Reprint from Generics, Vol. 47, No. 10, October, 1962. Printed in U.S.A.

RECOMBINATION STUDIES OF LACTOSE NONFERMENTING
MUTANTS OF ESCHERICHIA COLI K-12'?

ANN COOKS anp JOSHUA LEDERBERG
Department of Genetics, Stanford University School of Medicine, Palo Alto, California

Received May 15, 1962

TH fermentation of lactose is a diagnostic character of Escherichia coli; how-
ever, nonfermenting mutants can readily be obtained. These mutants may
show losses or changes in the accumulation of lactose or other @-galactosides in
the cell, or in the formation or activity of the enzyme, £-galactosidase, which
splits lactose to the constituent monosaccharides. Such mutants have played an
important part in the analysis of gene action (Jacos and Monon 1961).

A comprehensive model of the action of the Lac region has been formulated by
Jacos, Perrin, SANcHEz and Monop (1960) (see Figure 1). The Lae region is
an “operon” comprised of four elements:

y, responsible for the structure of the relevant permease;

z, controlling the structure of £-galactosidase;

i, which determines whether the permease and enzyme will be produced only
in the presence of an inducer (7+), or synthesized constitutively (i-);

o, the operator coordinating the other elements of the region.

The first lactose mutants analyzed in this laboratory were designated Lac,-—
Lac; without regard to their specific physiological effects (LEDERBERG, LEDER-
BERG, ZINDER and Livery 1951). Lac, and Lac, were the most thoroughly in-
vestigated. Crossing analysis (J. LepErBerG 1947; Cavatur-Srorza and JinKs
1956; Wotioan, Jacos and Hayes 1956; Harrman 1957) located Zac, between
V, and V,, loci controlling resistance to the phages T, and Ts, respectively (Figure
1). On the basis of cis-trans tests of heterozygous diploids, several recombina-
tionally distinct mutations were classified in the same cistron in Lac, (LEDERBERG
et al, 1951). Others, which did not recombine, could be distinguished by their
reverse-mutability (E. M. Leperserc 1952). Lac, and Lac, were closely linked,
but complemented each other (LEDERBERG et al. 1951).

In the early studies (LevEnBERe et al, 1951), Lac, and Lac, were found to lack
8-galactosidase, i.e. to have a z~ defect. However, Lac; showed appreciable levels

1 This work supported by grant G6411 of the National Science Foundation, and grants
26295 and C4496 of the National Institutes of Health, United States Public Health Service,
Washington, D.C.

2 Submitted in partial fulfillment of requirements for the degree of Doctor of Philosophy in
Genetics at Stanford University.

3 Present address: Syntex Institute for Molecular Biology, Stanford Industrial Park, Palo
Alto, California.

Genetics 47: 1335-1353 October 1962.
1336 ANN COOK AND JOSHUA LEDERBERG

Hir,

Hfr. a Lac ee

Mtl

Xyl

Mal

 

Ficure 1.—Genetic map of E. coli. Distance between markers is expressed in minutes.
(Adapted from Taytor and Apetserc 1960; and Jacos and Monon 1961.)

of the enzyme. This anomaly appeared to be clarified by the discovery of the
permease system for the entry of 8-galactosides, Lac, being classified as a y~ defect
(RickENBERG, CoHEN, Burrin and Monop 1956; Rorman 1958).

Lac; (LeperBenc et al. 1951) affects the fermentation of glucose and maltose,
as well as lactose. Both its physiology and genetics remain obscure (PARDEE,
JAcos and Monon 1959). The same is true of Lac;, which fails to split maltose or
to ferment gluconate. Lac, is recorded as producing “reduced but significant
amounts” of both B-galactosidase and permease (Parpee et al, 1959).

Parvee ef al. (1959) located ten y~ mutations clustered at the left of the Lac
region toward P, and several z~ mutations to the right of y. They could order eight
of the z~ mutations in relation to each other and an i mutation by three-point
crosses between pairs of z~ mutants, one member of each pair also being i-. The
z+ progeny were scored for the presence of i-. Although the method is sound if
all the z~ mutations lie on one side of 7, there was one exception which could not
be mapped by this procedure. In addition, the method is highly laborious at best,
and results can be obscured by high rates of coincidence. Deletion analysis
(McCiintockx 1944; Benzer 1959), because of its greater efficiency, should
permit ordering much larger numbers of mutations, as well as providing an
RECOMBINATION ANALYSIS 1337

independent estimate of their order to compare with more conventional types of
mapping.

This paper describes an attempt to map a series of Lac mutations by deletion
analysis. The procedure entails the isolation of a series of overlapping deletions,
used as testers against point mutations to be mapped.

Terms and Abbreviations

The following terms and abbreviations will be used.
ONPG o-nitropheny]-8-p-galactopyranoside
TMG methyl-8-p-thiogalactopyranoside
TDG £-p-galactosyl-8-p-thiogalactopyranoside
Permease refers to galactoside permease.
Genetic loci:
Nutritional requirements (~ indicates dependence)
M, methionine P, proline LE, leucine
T, threonine Th, thiamine Ad, adenine
Pur, purines (optimal growth with thiamine and adenine or guanine)
Fermentation markers (- indicates inability to ferment)

Ara, arabinose Gal, galactose Lac, lactose
Mal, maltose Mtl, mannitol Xyl, xylose
Resistance markers (" indicates resistance)
V, phage T, V, phage T, S streptomycin
Sexual compatibility characteristics
F- female, or recipient F+ infective, wild-type male

F’ infective male showing high frequency of recombination (Jacop and
ADELBERG 1959).
F’, injects chromosome in the order Ara, Lac, Gal,. (Hrrota and SNEATH
1961). Hfr male showing high frequency of recombination, not infective
Hfr, (Cavaii-Srorza 1950) injects chromosome in the order Pur, or V4,
LacPLT.
Hfr, (Hares 1953) injects chromosome in the order T ZL Lac Gal (SKaar
and Garen 1956).
The only infertile combination is F- x F-.

EXPERIMENTAL DESIGN, MATERIALS AND METHODS

Steps in the program: 1. Isolation of a series of Lac mutants. 2. Systematic
intercrossing, screening (qualitatively, for the most part) for the occurrence of
Lac* recombinants. 3. Tabulation of results to: a. distinguish the deletions from
point mutations b. map the point mutations within specified deletions.

Media: The media employed included Davis mmimal medium, complete
medium (YZ broth), minimal eosin-methylene blue medium without succinate
(EM), and complete eosin-methylene blue medium (EMB). These are defined
in J. Leperserc (1950a) as A, B, C, and D, respectively.

