JOURNAL OF BACTERIOLOGY, Nov. 1968, p. 1750-1759Copyright @ 1968 American Society for Microbiology

Partial Exclusion of the Nocardia erythropolisChromosome in Nocardial Recombinants

JAMES N. ADAMSDepartment of Microbiology, School ofMedicine, University ofSouth Dakota,

Vermillion, South Dakota 57069

Received for publication 8 August 1968

Segregation of genes specifying nutritional requirements and inhibitor resist-ance was analyzed in recombinants from crosses of Nocardia erythropolis withN. canicruria. With all employed characters as both selecting and nonselectingmarkers, a single linear linkage group was depicted with genes in the followingorder; leu-6 gly-4 purB2 tetA9 his-i eryA7 strB2 ser-J arg-1. The relative frequencyof crossovers between gene pairs was found to decrease as the map is read fromleu-6 to the right. The phenotype for arg-l was found infrequently in recombinants.This fact, coupled with the observed decrease in crossover frequency dependentupon map position, suggested two alternative explanations to account for the ob-served results. (i) Zygote formation among these nocardiae is a consequence ofpartial, oriented chromosome transfer from N. erythropolis to N. canicruria andis similar to the mechanism of polar transfer reported in matings of Hfr by FEscherichia coli. (ii) Being of diverse origin, the chromosomes of N. erythropolisand N. canicruria are not completely homologous, and decrease in crossover fre-quency dependent upon map position results from such chromosomal heterology.In this hypothesis, the infrequent occurrence of arg-J among recombinants isconsidered to be an artifact resulting from the action of a previously unrecognizedsuppressor gene for arg-J present in N. canicruria.

As a result of recent examinations of the no-cardiae, some aspects of their genetic makeupand their means for transmitting genetic informa-tion from parent to progeny have been specified(1, 2, 5). The nocardiae, like both Escherichia coliand Streptomyces coelicolor, appear to be func-tionally haploid. Our continuing observationshave not yet shown nocardial genes to be circu-larly permutated on the linkage group. However,apparent differences in the topology of theNocardia erythropolis and N. canicruria chromo-somes have been noted (5), and mating types,which did not appear to involve episomes (1, 2),have been described. Thus, although the heredi-tary mechanisms of these organisms are by nomeans completely understood at present, thenocardiae appear to be genetically dissimilar insome aspects to the eubacteria and the strepto-mycetes.

In this investigation, the segregation phe-nomena and linkage previously noted amongthese nocardiae were substantiated, and addi-tional information concerning the nocardialchromosome is reported. Because of the existenceof mating factor control of fertility (1, 2) in thenocardiae, it is quite time consuming to synthe-

size fertile strains with appropriate markers invarious configurations of coupling and repulsion.Thus, a simplified determination of linkage andsegregation, by use of such classic genetic method-ology which enables one to observe the behaviorof the nocardial chromosome, is not easily avail-able. To overcome this difficulty, strains with avariety of unselecting inhibitor resistance markers(5) were developed. The subsequent addition ofauxotrophic genes to these strains and the deter-mination of segregation of resistance and auxo-trophic markers in crosses of the initial mutantsand their successive derivatives allowed theanalysis of markers in alternative phases ofcoupling or of repulsion. These techniques wereused in the present study and it was establishedthat the nocardiae are among those bacteria whichexhibit conjugational phenomena. Features ofnocardial recombinational behavior which appearto exclude a portion of the N. erythropolis genomefrom recombinants were also recognized.

MATERIALS AND METHODS

Strains used. The strains of N. erythropolis and N.canicruria used in these investigations and theirgenotypes and phenotypes are listed in Table 1.

1750

Vol. 96, No. 5Printed in U.S.A.

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

GENE EXCLUSION IN NOCARDIAL MATING

Mutants derived from N. erythropolis JA-SD 2 andmutant JA-SD 3-64 were induced with ultravioletirradiation (1, 2). JA-SD 3-63 was obtained aftertreatment of JA-SD 3-48 with N-methyl-N'-nitro-N-nitrosoguanidine (Aldrich Chemical Co., Milwaukee,Wis.) by use of a method like that reported for E.coli (4). For easy recognition of strains derived indifferent laboratories, previously designated strainnumbers (1, 5) are herein prefixed with "JA-SD"; allnewly isolated strains from this laboratory will be soidentified. Other strain codification is as previouslyreported (1, 5) and conforms to the recommenda-tions of Demerec et al. (6).

Mating. The standard mating procedure was car-ried out as follows. The parental types were grown onplates of Nutrient Broth (Difco), solidified with1.5% (w/v) agar, for 2 or 3 days. Suspensions ofthese strains containing ca. 5 X 108 cells/nil wereprepared in sterile saline. Three sets of nutrient agar(NAg) plates were prepared. One set was inoculatedwith 0.1 ml of the N. erythropolis parental type, thesecond set was inoculated with 0.1 ml of the N.canicruria parental type, and the third set was inocu-lated with 0.2 ml of a 1:1 mixture of both parentaltypes. The crosses and controls were incubated for 3days. After incubation, saline suspensions containingca. 5 X 108 cells/ml were prepared from each set. Sam-ples (0.1 ml) of these suspensions, or appro-priate dilutions, were inoculated onto various mediain order to allow selection of recombinants fromcrosses. Parental controls were inoculated in parallelon selective media and permitted the recovery andestimation of revertant frequency among the controlpopulations. Total counts of all populations weredetermined on NAg. Selective plates were incubatedfor 6 to 8 days.

