JOURNAL OF BACrERioLOGY, Sept. 1968, p. 768-776Copyright © 1968 American Society for Microbiology

Coresistance to Neomycin and Kanamycin byMutations in an Escherichia coli Locus

that Affects RibosomesD. APIRION AND D. SCHLESSINGER

Department ofMicrobiology, Washington University School ofMedicine, St. Louis, Missouri 63110

Received for publication 24 May 1968

Mutant strains resistant to neomycin or to kanamycin sulfate were isolated fromEscherichia coli K-12. Nine mutants were analyzed; all were resistant to both anti-biotics (about 150 and 100 Ag/ml, respectively), and were designated nek. In themutant strains, the ribosomes are changed from those of the parental strain; forwhen they were used in assays for polypeptide formation directed by polyadenylicacid or polycytidylic acid, coding fidelity in presence of the drugs was increased andinhibition of synthesis by the drugs was lessened. Mating experiments and transduc-tion tests showed that all of the nine nek mutants are either closely linked or allelic,and the nek locus is closely linked to two genes-str (streptomycin) and spc (spec-tinomycin)-known to affect the 30S ribosome. The two nek mutants tested wererecessive to the sensitive, wild-type allele. When the nek mutants were comparedto the parental strain, pleiotropic effects of the nek mutations were observed. Resist-ance to low levels of streptomycin and spectinomycin was increased, whereas resist-ance to chloramphenicol was decreased. Also, the mutants were less able to adaptto high concentrations of lincomycin, and could no longer show phenotypic suppres-sion of an arginine requirement by neomycin or kanamycin. Such pleiotropic effectsare suggested to be the rule for mutations in genes that participate in the biosynthesisof a cellular organelle.

Because there appear to be at least 50 ribosomalproteins in Escherichia coli (9, 20), there is prob-ably a comparable number of structural genes.Though it is not now clear whether any knownmutation falls in a ribosomal structural gene (4),several approaches have permitted the isolationof mutants with altered ribosomes. One approach,exemplified here, involves the selection of mutantsthat are resistant to antibiotics that block proteinsynthesis.Neomycin and kanamycin, like streptomycin,

are aminoglycoside antibiotics. They share withstreptomycin the ability to "cure" (i.e., pheno-typically suppress) certain mutations in strepto-mycin-sensitive strains (10, 22); again, likestreptomycin, they lose this ability in strepto-mycin-resistant strains (5). These observations,as well as a study of an E. coli mutant (very likelya multistep mutant) resistant to kanamycin (17),suggested that mutants resistant to these anti-biotics have modified ribosomes.

Strains resistant to neomycin or to kanamycinhave been isolated; this paper presents the analy-sis of nine such mutants. All were resistant to both

neomycin and kanamycin, and were mapped nearthe streptomycin locus. The strains tested hadribosomes that differed from those of the parentalwild-type strains. These mutants have been desig-nated nek.

MATERIALS AND MFrroDsGenetic techniques and incorporation tests were as

described previously (3, 4), except that the ammoniumacetate wash of ribosomes was omitted.

In crosses between two nek mutants (see Results),the selective markers were chosen in such a way thatthe portions of the chromosome likely to bear thenek mutations would necessarily be included inthe zygotes.The levels of resistance to antibiotics were gauged

by replicating colonies from broth-agar onto the samemedium containing the relevant antibiotics. Selectionof mutants was also carried out on broth-agar platessupplemented with antibiotics.

All antibiotics were commerically obtained, exceptfor spectinomycin sulfate, which was a gift from G. B.Witfield, The Upjohn Co., Kalamazoo, Mich. Strepto-mycin sulfate was obtained from Eli Lilly & Co.,Indianapolis, Ind., neomycin sulfate and lincomycinhydrochloride, from The Upjohn Co.; kanamycin

768

Vol. 96, No. 3Printed in U.S.A.

on February 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CORESISTANCE TO NEOMYCIN AND KANAMYCIN

sulfate, from Bristol Laboratories, Inc., Syracuse,N.Y.; and chloramphenicol, from Parke, Davis & Co.,Detroit, Mich. Different lots of chloramphenicol showbacteriostatic potency that varies by as much as a fac-tor of 4; in general, the pure product, containingno additives, gives more reproducible results. Forall lots, the relative resistance of various strains isthe same. For all the other drugs tested, results wereinvariable from batch to batch.