Supplemented Davis minimal agar contained the amino acids at double the
1338 ANN COOK AND JOSHUA LEDERBERG

concentration specified in J. LeperBERc (1950a). 1.6 percent agar and 0.5 percent
glucose were used.

0.2 percent sugar was routinely added to Davis minimal broth after autoclaving.

200 wg per ml of streptomycin sulfate was incorporated when necessary.

Penassay broth, or Difco Bacto Antibiotic medium 3, was employed as a
standard complete liquid medium.

Cultures were routinely incubated at 37°C. Those in liquid media were gen-
erally not aerated.

Cultures were stored at room temperature in corked stab tubes containing 0.8
percent Difco nutrient broth and 0.8 percent Difco Bacto agar.

Strains: Most of the Lac mutations studied were induced by ultraviolet irradia-
tion. The parent strain, W3787 (Hfr, Lact M~ P- V,"), was streaked out on
EMB-lactose agar. Several Lac+ single colonies were isolated to Penassay broth,
and incubated overnight. Samples of 0.1 ml were spread on plates of EMB-
lactose agar, a maximum of 20 plates from each isolate. The plates were then
irradiated with ultraviolet at a dose that would leave 100 to 300 colony formers
per plate. Mutants were identified by their pale color on the EMB-lactose. They
were purified by at least two successive single colony isolations on EMB-
lactose agar.

Nineteen Lac- mutants were isolated during experiments with nitrous acid.
Concentrations of 0.025 m or 0.05 m of KNO, in citrate or acetate buffer at pH
values between 4 and 6 were used. It is questionable, however, how many of the
mutations the nitrous acid induced, as six mutants appeared in the citrate control
experiments.

Lac11D3 and Lac14D7 were collected by spreading Lac~ stocks having the
mutable markers, Lact? and Lac14, respectively, on EMB-lactose agar, and
choosing Zac colonies that had no secondary Lac+ colonies after three to four
days incubation.

Each mutation, including those previously described at the loci Lac, through
Lac,, was arbitrarily assigned an identifying Lac number. When possible, the
same number used as the locus designation was employed as the Lac number,
e.g. Lac, and Lac2 refer to the same mutation. When more than one mutation had
been described at the same locus, however, each was renumbered (Coox 1958).

For the study of recombination patterns, F- Lac~ prototrophs were derived
from the Hfr, P- M- V," Lac- mutants by crossing them with either Y10 (F-
T- L- Th-), or its streptomycin-resistant offspring, W1394.

In some cases, stocks having particular Lac markers in different genetic back-
grounds were necessary. These stocks and their manufacture will be described in
connection with the experiment for which they were used.

Table 1 lists some of the more important foundation stocks.

Crosses: Simplified procedures were required to screen Hfr, Lac- X F- Lac-
crosses for the formation of Lac+ recombinants.

Mixed cultures: one-half ml samples of overnight broth cultures of each parent
RECOMBINATION ANALYSIS 1339
TABLE 1

List of sorne foundation stocks

 

Y10 E-T- L- Th W4506 F- Pur V,"

W1394 F-T-L-Th S* W4607 F’, Gal,

Wi895 Hfr, M- W4983 F-M-V 7 S* Lac59 Suppressor +
W2612 F-T-L-Th S* Gal- W4984 FY, M-V,," St Lac59 Suppressor +

W3787 Hfr, P-M-V,7

 

were mixed in five ml of Penassay broth, incubated for two to three hours, then
samples spread on selective media.

Plate crosses: 10-* to 107+ ml of broth cultures were mixed and spread directly
on plates of selective medium.

Spot tests: drops of overnight broth cultures were deposited upon plates of suit-
able selective medium with a large platinum loop, or a fine-tipped inoculating
pipette, and allowed to dry on the agar. Each spot was then inoculated with
a drop of the other parent culture. A loop contained approximately 5 x 10-3 ml,
or 5 X10° cells, while a drop from a pipette was ten times larger.

Infection of F- with F'; To infect an F- strain with F’ (e.g. F’,), overnight
cultures were grown in Penassay broth of the F- to be infected, and of an F’
strain which differed from it by at least one fermentation marker. One ml of the
F’, and 0.1 ml of the F- were then mixed in five ml fresh Penassay broth, and
incubated overnight. The mixture was streaked out on medium which would
distinguish the two stocks. Appropriate colonies were picked to broth and tested
for fertility by spot testing against an F- with an auxotrophic marker in the
region the F’ transfers with high frequency. This procedure is essentially the
same as that for the transmission of F (LepERBERG, CavaLir and LEDERBERG
1952). Under the conditions described, at least 50 percent of the F- culture be-
comes FE” (Hirota, unpublished).

T, resistance: Strains were tested for resistance by the cross-streaking method
(Demerec and Fano 1945) on EMB-agar without sugar. Bacteria were inoculated
either by a loop from broth cultures, or by replica plating (LEDERBERG and
LEDERBERG 1952).

ONPG tests: B-galactosidase activity was tested for on agar and in broth (J.
Lepverserc, 1950b).

The medium for tests on agar was Davis minimal supplemented with 0.8
percent NZ Case (Sheffield Chemical Co., Norwich, Conn.) and either 0.5 percent
glucose (noninducing medium) or 0.5 percent lactose and 0.5 percent glycerol
(inducing medium). Small drops of the strains to be tested were put on the plates
and incubated overnight, then exposed for one to two hours to chloroform vapor.
The plates were righted, and each colony flooded with a drop of a solution of one
mg per ml of ONPG in 0.05 m sodium phosphate buffer, pH 7.5. Tests were read
when the yellow color developed in control areas. In general, the presence of
revertants in the cultures was looked for by spotting them on EMB-lactose or
1340 ANN COOK AND JOSHUA LEDERBERG

EM-lactose at the same time that they were inoculated onto the test medium.

Strains not giving positive reactions on agar with lactose as an inducer, were
grown in YZ, broth with {0-? m of TMG. The cultures were centrifuged and the
pellets resuspended in 0.05 m sodium phosphate buffer, pH 7.5. A few drops of
toluene were shaken with the cells. An ONPG solution was added to give a final
concentration of 3 X 10-* m. The mixtures were incubated at 37°C for ten minutes
and then compared to the control, W3787, by eye.