Segregation analysis. Diagnosis of unselectedcharacters was determined as follows and differedsomewhat from methods used previously (5). Withsterile flat toothpicks, recombinant colonies werepicked from selective media and were subcultured onan NAg plate; there was a total of 30 colonies perplate. These master plates were incubated for 24 hrand were replicated with sterile velveteen pads to amedium identical to the selective medium from whichthe recombinants were isolated (control) and todifferential media. Total patch growth on differentialand control media was used as the growth criterion.Parental types resulting from possible carry-over fromthe selective media and growing on these completemedia were thereby eliminated from scoring. Newmating rounds during the 24-hr incubation period onNAg (1) were previously shown to occur too infre-quently to contribute new recombinants in largeenough proportions to introduce serious error in thesegregation patterns for unselected markers.

Selective media and differential media (1, 5) forthe determination of auxotrophic characters wereprepared with our previously described minimalmedium (MM) supplemented with 10 ,ug of therequirement(s) per ml. Differential media for thedetermination of inhibitor resistance were preparedby adding 50 ug of streptomycin per ml, 5 ,ug oferythromycin per ml, or 3 ,ug of tetracycline per ml

to NAg. Inhibitor-containing selective media wereprepared by adding inhibitors, at concentrations likethose used in the differential media above, to MM orto growth factor-supplemented MM.The use of a complete medium for the preparation

of master plates for replication of recombinants todiagnostic media and the use of inhibitor containingNAg for diagnosis of recombinant resistance charac-ters substantially reduced the time necessary forsegregation analysis. With the described technique,selection and segregation analysis was completed in12 to 13 days. This may be compared with previousinvestigations in which analysis required 20 to 22days. In those reports (5), media homologous tothose used to select recombinants were used formaster plate preparation and for preparation ofinhibitor diagnostic media.

All incubations were carried out at 30 C.

RESULTS

In N. canicruria JA-SD 3-48, we previouslyestablished a linear linkage map, purB2 tetA9eryA7 strB2, in a single linkage group (5). Toexpand this map and to gain further knowledgeof nocardial recombination, additional genesspecifying auxotrophy were incorporated in thisstrain and in N. erythropolis strains (Table 1).The effect of the successive incorporation of thehis-i, arg-i, and ser-i loci in N. erythropolis onthe segregation of unselected resistance charac-ters in prototrophic recombinants from crossesof these strains with JA-SD 3-48 is shown inTable 2. To determine gene location from suchdata, the following assumptions must be made:(i) nocardial genes are distributed on a linearchromosome (5); (ii) crossovers occur betweenhomologues; (iii) the frequency of crossoverbetween two genes is proportional to the distancebetween the genes; (iv) crossovers in two regionsoccur with a probability that is the product ofthe probabilities of crossover in the individualregions. With these assumptions, the resistancephenotypes of prototrophic recombinants fromthe cross of JA-SD 2-10 with JA-SD 3-48 readilyspecify a location for the his-l+ gene. Three majorsegregant class types were observed in proto-trophic recombinants from this cross (Table 2).The most satisfactory linear arrangement ac-counting for such segregation situates his-l+between tetA9 and eryA7, as shown in the linkagemodel of Table 2. The most frequent phenotypicclasses, Tet-R Ery-R Str-R and Tet-S Ery-RStr-R, were considered to be the products ofsingle crossover events. The third most frequentclass, Tet-R Ery-R Str-S, resulted from crossoverevents in two regions. Other phenotypes were oflow frequency and appeared to arise from moreinfrequent multiple crossover events. Five of theeight possible phenotypes were recovered. Al-

1751VOL. 96, 1968

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

TABLE 1. Strain characteristics and designations

Mutant loci and mutation sitesb Relevant phenotypeCStrain' Derived from

arg gly his leu ser purB eryA strB tetA Arg Gly His Leu Ser Pur Ery Str Tet

JA-SD 2 + + + + + + + + + + + + + + + S S SJA-SD 2-10 JA-SD 2 + + 1 + + + + + + + + + + +S S SJA-SD 2-31 JA-SD 2-10 1 + 1 + + + + + + + + + +S S SJA-SD 2-83 JA-SD 2-31 1 + 1 + 1 + + + + + + - +S S SJA-SD 3 + + + + + + + + + + + + + + + S S SJA-SD 3-3 JA-SD 3 + + + + + 2 + + + + + + + + - S S SJA-SD 3-48 JA-SD 3-3 + + + + + 2 7 2 9 + + + + + - R R RJA-SD 3-63 JA-SD 3-48 + + + 6 + 2 7 2 9 + + + +- R R RJA-SD 3-64 JA-SD 3-63 + 4 + 6 + 2 7 2 9 +- + - R R R

a Strains of Nocardia erythropolis are prefixed with Ja-SD 2; strains of N. canicruria are prefixedwith JA-SD 3.