RESULTSIsolation and characteristics of mutants. E. coli

strains N141 or AB312 (see Table 1) were grownin broth, treated with N-methyl-N'-nitro-N-nitro-soguanidine (nitrosoguanidine), regrown in freshmedium for 4.5 hr, and plated on broth-agarcontaining different levels of neomycin sulfate orkanamycin sulfate or, as a control of the efficiencyof mutagenesis, on broth-agar containing 200,Mg of streptomycin per ml. Wild-type strains,tested by replication, are resistant to about 5Mug of streptomycin per ml, 5 Mug of kanamycin perml, and 20 Mg of neomycin per ml. The highestconcentrations of kanamycin and neomycin onwhich single-step mutants have been isolable are100 and 150 ,ug/ml, respectively. When eitherantibiotic was present at a concentration of 200,ug/ml, no colonies were produced by a number ofcells adequate to yield about 800 streptomycin-resistant mutants. In several trials in which either80 MAg of kanamycin per ml or 150 MAg of neomycinper ml was used, the spontaneous frequency ofmutants resistant to kanamycin or neomycin wasfound to be very low, of the order of 10-1". Thisvalue is equivalent to about one-tenth the fre-quency of mutation to high-level streptomycinresistance, as determined by plating samples of thesame culture on broth-agar plates containing 200Mg of streptomycin per ml (see also 15).

All mutants isolated on either drug are resistantto both drugs, and have been therefore designatednek (for neomycin and kanamycin). The strainsused in this work are listed and described inTables 1-3.

Pleiotropic effects. Comparison of the nekmutants with the parent strain revealed threeeffects of the nek mutations.

(i) Changes in resistance to other antibiotics.Eight nek mutants were tested by replication forresistance to streptomycin and spectinomycin; allshowed an increased resistance to streptomycin,and most also showed an increased resistance tospectinomycin (Table 3). However, the levels ofresistance were much lower than those usuallyobserved for mutants isolated as str-HR or spcr(5, 8, 15; also, cf. 17).The resistance of nek mutants to several other

antibiotics that affect ribosome function was also

checked. The nek mutants were no more resistantto lincomycin or erythromycin than the parentalstrains. However, they were appreciably moresensitive to chloramphenicol: wild-type strainswithstand about 10 Mg of chloramphenicol perml, but nek strains are able to resist less than 10Mug/ml (Table 3). This characteristic of the nekmutants enables one to select a few nek+ cellsfrom among a large population of nek cells. Thiswas successfully done with all three nek mutantstried; the results with one of them, N733, are givenin Table 4. The selection pressure is sufficient topermit the isolation of rare Nek-S (neomycin-kanamycin sensitive) recombinants from a crossoftwo nek strains; the results obtained in one suchcross are given in Table 5. In this cross of nekstrains N743 and N751, portions of the matingmixture were plated either on medium that per-mitted growth of ad+ his+ recombinants, or in thesame medium supplemented with chlorampheni-col. (Comparable numbers of the two strains wereplated independently, but no revertants werefound.) Colonies that grew on the medium con-taining chloramphenicol were streaked out; fivecolonies from each streak were tested for theircharacteristics. Most of the streaks gave rise tonek+ colonies, identified by their wild-type sensi-tivity to neomycin and kanamycin. The frequencyof recombination between the two unselectedmarkers mal and argG was measured among asample of ad+ his+ recombinants (all Nek-R) andamong an equivalent sample of ad+ his+ nek+recombinants (50 colonies in each case). In thefirst instance, the recombination frequency was28%; in the second, it was 37%. It therefore seemsvery likely that the nek+ colonies isolated in thiscross are true recombinants.

