt act together in formation of a prc)pc74y fi)ldcicl protein. The region around residue I67 seems to l)ts part icularly important, tvxaust: the rrgion itcatS t opt tier, with a region around residue 110 and thal around residue 250. The importance of the region in drt.er,mination of the function and structure of t hr protein is shown by the observation that thtl WC-AI mutation that alters residue 160 (Kawasliinra it nl.. 1984) prevtlnt,s t!he change of protein structurth caused 1)~ binding of ATP (Ogawa f,t al,, 197!): Wabiko uf ccl.. 1983: Rusc.he rt al.. 1985). 7’lle abilities for rrcvmbination and for u.v. resistance> are lost b>- a mut,ation at residue 163 (Mut~~lc~h B

Chimera

Analysis

of Struct~urus

Bryant, 1991). It was reported that t’he protein is susceptible to cleavage by trypsin at a site(s) between residues 150 and 180, and it is more so in the presence of ATP than in its absence (Kobayashi rt ccl., 1987). It seems likely that a long loop predicted from the primary sequence to be formed by residues 145 to 164 (Blanar et al.. 1984) is exposed more to t)rypsin in the presence of ATP. The analysis of the structure of the RecA protein by the chemical titration of cysteintx residues showed that while all three residues at positions 96. 116 and 129 can be t,itrated in the absence of ATP. only residue 116 can be titrat’ed in its presence (Kuramitsu et (II., 1984). Tn ot,her words. the region containing these residues forms different structures in the presence and absence of ATP. Formation of a loop l)y residues 100 to 113 is predicted from the primary sequence (our unpublished results). The loop sequence 251 to 253 is also in a predicted formed by residues 251 to 256 (Klnnar it nl., 1984). (‘lustering of mutations that affect the functions for u.v. resistance and recombination around thesca critical sequences (for reviews. see R0c.a & Cox. 1990: Kowalczykowski, 1991h; our unpublished results) show the importance of these looped regions for the functions of the protein.

The pattern of loss of u.v. resistance by formation of Types 11 and ITT chimeras shows t’hat sequence 31 to 53 and sequence 102 to 105 are a pair of critical sequences (Fig. 7(c) and (d)). (Chimeras w&h a junct,ion in the region between these critical sequences are fbirlv acti \e7 in the recombination function. Relatively high U.V. resistance of these chimeras is probably due to recombinational repair of the damage. i$‘e showed previously that the presence of a RNA protein species t’hat, is deleted for 2% amino a(+1 residues at the S terminus inhibit,s the activity of the wild-type protein, while one deleted for 41 t&dues is partially and one missing 59 residues is completely lacking this activity (Horii rt ul., 199%: Fig. 7(e)). The protein that’ is missing 32 amino acid residues from the N terminus forms oligomers poorly as shown by a gel filtration analysis (S. Kuramitsu. personal communication). These results show that t,he x-terminal region around residues 29 to 5!) is involved in oligorner formation. We a,lso shorn- that, the sequence 102 to 105 and rcasidue 1 1.3 are c.riticAal in oligomer formation. probably forming a srnall functional region 102 to 113 (Fig. i(c)). Therr>fore. oligomer formation may be a fun&ion for which the region containing t)he sequence 31 to 53 and that containing the sequence 102 to 1 13 c&o-operate. Thrbir interaction rventuall! a&& int.~~r-molPc,alar interaction of the RecA proteins. Thtl rff~~~~~ of cahimrra forma.tion in the region on 11.~. resistanc*r and on rrc~ombination may be secondary to thtl effect. on oligomer format,ion and inhibition of the rcicombinat,ion flmc%ion may need more tight inhibition of oligomer forrnat’ion than that of the function for II V. resistance.

oj

RrcA Protein (v)

659

General

comments

R#ecently, the X-ray crystallographic structures of frthe and ADP-bound forms of E. roli RecA protein were largely determined (St,ory et nl.. 1992). The above prediction of looped structures for the regions including critical sequences from the primary sequences turns out to be reasonable. *As described above. t)he structures of the prot,ein around the regions are altered by binding of ATP. Structural flexi bi1it.y of these looped regions mu\- facilitate proper folding of the protein. Tt is not known how the functional structure of the protein is formed by binding of ATP. The analysis described here shows that a pair of critical sequences in a functional protrin is not necessarily derived from the same species. In other words, some combinations of sryuencrs from different species allow formation of functional proteins. This may explain how members of protein fatnilies with different combinations of critical sequences can evolve from a common ancestor without involving formation of a non-fun&onal intjermediate. One of the paired srqurnc*cs c~ultl catlange to a compatible sequence in a process of evolution. and then a new combination c~ould evolve by further change in the ot)her sequer1c.e. Such a pro(*t’ss could be selectively nearly neutral. The present chimera analysis depentls on genes that are similar enough in their nucleotidr sequences to perrnit homologous re~otnbination. but different enough in t,he amino acid stquencses of the prot,cins they specify. It is not known how often these criteria are met by pairs of homologous genes. How~rr, successful isolation of both active atld inactive chimeras in various syst,rms wit bout part icular effort for selecting pairs of penes ~3uld mean that chances for getting useful pairs are not small. In addition. the ever-increasing information on amino acid sequences of proteins expands the possibilit,y of finding alternative pairs. For r~satnple, as of November 1991. the nucleoCde sequen(aes of t 8 difl+rent bacterial recA genes have bwn rrportrd (I>l~H,J,‘~~lRl,,‘C~rnUank Nuclroti& Data Bases) Cj:r t,hank Drs M. Hrenner. Nl. C+rllrrt. I. Hrrskowitz, T. Horli. T. Ttoh, Iv. Kleckner. Y. Sakakibara and R. IVeist)erg for their help in preparation of the manuscript. We acknowledge support from the Hayashi Mrmori;~l Foundation for Female h’atural Scientists. and H (it-ant-in-.Aid from the Ministry of F:drw;rtion. Srirllw at1d (‘ul t11 w of *Iq’“Il. References Rlanar, 11. A.. Knrllrr, I).. (‘lark. A. .I.. Karn. A. E.. (‘ohen. F. E.. Langridge. R. & Kuntz. 1. 1). (1981). I\ tnoclel fbr the core structure of the Esrhrrirhin co/i l