Permease tests: Single colonies from streakings on EMB-lactose were inoculated
into YZ broth containing 10° m TMG. (Dr. Metvin Coun kindly supplied the
thiogalactosides, together with advice on their use.) After eight to ten hours’
growth at 30°C with shaking, the cultures were diluted 1 to 4 with fresh medium
of the same composition. After 60 minutes incubation at 37°C with shaking, the
cultures were centrifuged ten minutes in the cold at 17,300 X g, and washed with
five ml of cold Davis minimal containing 10-* m glucose, and 10-* m mercapto-
ethanol. The pellets were resuspended in the same medium. Cell densities were
adjusted to 2-5 < 10® viable units per ml. Drops of the cell suspensions were
spotted on EMB-lactose for reversion controls, One ml of each cell suspension was
added to 0.1 ml of 10°? wm cold TMG containing enough C** labeled TMG to give
3 x 10° counts per minute per ml. After 15 minutes incubation with shaking at
37°C, the suspensions were filtered through Millipore membranes (HA), and
washed with ten ml of cold Davis minimal into which 10-* m glucose, and 10-* Mm
mercaptoethanol had been incorporated. Each filter was glued to a planchet with
rubber cement, dried and counted in a windowless gas flow counter for a period
of two minutes.

EXPERIMENTAL RESULTS

An Hfr, P- M-V,," stock of K-12 was chosen to provide the source of Lac
mutants because Hfr, transfers the Lac region with high frequency (SkaarR and
GarEn 1956), and because P and V, ere the outside markers closest to Lac
(Figure 1). From each Hfr Lac mutant stock, a corresponding F- Lac” prototroph
was derived by crossing. Matings were then made between different Hfr and F-
stocks on EM-lactose medium, and each cross scored for the production of Lact
prototroph recombinants. Mutants were also inoculated onto EMB-maltose,
EMB-glucose, and EMB-galactose to test for the fermentation of these sugars.
Results of the recombination tests suggested the following grouping for the Lac
mutations studied.

(1) Short deletions: A deletion is defined as a mutant which fails to recombine
with two or more recombinationally distinct mutants. When tested for reversion
by spreading 10° cells from fresh broth cultures on each of ten EMB-lactose
plates, no Lac+ colonies or papillae appeared even after seven days’ incubation.

(2) Long deletions: These have not been observed to yield Lact progeny with
any of the other Zac mutations (with the possible exception of the Nr group. see
below). When tested for reversion as outlined above, they also produced no Lact
clones.
RECOMBINATION ANALYSIS 1341

(3) Point mutations: These are mutants that cannot be recognized as deletions
by their recombination patterns (Figure 2). In general, they recombine with
every other point mutant. A few have been found which fail to recombine with
one another, but react identically with all other testers. Future tests against ad-
ditional markers could indicate the reclassification of a point mutant as a short
deletion. Most of the point mutants in the present classification are revertible,
however. The point mutants have been divided into three subgroups on the basis
of their recombination with the short deletions, Lac/1D3 and Lac39. Group A
members recombine with Lac39. Group B members recombine with Lac/1D3.
Group C members recombine with both Lac3? and Lac11D3.

(4) Nr mutations: These give anomalous reactions with certain other Lac
mutations. When the F- parent contains an Nr mutation, and the Hfr the other
Lac mutation, no Lac+ progeny are recovered (less than 10-7 Lac+ recombinants
per viable unit of the limiting parent). When the Hfr contributes the Nr muta-
tion to an F~ bearing the other mutation, Lac+ progeny appear with a frequency
of one per 10* cells. Most of the cases observed involved a long deletion as “the
other mutation.”

In general, the mutants that gave negative fermentation reactions on EMB-
glucose, EMB-maltose, or EMB-galactose, grew too well, or produced too much
color on EM-lactose to enable the determination of their recombination charac-
teristics by this method, Nine mutants formed pale colonies on all three sugars,
and thus resembled Lac;.

After the mutants had been divided into groups, the members of each group
were tested more extensively with each other for recombination. Those not in-
cluded were too revertible, produced too much color on EM-lactose, or were no
longer Hfr. Only a few of the long deletions were investigated. Drop tests were
made of all possible combinations within each group, and each series included
the standard testers, Lac85 (a long deletion), Lac11D3 (a short deletion),
Lact4D7 (a short deletion identical to Lacf1D3 except that it does not recombine
with Lac91), Lac39 (a short deletion distinct from both Lact1D3 and Lac14D7),
and Lac91 (the Group C point mutation that distinguishes Lac1/7D3 from
Lact4D7). Each series was repeated at least three times, using overnight broth
cultures inoculated from single colonies or directly from storage tubes for com-
parison. Syntrophy of the F- Lac” prototroph with an F- M~ stock, reversion of
the F- Lac” prototrophs, and their ability to give Lact progeny with W3787
(Hfr, P- M- V," Lact+) were looked for in each case by similar drop tests.

The results of all the crosses appear in Figure 2. It can be seen that the tests
failed to differentiate among the long deletions. As none of the deletions yielded
Lact progeny when crossed with Lac11D3 or £ac39, only one of them, Lac85,
was tested extensively against individual members of Group A and Group B. It
should be noted, however, that the preliminary tests to group the long deletions
involved spot tests with two Group A point mutations and two Group B point
mutations. In addition, the long deletions Lac61, Lac72, Lac?74, Lac75, Lac81,
and Lac86 were spotted at least once against 14 members of Group A and eight
members of Group B. The results were uniformly negative. It was thought, how-
1342 ANN COOK AND JOSHUA LEDERBERG

ever, that the Group C mutations might well be scattered along the Lac region,
with some located at its ends. In this case, they might react differently with the
various deletions if the latter varied in length. Fifteen long deletions were drop
tested three times with members of Group C, as well as the standard testers. As
none of the crosses gave rise to detectable numbers of Lac+ recombinants, the
long deletions all appeared to be identical. Neither could they be used to deduce
the linear order of Group C mutations.

There were several sets of mutations, for example, Lac96, Lac53, Lac16 and
Lact94, that failed to recombine with each other. Since members of such a set
showed identical recombination patterns in crosses with other stocks, they were
considered to be point mutations, recurrences at a single site.

Figure 2 also illustrates the fact that only three short deletions, Lac39,
Lact1D3, and Lact4D7, were discerned (possibly four, if Lac215 is a short dele-
tion; see “Anomalous recombination”); Because of this, Bewzer’s method of
locating point mutations from their behavior with overlapping deletions was not
applicable to the system.

