BIOSYNTHESIS OF /3-D-GALACTOSIDASE CONTROLLED BY
PHAGE-CARRIED GENES. I. INDUCED /3-D-GALACTOSIDASE
BIOSYNTHESIS AFTER TRANSDUCTION OF GENE z+ BY PHAGE*

BY H. R. REVEL, S. E. LURIA, AND B. ROTMAN~

MASSACHUf3ETTS INSTITUTE OF TECHNOLOGY, AND GENETICS DEPARTMENT,
                STANFORD MEDICAL SCHOOL

Communicated October 23, 1961

Transduction of the lac+ character to lac- strains of the Escherichia co&Shigella
dysetieriae group of organisms can be performed by lysates of phage Pl released by
Zac+ bacterial donors.`* 2  The production of high-frequency transducing (HPT) ly-
sates has been reported.3n 4  The transducing activity in these lysates is embodied
in defective phage particles Pldl, which carry the entire lac genetic region of the
donor bacterium.  The relations between Pl and Pldl and between these phages
and the host bacteria resemble in many (although not in all) respects the relations
between phage A, its gal transducing derivatives Xdg, and their bacterial hosts6~ 6
The well-known properties of the cellular components-+-D-galactosidase,'
galactoside-permease,* and acetylaseg--controlled by the lac genes, and the remark-
able body of information available concerning the structure, organization, and func-
tional regulation of these genes,`O make it desirable to study the function of these
genes when they are part of phage-related elements, which can assume various types
of relation to their bacterial host, and to compare it to the function of the same genes
in chromosomal structures, that can be transferred by mating.`lB l2  The present
paper describes the initiation of function of a phage-associated z+ gene-the genetic
determinant of galactosidase-upon entering a bacterial cell. The following
papers 13. l4 report some interactions between phage associated genes, their chromo-
somal counterparts, and the phage-immunity control system.
Materials and Methods.-Many of the strains, media, and procedures have been described
previously.3* 4 Table 1 summarizes the relevant properties of the bacterial strains used.  The
properties of the lac genes are classified as states of the four genes:  i (repressor), o (operator),

                         TABLE 1

                  PROPERTIES OF BACTERIAL STRAINS
             Strain                              Properties*
E. coli K-12 3.000             i+o+z+y+
<,    `I 3.300               i-o+z+y+
II  :: 3.320               i -093 +y +
`I      2.000 oc              i +oOz +y +
<,   (' 2.050, 3.050, 3.ouo            `+ + - +
t,    " 2.0SOR        t+Z+Z+i+ (revertant from 2.OSO)
"  1: W4032(X)                la&"
(`      W4032 = W4032X'        From W4032(X) ; X-cured by UV
,`    " W4032 lac+              Stable lac+ transductant from W4032
,I   " W4032-W-1, 2.OSO-W-1      Strains carrying prophage Pldl i+o+z+y+
,`                     Strain carrying prophages Pl and Pldlw i+o+z+y+
"  1; w403`2-w-5
      W403213-4, 2.050-13-4., 3.OSO-
       13-41, 3.050-13-42       Strains carrying prophage Pldl i-o+z+y+
!. clysenwiae Sh                ~+o+z-yad
        Sh I-4
,`     `,    Sh 31-20            Sh carrying prophage Pldl i+o+z+y+

<I     ,I                   Sh carrying Pldlw i +o+t +y +
        Sh 125-13-4         Sh carrying Pldl i-o+z+y*
* More complete information on the propert.ies and WWCBR of these strains is given elsewhere."* 4
derivatives were prepared 8s needed.                                    Pl-lysogenic

