JOURNAL OF BACTERIOLOGY, Dec. 1978, p. 994-1001 Vol. 120021-9193/78/0136-0994$02.00/0Copyright © 1978 American Society for Microbiology Printed in

Kasugamycin-Dependent Mutants of Escherichia coliERIC R. DABBS

Max-Planck-Institut fiir Molekulare Genetik, Abteilung Wittmann, Berlin-Dahlem, Germany

Received for publication 27 September 1978

Kasugamycin-dependent mutants have been isolated from Escherichia coli B.They were obtained through mutagenesis with ethyl methane sulfonate or nitro-soguanidine in conjunction with an antibiotic underlay technique. In the case ofnitrosoguanidine, dependent mutants were obtained at a frequency of about 3%of survivors growing up in the selection. In the case of ethyl methane sulfonate,the corresponding value was 1%. Nineteen mutants showing a kasugamycin-dependent phenotype were studied. In terms of response to various temperaturesand antibiotic concentrations, they were very heterogeneous, although most fellinto two general classes. Genetic analysis indicated that in at least some cases,the kasugamycin-dependent phenotype was the product of two mutations. Two-dimensional gel electropherograms revealed alterations in the ribosomal proteinsof seven mutants. One mutant had an alteration in protein S13, and one had analteration in protein L14. Three showed changes in protein S9. Each of twomutants had changes in two proteins, S18 and Lii. Three of these mutantsadditionally had protein S18 occurring in a partly altered, partly unaltered form.

36, No. 3

U.S.A.

The use of antibiotics has been a major ap-proach to obtain mutationally altered ribosomes.Ribosomally targeted antibiotics which havebeen studied in some detail include, among oth-ers, streptomycin and related antibiotics (e.g.,bluensomycin and dihydrostreptomycin), spec-tinomycin, and kasugamycin. Mutants of Esch-erichia coli resistant to streptomycin have beenshown to have an alteration in small-subunitprotein S12 (18). Mutants resistant to spectino-mycin have been found to be altered in proteinS5 (3). Distinct from these is kasugamycin,where resistance has been shown to be due to achange in the RNA moiety rather than in theprotein moiety of the ribosome (11). When an-tibiotic-dependent mutants are obtained, theycan give further insight conceming functionalinteractions between ribosomal componentssince from them antibiotic-independent second-ary mutants can be obtained which, in the casesstudied, are mutants with changes in a compo-nent(s) other than that originally altered. Thebest known case involves dependence on strep-tomycin, where the lesion due to the protein S12alteration is relieved by alterations in protein S4or S5 (10).Mutants dependent on other antibiotics, such

as spectinomycin or kasugamycin, are not ob-tained when cells of E. coli are spread on anti-biotic-containing plates and incubated. In thecase of streptomycin, when the underlay tech-nique (2) is used for bringing the cells into con-tact with antibiotic more gradually, 90 to 95% of

994

survivors are streptomycin dependent, versus 20to 50% when direct spreading on antibiotic-con-taining plates is used. The technique has beentried with another antibiotic, and the successfulisolation of spectinomycin-dependent (Spcd)mutants of E. coli has recently been reported(5).Mutants of E. coli resistant to kasugamycin

have lost activity of a methylase that acts ontwo adjacent adenine residues near the 3' ter-minus of the 16S RNA (11). This part of theRNA is essential to ribosomal function (4) andis near a functional site of the 30S subunit duringinitiation complex formation (22). Kasugamycinspecifically inhibits initiation (20). The factorIF-3 (23) and ribosomal proteins Si (15) and S18and S21 (21) are in the vicinity of this part ofthe 16S RNA. As a further probe of this ribo-somal neighborhood, it seemed worthwhile toattempt to isolate kasugamycin-dependent(Kgd) mutants.Ksgd mutants have in fact been obtained as 3

to 4% of the survivors under appropriate condi-tions. The properties of these mutants are de-scribed in this paper.

MATERIALS AND METHODSBacterial strains. Strain L44 (E. coli B argF leu

from L. Gorini) and its derivatives were used through-out this work. LR was a spontaneous rifampin-resist-ant (Rifr) mutant of L44. LT was a thr derivative ofL44 obtained by ethyl methane sulfonate (EMS) mu-tagenesis and penicillin selection. LTK was a sponta-

on August 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

KASUGAMYCIN-DEPENDENT MUTANTS 995

neous kasugamycin-resistant (Ksgr) mutant of LT.Several F- strains of E. coli K-12 were also used in

an attempt to isolate Ksgd mutants, since geneticanalysis is easier in such strains than in E. coli Bstrains. Such attempts were unsuccessful.