Nineteen Lac~ mutants obtained after exposing W3787 to nitrous acid, and
six from the citrate buffer controls, were briefly examined for short deletions.
Twenty-two of them, or 88 percent, behaved as long deletions, an increase of 46
percent above the proportion obtained with ultraviolet light. The rest were point
mutations, one fitting the Group B criteria (but not included in Figure 2), and
two, Lac334 and Lac337, belonging to Group A. Five of the long deletions,
Lac335, Lac340, Lac341, Lac344, and Lac350, were included in the tests repre-
sented in Figure 2. They appeared to be identical to the long deletions obtained
after ultraviolet treatment of W3787.

The first group of Lac” mutants obtained after ultraviolet irradiation of W3787
included 22, or 37 percent, which behaved as long deletions (Table 2). Seven
different Lact single colonies had provided the inocula for the irradiations that

TABLE 2
Incidence of different types of mutations among Lac~ mutants of W 3787 (Hfr, P-M-V,")

 

 

Series 1 Series 2
(W4001-W4060) (W5001-W5233)
Number Percent Number Percent
Type of mutation
Long deletion 22 37 100 43
Short deletion 0 0 0 0
Point mutation 23 38 95 41
Group A 10 17 52 22
Group B 7 12 26 11
Group C 6 10 17
Anomalous mutations (Vr) 5 8 14 6
Not classified
Not Hfr 1 2 6 3
Revertible, or able to
grow on E-.M-lactose 9 15 it 5

Other 0 0 7 3

 
RECOMBINATION ANALYSIS 1343

furnished the 60 mutants in this series. All seven of them gave rise to at least one
deletion mutant. Among the second set of mutants similarly collected, were 100,
or 43 percent, deletions. Forty-two of the 48 single colonies from which these 233
mutants were derived contributed deletions. In order to determine whether this
same high proportion of Lac deletions was characteristic of other K-12 derivatives,
three were irradiated with ultraviolet, according to the same procedure used with
W3787. Two were F- prototrophs unrelated to W3787, and one was the Hfr,
progenitor of W3787. More than 20 mutants were obtained from each, and tested
to see whether any were long deletions. None were found. The probability of
such a result being due to chance is less than one percent in each case (Table 3),
indicating that W3787 has a special propensity to yield long deletion mutants.

Tests for B-galactosidase

All of the mutants, and most of their F- Zac~ prototroph derivatives, were
qualitatively screened for the production of active @-galactosidase. Cultures were
first tested as colonies on agar, with 0.5 percent lactose as inducer. All members
of Group A responded positively. Lac?/D3, the short deletion which covers the
segment to which these mutations belong, did not hydrolyze ONPG in these con-
ditions. All of the long deletions tested, as well as the short deletions, Lac14D7
and Lac39, failed to give positive reactions. Group B and C point mutations in
general gave negative responses. The exceptions were Lac22, Lact55, and Lac277,
all of Group C. In each of these cases the reaction was less pronounced than that
of the Lact control, W3787. Table 4 illustrates some of the results.

Because lactose is considered a poor inducer for bacteria defective in permease,
cultures which showed no activity on agar with lactose induction were tested in
broth with TMG as inducer (Table 4). Lac85, Lact11D3, Lact14D7, Lac39, and
most of the members of Groups B and C still failed to give detectable reactions
after ten minutes’ incubation with ONPG. (Some did show color after 24 hours).
Three mutants which had previously been negative, Lac60, and Lac299 of Group

TABLE 3

Incidence of long deletions among Lac- mutants of four E. coli K-12 stocks

 

 

Number of Number of

independent mutants not
Source long deletions deletions Total
W3787 (1st series) 7 38 45
‘W3787 (2nd series) 42 133 175
W4955 0 23 23
W3100 0 30 30
W1895 0 27 27

W3787 (1st series) compared to W3787 (2nd series)

Observed x? == 1.37 P — 20-30 percent
W53787 (pooled) compared to W4955 P—0.42_ percent
W 3787 (pooled) compared to W3100 P=0.09 percent

W3787 (pooled) compared to W1895 P=0.17 percent

 
1344 ANN COOK AND JOSHUA LEDERBERG
TABLE 4

Examples of results of tests for B-galactosidase and permease

 

 

Reaction with ONPG TMG accumulation
Lactose TMG (Percent of control)
Lac designation induced induced (Lac39)
Long deletions
Lac72 — not tested 114
Lac85 _— —_ 4
Short deletions
Lact1D3 — — 4
Lac39 _ — 100
Point mutations
Group A
Lac17 + not tested
Lac2t + not tested 1
Lact29 + not tested 15
Lac269 + not tested 1
Group B
Lac60 — x 144
Lac89 — — 14
Lac97 — — 39
Lac212 -— — 13
Lac264 — — 6
Lac299 — ba 132
Group C
Lac22 na + 16
Lac88 — + 13
£ac155 & x 18
Lact65 _— — 5
Lac256 — — 90
Lac277 ad — 32
Nr Group
Lac43 —_ — 12
LacSt — # 11
Lac59 + + 11
Lac66 — — 12

 

B, and Lac88 of Group C, responded weakly. On the other hand, one mutant of
Group C, Lac277, would have been classified as negative on TMG, but produced

slight but perceptible color on lactose. Lac22 and Lac155 reacted weakly in the
presence of either inducer.

Tests for permease

Each series of tests included W4991, which has the original Lac, marker
(Lac3?). RickENnBERG et al. (1956) and Rorman (1959) have shown that an-
other strain (W2244) with the same Lac- mutation has the capacity to accumu-
late TMG after induction with the same compound. Twelve separate determina-
tions on induced cultures of W4991 averaged 1788 cpm (counts per minute) per
mi of culture (corrected for 81 cpm nonspecific radioactivity). Using 250 yg per
10° cells for the dry weight of E. coli (Roserrts et al. 1955),this is equivalent to
RECOMBINATION ANALYSIS 1345

82 » moles of TMG accumulated per gram dry weight of cells. The range was 30
to 142 » moles per gram dry weight of bacteria. Although these values compare
with those in the literature, the wide range found here made it difficult to establish
limits for classifying mutants as RicKENBERG ef al. (1956) have done. Some
general observations may be made regarding the results, however.

Group A mutants showed the least amount of accumulation of TMG, and the
greatest uniformity of responses within a group. Lact/D3 also showed reduced
accumulation (Table 4).

It was expected that all Group B mutants would accumulate TMG, since
Lac39 does. As Table 4 illustrates, however, responses were extremely hetero-
geneous, with some mutants showing no more accumulation than members of
Group A. Lac89, Lac264, and Lac212, which gave low values, were tested for
reversion by spreading EMB-lactose plates with 0.1 ml of overnight Penassay
broth cultures and incubating them for several days. Lact papillae appeared on
all the plates, indicating that the markers tested are probably not short deletions.