1956


VOL. 47, 1961   BIOCHEMISTRY: REVEL, LURIA, AND ROTMAN         1957

z (galactosidase), and y (permease-acetylase) according to the scheme of Jacob and Monod.10
The i- strains are normal constitutives; oc strains are constitutive dominants; o0 strains are
operator negative.
The Zac+ HFT lysates were all derived from strain Plkc of phagei (here called Pl for brevity's
sake) and contained mixtures of normal Pl and of transducing Pldl particles; the transducing
particles are classified as Pldl i+.z+, Pldl i-a+, etc., according to the properties of the Zuc genes
they carry; these correspond to the properties of the Zac+ donor from which the transducing phage
were derived.  Pldlw i+z+ is a variety of Pldl phage used in many experiments.*  Low-frequency-
transducing (LFT) lysates were lysates of phage Pl grown on normal lac+ bacteria.*
The HFT lysates were produced by collecting bacteria from a growing culture, treating them
with ultraviolet (UV) light (40 set at 50 cm from a GE "germicidal" bulb), superinfecting with
active Pl at a multiplicity of infection (m.0.i.) of about 2, and incubating 2.5 hr in LB broth at 37"
before centrifuging and filtering through Millipore membranes.  The phage was concentrated and
freed of any galactosidase by centrifuging twice at 16,000 X g for 1 hr.  The phage pellet was re-
suspended in TGA medium minus glycerol, supplemented with 0.1 per cent gelatin.

TGA medium (NaCl, 0.08 M; KCl, 0.02 M; NH&I, 0.02 M; MgClr, 10-a M; CaClz, 2 X
lo-' M; FeCh, 2 X 10-O M; NazSO,, 2.5 X lo-* M; Tris (hydroxymethyl) aminomethane,
0.12 M; K*HPO,, 2 X lo-* M; pH 7.5, supplemented with 0.2 per cent glycerol, 0.5 rgjml
thiamine, and 0.1 per cent tryptone or 0.1 per cent casein hydrolysate) was the growth medium.
The doubling time for E. coli strains in this medium is 56 to 60 min at 37'C.  CaCh was added at a
concentration of 2.5 X 10-S M for adsorption of phage.
Reducing buffer for galactosidase assay:  sodium phosphate, 10-l M; MgSOh, lo-" Af; MnSO,,
2 X IO-4 M; mercaptoethanol, 10-r M; pH 7.0.
Optical density (O.D.) of experimental cultures was read at 500 rnr in a Zeiss spectrophotom-
eter.  For E. coli bacteria in TGA medium, an O.D. of 0.18 corresponds to 1 X 108 cells/ml.
Transductions were assayed by plating suitable dilutions of phage-bacterial mixtures on EMB

lactose agar plates.  An excess of Zac- bacteria was added to provide a uniform background.
For galactosidase assay, the o-nitrophenyl+D-galactoside (ONPG) method"1 16, 16 was used.
One-ml aliquots of a growing culture (<4 X 108 bacteria/ml) were added to l-ml samples of reduc-
ing buffer containing 0.01 per cent sodium desoxycholate and 0.02 ml toluene.  The tubes were
stoppered, shaken vigorously, and incubated with shaking at 37" for 30 min.  After equilibration
at 28"C, 0.6 ml of M/75 ONPG in M/4 phosphate buffer (pH 7.0) was added and the tubes were
incubated until a visible yellow color had developed.  The reaction was stopped by addition of
1.3 ml 1 M sodium carbonate and the optical density was measured at 420 rnw in a Zeiss spectro-
photometer.  A correction for turbidity at 550 mp was made.11  One unit of enzyme is defined as
the amount of enzyme that releases 1 mpmole of o-nitrophenol ( = ONP) per minute at 28", pH
7.0; a solution of 1 mpmole of ONP per ml has an optical density of 0.005 under the above condi-
tions (final pH - 10.2; 10 mm light path).
The inducers for galactosidase were methyl-thio+D-galactoside (TMG) or isopropyl-thio-p-D-
galactoside (IPTG).  ONPG, TMG, and IPTG were products of Mann Biochemical Co.  (The
initial sample of IPTG was kindly provided by Dr. J. Monod.)
The studies with the fluorogenic substrate 6-hydroxyfluoran+D-galactoside (GHFG) were
done by the method of Rotman"

Results.-Production of fl-galactosidase after introduction of Pldl z+ into z- bac-
teria in the presence of inducer:  A typical experiment consists of mixing a trans-
ducing phage preparation with log-phase z- bacteria at about 2 X 108 per ml in TGA
medium plus CaClz at 37V, allowing 2 to 15 min for phage adsorption, and diluting
1:lO in prewarmed medium (without added CaClJ containing 10m3 M TMG or
2 X lo-* M IPTG as inducers.  Incubation is continued in a shaker bath at 37'C
using shallow layers of culture in Erlenmeyer flasks.  Samples for assays of total
bacteria, of Zac+ transductants, and of galactosidase are taken at intervals.  Free
phage was also measured in some experiments.  The results of an experiment of this
kind, in which the recipient bacteria were E. coli i+z-, are shown in Figure 1.  In
this experiment, the ratio of Pl-plaque formers to bacteria was 0.2.  It is seen that