Media. The media used were as described in refer-ence 6. Rich medium used in the preparation of P1lysates, and for transductions, contained 3 g of NaClper liter instead of buffering salts, since these saltsincluded sodium citrate.EMS mutagenesis was carried out by washing ex-

ponentially growing cells of strain L44 in buffer, sus-pending them in buffer at a concentration of about 5x 10O cells per ml in the presence of 1% EMS (vol/vol),and incubating them at 37°C for 60 min. Afterwards,cells were washed in buffer and then used in theselection as described below. N-methyl-N'-nitro-N-ni-trosoguanidine (NTG) mutagenesis was done by wash-ing exponentially growing cells in tris(hydroxy-methyl)aminomethane-maleate buffer at pH 6.2 (1),suspending them at a concentration of about 5 x 108in the same buffer, and then adding 0.1 volume of asolution of 2 mg of NTG per ml to give a final concen-tration of 200 ,ug of NTG per ml. The suspension wasincubated for 10 min at 370C, after which it was cooledin an ice-water slurry. Cells were washed twice in pH7.0 buffer and then used in selection. EMS was ob-tained from Sigma Chemical Co., and NTG was fromEGA Chemie KG. Kasugamycin sulfate was obtainedfrom Boehringer Mannheim GmbH.

Selection of Ksgd mutants. A variation of thestreptomycin underlay technique (2) was used inwhich kasugamycin was employed instead of strepto-mycin. Cells suspended at about 5 x 108/ml in buffer,either with or without prior mutagenesis, were spreadin a series of dilutions on rich plates. A 0.1-ml portionof suspension was spread per plate, and the plateswere incubated for 3 h at 37°C. A 0.15-ml amount ofa stock solution of kasugamycin containing 200 mg ofantibiotic per ml was introduced under the agar, usinga sterile spatula. The plates were stored for 36 to 48 hat 4°C to allow diffusion of the antibiotic. Since platesheld about 30 ml of medium, the final kasugamycinconcentration in the plates was about 1 mg/ml. Theplates were returned to 37°C and incubated for 4 to 5days. Colonies appearing were patched onto platescontaining 1 mg of kasugamycin per ml, incubated at37°C for 2 days, and then streaked for single colonieson plates also containing 1 mg of kasugamycin per ml.A large and a small colony from the streak of eachmutant was selected and spotted on plates containing0, 50, 100, 400, and 1,000 ug of kasugamycin per ml.These plates were incubated at 30, 37, and 42°C.Putative Ksgd mutants, designated MV1 throughMV35, were restreaked and spot-tested again.

Transduction. Transductions were carried outwith phage Plvir essentially according to the methodof Lennox (16). For strains with a generation time ofabout 20 min, plate lysates were incubated for 6 h at37°C. For slower growing strains, plates were incu-bated proportionally longer according to their gener-ation time. Afterwards, plates were selected which hadcomplete lysis but the lowest concentration of inputphage. P1 lysates of Ksgd strains were dialyzed against2 x 50 volumes of medium to remove antibiotic. Intro-

duction of antibiotic resistance markers into a strainrequired preincubation of recipient cells to allow phe-notypic expression before they were exposed to theantibiotic. Recipients with growth rates of about 20min were preincubated for 2 h, this time being pro-longed proportional to the generation time. Antibioticwas then introduced under the agar, and a procedurethe same as that described above was followed.Ribosomal proteins. Mutants were grown in 40

ml of rich medium to stationary phase. Ribosomalproteins were prepared from frozen cells as describedin reference 7. Samples of 0.6 to 1.0 mg of protein wererun on a scaled-down version of the two-dimensionalpolyacrylamide gel system described by Kaltschmidtand Wittmann (14). The first dimension was 9.5 cm,and the second was 9.0 cm. Both dimensions were runat 4°C for 18 h at 80 V and 16 h at 80 V, respectively.