Group C mutants also showed heterogeneous responses (Table 4). Since there
is no single deletion that spans the Group C mutations as Lac39 does the Group
B members, such a result was not so surprising.

Seven long deletions were examined. All were negative in accumulation tests.
In general, their values were slightly higher than those of the Group A mutants.
The highest value was registered by Lac72, reported elsewhere (Luria, ADAMS
and Tinc 1960) tober yz.

Anomalous recombination

Although in general, a cross in which one of the parents contained a long Lac
deletion produced no Lact progeny unless the other parent was Lact, there
were some peculiar exceptions. As mentioned before, a small number of muta-
tions gave about 10~* (one per 10‘ cells of the limiting parent) recombinants on
EM-lactose when the Hfr was crossed to an F- carrying a long Lac deletion. In
contrast, when the Hfr transferred the deletion, less than 10-7 Lact offspring
were recovered. Of the 293 mutants of W3787, 19 showed this property, and so
did Lac2. Crosses of these mutations with point mutations revealed no anomalies,
as all of them produced Lact recombinants.

The hypothesis proposed to account for this phenomenon is that in Y10 and
its S’ derivative, W1394, there is a factor, or factors, which may be called a sup-
pressor, to which certain of the Lac mutations are sensitive. Bacteria which have
both the suppressor and one of the Lac mutations which responds to it, are able
to grow on EM-lactose, and to produce dark colonies on EMB-lactose. Bacteria
which have only the Zac” mutation, but not the suppressor, do not grow on EM-
lactose, and produce only white or pale pink colonies on EMB-lactose. As the
suppressor does not noticeably affect the expression of the majority of the Lac
mutations, especially the long deletions, there may be F- Lac” prototroph strains
derived from crosses of the Hfr Lac” mutants and Y10 or W1394, which include
the suppressor. If the suppressor is located on that portion of the genetic map
1346 ANN COOK AND JOSHUA LEDERBERG

between M and the end of the chromosome segment donated by Hfr, (on the M,
S, Gal... segment), the number of such strains may be quite large, as the nature
of the crosses precluded the Hfr parent from contributing the alleles in this region
very frequently (RicuTer 1957).

Accordingly, if an F- Lac” having both a long deletion and the suppressor is
crossed with an Hfr having a mutation sensitive to the suppressor, Lac+ progeny
will arise which include the suppressor and the sensitive mutation. In the reverse
cross, the F~ parent has no suppressor, only the suppressor-sensitive mutation
(otherwise it would not be Lac-), while the Hfr has the deletion and no suppres-
sor. No Lac* offspring appear since the deletion affects some or all of the genetic
material altered by the suppressor-sensitive mutations.

Testable predictions based upon this theory include: (1) at least two types of
recombinants able to grow on EM-lactose should occur in a cross between Y10 or
W1394 and an Hfr with a suppressor-sensitive mutation, a true Lac*, and a sup-
pressed Lac”; (2) Lac+ recombinants from crosses involving a deletion and an-
other Zac- mutation should contain both the Lac” mutation and the suppressor,
i.e. Lac” progeny should be recoverable from a cross between such Lact recom-
binants and a Lact strain which has no suppressor.

Evidence for a suppressor

In the experiments to be described, Lac59 and Lac66 were chosen as examples
of mutations participating in the anomalous reactions. Lac85 represented a
typical long deletion, and Lac89 a normal point mutation.

Usually from a cross involving W1394 (F- T-E- Th: 8”) and a Lac~ mutant of
W3787 (Hfr, P- Lact M- V,"), a large number of the recombinants able to grow
on Davis minimal plus streptomycin with or without proline, were Lac. For
example, if the Hfr was Lac85, or Lac89, more than 50 percent of the recom-
binants were Lac~ (Table 5). In addition, most of them were V,”, When the Hfr
was Lac59, however, less than one percent of the recovered progeny were Lac-.
Instead, about 50 percent were phenotypically intermediate between the Lact
F- and the Lac- Hfr. Most of the intermediates, Lac*, were V,” (Table 5).

To select recombinants which received the terminal portion of the chromosome
from the Hfr, a Gal derivative of W1394 (W2612) was used as the recipient. If
the cross was plated on EM-galactose, or EM-galactose plus methionine (Figure
3), the proportion of Lac" progeny approached 30 percent.

TABLE 5

Incidence of different types of recombinants from crosses of W1394 (F- T-L- Th- S") with
various Hfr, Lac P- M- V,," stocks. Selection on Davis minimal plus proline

 

 

fac Lact Fact Total number
examined
ver Vs V," V,° Vv," Ve
W1394 x W4045 (Lac85) 45 5 4 34 0 0 88
W1394 x W4049 (Lac89) 53 5 6 28 0 0 92
W1394 x W4019 (Lac59) 0 G 2 33 57 3 95

 
RECOMBINATION ANALYSIS 1347

Hfr, WHOld ver Lacs9 FP OOo@ vw s? @

 

- i. en
Fo wo612 Veo ot @ LUT Tho + 8! Gal

 

Figure 3.—Diagram of W4019 x W2612 plated on EM-galactose + methionine. Circles
indicate selected markers. If the same cross is plated on EM-galactose, the M* from the F- will
also be selected.

This supported the idea that the Lac’ phenotype was associated with a factor
on the distal portion of the Hfr chromosome. If this factor was the (inactive)
allele of a suppressor contained in W1394, the intermediate phenotype might
characterize recombinants which received Lac59 from the male, but retained
the suppressor from the female. If this be so, Lac” offspring should appear in a
mating of a clone of intermediate phenotype with a Lac+ which does not have
the suppressor.

W4506 (F- Pur-) gave Lac’, but no Lac* recombinants in crosses with W4019
(Hfr, Lac59 P-M-V,"). In interrupted mating experiments (Woriman and
Jacos 1955) Lac” appeared to enter about 17 minutes after the start of mating,
and about 15 minutes after the entry of the Pur marker, as expected if the Lac”
phenotype depended upon a factor in the region between V, and P. W4506 could
thus be used as an indicator for the presence of Lac59 in Lac* recombinants.