BIOCHEMISTRY: REVEL, LURIA, AND ROTMAN   PROC. N. A. s.

p -Goloctosidase

Optical Density

Transductions

??
  o                       ?              ?
           ??            120            180

             Minutes After Dilution

I.6

   A
  .E
  2
   (u
I.4 O
  0
  .o
  `;
  0









1.2

FIG. l.-Bacterial growth, luc+ transductions, and galactosidaae synthesis after
infection of E. coli 2.050 i+z- bacteria with an HFT lysate containing Pldlw i+z+.
The cell-phage mixture, after incubation in TGA medium + CaCla for 15 min., was
diluted at time 0 in TGAmedium with 10-a M TMG.

the bacterial population as a whole grew exponentially, while the transductions,
scored as Zuc+ colonies, remained more or less constant for at least 3 hr.  Gala&s
idase was present from the time of the earliest measurement (20 min) and increased
at first at an accelerated rate, then linearly for about 3 hr.

Essentially similar results were obtained in experiments of this type (in the pres-
ence of an inducer) whether the inducer was added at dilution time, as in Figure 1,
or at infection time; whether the adsorption period was 2 or 15 min; whether the
transduced genes were i-z+, i+z+, or i+o%+; and whether the recipient lac- bac-
teria were E. coli i+z-, i-z-, i-z-y- (= ZacdCi), or i-o".  Quantitative differences in
the rate of enzyme production reflect differences in phage-adsorption rate among
different bacterial strains (Fig. 2).


Vm. 47, 1901   BIOCHEMISTRY: REVEL, LURIA, AND ROTMAN

1959

I          I          I




                 2.340

Sh

Minutes After Dilution

PIG. 2.-Production of galactosidaae after infection of different Zac- recipient strains with a
lysate of Pldlw i+z+ in the presence of 10-a M TMG.  The adsorption of pha e Pl, measured for
each strain under the conditions of this experiment, wae ae follows: strains 2.0 ib ,2.340, and 3.320:
5040%; strain W4032; 2040%; strain Sh: 7545%.

When the recipient is S. dysenteriae Sh there is usually a faster rate of enzyme
production in the linear phase, due to better phage adsorption, and a drop in enzyme
production rate at about the time of lysis (see Fig. 2), probably correlated with
death of some of the bacteria.

The results of this experiment are the same if the adsorption mixture, instead of
being diluted to stop phage adsorption, is centrifuged at room temperature and the
bacteria are resuspended in calcium-free medium with inducer.  With all recipients,
only a very small fraction, from 1 to 7 per cent, of the galactosidase formed after
infection is released into the medium; this release is fully accounted for by the
minority of bacteria that are lysed due to infection with active phage Pl.
Two aspects of enzyme production require explanation: the initial phase of ac-
celeration and the later prolonged phase of linear synthesis.
Phase of accelerated synthsis:  Except for minor differences, this phase is present
in all combinations of phage and recipient types.  It is not due to a delayed effect


1960

BIOCHEMISTRY: REVEL, LURIA, AND ROTMAN

PRO% N. A.. 8.

   TMG added at 62'
                  I
60         120        180

Minutes After Dilution

FICA 3.-Dissociation of inducer function from the acceleration
Phage Pldlw i+z+ on E. coli 2.OSO i+6- recipient.  A, TMG adde s hose of ensymc Hynthesis.
B, TMG added at 62 min.                     at time 0, ILB in Pigure 1..

of inducer, because if the inducer is added at 60 or 90 min after infection the linear
phase begins almost immediately, as shown in Figure 3.  The 3- to ii-mm delay
corresponds to the delay observedls for initiation of response to inducer.  The
steps needed to attain the final linear rate of enzyme production do not require
protein synthesis: if infection is done in the presence of 5-methyltryptophan, which
stops protein synthesis completely, a later addition of tryptophan causes an almost
immediate start of linear enzyme production (Fig. 4).  Similar experiments with
chloramphenicol instead of 5-methyltryptophan gave ambiguous results, probably
because of slow recovery of protein synthesis after removal of the drug by washing.
Mitomycin treatment gave rapid loss of enzyme-synthesizing capacity; other in-
hibitors of DNA or RNA synthesis have not been tested.