6 Descriptions of mutant loci, mutation sites, and relevant phenotypes are, insofar as possible, inconformity with the recommendations of Demerec et al. (6).

c Symbols: +, synthesized; -, required: R, resistant; S, susceptible; Arg, arginine; Gly, glycine;His, histidine; Leu, leucine; Ser, serine; Pur, purine; Ery, erythromycin, 5 ,ug/ml; Str, streptomycin,50 jug/ml; Tet, tetracycline, 3 ,ug/ml.

TABLE 2. Comparison ofsegregation ofunselected resistance markers in prototrophs from crosses ofJA-SD3-48 with JA-SD 2-10, JA-SD 2-31, or JA-SD 2-83

Recombinant phenotypeFraction of prototrophs exhibiting phenotype in crosses of JA-SD 3-48

(purB2 teIA9 eryA7 strB2) with

Tet Ery Str JA-SD 2-10- JA-SD 2-31b JA-SD 2-83CTet Ery Str ~~~~~~~(isi)(his-i erg-i) (his-i erg-i ser-1))

R R R 0.841 0.831 0.800S R R 0.126 0.112 0.108R R S 0.027 0.044 0.058S S R 0.003 0.006 0.008S R S 0.003 0 0.008R S R 0 0 0.017S S S 0 0.006 0

No. of prototrophs tested 301 160 120

Linkage modeld + + his-i + + (ser-l arg-1)purB2 tetA9 + eryA7 strB2 ( + + )

a Prototrophs selected on MM at a frequency of 3.8 X 10-5.b Prototrophs selected on MM at a frequency of 1.6 X 10-6.c Prototrophs selected on MM at a frequency of 2.3 X 105.d Linkage model not to relative scale. Order of loci in parentheses not determined from this data.

though all recoverable phenotypes were used tocalculate relative distances between genes, pheno-types other than the three major types wereusually too infrequent to be significant for thispreliminary mapping of his-l.

Since the placement of his-i appeared to bequite straightforward, the arg-i gene was suc-cessively incorporated into the his-i-containingstrain in order to produce JA-SD 2-31. Ser-l wassuccessively incorporated into this strain in orderto produce JA-SD 2-83. The effect of these addi-

tional auxotrophic markers of N. erythropolis onsegregation patterns of the unselected resistancecharacters in recombinants obtained from mat-ings of these strains with JA-SD 3-48 are pre-sented in Table 2. The successive establishment ofarg-i and ser-I in derivatives of N. erythropolisJA-SD 2-10 did not alter dramatically the patternof segregation of the unselected resistancemarkers. Therefore, these additional auxotrophicgenes can be considered to be distal to the threeresistance loci and to his-i or very closely linked

1752 ADAMS J. BACTERIOL.

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

GENE EXCLUSION IN NOCARDIAL MATING

to his-i. As presented in the linkage model ofTable 2, ser-l and arg-l were placed to the rightof strB2. (This location will be further verifiedbelow.) Had either one or both of these genesspecifying auxotrophy been placed to the left ofpurB2, prototrophic recombinants must thenoccur as a consequence of two crossover events.Although this possibility is not ruled out by theseresults, it will be shown that this hypothesis is nottenable. However, there is a most interesting con-sequence of the location of these markers to theright of strB2. According to the linkage model inTable 2, the frequency of the Tet-R Ery-R Str-Sphenotype, postulated to result from two cross-over events in the mating of JA-SD 2-10 withJA-SD 3-48, should have decreased with theaddition of either arg-l or ser-J to JA-SD 2-10.This is a consequence of the additional crossovernecessary between strB2 and arg-I or ser-1 arg-Jin order to incorporate arg-l+ and ser-l+ arg-l+loci in the selectable prototrophs of the Str-Sphenotype. The frequency of this phenotype(Table 2) was altered slightly upwards. Thisstriking result will be discussed below.The use of N. canicruria JA-SD 3-48 or N.

erythropolis JA-SD 2-10, or both, which containsingle genes for auxotrophy, leads to possibleerror in segregation patterns and consequenterror in mapping. A reversion or suppressormutation in either strain would allow the re-covery of prototrophs which are indistinguishablefrom recombinants. To eliminate such error andto confirm the position of his-i in relation to the

unselecting resistance genes, two further series ofexperiments were undertaken. In one series,additional auxotrophic markers were incorpo-rated into the N. canicruria inhibitor-resistantstrain. In the other series, inhibitor resistance andprototrophy were used to select against N.erythropolis JA-SD 2-10. The effects of the suc-cessive addition of leu-6 to JA-SD 3-48, in orderto produce JA-SD 3-63, and the addition of gly-4to this strain, in order to produce JA-SD 3-64, onsegregation in crosses of these strains with JA-SD2-10 are presented in Table 3. Patterns of re-

sistance in these prototrophic recombinants,like the patterns in prototrophs observed in mat-ings of the single auxotrophs, were generallyunaffected by the incorporation of these addi-tional genes in derivatives of N. canicruria JA-SD 3-48. Thus, leu-6 and gly-4 can definitely beconsidered to be distal to the purB2, his-i, andresistance loci. (This placement will be furtherverified below.) A model representing the loca-tion of these genes is shown in Table 3. It is notpossible to determine from these data only theproper ordering of the leu-6 and gly-4 markers,but the results from these matings further estab-lished the location of his-i recognized above.

