Independent functions of viral protein and nucleic acid in growth of bacteriophage

By A. D. Hershey and M. Chase, 1952

Annotated on Fermat's Library with 16 comments. Text and annotations as Markdown Original paper

This seminal paper by Hershey and Chase confirmed that DNA, and not...
Bacteriophages are viruses that infect bacteria and they resemble R...
Among other things, these data show that the "sulfur" (synecdochic ...
Throughout the paper, Hershey and Chase frequently refer to sulfur ...
This is interesting because it posits a custodial relationship betw...
In order for the phage to multiply, it must shed its protein coat. ...
This is sensible: phage DNA is no longer protected after it is inje...
"DNAse" is a common, tempting typo.
This experiment is kind of unsettling because they test whether or ...
Again, "sulfur" stands for "protein" for the purposes of this study...
These are the classic "Waring blender" experiments we learn about i...
Notice how exhaustively they eliminate a role for the radiolabeled ...
Even though they show isotopic phosphorous passes from parent phage...
To be completely aboveboard, I wrote the preceding comment prior to...
This is the result with which we are all familiar: sulfur (protein)...
They understand that their results imply that DNA is the hereditary...
INDEPENDENT FUNCTIONS OF VIRAL PROTEIN AND NUCLEIC
ACID IN GROWTH OF BACTERIOPHAGE*
B~ A. D. HERSHEY AND MARTHA CHASE
(From the Department of Genetics, Carnegie Institution of Washington, Cold Spring
Harbor, Long Island)
(Received for publication, April 9, 1952)
The work of Doermaml (1948), Doermann and Dissosway (1949), and
Anderson and Doermann (1952) has shown that bacteriophages T2, T3, and
T4 multiply in the bacterial cell in a non-infective form. The same is true of
the phage carried by certain lysogenic bacteria (Lwoff and Gutmann, 1950).
Little else is known about the vegetative phase of these viruses. The experi-
ments reported in this paper show that one of the first steps in the growth of
T2 is the release from its protein coat of the nucleic acid of the virus particle,
after which the bulk of the sulfur-containing protein has no further function.
Materials and Methods.--Phage T2 means in this paper the variety called T2H
(Hershey, 1945); T2h means one of the host range mutants of T2; UV-phage means
phage irradiated with ultraviolet light from a germicidal lamp (General Electric
Co.) to a fractional survival of 10 -5.
Sensitive bacteria means a strain (H) of
Escherichia coli sensitive to T2 and its
h mutant; resistant bacteria B/2 means a strain resistant to T2 but sensitive to its
h mutant; resistant bacteria
B/2h means a strain resistant to both. These bacteria
do not adsorb the phages to which they are resistant.
"Salt-poor' broth contains per liter 10 gin. bacto-peptone, 1 gm. glucose, and 1
gin. NaC1. "Broth" contains, in addition, 3 gin. bacto-beef extract and 4 gm. NaCl.
Glycerol-lactate medium contains per liter 70 m~ sodium lactate, 4 gin. glycerol,
5 gin. NaC1, 2 gin. KCI, 1 gin. NI-I_aCI, 1 mm MgC12, 0.1 m~t CaC12, 0.01 gm. gelatin,
10 rag. P (as orthophosphate), and 10 mg. S (as MgSO4), at pH 7.0.
Adsorption medium contains per liter 4 gin. NaC1, 5 gm. K~SO4, 1.5 gin. KH~PO,,
3.0 gm. Na~-IP0,, 1 mE MgSO,, 0.1 m_~ CaC12, and 0.01 gm. gelatin, at pH 7,0.
Veronal buffer contains per liter 1 gm. sodium diethylbarbiturate, 3 mE MgS04,
and 1 gin. gelatin, at pH 8.0.
The HCN referred to in this paper consists of molar sodium cyanide solution
neutralized when needed with phosphoric acid.
* This investigation was supported in part by a research grant from the National
Microbiological Institute of the National Institutes of Health, Public Health Service.
Radioactive isotopes were supplied by the Oak Ridge National Laboratory on alloca-
tion from the Isotopes DivisiOn, United States Atomic Energy Commission.