RESULTSIsolation of mutants. Selection of Ksgd mu-

tants is shown in Table 1. The best conditionsfor selection of mutants involved an incubationtemperature of 37 or 420C, with NTG as muta-gen. Attempts to isolate Ksgd mutants sponta-neously were unsuccessful. EMS mutagenesiscould be used but was less effective than NTGmutagenesis. Further selections have been sub-sequently done; they also resulted in about 1%Ksgd mutants for EMS and between 3 and 4%for NTG.Phenotype ofmutants. Mutants were tested

for their antibiotic response at various kasuga-mycin concentrations and various temperatures(30, 37, and 42°C). The L44 parent strain wasable to grow at 100 jig, but not at 300 ,ug, ofkasugamycin per ml at all three temperatures.Most mutants fell into one of two classes. In

mutants of class I (mutants MV 3, 7, 8, 11, 12,21, 22, and 27), the maximnum and minimumconcentrations of kasugamycin at which cellswould grow were independent of temperature.In mutants of class II (mutants MV 4, 10, 15, 16,17, 34, and 35), these parameters correlatedstrongly with temperature. Class II mutantsgrew at 30°C in the absence of kasugamycin butwere inhibited by the presence of 300 ,ug ofkasugamycin per ml. At 420C, they required 200,ug of kasugamycin per ml to grow, and they alsogrew on 1,000 ug/ml. The situation at 370Cvaried from mutant to mutant. Some mutantswere Ksgr at this temperature, and others wereKsgd. Even within each class, the mutants wereheterogeneous in terms of exact response totemperature and antibiotic concentration.Four mutants could not be placed in either

class I or class II. Mutants MV2 and MV29 grewat neither 30 nor 420C, whatever the kasuga-mycin concentration. Mutant MV1 grew only at300C or below 370C and was Ksgd within thisrange. Mutant MV9 was Ksgd at 300C and Ksgr

VOL. 136, 1978

on August 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

996 DABBS

at 420C. For a more precise idea of the growthresponse of Ksgd mutants, the growth rates oftwo mutants were determined at a variety ofantibiotic concentrations (Fig. 1).Ribosomal proteins ofKsg d mutants. The

ribosomal protein patterns of the 19 Ksgd mu-tant strains were analyzed by using two-dixnen-sional polyacrylamide gel electrophoresis (14).Seven mutants showed reproducible alterations.Portions of the electropherograms of the ribo-somal proteins of these mutants are shown inFig. 2. The acidic half of the gels is not shownsince no mutants showed alterations in proteinslocated in this region. The electropherograms ofthe proteins of these mutants can be comparedwith a control 70S pattern in which all theproteins are identified, as shown in reference 12.

Electropherograms of ribosomal proteins of E.coli K-12, as shown in reference 12, are the sameas for E. coli B (of which L44 is a derivative),except for two proteins. Protein S5 is more basicin E. coli B and, hence, comigrates with proteinL6 and not Lll. Protein S7 is smaller in E. coliB and is located on the gel at a position belowthat seen in electropherograms ofribosomal pro-teins of E. coli K-12. Electropherograms of pro-teins from 30S and 50S subunits ofxibosomes ofE. coli B are shown in reference 24. The proteinalterations of mutants MV2 and MV29 wereindistinguishable. The alteration in mutant MV9appeared to be the same as that in mutantMV35. Therefore, mutants MV29 and MV35 arenot illustrated.Mutants MV2 and MV29 had two alterations

TABLE 1. Isolation ofKsgd mutants under varying conditionsPlate incuba- No. of Col-

Pretreatment tion temp niestested Ksgd Mutants %Ksgd(OC) mst

None 37 94 0 0None 42 112 0 0EMS 37 120 2IEMS 42 155 21 MV7, 15, 17, 21 1.3

NTG 30 169 0 0NTG 37 205 71 MV1, 2, 3, 4, 8, 9, 10, 11, 12, 16, 22,NTG 42 202 81 27,29,34,35

a Ksgd refers herein to a mutant having a kasugamycin-dependent phenotype at at least one of the threetemperatures tested. Pooled results from two separate experiments.

u 10 100 1000 10000KASUGAMYCIN CONCENTRATION (pg/ml)