(ii) Abolition of phenotypic suppression. In theparental strain N141, the arginine requirementcan be replaced if the medium is supplemented bylow, nonlethal levels of streptomycin, kanamycin,or neomycin (2 Mg/ml; these tests were carriedout as described in reference 5). This phenotypicsuppression (curability) by kanamycin and neo-mycin was lost in four nek mutants (N71 1, N712,N733, and N735), and was partially retained inthe fifth mutant examined (N734). Phenotypicsuppression by streptomycin was retained in threeof these five nek strains, partially retained inN712, and abolished in N735 (see Discussion).

(iii) Lessened adaptability to lincomycin.Another difference between the response of nekstrains and wild-type E. coli was observed in theirresponse to phenotypic adaptation to high levelsof lincomycin (4). Whereas the parental strainN141 adapted, growing up after about 24 to 48hr, the nek mutants adapted only after a longer

VOL. 96, 1968 769

on February 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

770

00U)

APIRION AND SCHLESSINGER

0~~~ ~0 0I: $.4 ,y. >4'O >00 Cd 0

EE < CcE E'oE 3

0 o < o o mmY m

r.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~[1

-,Za ~ ~ ~ -

++++++++ ++I+++ +

+ ++ ~+±±+ ±~~+± +

+ ±+±+ + +++ +

. +II ++

±+ ++±e++ Q+± 0

~ +±+-+±+± ±+ ±.~ I+I++I+ ++

z I+I±I+I++++

~+++++++++ ±++I +

".+ ++r1r1±c1+±II

; f++i±+i±++ ±÷±++ +

"+ ++I++ +

±++III+ +++ +

"SI+++i++++ ++II

0

._

CU)

00 _-4 -.

en -m en en o 8 CeZZ<:<<Z==

ZZm

J. BACTERIOL.

i

1.I

c

-0"zt tttn r- 4, eq,--I r- l--qcn = 0-. u eqZ < .4 -, Z

on February 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CORESISTANCE TO NEOMYCIN AND KANAMYCIN

TABLE 2. Genzotypes of (nek) neomycin- and kanamycin-resistan2t strainsa

ilv pro; sir spc Matingilvpro ~type

+2

2222

++

rr

sssssrsss

s

r

sssssss

ss

HfrF-

F-F-F-F-F-HfrHfrHfrF-

Sourceb

Kan from AB312From cross N709 XN304

Kan from N141Kan from N141Neo from N141Neo from N141Neo from N141Neo from AB312Kan from N2Kan from N2From cross N741 XN1304

a For symbols and parental strains, see Table 1.bThe parental strains resist about 5 Ag of kanamycin per ml and 20 Ag of neomycin per ml. All of the

mutants are resistant, in replication on broth-agar plates, to about 100 /g of kanamycin (Kan) per mland 150,g of neomycin (Neo) per ml, and were initially isolated, after nitrosoguanidine treatment ofthe relevant parental strain, as colonies that grew on plates containing one or the other drug, as in-dicated.

TABLE 3. Response to antibiotics of nek strainsa

Strain nek allele Strepto- Spectino- Chloram-mycin mycin phenicol

N141 nek+ 2-5 30-50 10N709 nek759 > 5,Ooob 30-50 <10N711 nek73, 30-40 30-50 <10N712 nek712 30-40 >200 <10N733 nek7 3 3 30-40 > 200 <10N734 nek734 30-40 50-100 <10N735 nek735 30-40 50-100 <10N737 nek737 >5,000b 50-100 <10N741 nek74, 30-40 50-100 <10

a Tests were performed by replicating coloniesfrom nutrient broth-Casamino Acids-agar ontothe same medium supplemented with variousconcentrations of the antibiotics; plates werescored after 1 and 2 days. The numbers listed arethe range of antibiotic concentration (,mg/ml)at which colony formation is blocked.bThese two strains carry a strr allele.

lag of 3 to 5 days, or, in the case of one strain(N734), could not adapt at all (Table 6).The mutation frequency of N734 to lin (high

resistance to lincomycin; 4) was found to be about10-9, reduced by a factor of 103 in comparisonwith the parental strain N141. This incompatibil-ity of a nek mutation and lin mutations may berelated to the failure of N734 to adapt to highlevels of lincomycin.