VOL. 47, 1961   BIOCHEMISTRY: REVEL, LURIA, AND ROTMAN

J J    J I         I
    60         120        180

1961

Minutes After Dilution

FIG. 4.-The effect of B-methyltryptophan on the early enzyme synthesis.
on E. coli 2.340 i-z- recipient.                       Lysate Pldlw i+z+
pg/ml DL-5-meth@ryptophsn. At the time of mixing bacteria with
               L-tryptophan, 100 #g/ml, was ad B hage, culture A received 100
                                    ed to both cultures A and B
(control) at 92 mm.  TMG was present from time 0 in both cultures.  The number of stable
lac + transductions was the same in both cultures.

The accelerated phase reflects either a delay or a heterogeneity in the accomplish-
ment of some early reactions; these may he some of the early steps in phage-cell
interaction.  The latent period for growth of phage Pl at 37"C in TGA medium is
about 60 min; but no information is available as to the early synthesis of Pl-phage
materials.  An attempt to suppress enzyme production by treatment in a Lourdes
homogenizer (blendor technique) at 5 min after infection gave negative results,
which, therefore, neither support. nor exclude the possibility of delayed injection
of genetic material by phage particles.  The hypothesis of a multiplication of the
transduced genetic elements will be discussed later.

Phase of linear synthesis: The most immediate question concerns the relation
of linear galactosidase synthesis to the cells that give rise to Zuc+ transductant colo-
nies.  On the one hand, linear synthesis suggests either a constant rate of function
of nonmultiplying genetic elements (abortive transduction1ga 2"), or some other
limitations in the amount of enzyme that can be made per unit time.  On the other
hand, the roughly constant number of Zac+ colonies in platings done over a period of
several hours (see Fig. 1) indicates either a nonmultiplying group of cells or the
segregation of kc- cells from a steady number of kc+ bacteria.21  Segregation with
occasional late integration is suggested by the finding that, when few infected bac-
teria are plated on EMB lactose agar so that isolated colonies can be observed, most
of the kc+ transductants appear late, as small papillae within well developed
luc- colonies.

A quantitative relation between number of transductions and amount of enzyme
can be obtained from experiments such as that of Figure 1.  In that experiment,
for example, the amount of enzyme synthesized in 60 min (corresponding to one
doubling time) during the linear phase is about 15 units, corresponding to the pro-


1962        BIOCHEMlSTRY: REVEL, LURIA, AND ROTMAN   PROC. N. A. S.

duction of 6 X 10" fully induced lac+ cells.  This is twenty times higher than the
number of Eat+ transductants measured at 0 time.  A series of results, with dif-
ferent combinations of lysates and recipient strains and with different ratios of lysate
to cells, is given in Table 2.  The excess of enzyme over transductions is present
in all cases.  Inspection of Table 2 reveals that there is no correlation between



                         TABLE 2
RATES OF LINEAR SYNTHESIS OF GALACI-OSIDASE AND NUMEEHS OF STABLE TRANSDUCTIONS FOR
               VARIOUS DONOH-RECIPIENT COMBINATIONS*

Expt.
no.
29
I`
(`
I`
II
`I
36
`I

45
"
1`

69
`I
92
I`

53
I`
,`

HFT

1ysate
type
w-5 :

Pldlwi+z+

W-l :
Pldli+z+

13-4:
Pldl i-z+

Pldl
i+O%'

Recipient
  strain
E. coli i+z-
" I`
11 laed.d
,I "

Shi~ella i ?-
E. coli i-z-
Shigella i +z-
E. wli i+z-
`, .- -
Shigellai+z-
E. coli i+z-
I,   i-z-
I`   i+z-
Shigella i+z-
E. coli i+z-
`I   llI&
Shigella i +z-

-

-
-

M.&i.
it05
-1.5
0.05
-1.0
0.05
-2.0
0.1
0.1
0.2
:::
% :'
--0:25
~0.25
0.13
0.11
0.15

8.3 x 10"
6.3 x IO"
1.4 x 1.0*
1.8 X 10'
1.0 x 10'
1.3 x 1w
3.4 x 10"
2.0 x 10'
;:; x" ::
6.2 x 10"
1.1 x 106
6.8 x 104
1.2 x 106
1.0 X 10"

1.78 :: ii"
1:9 x 10:

Enzyme
units/ml/hour
  2.6
  1.3
  0.8
  0.8
  3.7
z
  8.4
  4.5
  1.1
  4.2
  3.8
  6.0
  6.8
  5.5
  0.6

0.6
1.3

EXCl?W
enzymet
13
8.3
228
18
14s
ii
1.64
4.8
3.4
27

3.6
i:

* hoc+ transductions were measured between 0 and 30 min after the end of phsge adsorption.  The amounts
of enzyme are those formed in 1 hr during the linear phase of synthesis (see Pig. 1).  Both values are normalized
to 1 ml of infected cell suspension.

t The ~~exce~s enzyme" values are the ratios of the amounta of enliyme produced per hour (normalized to the
amount of encyme contained in one full
             B induced a+ cell = 2.5 X 10" units) to the number of lac+ transductions.
Note the large excess values in trans uctions to recipients like S/&wZZa or E. coli lo&l, which require helper
phege to give atable transductions.

numbers of Zac+ transductants and amounts of enzyme.  This is particularly evi-
dent when one compares E. coli a- recipients, in which transduction is frequent and
occurs by integration of the transduced genes into the recipient chromosome, with
recipients such as S. dysenteriue or E. coli laP', in which transduction is rare and
occurs by lysogenixation, with a requirement for active phage Pl as helper.3  The
enzyme levels are roughly similar with all recipients, as though similar numbers of
cells received the lacf genes and made enzyme, but different proportions of them
gave lac+ transductant colonies.

The hypothesis that the excess of enzyme over transduction is due to abortive
transductions, that is, to the presence of bacteria that have received Pldl phage
particles whose genes can function but do not become established, appears to ac-
count for all the above findings.  An alternative explanation would be that only
those bacterial cells that will give complete transductants are making enzyme, and
that the enzyme levels are exceptionally high because the incoming Eat+ element,
being partly phage, multiplies vegetatively producing many copies of the Zac+
genes per cell.  This explanation is hard to entertain, because if the amounts of
galactosidase observed in the experiments of Table 2 were made in the relatively
few lac+ transductant cells this would require an enormous increase in cellular


VOL. 47, 1961   BIOCHEMISTRY: REVEL, LURIA, AND ROTMAN         1963

protein; the galactosidase in fully induced normal E. COG cells already represents
over 5 per cent of the total protein.7

Yet, the hypothesis that multiplication of the lac+ genes is responsible for the en-
zyme excess must be entertained seriously because of another set of observations,
illustrated in Figure 5.  This refers to transduction of Pl-lysogenic recipients and

8 t  o Pi dl, i+z+ on E. coli 2,OSO

0 Pi dl, i+z+ on E.coli 2.OSO (Pi)
m Pi dl i+z+ on E.coli 2.OSO   /

0 Pi dl i*z+ on E. coli 2.050 IPi)

6-

30         60        90        120

   Minutes After Dilution

1









I!









I

FIG. S.--Synthesis of galactosidase in PI-sensitive and Pl-lysogenic recipients. The donor
phages were Pldlw and Pldl, both i+z+.
CLOSO(Pl), both A-.            The recipient bacteria were E. cdi 2.050 and
          The rates of adsorption of Pl to these two straina are identical.

will be discussed more fully in the following paper.  The rate of enzyme production
with lysogenic recipients is lower than with Pl-sensitive ones, the ratio varying with
the type of phage lysate and the nature of the recipient strain from 1:2.3 to 1: 67
(see ref. 13, Table 3). It is known 22 that, at least with some of the temperate
phages, a phage that superinfects a bacterium already lysogenic for a related co-
immune prophage fails to multiply vegetatively.  Thus, the lower enzyme levels
observed with lysogenic recipients might `reflect repressed multiplication of the
transducing phage elements.


1964       B1OCHBMISTRY: REVEL, LURIA, AND ROTMAN   PROC. N. A. S.

The number of bacteria with galactosidase after transduction:  To settle this ques-
tion, enzyme studies on individual cells were done using a modification of a recently
devised fluorogenic assay.17 This consists in observing microscopically the fluores-
cence of 6-hydroxy-fluoran (6HF) released by the action of galactosidase on the
nonfluorescent substrate GHFG.  A mixture of bacteria and substrate in a suitable
buffer is prepared and dispersed in microdrops under silicone oil by spraying a fine
mist of suspension. Alternatively, bacteria are distributed into microdrops by
means of a micromanipulator. The fluorescence developed upon incubation in
single drops of appropriate size is observed microscopically with appropriate light
filters or can be measured with a photomultiplier attachment. The number of
bacteria present in a microdrop can be counted with dark field illumination.