FIG. 1. Growth rate of kasugamycin mutants at different antibiotic concentrations. Growth rate determi-nations were made with cells growing in rich medium at 37°C. The fastest doubling times (all at 37°C) were:for MV10, 65 min; for MV29, 60 min; for MV200, 35min; for L44, 22 min. These values are taken as 100% foreach mutant. Growth rates ofmutants were measured by optical density of cultures at 650 nm. Ksgd mutantswere pregrown in rich medium containing 200 pg of kasugamycin per ml, washed twice in buffer, and thenallowed to grow for 90 min in new medium before readings were taken to allow for adjustment ofendogenousantibiotic level. Symbols: Ksgd strains MV10(0) andMV29 (U); Ksg' mutant MV200 from the same selection(A); L44 (Ksg' parental strain) (0).

J. BACTERIOL.

on August 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

a

AMd

U.=._

'aw.

..

dc

.4

.0

........ 'W

._

A.iik

e

:a

FIG. 2. Portions oftwo-dimensional gel electropherograms of 70S ribosomalproteins ofKsgd mutants ofE.coli. Arrows indicate theproteins altered by mutation; the circle indicates theposition ofthe wild-type protein.(a) MV2 (similar to MV29); (b) MV9 (similar to MV35); (c) MV21; (d) MV27; (e) MV34.

997

OS .i

Sx Q __F-.01A.v

b

*434.:

4:

;...mk

,:.q

.-dddffibw...M

"MM,......

/A

IRW.Iddiml.w

.,Ammk:

IF,

on August 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

998 DABBS

in their ribosomal protein patterns. One differ-ence was in ribosomal small-subunit protein S18and the other was in ribosomal large-subunitprotein Lli (Fig. 2a). These two mutants werederived from the same batch of mutagenizedcells, although they were selected on differentplates. They had a different kasugamycin re-sponse and gave rise to kasugamycin-independ-ent revertants at frequencies which differed bya factor of five; however, it could not be excludedthat they shared a common mutational origin asregards the protein alterations. In mutants MV2and MV29 the Lll spot was no longer visible onthe two-dimensional gels of 70S ribosomal pro-teins. This was because in gels of 70S proteins,the Lll spot was merged with the combinedS5/L6 spot normally located to the right of theLll spot. If 50S proteins alone were subjectedto electrophoresis, then the L1i spot was distinctfrom L6. Ribosomal protein Lii was thereforemore basic in these mutants.The three mutants which had alterations in

ribosomal small-subunit protein S9 had separatemutational origins. Mutant MV21, obtainedthrough EMS mutagenesis, had a more acidicprotein, S9 (Fig. 2c). Mutants MV9 and MV35were both isolated from selections in which thecells had been treated with NTG; however, theywere obtained in separate selections from inde-pendently mutagenized batches of cells. In thecase of mutants MV9 and MV35, the electro-pherograms did not show a shift in the positionof protein S9, but rather its apparent absence(Fig. 2b). It should be pointed out that on two-dimensional polyacrylamide electropherogramsthe spots for proteins S9 and Sii coincide; how-ever, protein Sii occurs as a double spot withthe second one below, and somewhat to theright, of the S9-S11 spot (25).

In mutants MV9 (Fig. 2b) and MV35, thecomposite spot of proteins S9 and Sii-nor-mally very prominent-was much fainter. How-ever, the lower spot of protein Sii was presentas usual, indicating that it was protein S9 thatwas absent. 30S subunit electropherogramsfailed to show an extra spot, excluding the pos-sibility that protein S9 has moved so that itsposition was masked by a 50S subunit protein,nor were any 30S protein spots more stronglystaining. Electropherograms of whole-cell ex-tract showed the relatively faint S9-Sii spotalso, eliminating the possibility that protein S9was more loosely bound and was washed offduring ribosome preparation.There was one difference between the ribo-

somes of mutants MV9 and MV35. Electropher-ograms of the ribosomal proteins of mutantMV35 were similar, whether the ribosomes had

been washed in 500 mM NH4Cl buffer or not.However, when ribosomes of mutant MV9 werewashed in 500 mM NH4Cl buffer there wasextensive change in the 30S subunit ribosomalprotein pattern: protein S14 was undetectable,and proteins S2, 3, 5, 7, 10, 13, and 21 werepresent in much reduced amounts. However,proteins S4, 8, 11, 12, 15, 16, 17, 18, 19, and 20were present in normal amounts. The 50S ribo-somal protein pattern was unaltered. Appar-ently, mutant MV9 harbors a mutation whichconfers considerable instability on the ribosomesof this strain.The type of alteration in protein L14 seen in