Further confirmation of the mapping of his-iof N. erythropolis was obtained from experimentsin which the His+ phenotype and a single in-hibitor resistance marker were simultaneouslyselected and segregation of the remaining unse-lected resistance markers was determined. Con-traselection for auxotrophic markers of N.

TABLE 3. Comparison ofsegregation ofunselected resistance markers in prototrophs from crosses ofJA-SD2-10 with Tet-R Ery-R Str-R strains JA-SD 3-48, JA-SD 3-63, and JA-SD 3-64

Recombinant phenotypes Fraction of prototrophs exhibiting phenotype in crosses ofJA-SD 2-10 (his-I) with

Tet Ery Str JA-SD 3-48a JA-SD 363b JA-SD 3-64cTet Ery Str (purB2) (purB2 leu-6) (purB2 leu-6 gly4)

R R R 0.841 0.809 0.830S R R 0.126 0.139 0.132R R S 0.027 0.043 0.032S S R 0.003 0 0S R S 0.003 0.009 0.009R S R 0 0 0S S S 0 0 0

No. of prototrophs tested 300 115 106

Linkage modeld + + + + his-i + +(leu-6 gly-4) purB2 tetA9 + eryA7 strB2

a Prototrophs selected on MM at a frequency of 3.8 X 10-5.b Prototrophs selected on MM at a frequency of 3.5 X 10-6.c Prototrophs selected on MM at a frequency of 1.1 X 10-6.d Linkage model not to relative scale. Order of loci in parentheses not determined from this data.

1753VOL. 96, 1968

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

canicruria was also carried out (Table 4). Com-pared to MM, the incorporation of a singleinhibitor in MM resulted in fluctuation in theratios of resistant to sensitive phenotypes for thetwo unselected resistance markers in the recom-binant prototrophs. Selection on streptomycin-containing media produced the greatest observedfluctuation. However, the average proportions ofsensitive and resistant phenotypes among theprototrophs selected on inhibitory media, calcu-lated as the average from all three inhibitor-containing selective media, were not appreciablydifferent from the ratios noted when inhibitor-free MM was used for selection. Thus, thesefluctuations were not the direct result of linkagebut reflected selective pressures of the media.The foregoing experiments can be used to

locate auxotrophic genes. These experimentspermit the use of alternative alleles specifyingauxotrophy or prototrophy in a manner similarto classic segregation analysis techniques in whichmating of strains with alleles in coupling and re-pulsion phases are carried out. For example, if anarg-i gene is added to a his-i-containing strain,two strains are available: his-i arg-i+ and his-iarg-i. The mating of these strains effectivelyserves to place the alleles arg-i and arg-i+ incoupling and repulsion with the his-i gene (al-though not with the his-i+ gene) and permitsmapping of the genes.

After preliminary mapping of the loci in thepreceding manner, the auxotrophic genes wereused as nonselecting characters. By incorporationinto the selective medium of only one of two ormore nutrient requirements specified by the multi-auxotrophic strains, genes determining auxo-

trophy could be treated as nonselecting markers.Therefore, we could compare mapping of thesegenes by nonselective means with mapping ob-tained by use of selective methods. The results ofthe mating of N. erythropolis JA-SD 2-31 with N.canicruria JA-SD 3-63 are shown in Table 5. Inthis cross, four auxotrophic and three inhibitorresistance genes were available for analysis. Thedistribution of resistance phenotypes among thefully prototrophic recombinants from this crosswas like the patterns observed in crosses ofJA-SD2-31 or JA-SD 3-63 with other strains, as pre-sented in Tables 2, 3, and 4. These data alonetended to substantiate the location of the leu-6,purB2, his-i, and arg-i loci. However, determina-tion of the patterns of segregation among the re-sistance markers in relation to unselected auxo-trophic phenotypes was the prime objective ofthese experiments. By use of media selecting forthe four new pairwise combinations of parentalauxotrophic characters, but not allowing recoveryof parental diauxotrophs, about one half of the 72theoretically recoverable phenotypes were ob-served. Many of these appeared in only one ortwo recombinants. For tabular clarity, only themost frequent phenotypes necessary for geneplacement are specifically identified in Table 5.When the auxotrophic genes were mapped byselective means, leu-6 and arg-i were placed at theends of the chromosome, with purB2, his-i, andresistance genes located between them. Conse-quently, a single crossover in the mating of JA-SD 2-31 with JA-SD 3-63 should produce bothPur and His- recombinants of various resistancephenotypes. The largest proportion of auxo-trophic phenotypes should be found on selective

TABLE 4. Comparison of segregation of selected and unselected resistance markers in prototrophs fromcrosses of JA-SD 2-10 with Tet-R Ery-R Str-R strains JA-SD 3-48, JA-SD 3-63, or JA-SD 3-64

Fraction of recombinants exhibiting resistance phenotypeafrom crosses of JA-SD 2-10 with

Selection on b Phenotype -

JA-SD 3-48 JA-SD 343 JA-SD 364 Avg ofall crosses

MM Ery-R Str-R 0.126 0.139 0.132 0.132Tet-R Ery-R 0.027 0.043 0.032 0.034

MM plus tetracycline Tet-R Ery-R 0.031 0.045 0.053 0.043MM plus erythromycin Ery-R Str-R 0.171 0.225 0.139 0.178

Tet-R Ery-R 0.016 0.044 0.072 0.044MM plus streptomycin Ery-R Str-R 0.154 0.117 0.300 0.190

Avg from all selective Ery-R Str-R 0.163 0.171 0.190media Tet-R Ery-R 0.024 0.045 0.063

a A minimum of 106 recombinants were analyzed from each medium for determination of resistancephenotype.

b Recombination frequencies on MM as in Table 3. Recombinant recovery on inhibitor-containingmedia was approximately 25% lower.