Control experiments showed that when intact, fully induced lacf cells are used,
only about 10 per cent of them give a positive reaction; but a brief treatment with
toluene (5 min at room temperature) increases the proportion of positive reactions
to between 50 and 100 per cent, without releasing enough free enzyme to give any
diffuse background fluorescence.

Accordingly, a series of experiments similar to that shown in Figure 1 were done,
in which, in addition to the other measurements, a sample was taken near the 100th
minute after infection and used for a count of the number of cells with enzyme.
The lOO-min time was chosen because by then the linear rate of synthesis is estab-
lished, a substantial amount of enzyme has been synthesized, and yet the recipient
cells have multiplied at most by a factor 3.2 (1.7 generations).  The results are
shown in Table 3. It is clear that in every experiment there was a substantial

TABLE 3

THE NUMBEHS OF CELLS WITH GALACTOSIDASIS AT 100 MINIJTES AWBR INFECTION
            Recipient     Transduction8 = o
                           Tots1 cells     Cells with enzyme = b
Experiment         strain                          Total oelle         b:o
fg:;      E. coli i+z-       4.0 x 10-a       2 x 10-z        5
            `r i-~-       2.0 x 10-a       2 x 10-p
6-28-5        " i-z-       2.8 X lO+
::$I:          "  i+z-
            I` i-~-   66:; 2 :;r:   "-i :: ::"
                8 X lo-:  7;io

6-30-7          "   i+z-    ;:z z ;w:       3-4 x 10-a   8%
0-30-7         " i-z-                              (50)
7-l-9         " i-z-     2-4.0 X lo-    `; "x ;;I;)       5-10
HFT lysatee containing Pldl i +z + were used in all experiments.  The values in parentheses me based
on & poor count and may be too high.

excess, from 5- to 16-fold, of cells with enzyme over Zuc+ transductants.  Even if
each of the potential transductants had given rise on the average to three cells with
enzyme (of which only one would ultimately give a Zac+ clone since the number of
lac+ colonies does not iucrease for 4 hr or longer) this could not account for the
number of cells with enzyme found at 100 min.  In fact, since the scoring is less
than 100 per cent effective, the exce= of positive cells over a complete transduc-
tions is probably even greater.

We conclude that most of the enzyme produced during the phase of linear syn-
thesis is made in cells that do not register as Eat + transductants in platings because
they do not give rise to clones of lacf cells, at least under the conditions of our ex-
periments.  These are the equivalent of abortive transductants.
Ultraviolet irradiation of HFT transducing l?/snle:s:  Small doses of TTV irradiation


VOL. 47, 19Gl   BIOCHEMISTRY: REVEL, LURIA, AND ROTMAN          1965

increase the number of complete transductants for the Zac+ marker* as well as for
other genetic markers,%** 24 while reducing the number of abortive transductions.
Several experiments with UV-treated lysates were done in an attempt to study the
inactivation of the genetic element responsible for galactosidase production in
newly infected z- bacteria.  These experiments did not fulfill the expected goal.
It was found that the apparent rate of inactivation of phage Pl, measured by pre-
adsorption and plating on sensitive bacteria, varied greatly with the bacterial host
strain.  For a given strain, the plaque-forming ability of phage Pl and the galac-
tosidase-forming capacity of the infected cells (measured from the rate of enzyme
production in the linear phase) were closely parallel.  This result suggests that some
UV-sensitive, host-specific key function of the phage is needed to initiate both in-
fection by active Pl and galactosidase production by PM.  The previously re-
ported results? on the effect of UV on the frequency of transduction by Pldl on
various hosts lose much of their potential significance since they reflect mainly the
UV-sensitivity of the host-specific phage functions.

Transduction by LFT lysates:  The study of galactosidase production after in-
fection with LFT lysates is hampered by their low transducing activity, because
measurable enzyme levels are obtained only with concentrations of lysate that give
high multiple Pl infection. This in turn depresses enzyme synthesis, probably
through damage to cells and early lysis.  Using an LFT lysate with an unusually
high transducing titer (3.8 X lo6 2ac+ colonies for 2.2 X log active Pl per ml),
it was found that the amount of galactosidase produced was again four-fold in
excess over that corresponding to the luc+ transductants.  Hence, with LFT as
well as with HFT lysates, there appears to be extensive abortive transduction.