mutant MV27 (Fig. 2d) was similar to that ob-served in streptomycin-independent revertantsof a novel streptomycin-dependent (Strd) mu-tant (7; E. Dabbs, Mol. Gen. Genet., in press).The type of alteration in protein 813 in mutantMV34 (Fig. 2e) likewise resembled that obtainedin the system employing a novel Strd mutant.An alteration in protein S13 was the only alter-ation, apart from changes in protein S5, seen inelectropherograms of ribosomal proteins of nineSpcd mutants obtained as described in reference5 (E. Dabbs, unpublished data).

In addition to the changes described above,several Ksgd mutants showed an alteration inprotein S18 different from that seen in mutantsMV2 and MV29 (Fig. 2a). Mutant MV27 (Fig.2d) had wild-type protein S18. In mutants MV9and MV34, and to a lesser extent in mutantMV21 (Fig. 2b, e, and c), protein S18 occurredas a multiple spot.Ksgr mutants obtained in the same selections

as those in which Ksgd mutants were isolatedwere also checked for ribosomal protein altera-tions. Twenty each obtained from EMS- andNTG-treated cells were tested. None of the 40Ksgr mutants checked showed any alteration inribosomal protein pattern. This supports theview that the alterations in ribosomal proteinsseen in the Ksgd mutants were connected withthe Ksgd phenotype.Genetic studies. Since the best characterized

locus for Ksgr mutants is the ksgA locus, locatedbetween thr and leu on the chromosome (19),and since NTG is known to produce closelylinked double mutations (9), an initial screeningwas made for location of the Ksgd mutation inthis region.Phage Pi lysates grown on seven Ksgd mu-

tants (with a variety of phenotypes) were usedas donors of Thr+ to strain LTK in the presenceof 1,000 ,ug of kasugamycin per ml. Thirty to 36transductants were tested for their ability togrow on plates containing varying concentra-tions of antibiotic. No transductants showed the

J. BACTERIOL.

on August 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

KASUGAMYCIN-DEPENDENT MUTANTS 999

Ksgd phenotype at any temperature.Several of the mutants gave transductants

that were heterogeneous in morphology and ka-sugamycin response, however. Since, in the caseof Spcd mutants (5), all mutants analyzedshowed the combination of a spectinomycin-re-sistant (Spcr) mutation and a genetically sepa-rable second mutation conferring dependence,transductions were done to determine if the Ksgrmutation of the ksgA locus was present. P1lysates of six Ksgd mutant strains were used asdonors of Thr+ to strain LT, and the transduc-tants were screened for their response to kasu-gamycin. Mutants MV17 and MV29 gave Ksgrtransductants (resistant to 1,000 ,ug of antibioticper ml versus 100 ,ug/ml for the kasugamycin-sensitive [Ksg5] parent) as 9 and 16% of the total(48 transductants tested per mutant). MutantsMV9, MV10, MV11, and MV21 gave no Ksgrtransductants. Subsequently, it became clearthat although MV9 and MV11 required kasu-gamycin to grow, the maximum kasugamycinconcentration at which they would grow wassimilar to that of the Ksg8 parent. MutantsMV10 and MV21, on the other hand, were ca-pable of growing on 1,000 ,ug of antibiotic per ml.The possibility that there was a lesion locatedbetween thr and ksgA in mutants MV10 andMV21 that prevented cotransduction betweenthese markers could not be excluded on the basisof experiments done in this work.

In an Spcd mutant previously studied (5), thelocus of the mutation causing the Spcd pheno-type was located near rif on the E. coli chro-mosome. Two Ksgd mutants, MV2 and MV29,showed an alteration in protein Lii. Since thegene for protein Lii is near rif (17), whether thealtered protein Lii was responsible for the Ksgdphenotype was determined. Several other Ksgdmutants were also tested at the same time. SixKsgd strains were used as recipients, and an Rifrderivative of strain L44, LR, was used to donatethe Rifr mutation into them. In the case ofmutant strains MV9, MV10, MV11, and MV27,no alteration was found in the kasugamycin phe-notype of the transductants compared with thestrains from which the transductants were de-rived. Mutant MV29, with the altered proteinL11, also kept its Ksgd phenotype when thealtered protein Lii was transduced out but be-came able to grow at 300C as well as at 370C.Only in the case of mutant MV21 were thetransductants predominantly Ksgr rather thanKsgd in phenotype (20 of 20 and 16 of 20 in twodifferent experiments), indicating that the Ksgdphenotype in this mutant involved a mutationnear the rif locus. A lysate of an Rifr transduc-tant of mutant MV21, in which the Ksgd phe-