From a mating of W1394 (F- 7- L- Th: 8") with W4019 (Hfr, Lac59 P- M-
V.") an F- Lac* M- S’ (W4983) was chosen, After purification on EMB-lactose
agar, it was infected with F’, by growth in mixed culture with W4607 (F’, Gal,).
The repurified F’, Lac* M- S" (W4984) was crossed with W4506 and selection
made on Davis minimal (Figure 4). Of 94 colonies picked and tested, 32 were
Lac* (like W4506), 42 were Lac* (like W4984), and 20 were Lac-. Accordingly,
W1394 may be said to be Lact Sut, W4019 to be Lac59 Su-, and the intermedi-
ate type to be Lac59 Sut+. The phenotype of the other possible recombinant,
Lac* Su, is expected to be Lact, like W1394.

The proportion of Zac” among offspring of the W4984 x W4506 cross may
argue that the suppressor is fairly close to Pur. Unfortunately, the cross of
W2612 x W4019 did not readily permit locating the suppressor. There appeared
to be interaction between Gal and some factor near S, so that when selection was
made for Gal+ S” prototrophs on EM-galactose plus streptomycin, the colonies
which grew were not Galt on EMB-galactose, but of intermediate phenotype.
Since a few S’ Gal+ prototrophs (about five percent) have been obtained, the

Fi, wig8h . taco @ sy ow

Fo Wh506 + Pur SS

Ficure 4.—Diagram of W4984 x W4506 plated on Davis minimal. Circles indicate selected
markers. Note that the direction of transfer of F’, is opposite to that of Hfr,, and that the point
of origin of F’, is just before L. The distance between Pur and S is very great (cf. Figure 1).
1348 ANN COOK AND JOSHUA LEDERBERG

factor cannot be S" itself. Since Lac+, Lac*, and Lac~ phenotypes have all been
observed among the Gal+ offspring, it seems unlikely that Lac59 or its suppressor
is to blame.

The most important prediction of the suppressor idea is that anomalous lactose
positive progeny contain a Lac’ mutation and the suppressor, Lactose positive
phenotypes were tested by infecting them with F’, and mating them with an S”
derivative of W4506 (F- Pur-). None of the progeny of control crosses, W4145
(F- Lac85) x W3787 (Hir, Lact P- M-V,"), or W3133 (F- Lact1D3) x W4991
(Hfr, Lac39 P- M-V,"), sired anything but Lac*. All of the progeny tested from
anomalous crosses, nine from W4019 (Hfr, Lacd9 P- M-V,") x W4145 and ten
from W4026 (Hfr, Lac66 P- M-V,") x W4145, produced Lac- recombinants
with the Zac+ F-. In other words, the results substantiated the hypothesis.

Clones whose genotype included a suppressor and Lac59 or Lac66 were un-
stable. Colonies formed by them on EMB.-lactose often contained lighter or darker
colored sectors. In the case of Lac59, the colonies darkened gradually over a period
of two days or so, and sectors could be detected with certainty only after about 36
hours at 37°C. Colonies from cultures having Lac66 and its suppressor, however,
darkened rapidly. Sectors appeared by 24 hours and then were obscured by dark-
ening of the entire colony. Occasionally, whole colonies were found having either
a Lac” phenotype, or a very dark color almost like Lact. Lac” sectors and segre-
gants appeared in F- as well as F’, cultures with a frequency of approximately
one per 1000 viable units. These lactose negative sectors appear to contain new
suppressor-resistant Lac” mutants, which behave like long deletions. At least
they failed to recombine with Lac11D3, Lac39, and 16 different point mutants
from Groups B and C. They must also have had the suppressor, as they gave posi-
tive recombinants with both Lac59 and Lac66. When recovered from recombin-
ants of crosses with W4145 (F- Lac85) as the recipient, such Lac” behaved like
Lac85 in the tests employed, but when they occurred in clones derived from Lact
recipients, they resembled neither parental marker.

From the foregoing, it is evident that W4019 (Hfr, Lac59 P- M-V,,") typifies
a series of Hfr mutants having Lac- mutations which are sensitive to a suppressor
carried by W1394. It is not known whether W1394 has one suppressor which is
active upon all of the Nr mutations, or whether there exists a series of suppressors,
each one specific for a particular mutation. In connection with this, Lac215 can
be noted (Figure 2). This mutation does not recombine or behave abnormally
with the suppressor-carrying long deletion stocks. It does give a nonreciprocal
pattern of recombination with Lac139, Lac284, Lac229, however. This suggests
that the F- Lac215 derivative of W1394 is carrying a suppressor of different
specificity from the one characteristic of the long deletion stocks.

Other characteristics of suppressor-sensitive Lac mutations

The fact that W4506 (F~ Pur-) behaves normally with the sensitive mutations
enables one to use it to obtain F- Lac” prototrophs for recombination tests. Such
derivatives were made for Lac85, Lac61, Lac? 2, Lac74, Lac81, Lac86 (all long
deletions) , Zacf/D3, and Lac39 (both short deletions), and Lacd9 (an Nr muta-
RECOMBINATION ANALYSIS 1349

tion). These were drop tested on EM-lactose with the suppressor-sensitive muta-
tions, Lac2, Lac43, Lac45, Lac51, Lac59, and Lac66. Figure 5 presents the results.
The lesions in the Lac region seem to be located within the area covered by the
long deletions, with Lac45 and Lac59 at the same site within the Lacf1/D3 seg-
ment, The other suppressor-sensitive mutations act as different Group C muta-
tions, recombining with both Lacf7D3 and Lac39, but not with the long deletions.
Preliminary examination of other Nr mutations suggested that three of them,
Lac230, Lac231, and Lac237, might occupy the same site as Lac2. Tests were
not rigorous enough to warrant including the results in Figure 5.

A derivative of W4506 carrying Lac215 was also made, and tested against
Lac139, Lac284, and Lac229 (cf. Figure 2). No recombinants appeared in any
of the three crosses, indicating that Lac215 overlaps these other mutations.

Tests for 8-galactosidase and permease were made of some of the Nr mutants.
Lac45 and Lac59 both gave positive reactions with ONPG. Lac2, Lac27, Lac43,
Lac51, and Lac66 were all negative. All of the mutants tested, Lac43, Lac51,
Lac59, and Lac66, showed reduced accumulation of TMG (Table 4). Lac2 was
not tested, but according to the results of RickENBERc ef al. (1956), it should be
permease positive.

DISCUSSION

Studies of the problem of coding proteins from nucleic acids necessitate order-
ing large numbers of mutations. For this, the most efficient method employs over-
lapping genetic deletions (BENzER 1959). For the procedure, however, a suitable
set of overlapping deletions represents a sine gua non. Plainly, the Lac mutations
studied here do not meet this requirement, Deletions obtained from W3787 were
all too large to be of use in mapping. That they did not recombine with any of
the point mutations, together with the fact that those tested were deficient in both
B-galactosidase and permease means they are probably y~ z-. Lurra, Apams and
TinG (1960) used one of them, Lac72, and reported it to be i- yz. On the basis

of the gene order given by Jacos e¢ al. (1960), one might also expect it to include
o, between i and z.