Occurrence of altered ribosomes in nek mutants.Ribosomes from nek strains N711 and N733 werecompared with those from the parental nek+

TABLE 4. Reconistruction experiment for selectionof nek+ cells from among nek cellsa

No. of cells Total cells Found amongper plate plated No. of C tap-R

Cap-R colonies testedcolonies

AB774N733 ~~~observed(ABek7+) (nek) AB774 N733 AB774 N733

108 107 324 3 X107 284 70 0

a Two strains were separately titered on mini-mal medium agar supplemented with all the re-quirements necessary for growth. Cells of the twostrains were then mixed and plated together on thesame medium supplemented with 20,Ug of chloram-phenicol per ml, for the selection of nek+ cells.Colonies growing on the plates supplementedwith chloramphenicol were isolated (Cap-R)and streaked out, and some individual coloniesfrom each streak were assayed for their genotype.AB774 was characterized by being Mal+ and Nek-S; N733 was characterized by being Mal-and Nek-R.

strain N141 in their response to artificialmessenger ribonucleic acids (mRNA), includingpolyuridylic acid (poly U), polyadenylic acid(poly A), and polycytidylic acid (poly C). Theresults are summarized in Table 7.

All the experiments were run in duplicate andwere repeated three to five times. The mean ofeach set of experiments with ribosomes from amutant strain was compared to the relevant meanobtained with the ribosomes from the parentalstrain, in the t (Student) test, in order to gauge the

Strain nek lacallele thr leu

+I+

xyl

+

4

111114

gal

2

22222+

N709N710

N711N712N733N734N735N737N741N743N751

mal

141141141141141

++

741709

711712733734735737741743741

arg

3

33333

++

+

G

his

4

44444+++

++

+

+++++

++

+

+

+±

VOL. 96, 1968 771

on February 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

APIRION AND SCHLESSINGER

TABLE 5. Selection of nek+ recombinants in a cross between two nek mutantsa

Recombinants/ml No. of Cap-R N f Nek-SCross nek alleles colonies NolofNiekS

Ad+ His+ Ad+ His+ Cap-R analyzed

N743 X N751 nek743 X nek741 0.51 X 106 2,390 64 49

v The cross was performed by "prolonged mating" (4). Portions were withdrawn, diluted, andplated on minimal medium plates supplemented with L-arginine and L-leucine (Ad+ His+ selection),or in the same medium with the addition of 10 ,ug of chloramphenicol per ml (Ad+ His+ Cap-R selection).Colonies that grew up on plates containing chloramphenicol were streaked out on broth-agar and re-tested for their other characteristics.

TABLE 6. Adaptation of nek mutants to growth inpresence of 1,000 ,Ag of lincomycin per mla

Strain nek allele Lag beforegrowth (days)

N141 nek+ 2N711 nek711 3N712 nek752 3N733 nek733 5N734 nek734 b

N735 nek735 3

a From cultures grown exponentially in broth,about 500 cells were inoculated into 10 ml of brothcontaining 1,000 Mg of lincomycin per ml; growthwas recorded as positive when cultures showedturbidity corresponding to about 5 X 108 cells/ml.Colonies from the adapted cultures, if grown onmedium lacking lincomycin, are once moreindistinguishable from the original strains.

b Incapable of adapting; after 7 days, even theinput cells were no longer recoverable.

probability that the two means represented deter-minations on samples of the same population(Table 7). When this probability was less than0.05, it was concluded that the two means aresignificantly different. The primary conclusionfrom the results presented in Table 7 is that theribosomes of the two nek strains tested are differ-ent from those of the parental strain. (In allexperiments, changing the source of the superna-tant enzymes did not change the results; one suchset of experiments is presented in Table 7.)The results given in Table 7 also indicate that,

when poly U is the mRNA employed, mutantand wild-type ribosomes function with no signif-icant difference. They are equally inhibited bydrugs (Table 7), and also miscode for isoleucine tothe same extent (unpublished data). When poly Ais the messenger used, the ribosomes from nekmutants are significantly more resistant to theinhibition of lysine incorporation induced byneomycin, by kanamycin, or by streptomycin.When poly C is the mRNA, miscoded incorpora-tion of histidine, threonine, or serine induced by