notype was still present, was used to donate theRif' marker to strain LTK (spontaneous Ksgrderivative of L44). Kasugamycin was present inthe medium. A majority of the transductants (9of 12 and 24 of 40 in two different experiments)were Ksgd. The Ksgd phenotype of the transduc-tants was similar to, but not identical with, thatofmutant MV21. Therefore, the Ksgd phenotypecan be due to the combination of a Ksgr muta-tion and a genetically separable second mutationconferring dependence.Two strains showed alterations in proteins,

S13 and L14, whose genes are located in themain cluster of ribosomal genes (13). To deter-mine if alterations in these proteins were in-volved in the mutant phenotype, transductionswere attempted using a spontaneous strepto-mycin-resistant (Strr) (rpsL) derivative of L44.However, no transductants could be obtainedwhen attempts were made to introduce this mu-tation into Ksgd strains. Possibly this observa-tion was connected with the fact that an attemptto isolate Ksgd mutants from strain Ll-401 (anStrr derivative of strain Li, the strain from whichthe strain L44 used in this work is derived) wasunsuccessful (228 colonies tested). It may bethat an Strr mutation and a Ksgd mutation can-not coexist in a strain.The involvement of the ribosomal protein S9

alterations in the kasugamycin phenotype of thestrains could not be determined since the loca-tion ofthe gene for S9 is not yet known; however,the independent occurrence of such mutationsin three separate selections strongly points to aconnection.

DISCUSSIONKsgd mutants were not difficult to isolate un-

der appropriate conditions. They resembledSpcd mutants and differed from Strd mutants inthat genetic analysis indicated that in at least asizable fraction of the mutants the antibioticphenotype was the product of more than onemutation. The phenotype of Ksgd mutants wasthe result of an antibiotic resistance mutation(which in some cases was likely to be at theksgA locus) together with a further mutation(s)conferring dependence.Two other loci for mutations altering the ka-

sugamycin response of E. coli, ksgB (19) andksgC (27), have been described. Ribosomes frommutants harboring a ksgB mutation do not showkasugamycin resistance in vitro, in contrast towhat is found with ribosomes from mutants har-boring a ksgA mutation (19). It may be thatksgB mutants have a mutation altering thepermeability of cells to kasugamycin, and theyare probably distinct from the mutants with

VOL. 136, 1978

on August 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

1000 DABBS

altered ribosomal proteins described here. TheksgC mutation has been linked to an alterationin the electrophoretic mobility on gels of ribo-somal protein S2; no alterations in protein S2were detected in any of the Ksgd mutants de-scribed here.Dependent mutants arose as 1 to 4% of all

Ksgr strains after mutagenesis. The Ksgr muta-tions present in some strains could be trans-ferred to another genetic background, and theresulting strains grew well. In the two casesamenable to study, the mutation causing analtered protein Lll in mutant MV29 and themutation causing kasugamycin dependence inMV21, the Rifr marker could be used to transferthese lesions back into strain L44. Most Rifrcolonies obtained in such selections grew poorlyand showed faster growing secondary mutantsarising within the colonies. Thus, it might bethat mutations induced in genes for ribosomalcomponents by the mutagenesis were stabilizedby further mutations, for example, to kasuga-mycin or spectinomycin resistance, but thatthese combinations were only viable within acertain rang0 of antibiotic concentrations.

Since in the case of both Spcd and Ksgd mu-tants, mutations near rif could confer depend-ence on a strain containing a resistance allele, itwould be interesting to interchange these mu-tations and see whether they confer dependenceon an antibiotic other than that involved in theirisolation. Since mutations altering the strepto-mycin phenotype of an Strd strain have alsobeen mapped close to rif (6), a cross-comparisonwould also be worthwhile in this case.