1754 ADAMS

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

GENE EXCLUSION IN NOCARDIAL MATING

TABLE 5. Segregation patterns ofunselected resistance and auxotrophic markers in recombinants from thecross of JA-SD 2-31 with JA-SD 3-63

Fraction of recombinants exhibiting phenotype when selected on MM plusPhenotype -

Pur-His4 Pur-Argb His-LeuC Arg-Leud

Prototrophic ............................. 0.032 0.014 0.347 1.000Pur Tet-R Ery-R Str-R.0.758 0.977 -fHis- Tet-S Ery-S Str-S.0.124 0.571His- Tet-S Ery-R Str-R.0.016 _ 0.082Otherso.................................. 0.070 0.009 0 0

No. of recombinants tested 186 222 245 361

Linkage model" + + + his-i + + arg-ileu-6 purB2 tetA9 + eryA7 strB2 +

The purine used for supplementation was adenine. Recombination frequency 3.6 X 104.b Recombination frequency, 5.6 X 104.c Recombination frequency, 1.1 X 105.d Recombination frequency, 1.0 X 10g.e Prototrophicrecombinants selected on MM exhibited segregation patterns for inhibitor resistance

characters like those patterns observed with crosses of JA-SD 2-31 and JA-SD 3-63, as shown in Tables2 through 4. Segregation patterns of prototrophs on supplemented medium were like those observed onMM.

f This phenotype was not detectable on the indicated selective medium.g This group contains the total number of recombinants exhibiting phenotypes unlike others listed

in the table. These phenotypes usually occurred in only one or two recombinants.hs The linkage model is not to relative scale.

media containing Pur or His, or both. Thisanticipated result was observed (Table 5). Theaverage phenotypic ratios important to the order-ing of the auxotrophic loci were calculated fromthe results from appropriate selective media(Table 5) and are as follows: Pur- to Pur+, 40: 1;His- Ery-S Str-S to His- Ery-R Str-R, 5:1; Histo His+, 2.6:1. These ratios are in accord with theordering of the linkage model in Table 5. It isalso apparent from the fluctuation of these ratioson specific media (e.g., compare Pur- to Pur+ onPur-His medium with Pur- to Pure on Pur-Argmedium) that the addition of different nutritionalsupplements to the media established selectivepressures.A most significant finding from these data

(Table 5) was the lack of the Arg- phenotypeamong recombinants. From the analysis of mat-ings of JA-SD 2-10 with JA-SD 3-48, JA-SD3-63, or JA-SD 3-64, two crossovers were postu-lated to account for the production of the Tet-REry-R Str-S phenotype noted in approximately4% of the prototrophic recombinants from thesecrosses. The frequency of this class type amongprototrophs from crosses of JA-SD 2-31 withJA-SD 3-63 was essentially identical. However,arg-i in strain JA-SD 2-31, derived from JA-SD2-10, was mapped at the right of strB2+. There-fore, in matings of JA-SD 2-31 with JA-SD 3-63,

a greater frequency of Tet-R Ery-R Str-S Arg7progeny, compared with Tet-R Ery-R Str-S Arg+,should be recovered, since the former phenotypeshould result from only two crossovers whereasthe latter should result from three crossovers. Asshown in the results of Table 5, the Arg- pheno-type was not observed in these experiments. It ispossible that arg-i was erroneously located in thelinkage model of Table 5. Consequently, Arg- re-combinants should not be recovered in detectablefrequencies. This does not seem likely. If arg-iwere to the left of his-i, single crossovers shouldproduce Arg- recombinants and arg-i shouldoccur frequently. If arg-i were very closely linkedto the left of his-i, then His- Ery-S Str-S recom-binants should result from three crossovers andshould occur infrequently., However, they werefound to occur frequently: If arg-i were closelylinked to the right of his-i, two crossovers wouldbe necessary for the production of Arg- recom-binants or His- Ery-S Str-S recombinants and ifone is detected both should be detectable. Also, inthis instance, His- Ery-R Str-R recombinantswould result from single crossovers and the ratioof His- Ery-R Str-R to His- Ery-S Str-S shouldbe high. However, this is the reverse of the ob-served results (Table 5). None of these al-ternatives, therefore, is satisfactory for relocatingthe arg-i gene on the linkage model in Table 5.

VOL. 96, 1968 1755

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

However, for any location of arg-i on the map,the detectable production of arg-l-containing re-combinants would be a consequence. If, on thecontrary, arg-i is not linked, a 1 to 1 ratio ofarg-i+ to arg-i alleles should occur in recom-binants. From these considerations, it must beconcluded that arg-i is linked to the other genesbut this allele is generally excluded from recom-binants.