neomycin or kanamycin is reduced in the nekstrains.Formal genetics of mutants. The nek mutations

occur at a chromosomal site very closely linkedto the well-known str and spc loci. This locationwas inferred from crosses of four nek mutants.For example, in the cross N709 X N302 (strr nekX spCr), of 54 SpCr strr recombinants, 38 were nek.Similarly, in the cross N737 X N302 (strr nek XSpCr), among 161 strr spCr recombinants, 147were nek, and of 145 Spcr nek colonies isolated,all were strr. In the cross N53 X N711 (strr spCrX nek), 100 ilv+ thr+ leu+ recombinants wereisolated, and among them the various classes ofrecombinants with respect to drug markersshowed 7% recombination between the str andspc markers, 1% between str and nek markers,and 8% between the spc and nek markers. Asimilar result was obtained in the cross N53 XN735 (strr SpCr X nek).Close linkage of nek to the str and spc loci has

been also confirmed in a series of transductionexperiments. The results of one such experimentwere as follows.When N710 (nek strr spcr mal+) was the donor,

and N141 (nek+ str8 spcs mal-) was the recipient,of 30 nek transductants, all were mal-, 20 werespcs strP, 4 were SpCr strr, 4 were spcs strr, and 2were spCr strs.The close linkage of various nek mutations with

the str and spc loci suggests that they all fall withina very small chromosomal region. Corroborationwas obtained by direct tests in mating experi-ments. Each of the two nek Hfr strains (N709 andN737) was crossed with three different nek F-strains (N711, N733, N735). In each case, 224thr+ leu+ ilv+ arg+ recombinants were selected;among all of these, only a single nek+ recombi-nant was found (in the cross N709 x N735). Asimilar result was obtained with two other nekF- strains, in crosses of N709 X N712 and N737X N734; in each case, 100 recombinants wereanalyzed, and only in the first cross was a singlenek+ recombinant colony detected. Thus, all

772 J. BACTERIOL.

on February 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CORESISTANCE TO NEOMYCIN AND KANAMYCIN 773

z2WI)

0 0

vv't Co

U)~~ ~ ~ ~ ~ ~~)W

eq

W)soON

olNr- I'*cn~ ~ A$-> A A:3m en

o) tnIQ _NC

vv

I mNol- VV-U)C)

0tr eowA

knO v~~t 0omC'

vvo N A A

en~~~N °

vv

__ o~~~r ~ o

.n00 "t --

cn

O

o0

"%o0 1)%C 8c A A_" " C1

!j o) "0oo n

z o

04 ~ ~ 0

_ ClN Cl ~N V

0) 0 0W/

C1 _14 vo

-F)0 ~'/~)

O0N Nt:j

0C1~~~Cl ~~OCC____~~~~~.OCl- ~~~~~-C WI

z- r-'

AA

A A

88Ilt tn lq A A

cno.n-

lt tnvnAA_Ir)OIr)ClN OcrNCl i

C-

W)'I enY >r

w

*--4O

'I 7- en

U)-(- -

+I

N

zz

-4 enN Nt-ZZo.

VOL. 96, 1968

U)uzr

Q

0

S

r..2

c°

3.-

H

r.:D .

._ CAo 9)

u UX0 00.0

0

on February 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

APIRION AND SCHLESSINGER

seven nek mutations analyzed fall in a single locus.Two more nek mutants, N741 and N743, weresubsequently analyzed and shown to be allelic(by the same criteria) with the seven nek mutantsmentioned here.