Discussion.-Abortive transduction was recognized originally with genes con-
trolling bacterial flagellation, by the occurrence of motile cells that gave rise to
clones consisting of nonmotile descendants plus one line of motile cells (unilinear
transmission of the motility trait). I9  Later, abortive transduction was invoked in
transduction of nutritional markers to explain the appearance of unsupplemented
media of small colonies containing large numbers of genetically auxotrophic bac-
teria plus one genetically autotrophic ce11.20  The enzyme produced by a single
autotrophic cell in each generation appeared to be sufficient, when distributed
among the descendant cells, to supply the synthetic requirements for several
generations of growth.  The Zac case resembles the nutritional mutant situation in
that the gene studied controls a well-defined enzymatic reaction, and the motility
situation in that the function of the transduced gene can be recognized directly
on individual cells.

The linear rate of production of galactosidase can be explained by the scheme
of Figure 6, in which it is assumed that enzyme is produced, at a constant rate per
unit tune, only in those cells that have the (nonmultiplying) z+ gene.  Since galac-
tosidase is very stable under the conditions of these experiments, measurements of
the actual amounts of enzyme in individual cells at various times after transduc-
tion, using the fluorogenic method, I7 should make it possible to check both the
distribution function of enzyme at cellular division and the actual rates of en-
zyme synthesis per cell early after transduction.  These in turn may reveal whether
any multiplication of the incoming genes occurs during the early phase of acceler-
ated enzyme synthesis.


1966       BIOCHEMISTRY: REVEL, LURIA, AND ROTMAN   PROC. N. A. S.

The occurrence of abortive transduction shows that enzyme production after
transduction is not correlated with integration of the newly entered genes into the
recipient chromosome.  This situation differs from the transfer of the maZ+ genes
as free DNA in transformation of pneumococci,26 where the early rate of enzyme
synthesis corresponds closely to the number of bacteria in which integration occurs
and which yield stably-transformed clones.  This finding may reflect the remark-
able speed of integration of genes transferred as free DNA26 rather than an inability

  ENZYME
(Arbitrary Units.1

1.5


2.0

2.5

  shading = p - galactosidase
FIG. 6.-Linear inheritance of nonmultiplying z+ gene.

of such genes to function from a nonchromosomal location.  It is not excluded,
however, that the presence in transducing particles of some phage genes is required
for the autonomous function of the transduced genes.  The early accelerated phase
of enzyme production after transduction may reflect the need for a sequential
initiation of the function of such phage genes.  The biosynthesis of enzymes con-
trolled by transduced genes may resemble in this respect the biosynthesis of en-
zymes and other proteins controlled by phage genes, which also exhibits a character-
istic sequential progression.n

It is interesting to compare the course of galactosidase production after transduc-
tion with that after mating Zac+ male bacteria with Zac- females.", I2  In mating,
the z+ gene from the donor appears to function immediately at maximum rate after
entering the recipient cell, without an early acceleration phase.  This again suggests
that the phase of acceleration observed early after transduction reflects some specific
requirements of genes entering a part of a phage.

The only other system of high-frequency transduction available, the X-Xdg
system, provides an opportunity to study the synthesis of the galactose-utilizing
enzymes after introduction of the corresponding genes by phage into gal- recipient
cells. Preliminary reportsZR indicate that the course of synthesis resembles that
of galactosidase described here; but the shortcomings of assay methods for these
enzymes, more cumbersome and less sensitive than those for galactosidase, have
apparently hampered progress of this work.  Actually, the Pldl transduction sys-
tem, because of a felicitous combination of circumstances, may offer a unique ma-
terial for a variety of studies on enzyme biosynthesis.  For example, P32-labeled
phage PldE can be used to study the effects of damage in the DNA of a gene at
various stages of its functional career.  Also, PldZ phage provides a source of a
DNA specialized in the control of an easily measurable enzyme for use in study of
enzyme synthesis in cell-free preparations.  Experiments along these lines are