It was interesting that several of these Ksgdmutants showed alterations in ribosomal proteinS18. This is a protein that other work (21) hasindicated is near the adenine residues that arethe target of the ksgA methylase. The multipleforms of ribosomal protein S18 in mutants MV9and MV34 may be due to an incomplete acetyla-tion of the N-terminal amino acid (26; K. Isono,personal communication), and the altered S18in mutants MV2 and MV29 may represent anunacetylated S18. Further work is necessary tosee if the S18 alterations are connected with theKsgd phenotype of these mutants.The reason why mutations affecting ribosomal

protein S9 were seen several times among theseKsge mutants will have to await a better under-standing ofthe ribosome. One explanation wouldfollow from the knowledge that the methylase,whose activity is impaired in ksgA mutant cells,acts late in assembly of the ribosome; it cannotuse naked 16S RNA as a substrate (22). It maybe that the ribosomal environment of the 3'-endof the 16S RNA is altered enough by the absence

of protein S9 from the ribosome that the meth-ylase fails to recognize its usual substrate. Itwould be worthwhile to study methylase activityand the degree of methylation of its usual sub-strate in these mutants.

Since several Ksgd mutants showed altera-tions in protein S9 or S18, and no Ksgr mutantsshowed such alterations, it would seem verylikely that the alterations in these proteins wereconnected with the Ksgd phenotype.Kasugamycin-independent secondary mu-

tants could readily be obtained from all Ksgdmutants. A preliminary screening of 20 sponta-neously isolated kasugamycin-independent mu-tants obtained from each of six Ksgd strains wasdone, using two-dimensional polyacrylamide gelelectrophoresis of ribosomal proteins. Severalantibiotic-independent mutants obtained fromstrains MV2 and MV29 showed further changesin protein Lii. This would suggest that thealteration in protein Lii was involved in theKsgd phenotype. Several kasugamycin-inde-pendent mutants isolated from mutant MV17 inseparate selections showed alterations in proteinL23.There were clear indications from these mu-

tants that the phenotype of E. coli mutants withrespect to kasugamycin (whose site of action isthe 30S subunit) was modified by mutationsleading to alterations in components of the 50Ssubunit. This would point to the close functionalinteraction between the subunits.

ACKNOWLEDGMENTSI thank H. G. Wittmann for critically reading the manu-

script. I also thank B. Schroeter for excellent technical assist-ance.

LITERATURE CITED1. Adelberg, E. A., M. Mandel and G. C. C. Chen. 1965.

Optimal conditions for mutagenesis by N-methyl-N'-nitro-N-nitrosoguanidine in Escherichia coli 112. Bio-chem. Biophys. Res. Commun. 18:788-795.

2. Bjare, U., and L. Gorini. 1971. Drug dependence re-versed by a ribosomal ambiguity mutation, ram,- inEscherichia coli. J. Mol. Biol. 57:423-435.

3. Bollen, A., J. Davies, M. Ozaki, and M. Mizushima.1969. Ribosomal protein conferring sensitivity to theantibiotic spectinomycin in Escherichia coli. Science165:85-86.

4. Borman, C. R., J. Sidikaro, and M. Nomura. 1971.Specific inactivation of ribosomes by colicin E3 in vitroand mechanism of immunity in colicinogenic cells. Na-ture (London) New Biol. 234:133-137.

5. Dabbs, E. R. 1977. A spectinomycin dependent mutantof Escherichia coli. Mol. Gen. Genet. 151:261-267.

6. Dabbs, E. R. 1977. Correlation between a novel pheno-type towards streptomycin and the binding of an addi-tional protein to the ribosome in mutants of Esche-richia coli B. Mol. Gen. Genet. 158:55-61.

7. Dabbs, E. R., and H. G. Wittmann. 1976. A strain ofEscherichia coli which gives rise to mutations in a largenumber of ribosomal proteins. Mol. Gen. Genet. 149:303-309.

J. BACTERIOL.

on August 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

KASUGAMYCIN-DEPENDENT MUTANTS 1001

8. Davis, B. D., and E. S. Mingioli. 1950. Mutants ofEscherichia coli requiring methionine or vitamin B12.J. Bacteriol. 60:17-28.