Test for dominance. To see whether nek+ ornek is dominant, they were put into the same cyto-plasm in a partial diploid. The nek+ allele wasintroduced from strain KLF41 into N711 (nek711)in one experiment, and into N735 (nek735) in asecond. In each case, 2 X 107 cells per ml of eachparental strain were slowly agitated for 5 min at37 C, and then left without further agitation for25 min at 37 C. The mixture was then plated onbasal medium containing maltose, L-isoleucine,L-valine, L-proline, L-arginine, L-histidine, andthiamine to select for colonies with the pheno-type Mal+ Leu+ Met+. These recombinantcolonies include both haploid cells and mero-diploids.Among the Mal+ Leu+ Met+ recombinants,

colonies sensitive to neomycin and kanamycinwere observed; a number were repurified bystreaking on rich medium, and were then grownup in broth and restreaked on eosin methyleneblue-maltose plates. Some of the colonies testedyielded Mal- segregants. A number of Mal+ andMal- colonies were then tested for resistance toneomycin and kanamycin, and colonies resem-bling the female parental strain were found. Froma merozygote isolated from the cross KLF41 XN711, 23 Mal+ Nek-R, 73 Mal- Nek-R, 14 Mal+Nek-S, and 3 Mal- Nek-S segregants were ob-served. Similarly, from a merozygote isolatedfrom cross KLF41 X N735, segregants included 9Mal+ Nek-R, 57 Mal- Nek-R, 6 Mal+ Nek-S,and 1 Mal- Nek-S colonies. Since both nek andnek+ alleles could be recovered in segregants ofNek-S partial diploids, it is concluded that thenek+ allele is dominant.

DISCUSSIONThe new locus described is located near the str

locus of E. coli and affects ribosomes. Thephenotypic characteristics of mutants in this locussuggest a general feature for mutations that affectribosomes: pleiotropy. Here we discuss some ofthe effects of the mutations.

Effects of neomycin and kanamycin on proteins.ynthesis in the mutants and the parental strain. Aswith most of the other known mutations thataffect E. coli ribosomes, the in vitro differencebetween nek+ and nek strains in amino acid incor-poration studies are relatively small, whereas invivo differences between these strains are striking(cf. 4). Also, as observed for mutants to strepto-mycin resistance, in vitro tests with homopoly-mers exhibit sharper differences in miscoding,

rather than in susceptibility to inhibition by thedrugs (Table 7). The binding site for streptomycinis probably distinct from that for neomycin andkanamycin, since the mutations to resistance tothe drugs fall in genetically distinct loci. However,the two sites must either overlap physically, or elsea change in one site must distort the other, since,for example, mutations to streptomycin resistanceabolish phenotypic suppression by all three drugs.Also, nek mutations confer limited resistance tostreptomycin as well (Table 3).

In some instances, the effects in vitro parallelthose in vivo. For example, streptomycin onlypartially inhibits polyphenylalanine synthesis onribosome from a sensitive strain, even at very highdrug concentrations, and this inhibition is totallyalleviated in resistant strains; cultures of str-HRmutants withstand any measurable level of strep-tomycin. In contrast, comparable levels of neo-mycin and kanamycin block polymerization ofphenylalanine and lysine nearly totally both inextracts of nek+ and of nek- strains (Table 7);even the small remaining incorporation can beabolished by higher levels of drug (unpublisheddata), and nek strains acquire only limited resist-ance to either neomycin or kanamycin. Anotherexample of parallelism in vitro and in vivo is pro-vided by the miscoding induced by antibioticswhen the messenger is poly C. While this miscod-ing is significantly reduced in the nek strains N711and N733 in the presence of neomycin or kana-mycin, the miscoding is not significantly reducedin the presence of streptomycin (Table 7). In acomparable way, in vivo in these two strains,neomycin and kanamycin cannot give phenotypicsuppression, but streptomycin does (see Results).However, the parallels in vivo and in vitro may befortuitous, as it is not clear to what extent inhibi-tion or miscoding in protein synthesis directed bya homopolymer is a relevant index for the sensitiv-ity of cells to drugs.The effects of neomycin and kanamycin are

probably exerted through an effect on 30S ribo-somes, for there is (i) tight linkage of the neklocus to other genes which are known to affectthe 30S ribosomes, the str and spc loci (6-8);(ii) cross-resistance of nek mutants to strepto-mycin; and (iii) similar behavior of neomycin,kanamycin, and streptomycin in phenotypic sup-pression.