 

 

Lac59
Lack5
Lace
Lac27
Lach3
Lac5l
Lac66
im
Nt D2 FAA A EM

— _ _— =™ \O

Qo eE wo _

Pee eee RRR BIIS SS
|| recombination Cd no recombination

Ficure 5.—Recombination of suppressor-sensitive mutations.
1350 ANN COOK AND JOSHUA LEDERBERG

The unusually high frequency of deletions encountered among the derivatives
of W3787 probably indicates the presence of a structural abnormality which
predisposes to large deletions, and may also make short deletions correspondingly
rare. Generally, deletion mutations occur more rarely than point mutations, but
at least two cases have been reported of loci unusually susceptible to deletions:
the region controlling T, resistance in Escherichia coli B (Gors, Kon and Hunt
1954; Yanorsky and Lennox 1959) and the CysC locus in Salmonella (DE-
MEREC 1960; DEMEREc e7¢ al. 1960). In the present case, the aberration does not
seem to be a property of W1895, the Hfr, M- progenitor of W3787; therefore, a
more rigorous comparison of the two stocks might lead to an elucidation of the
nature of the aberration.

The three short deletions available (Lacf1D3, Lact4D7, and Lac39) were not
obtained from W3787. Lac11D3 closely resembles Lac14D7, although they were
stable (Lac”) derivatives of two, mutable Zac; mutations, Lact1 and Lact 4,
respectively, The distinction between them is that Lacf4D7 does not recombine
with Lac91, while Lact1D3 does. Lac91 recombines with both of the original
Lac, mutations. Lac14 is so revertible that it has not been possible to determine
whether it occupies the same site as Lac//, but the two mutations originated in
different stocks, and have different reversion rates. Two other stable derivatives
of Laci11, and one of Lac14, had previously been tested (Cook 1958) and found
to be long deletions indistinguishable from each other or from other long dele-
tions, such as Lac85.

Group A point mutations failed to recombine with Lac/1/D3. As this is a two-
step mutation, and the two events need not have altered contiguous nucleotide
sequences, it would not have been surprising to find two sorts of mutations in
Group A, particularly since Lact1D3 itself decreased @-galactosidase activity as
well as permease activity. All members of Group A, however, showed reduced
permease activity, but made perceptible amounts of 8-galactosidase.

In addition to being derived from a Lac, mutant, Lacf1D3 does not recombine
with other Zac, mutants tested (e.g. Lact2 from W2241, cited by RicKENBERG
et al, 1956). In many cases the reversion rate of Lac, mutants approaches the
expected rate of recombination between them, making interpretation difficult.
However, it is fairly certain that Lac12 fits into Group A.

Lac39 is the Lac, prototype (LeperBenc et al. 1951; RicKENBERG et al. 1956).
Because it probably resulted from a single event, the deleted segment should span
an uninterrupted sequence of point mutations. Like Lac39 (i+ y+ z-), these
mutants should fail to make active @-galactosidase, but should have an intact per-
mease mechanism. Paradoxically, several mutants of Group B, located at different
sites, lack 8-galactosidase, and also have defective permease systems. Jacos and
Monop (1961) have described point mutations having reduced permease ca-
pacity and lacking @-galactosidase. They have not reported a case where such
mutants fail to recombine with a deletion having one of the activities missing in
the point mutants.

Of the original seven Lac loci, then, Lac, and Lac, can be most easily related
to the present system. The mutation at Lac, belongs to the group of mutations
RECOMBINATION ANALYSIS 1351

showing anomalous recombination. On the basis that such recombination is due
to a suppressor at a different locus, and not to crossing-over within the Lac region,
Lac, can be said to be in the Lac region. The finding of ParpEE et ai. (1959), that
phage 363 can transduce it jointly with other Lac mutations agrees with this.
Lac,, Lac;, and Lac, mutants grow too well on EM-lactose to tell by this method
whether they recombine with the long deletions or not, and thus whether they
are in the Lac region. The same was true of several mutants of W3787 with the
traditional Lac, phenotype.

While the original aim of comprehensive mapping of the Lac region has not
been accomplished, the deletion stocks offer useful material for the study of the
activity of the Lac gene; for example, the behavior of deletions in transductional
heterozygotes (Luria, Apams and Ting 1960; Reve., Luria and Rorman 1961;
Revet and Lurra 1961; Rever, Lurta and Younc 1961), or the discrepancy
between the biochemical properties of Lac39 and the point mutations it spans.

SUMMARY

Two hundred ninety-three Lac” mutants were collected after ultraviolet ir-
radiation of a stock of Escherichia coli K-12. Stocks for recombination analysis
were derived from them by appropriate matings. Drops of male and female
cultures carrying the mutations were paired on minimal lactose agar and scored
for the production of Lac*+ prototrophic recombinants. All of the mutations had
altered sites in one region located between the markers P and V,. Thirty percent
were long deletions. As the long deletions did not recombine with any of the
point mutations tested, they were indistinguishable from each other. The propor-
tion of deletions was significantly higher among Lac” derivatives of one stock of
K-12 (W3787) than among Lac~ derivatives of any of the other three stocks
tested,

The point mutations formed three groups on the basis of recombination with
two short deletions. Mutants belonging to one group all had active @-galactosidase,
but accumulated TMG only weakly. In these properties they resembled Lac, or
y mutants (RICKENBERG et al. 1956). Both the other two groups included some
mutants with reduced #-galactosidase activity, but normal TMG accumulation,
and some with both the f-galactosidase activity and TMG accumulating ability
diminished. A short deletion that retained the capacity to accumulate TMG
spanned several of the point mutations which changed both functions.

A fourth group of point mutations was found to be sensitive to the actions of a
suppressor factor, or factors, present in certain K-12 stocks.

As only three short deletions were available, no unique sequence for point
mutations could be obtained by the method of overlapping deletions.
1352 ANN COOK AND JOSHUA LEDERBERG
LITERATURE CITED

Benzer, S., 1959 On the topology of the genetic fine structure. Proc. Natl. Acad. Sci. U.S. 45:
1607-1620.
Cavatir-Srorza, L. L., 1950 La sessualita nei batteri. Boll. Ist. Sieroterap. Milan. 29; 281-289.

Cavatti-Srorza, L. L., and J. L. Jmnxs, 1956 Studies on the genetic system of Escherichia coli
K-12. J. Genet. 54: 87-112.