Further confirmation of the mapping of theauxotrophic markers in the order indicated in themodels in Table 2 through 5 were sought. Thiswas accomplished by adding (and then analyzing)a third auxotrophic gene to strains JA-SD 2-31and JA-SD 3-63. The results of the addition ofser-i to JA-SD 2-31 in order to produce JA-SD2-83 are shown in Table 6. Recombinants result-ing from the cross ofJA-SD 2-83 with JA-SD 3-63were selected on media supplemented with eachone of the nutritional requirements of JA-SD2-83, and the auxotrophic markers and resistancecharacters were scored. Thus, the most frequentauxotrophic phenotype could be determined di-rectly. Ifthe previous linkage models were correct,his-i and purB2 should be the only genes ex-pected to segregate with appreciable frequency.The additional auxotrophic gene could also beexpected to affect segregation of unselecting re-sistance loci and to permit the ordering of theauxotrophic loci. His- phenotypes were the mostfrequent single auxotrophic phenotype recovered(Table 6). In the cross of JA-SD 2-31 with JA-SD

3-63, the ratio of His- Ery-S Str-S to His- Ery-RStr-R was 5:1; in the cross of JA-SD 2-83 withJA-SD 3-63, this ratio was 1:18. The additionof ser-J in JA-SD 2-83 inverted the frequency ofthe inhibitor phenotypes of His- recombinants.This result was fully consistent with the position-ing of ser-i to the right of strB2+ between arg-iand his-l. Table 7 presents the effects of the addi-tion of gly-4 to JA-SD 3-63 in order to produceJA-SD 3-64. The significant result of the matingof JA-SD 2-31 with JA-SD 3-64 is the decrease inthe Purc to Pur+ ratio among recombinants(7:1), as compared with the mating of JA-SD2-31 with JA-SD 3-63 in which this ratio was 40:1.These data located gly-4 between leu-6 and purB2.Among the fully prototrophic recombinants ob-served in the crosses in Tables 6 and 7, segrega-tion of unselected inhibitor resistance markerswas, in general, like that observed in crosses ofJA-SD 2-10 with JA-SD 3-48. Thus, all data are

consistent with the linkage models presented inTables 2 through 7.With recovery fractions of unselected pheno-

types from the data in Tables 2 through 7, rela-tive distances between loci were calculated (5)and a linkage map (Fig. 1) was constructed forthe examined markers. Relative map distancesare expressed as percentiles of the examinedpopulations exhibiting crossovers in the postu-lated crossover regions. The major reference re-gions for map construction were between purB2and his-I and the inhibitor resistance loci.

TABLE 6. Segregation patterns of unselected resistance and auxotrophic markers in recombinants from thecross of JA-SD 2-83 with JA-SD 3-63

Fraction of recombinants exhibiting phenotype whenPhenotype selected on MM plus

Hisa Serb Ar_

Prototrophicd................................. 0.807 0.979 0.991His- Tet-S Ery-S Str-S ........................ 0.012 -His- Tet-S Ery-R Str-R ........ ............... 0.180 _Othersf............0..... 0.021 0.009

No. of recombinants tested 483 390 215

Linkage modelg + + + his-i + + ser-J arg-ileu-6 purB2 tetA9 + eryA7 strB2 + +

a Recombination frequency 1.6 X 10-.b Recombination frequency 1.3 X 10-5.c Recombination frequency 2.0 X 10-.d Prototrophic recombinants exhibited segregation patterns for inhibitor resistance characters like

those patterns observed in crosses of JA-SD 2-31 with JA-SD 3-63.e This phenotype was not detectable on the indicated selective medium.f This group contained the total number of recombinants exhibiting phenotypes unlike others listed

in the table. These usually occurred in only one or two recombinants.g The linkage model is not to relative scale.

1756 ADAMS

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

GENE EXCLUSION IN NOCARDIAL MATING

TABLE 7. Segregation patterns ofunselected resistance and auxotrophic markers in recombinants from thecross of JA-SD 2-31 with JA-SD 3-64

Fraction of recombinants exhibiting phenotype whenselected on MM plusPhenotype______________________________

Put' Leub Gly°

Prototrophicd ............................ .. 0.133 1.000 1.000Pur Tet-R Ery-R Str-R....................... 0.867 -6

No. of recombinants tested 120 199 199

Linkage modelf + + + + his-i + + arg-Jleu-6 gly-4 purB2 tetA9 + eryA7 strB2 +

a Recombination frequency, 7.7 X 10-4; purine used for supplementation was adenine.b Recombination frequency, 2.0 X 10-.- Recombination frequency, 1.7 X 10'.d Prototrophic recombinants exhibited segregation patterns for inhibitor resistance characters like

those patterns observed in crosses of JA-SD 2-31 with JA-SD 3-63.e This phenotype was not detectable on the indicated selective medium.f The linkage model is not to relative scale.

LEU-6 79.30 GLY-4 15.39 PURB2 2.04 TETA9 o.e SER-I 2.67 ARG-I

S~~~~~~~-- K$

I ~I I ItTETA9 o.33 HIS-1 0.19 ERYA7 o.os STRB2 o.oi SER-I

FIG. 1. Composite linkage map of examined genes in Nocardia erythropolis (his-i, ser-J, and arg-i) and N.canicruria (leu-6, gly-4, purB2, tetA9, eryA7, and strB2). The genes, illustrated by vertical bands, are placedaccording to their relative distances from each other. Values for map distances between loci are given as averagerelative percentages as accruedfrom all of the presented experimental crosses.