VOL. 47, 1961   BIOCHEMISTRY: REVEL, LURIA, AND ROTMAN          1967

in progress.  The following papers13* I4 report studies on the role of repression in
the function of Zac genes in phage Pldl.
Summary.--Upon introdubtion of the Zac+ genetic determinants of E. coli into
Zuc- bacterial cells by transducing particles derived from phage Pl, synthesis of
P-D-galactosidase occurs, first at an accelerated rate, then at a linear rate for a
period of several hours.  The phase of accelerated synthesis is attributed to the
initiation of function of phage genes in the transducing particles; the phase of
linear synthesis, to the production of enzyme under control of nonmultiplying
Zac+ genes.  Enzyme is produced by many cells which, although they have received
a functional transducing particle, fail to give rise to permanently transduced
cell lines (abortive transduction).  The counting of individual bacterial cells con-
taining @-D-galactosidase in a mixed population is described.

* Aided by grants from the National Science Foundation, G-8808, and the National Institute
of Allergy and Infectious Diseases, National Institutes of Health, E-3038 (C-l).  Work on this
problem was initiated by one of us (S.E.L.) in collaboration with J. Monod and F. Jacob during a
brief stay at the Pasteur Institute, Paris, in 1959.  Throughout this work, we have enjoyed and
benefited from communications and exchanges of materials with Drs. Monod and Jacob.  Their
professional cooperation and personal friendship have made this undertaking a most pleasant
experience.

t Present address: Syntex Institute for Molecular Biology, Stanford, California.
l Lennox, E. S., Virology, 1, 190 (1955).
* Adams, J. N., and S. E. Luria, these PROCEEDINGS, 44, 590 (1958).
s Luria, S. E., J. N. Adams, and R. C. Ting, Virology, 12, 348 (1960).
4 Franklin, N. Cl., and S. E. Luria, Virology, 15, 299 (1961).
6 Morse, M. L., E. M. Lederberg, and J. Lederberg, Genetics, 41, 142 (1956).
6 Arber, W., G. Kellenberger, and J. Weigle, Schweiz. 2. allg. Path. Bakt., 20, 659 (I 957).
7 Cohn, M., Bacterial. Reus., 21, 140 (1957).
* Cohen, G., and J. Monod, Bacterial. Revs., 21, 169 (1957).
0 Zabin, I., A. KBp&s, and J. Monod, Biochem. Biophys. Res. Comm., 1,289 (1959).
lo Jacob, F., and J. Monod, J. Mol. Biol., 3,318 (1961).
iI Pardee, A. B., F. Jacob, and J. Monod, J. Mol. Biol., 1,165 (1959).
I* Riley, M., A. B. Pardee, F. Jacob, and J. Monod, J. Mol. Btil., 2,216 (1960).
1s Revel, H. R., and S. E. Luria, these PROCEEDINGS, 47, 1968 (1961).
I* Revel, H. R., S. E. Luria, and N. L. Young, these PROCEEDINGS, 47, 1974 (1961).
I6 Lederberg, J., J. Bacterial., 60, 381 (1950).
I6 Novick, A., and M. Weiner, these PROCEEDINGS, 43, 553 (1957).
I' Rotman, B., these PROCEEDINGS, 47, 1981 (1961).
18 Pardee, A.`B., and L. S. Prestidge, Biochim. Biophys. Acta, 49,77 (1961).
19 Stocker, B. A. D., N. D. Zinder, and J. Lederberg, J. Gen. i'vficrobiol., 9,410 (1953).
m Ozeki, H., in Genetic Studies with Bactetia, M. Demerec et al., Carnegie Institute of Wash-
ington Publ. No. 612 (1956), p. 97.

11 In experiments of longer duration the number of lac+ transductions remained approximately
constant up to 5 hr.  The rate of linear enzyme synthesis often decreased slightly after 3 to 4 hr.
I2 Bertani, G., A&. in Virus Research, 5, 151 (1958).
*' Garen, A., and N. D. Zinder, Virology, 1, 347 (1955).
*' Arber, W., Virology, 11, 273 (1960).
s Lacks, S., and R. D. Hotchkiss, B&him. Biophys. Acta, 45, 155 (1960).
nr Fox, M. S., Nature, 187, 1004 (1960).
27 Jacob. F., Harvey Lectures, 54, 1 (1959).
18 Matsushiro, A., and K. Mizobuchi, Biken's Journal, 2, 299 (1959); P. Starlinger, personal
communication.