Genetic analysis. The crosses and transductiondata support the claim that all of the nine nekmutations tested are closely linked, and that thenek locus is closely linked to the str and spc loci.We conclude that it is less than 1 min from theseloci, assuming that I time unit = 20 recombina-tion units = 10% contransduction frequency.This correlation is constructed from data pub-

774 J. BACTERIOL.

on February 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CORESISTANCE TO NEOMYCIN AND KANAMYCIN

lished previously (12, 19). The relationship isdefinitely adequate, at least to a first approxima-tion, since all of the data obtained in experimentswith other markers in this chromosomal regionare consistent with it. The linkage data presentedin this paper concerning the str, nek, and spc,loci suggest that they are located in this order.However, the data are not conclusive.

Pleiotropic effects. Pleiotropy should be therule rather than the exception for mutations thataffect an organelle such as the ribosome, bothbecause a ribosomal protein must interact with avariety of other proteins and with ribosomalRNA to form the structure, and because the resul-tant structure has to function in the precise andelaborate way required for accurate protein syn-thesis. Thus, the better-known cause for pleio-tropy in bacteria, polar mutations in operons, isnot a prerequisite for pleiotropy caused by muta-tions in genes that affect ribosomes.

In the other reported case in which mutants inthe same locus confer a drastically changed re-sponse to two antibiotics, the lir locus, not all themutants shared the property [sensitivity to eitherlincomycin or erythromycin (4)]. In the presentcase, however, all of the nek mutant strains testedresist both of the antibiotics. This might result ifthe ribosomal binding sites of these two antibi-otics overlap completely, in contrast to those oflincomycin and erythromycin which do not.Alternatively, the nek mutations occur only atthe overlap of the ribosomal binding sites for ne-omycin and kanamycin. The rarity of the nek mu-tations might result from such a limitation. Inagreement with the second alternative is the factthat lower-level resistant mutants have been iso-lated that show increased resistance only to onedrug or the other (unpublished data).The fact that nek strains, unlike the parental

strain, adapt very poorly to high concentrationsof lincomycin is of particular interest. It has beenunclear whether adaptation to lincomycin occursthrough a change in the ribosome, or throughsome other process, such as a decrease in cellpermeability to the drug. Since nek mutationshave no obvious direct effect on cellular compo-nents other than the ribosome, the loss of adapt-ability in nek mutants supports the idea thatadaptation involves a modification in the ribo-some, possibly comparable to the one observedin ribosomes isolated from anaerobic cultures(12). Another possible case of adaptation througha change in the ribosome is the "phenotypicmasking" described by Gorini, Rosset, and Zim-mermann (11).

Perhaps the most interesting and useful ofthe observed pleiotropic effects is the increasedsensitivity of nek mutants to chloramphenicol.

The nek mutants probably affect the 30S ribo-some, but chloramphenicol binds specifically tothe 50S ribosome (21; Monro and Staehelin,personal communication). Thus, a mutation in oneribosomal particle may indirectly affect the func-tion of the other particle.The increased sensitivity of the nek strains to

chloramphenicol permitted a reconstruction ex-periment (Table 4), in which a few nek+ (sensi-tive) cells were selected from among a large num-ber of nek (resistant) cells. This finding openedthe way for the fine genetic analysis of the neklocus and made possible a "two-way selection"(1, 2) for a gene that affects the ribosome. Crossesbetween pairs of nek mutants have been per-formed, and indeed nek+ recombinants have beenobtained (Table 5).

ACKNOWLEDGMENTS

We are indebted to C. Mayuga and J. Henry forskillful assistance.

This investigation was supported by Public HealthService research grants HD-01956 and GM-10447,and by Training Grant 5T-AI-257, from the NationalInstitutes of Health.

LITERATURE CITED

1. Apirion, D. 1962. A general system for the auto-matic selection of auxotrophs from prototrophsand vice versa in micro-organisms. Nature195:959-961.