DIscussioNWe previously reported (5) that the purB2 and

resistance loci used in the present studies could bebest mapped in a single linear linkage group. Inthe present studies, we examined segregation, andthus linkage, of these and additional genes speci-fying auxotrophy in matings of N. erythropoliswith N. canicruria under conditions where allmarkers could be used to advantage as selectingand nonselecting characters. The most satisfactorymap which could be constructed from the ob-served data was presented as a single, linear array(Fig. 1). A circular map, like those typical of E.coli (10), S. coelicolor (8), and Salmonella typhi-murium (9), did not appear to be superior for theinterpretation of the segregation phenomena inthese matings of the nocardiae. In most of theseexperiments, three or more loci were used as un-selecting characters in a variety of selective con-

ditions. Thus, apparent linkage of a pair of genesunder one set of selective conditions and relativeindependence of the same gene pair under anotherset of selective conditions, behavior typifyingcircular linkage, could have been recognized.However, such behavior was not observed. Asmay be deduced from the experiments describedhere, the proposed linear linkage map was foundto have good predictive value within the limitsimposed by the selective medium pressures andthe inherent difficulty in determining absolutemap distances (1, 5) in these organisms. The ob-served relative distances were internally consistentin individual experiments and were reproduciblefrom experiment to experiment. These considera-tions support the value of the map, and it is con-sidered to be a reasonable depiction of the no-cardial chromosome as it is presently understood.This map, of course, is subject to future modifica-

1757VOL. 96? 1968

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.-

tion when further knowledge of nocardial heredi-tary phenomena is achieved.

Previously, it was hypothesized (1, 2, 5) that thefollowing series of events occur in nocardialmating. A heterogenomic zygotic element waspresumed to be formed between compatible pairsof N. erythropolis and N. canicruria. It was postu-lated that the total genome of both parents tookpart in the formation of the zygotic element. Inthis hypothesis, it was not necessary for thegenomes to be complete homologues, and someevidence was presented indicating that they werenot (5). After zygote formation, effective genomalpairing allowed production of new haploid com-binations between disparate genomes and re-sultant segregants were selectable on appropriatemedia.Two features of the present data demand re-

evaluation of the zygotic element, containingthe total genome of both strains, postulated toaccount for nocardial recombination. The moststriking feature is, of course, the apparent ex-clusion of the arg-i locus or, more generally, theexclusion of an unknown length of the N. ery-thropolis chromosome to the right of the his-igene from recombinants. Another feature ofspecial significance in the present report is theobserved decrease in relative crossover frequencybetween genes as the map (Fig. 1) is read fromleft to right. Within the limits of an individualmating, random crossovers between loci appearedto account for observed segregation patternsamong the selected recombinants. However, whenall genes were relatively placed on the compositelinkage map, a trend toward compulsory cross-over in specified regions was observed. Twomodels may be presented to account for these ob-servations.On the one hand, the possibility exists that

oriented chromosome transfer, similar to thatobserved in E. coli (7, 11), occurs in nocardialconjugation. With such a model, itcan be assumedthat the polarity of transfer is from N. erythro-polis to N. canicruria and that the origin of trans-fer is uniform in all mating pairs. These conten-tions are supported by the occurrence of uniformsegregation and the high proportion of unselectedN. canicruria alleles observed in nocardial recom-binants. In these nocardiae, as in E. coli, it couldbe expected that partial transfer occurs as a re-sult of random breaks in the donor chromosomeduring the transfer process. Under such condi-tions, at the termination of unit mating time, apopulation of partial zygotes would be formed ina nonsynchronously mating population. The totalgenome of N. canicruria would be present in themerozygote, but the N. erythropolis chromosomalcontribution to the merozygote would be variable.

The farther a gene is located from the origin oftransfer, the less likely that the chromosomalsegment carrying that gene would be found in themerozygote. Therefore, the total number ofmerozygotes having genes located on segmentsnear the origin of transfer would exceed the totalnumber of merozygotes having genes located onchromosomal segments more distal to the originof transfer. As a consequence, in segregationanalysis of recombinants formed from suchmerozygotic elements, there would appear to be agreater frequency of crossovers occurring in somechromosomal regions than in others. There wouldappear to be a decreasing crossover frequencyoccurring from the origin of chromosomal trans-fer to the terminus. Such a series of events wouldproduce a map like that presented in Fig. 1. Insuch a model, the observed exclusion of arg-i, al-though not absolute, would occur simply as a re-sult of the minimal probability of transfer of themarker most distal to the origin. Genomic ex-clusion would result from the transfer mechanismsand could be considered to be prezygotic as in thecase of E. coli (11).A second model, however, can account equally