9. Guerola, N., J. L. Ingraham, and E. Cerda-Olmedo.1971. Induction of closely linked multiple mutations bynitrosoguanidine. Nature (London) New Biol. 230:122-125.

10. Hasenbank, R., C. Guthrie, G. Stoffler, H. G. Witt-mann, L. Rosen, and D. Apirion. 1973. Electropho-retic and immunological studies on ribosomal proteinsof 100 Escherichia coli revertants from streptomycindependence. Mol. Gen. Genet. 127:1-18.

11. Helser, T. L., J. E. Davies, and J. E. Dahlberg. 1972.Mechanism of kasugamycin resistance in Escherichiacoli. Nature (London) 235:6-9.

12. Isono, K., J. Krauss, and Y. Hirota. 1976. Isolation andcharacterization of temperature-sensitive mutants ofEscherichia coli with altered ribosomal proteins. Mol.Gen. Genet. 149:297-302.

13. Jaskunas, S. R., A. M. Fallon, and M. Nomura. 1977.Identification and organization of ribosomal proteingenes of Escherichia coli carried on fus2 transducingphage. J. Biol. Chem. 252:7323-7336.

14. Kaltschmidt, E., and H. G. Wittmann. 1970. Ribosomalproteins. VII. Two dimensional polyacrylamide gel elec-trophoresis for fingerprinting ribosomal protein. Anal.Biochem. 36:401-412.

15. Laughrea, M., and P. B. Moore. 1978. On the relation-ship between the binding of ribosomal protein S1 to the30S subunit of Escherichia coli and 3' terminus of 168RNA. J. Mol. Biol. 121:411-430.

16. Lennox, E. S. 1955. Transduction of linked genetic char-acters of the host by bacteriophage P1. Virology 1:190-206.

17. Lindahl, L. F., S. R. Jaskunas, P. Dennis, and M.Nomura. 1975. Cluster of genes in Escherichia coli forribosomal proteins, ribosomal RNA and RNA polym-erase subunits. Proc. Natl. Acad. Sci. U.S.A. 72:2743-2747.

18. Ozaki, M., S. Mizushima, and M. Nomura. 1969. Iden-tification and functional characterization of the proteincontrolled by the streptomycin-resistant locus in E. coli.Nature (London) 222:333-339.

19. Sparling, P. F., A. Ikeya, and D. Elliott. 1973. Twogenetic loci for resistance to kasugamycin in Esche-richia coli. J. Bacteriol. 113:704-710.

20. Tai, P. C., B. J. Wallace, and B. D. Davis. 1973. Actionsof aurintricarboxylate, kasugamycin and pactamycin onEscherichia coli polysomes. Biochemistry 12:616-620.

21. Thammana, P., and C. R. Cantor. 1978. Studies onribosome structure and interactions near the m26A m26Asequence. Nucleic Acids Res. 5:805-823.

22. Thammana, P., and W. A. Held. 1974. Methylation of16S RNA during ribosome assembly in vitro. Nature(London) 251:682-686.

23. van Duin, J., C. G. Kurland, J. Dondon, and M.Grunberg-Manago. 1975. Near neighbors of IF3bound to 30S ribosomal subunits. FEBS Lett. 59:287-290.

24. Wittmann, H. G. 1974. Purification and identification ofEscherichia coli ribosomal proteins, p. 93-114. In M.Nomura, A. Tissieres, and P. Lengyel (ed.), Ribosomes.Cold Spring Harbor Laboratory, Cold Spring Harbor,N.Y.

25. Wittmann, H. G., G. Stdffier, L. Hindennach, C. G.Kurland, L. Randall-Hazelbauer, E. A. Birge, M.Nomura, E. Kaltschmidt, S. Mizushima, R. R.Traut, and T. A. Bickle. 1971. Correlation of 30Sribosomal proteins of Escherichia coli isolated in dif-ferent laboratories. Mol. Gen. Genet. 111:327-333.

26. Yaguchi, M. 1975. Primary structure of protein S18 fromthe small Escherichia coli ribosomal subunit. FEBSLett. 59:217-220.

27. Yoshikawa, M., A. Okuyama, and N. Tanaka. 1975. Athird kasugamycin resistance locus, ksgC, affecting ri-bosomal protein S2 in Escherichia coli K-12. J. Bacte-riol. 122:796-797.

VOL. 136, 1978

on August 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from