2. Apirion, D. 1965. The two-way selection of mu-tants and revertants in respect of acetate utiliza-tion and resistance to fluoroacetate in Aspergil-lus nidulans. Genet. Res. 6:317-329.

3. Apirion, D. 1966. Altered ribosomes in a suppres-sor strain of Escherichia coli. J. Mol. Biol.16:285-301.

4. Apirion, D. 1967. Three genes that affect Escher-ichia coli ribosomes. J. Mol. Biol. 30:255-275.

5. Apirion, D., and D. Schlessinger. 1967. The lossof phenotypic suppression in streptomycinresistant mutants. Proc. Natl. Acad. Sci. U.S58:206-212.

6. Cox, E. C., J. R. White, and J. G. Flaks. 1964.Streptomycin action and the ribosome. Proc.Natl. Acad. Sci. U.S. 51:703-709.

7. Davies, J. E. 1964. Studies on the ribosome ofstreptomycin sensitive and resistant strains ofEscherichia coli. Proc. Natl. Acad. Sci. U.S.51:659-664.

8. Davies, J., P. Anderson, and B. D. Davis. 1965.Inhibition of protein synthesis by spectino-mycin. Science 149:1096-1098.

9. Fogel, S., and P. S. Sypherd. 1968. Chemicalbasis for heterogeneity of ribosomal proteins.Proc. Natl. Acad. Sci. U.S. 59:1329-1336.

10. Gorini, L., and E. Kataja. 1965. Suppressionactivated by streptomycin and related anti-biotics in drug sensitive strains. Biochem.Biophys. Res. Commun. 18:656-663.

775VOL. 96, 1968

on February 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

APIRION AND SCHLESSINGER

11. Gorini, L., R. Rosset, and R. A. Zimmermann.1967. Phenotypic masking and streptomycindependence. Science 157:1314-1317.

12. Hayes, W. 1964. The genetics of bacteria and theirviruses, p. 592. John Wiley & Sons, New York.

13. Kwan, C. N., D. Apirion, and D. Schlessinger.1968. Anaerobiosis induced changes in anisoleucyl-tRNA and the 50 S ribosomes ofEscherichia coli. Biochemistry 7:427-433.

14. Maas, W. K., J. M. Wiame, and N. Glansdorff.1964. Studies on the mechanism of repression ofarginine biosynthesis in Escherichia coli. I.

Dominance of repressibility in zygotes. J. Mol.Biol. 8:359-364.

15. Newcombe, H. E., and R. Hawirko. 1949.Spontaneous mutation to streptomycin resis-tance and dependence in Escherichia coli. J. Bac-teriol. 57:565-571.

16. Silengo, L., D. Schlessinger, G. Mangiarotti, andD. Apirion. 1967. Induction of mutations tostreptomycin and spectinomycin resistance inEscherichia coli by N-methyl-N-nitro-N-nitro-soguanidine and acridine half-mustard ICR-191. Mutation Res. 4:701-703.

17. Tanaka, N., K. Sashikata, T. Nishimura, and H.Umezawa. 1964. Activity of ribosomes fromkanamycin-resistant E. coli. Biochem. Biophys.Res. Commun. 16:216-220.

18. Taylor, A. L., and M. S. Thoman. 1964. Thegenetic map of Escherichia coli. Genetics 50:659-677.

19. Taylor, A. L., and C. D. Trotter. 1967. Revisedlinkage map of Escherichia coli. Bacteriol. Rev.31:332-353.

20. Traut, R. R., P. B. Moore, H. Delius, H. Noller,and A. Tissieres. 1967. Ribosomal proteins ofEscherichia coli. I. Demonstration of differentprimary structures. Proc. Natl. Acad. Sci. U.S.57:1294-1301.

21. Vasquez, D. 1966. Binding of chloramphenicolto ribosomes, the effect of a number of anti-biotics. Biochim. Biophys. Acta 144:277-288.

22. Whitfield, H. J., Jr., R. G. Martin, and B. N.Ames. 1966. Classification of aminotransferase(C Gene) mutants in the histidine operon. J.Mol. Biol. 21:335-355.

776 J. BACTERIOL.

on February 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from