as well for the reported results. In this model, azygote consisting of the total genome of both or-ganisms is formed. As a result of incompletechromosome homology, the pairing process couldbe expected to occur in direct proportion to thelength of the regions of greatest topological ho-mology. Thus, a decrease in crossover frequencywould occur from left to right on the linkage map(Fig. 1) as a consequence of increase, in the samedirection, of chromosomal heterology. Suchheterology may well exist in these nocardialstrains and may result from nothing more ge-netically complex than the occurrence of inver-sions and deletions, or both, during the evolutionof these organisms which are of diverse origin (3).Apparent topological differences in the N.erythropolis and N. canicruria chromosomes werereported earlier (5) and support this contention.Should this model prove to be correct, the ex-clusion of arg-I would be more apparent thanreal and could be readily explained by the pres-ence in N. canicruria of a suppressor gene forarg-i. If an arg-J suppressor (sup) is closely linkedto his-I+, prototrophic recombinants, postulatedto be produced as a result of two crossover events(e.g., the crosses of Table 2 and 3), would begenotypically his-l+ sup arg-l+ and phenotypi-cally Arg+. In matings of a strain in which arg-lwas added to the right of his-i (Table 5), recom-binants would be formed as a consequence of thesame two crossovers, with the production of thegenotype his-l+ sup arg-i. This genotype would bephenotypically Arg+ and would be indistinguish-

1758 ADAMS

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

GENE EXCLUSION IN NOCARDIAL MATING

able from recombinants containing the arg-l+gene. Therefore, Arg+ phenotypes produced incrosses of JA-SD 2-31 would result from only twocrossover events rather than three. The incon-sistency of the additional crossover necessary toproduce Arg+ phenotypes in this cross would beeliminated. In exceptional instances, in which supwas eliminated from recombinants as a result ofcrossover between sup and the selected his-l+,arg-l would be expressed (Table 6). A model em-bodying these concepts is consistent with the re-sults of this report. Consequently, the exclusion ofarg-i in recombinants, as explained by this model,would be phenotypic but not genotypic.At present, a methodology is being developed

which will permit more precise determination ofthe time of occurrence of nocardial mating andrecombination. With such a methodology, acomparison of the changes in recombinant pheno-typic frequencies with duration of mating will bepossible and should produce results establishingthe validity or lack of validity of the orientedtransfer model. Such experiments would be com-parable to interrupted mating techniques usedwith E. coli (11). However, experiments are alsobeing conducted in which recombinants contain-ing genes mapping to the right of his-i butoriginating from N. erythropolis strains are beingderived. These recombinants can be expected tocontain a chromosome more homologous withthat of JA-SD 2-31 and JA-SD 2-83, to the rightof his-i, than strains derived solely from JA-SD3. When fertile matings of such strains are accom-plished, direct determinations of the effect ofchromosomal homology on segregation of un-selected N. erythropolis genes will be possible.The completion of these experiments should per-mit a rational choice between the models pre-sented here to account for phenomena observedin nocardial conjugation and recombination.

ACKNOWLEDGMENTSThis investigation was supported by Public Health

Service grants GM 12008 and Research Career De-velopment Award 5K03 GM 17214 from the NationalInstitutes of General Medical Sciences, FR 5421 from

the General Research Support Branch of the NationalInstitutes of Health, and Training Grant Al 00232from the National Institute of Allergy and InfectiousDiseases.

I thank J. M. McGuire, Eli Lilly & Co., for giftsof crystalline erythromycin and Susan McNeil andJane Faulstich for excellent technical assistance invarious phases of these studies.

LITERATURE CITED

1. Adams, J. N. 1964. Recombination betweenNocardia erythropolis and Nocardia cani-cruria. J. Bacteriol. 88:865-876.

2. Adams, J. N., and S. G. Bradley. 1963. Recom-bination events in the bacterial genus Nocardia.Science 140:1392-1394.

3. Adams, J. N., and N. M. McClung. 1962. Com-parison of the developmental cycles of somemembers of the genus Nocardia. J. Bacteriol.84:206-216.

4. Adelberg, E. A., M. Mandel, and G. C. C. Chen.1965. Optimal conditions for mutagenesis byN-methyl - N' - nitro - N - nitrosoguanidine inEscherichia coli K12. Biochem. Biophys. Res.Commun. 18:788-795.

5. Brownell, G. H., and J. N. Adams. 1967. Linkageand segregation of unselected markers inmatings of Nocardia erythropolis with Nocardiacanicruria. J. Bacteriol. 94:650-659.

6. Demerec, M., E. A. Adelberg, A. J. Clark, andP. E. Hartman. 1966. A proposal for a uniformnomenclature in bacterial genetics. Genetics54:61-76.

7. Hayes, W. 1953. The mechanism of geneticrecombination in Escherichia coli. Cold SpringHarbor Symp. Quant. Biol. 18:75-93.

8. Hopwood, D. A. 1967. Genetic analysis andgenome structure in Streptomyces coelicolor.Bacteriol. Rev. 31:373-403.

9. Sanderson, K. E. 1967. Revised linkage map ofSalmonella typhimurium. Bacteriol. Rev. 31:354-372.

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

11. Wollman, E. L., F. Jacob, and W. Hayes. 1956.Conjugation and genetic recombination inEscherichia coli K12. Cold Spring HarborSymp. Quant. Biol. 21:141-162.

VOL. 96, 1968 1759

on August 24, 2020 by guest

http://jb.asm.org/

Dow

nloaded from