Coox, A., 1958 Recombination studies of mutations affecting lactose fermentation in Escherichia
coli K-12. M.S. Thesis. Univ. of Wisconsin.

Demerec, M., 1960 Frequency of deletions among spontaneous and induced mutations in
Salmonella. Proc. Natl. Acad. Sci. U.S. 46: 1075-1078.

Demerec, M., and U. Fano, 1945 Bacteriophage-resistant mutants in Escherichia coli. Genetics
30: 119-136.

Demerec, M., E. L. Lan, FE. Barsrnper, T. Mryaxe, J. Isnrpsu, K. Mizosucut, and B. MaHLer,
1960 Bacterial Genetics. Carnegie Inst. Wash. Yearbook 59: 426-428.

Gots, J. S., W. Y. Kou, and G. R. Hunt, Jn., 1954 Tryptophan metabolism and its relation to
phage resistance in Escherichia coli. J. Gen. Microbiol. 11: 7-16.

Harrman, P. E., 1957 Transduction: a comparative review. Pp. 408-462. The Chemical Basis of
Heredity. Edited by W. D. McExroy and B. Grass. The Johns Hopkins Press. Baltimore,
Maryland.

Hayes, W., 1953 The mechanism of genetic recombination in Escherichia coli. Cold Spring
Harbor Symp. Quant. Biol. 18: 75-93.

Hirota, Y., and P.H.A. Snearu, 1961 F’ and F mediated transduction in Escherichia coli K-12.
Japan. J. Genetics 36: 307-318.

Jacos, F., and E. A. ApeLperc, 1959 Transfert de characterés génétique par incorporation au
facteur sexuel d’Escherichia coli. Compt. Rend. 249: 189-191,

Jacos, F., and J. Monon, 1961 On the regulation of gene activity. Cold Spring Harbor Symp.
Quant. Biol. 26: 193-209.

Jacos, F., D. Pearrn, C. Sancuez, and J. Monon, 1960 L’opéron: groupe de génes a expression
coordonée par un opérateur. Compt. Rend. 250: 1727-1729.

Leverrerc, E. M., 1952 Allelic relationships and reverse mutation in Escherichia coli. Genetics
37: 469-483.

Leprerzerc, J., 1947 Gene recombination and linked segregation in Escherichia coli, Genetics
32: 505-525.

1950a Isolation and characterization of biochemical mutants of bacteria. Methods Med. Res.
3: 5-22.
1950b The -galactosidase of Escherichia coli, strain K-12. J. Bacteriol. 60: 381-392.

Leperserc, J., L. L, Cavauir, and E. M. Leperserc, 1952 Sex compatibility in Escherichia
coli. Genetics 37: 720-730.

LepErserG, J., and E. M. Leperserc, 1952 Replica plating and indirect selection of bacterial
mutants. J. Bacteriol. 63: 399-406.

Leperperc, J.. E. M. Lepensenc, N. D. Zrinper, and E. R. Livery, 1951 Recombination
analysis of bacterial heredity. Cold Spring Harbor Symp. Quant. Biol. 16: 413-443.

Luria, S. E., J. N. Apams, and R. C. Trnc, 1960 Transduction of lactose-utilizing ability
among strains of E. coli and S. dysenteriae and the properties of the transducing phage
particles. Virology 12: 348-390.

McCuintock, B., 1944 The relation of homozygous deficiencies to mutations and allelic series
in maize. Genetics 29: 478-502.
RECOMBINATION ANALYSIS 1353

Parpeg, A. B., F. Jacos, and J. Monop, 1959 The genetic control and cytoplasmic expression of
8-galactosidase by Escherichia coli. J. Mol. Biol. 1: 165-178.

Reve., H. R., and S. E. Lurra, 1961 Biosynthesis of 8-p-galactosidase controlled by phage-
carried genes. II]. The behavior of phage-transduced z+ genes toward regulatory mecha-
nisms. Proc. Natl. Acad. Sci. U.S. 47: 1968-1974.

Rever, H. R., S. E. Luria, and B. Rotman, 1961 Biosynthesis of @-p-galactosidase controlled
by phage-carried genes. I. Induced #-p-galactosidase biosynthesis after transduction of gene
z+ by phage. Proc. Natl. Acad. Sci. U.S. 47: 1956-1967.

Reve, H. R., S. E. Lurra, and N. L. Youne, 1961 Biosynthesis of 8-p-galactosidase controlled
by phage-carried genes. III. Derepression of 8-p-galactosidase synthesis following induction
of phage development in lysogenic bacteria. Proc. Natl. Acad. Sci. U.S. 47: 1974-1980.

Ricuter, A. A., 1957 Genetic recombination in Escherichia coli K-12. M.S. thesis Univ. of
Wisconsin, Madison.

Rickenserc, H. W., G. N. Conen, G. Buttin, and J. Monon, 1956 La galactoside-perméase
d’Escherichia coli. Ann. Inst. Pasteur 91: 829-857.

Roserts, R. B., P. H. Anevson, D. B. Cowrz, E. T. Botton, and R. J. Barrren, 1955 Studies of
Biosynthesis in Escherichia coli. p. 5. Carnegie Inst. Wash. Publ. 607.
Rorman, B., 1958 Regulation of enzyme activity in the intact cell: the @-p-galactosidase of
Escherichia coli. J. Bacteriol. 76: 1-14.
1959 Separate permeases for the accumulation of methyl-8-v-galactoside and methyl--p-
thiogalactoside in Escherichia coli. Biochim. Biophys. Acta 32: 599-601.

Sxaar, P. D., and A. Garen, 1956 The orientation and extent of gene transfer in Escherichia
coli. Proc. Natl. Acad. Sci. U.S. 42: 619-624.

Tayzor, A. L., and E. A. Apeperc, 1960 Linkage analysis with very high frequency males of
Escherichia coli. Genetics 45: 1233-1243.

Wo.toan, E. L., and F. Jacos, 1955 Sur le méchanisme du transfert de matériel génétique av
cours de la recombinaison chez Escherichia coli K-12. Compt. Rend. 240: 2449-2461.

Woitman, E, L., F. Jacos, and W. Hayes, 1956 Conjugation and genetic recombination in
Escherichia coli K-12. Cold Spring Harbor Symp. Quant. Biol. 21: 141-162.

Yanorsky, C., and E. S. Lennox, 1959 Transduction and recombination study of linkage

relationships among the genes controlling tryptophan synthesis in Escherichia coli. Virology
8: 425-447,