39
The Journal of General Physiology
40 VIRAL PROTEIN AND NUCLEIC ACID Ilq BACTERIOPHAGE GROWTH
Adsorption of isotope to bacteria was usually measured by mixing the sample
in adsorption medium with bacteria from 18 hour broth cultures previously heated
to 70°C. for 10 minutes and washed with adsorption medium. The mixtures were
warmed for 5 minutes at 37°C., diluted with water, and centrifuged. Assays were
made of both sediment and supernatant fractions.
Precipitation of isotope with antiserum was measured by mixing the sample in
0.5 per cent saline with about 1011 per ml. of non-radioactive phage and slightly
more than the least quantity of antiphage serum (final dilution 1:160) that would
cause visible precipitation. The mixture was centrifuged after 2 hours at 37°C.
Tests with DNase (desoxyribonuclease) were performed by warming samples
diluted in veronal buffer for 15 minutes at 37°C. with 0.1 rag. per ml. of crystalline
enzyme (Worthington Biochemical Laboratory).
Acid-soluble isotope was measured after the chilled sample had been precipitated
with 5 per cent trichloroacetic acid in the presence of 1 mg./ml, of serum albumin, and
centrifuged.
In all fractionations involving centrifugation, the sediments were not washed, and
contained about 5 per cent of the supernatant. Both fractions were assayed.
Radioactivity was measured by means of an end-window Geiger counter, using
dried samples sufficiently small to avoid losses by self-absorption. For absolute meas-
urements, reference solutions of P~ obtained from the National Bureau of Standards,
as well as a permanent simulated standard, were used. For absolute measurements
of S s5 we relied on the assays (4-20 per cent) furnished by the supplier of the isotope
(Oak Ridge National Laboratory).
Glycerol-lactate medium was chosen to permit growth of bacteria without un-
desirable pH changes at low concentrations of phosphorus and sulfur, and proved
useful also for certain experiments described in this paper. 18-hour cultures of sensitive
bacteria grown in this medium contain about 2 × 109 cells per ml., which grow ex-
ponentially without lag or change in light-scattering per cell when subcultured in
the same medium from either large or small seedings. The generation time is 1.5 hours
at 37°C. The cells are smaller than those grown in broth. T2 shows a latent period
of 22 to 25 minutes in this medium. The phage yield obtained by lysis with cyanide
and UV-phage (described in context) is one per bacterium at 15 minutes and 16
per bacterium at 25 minutes. The final burst size in diluted cultures is 30 to 40 per
bacterium, reached at 50 minutes. At 2 × l0 s cells per ml., the culture lyses slowly,
and yields 140 phage per bacterium. The growth of both bacteria and phage in this
medium is as reproducible as that in broth.
For the preparation of radioactive phage, P= of specific activity 0.5 mc./mg, or
S 35 of specific activity 8.0 mc./mg, was incorporated into glycerol-lactate medium,
in which bacteria were allowed to grow at least 4 hours before seeding with phage.
After infection with phage, the culture was aerated overnight, and the radioactive
phage was isolated by three cycles of alternate slow (2000 6") and fast (12,000 G)
centrifugation in adsorption medium. The suspensions were stored at a concentration
not exceeding 4
i.tc./ml.
Preparations of this kind contain 1.0 to 3.0 × 10-" gg. S and 2.5 to 3.5 × 10-1~
ttg. P per viable phage particle. Occasional preparations containing excessive amounts
of sulfur can be improved by absorption with heat-killed bacteria that do not adsorb
A. D. HERSHEY AND MARTHA CHASE
41
the phage. The radiochemical purity of the preparations is somewhat uncertain, ow-
ing to the possible presence of inactive phage particles and empty phage membranes.
The presence in our preparations of sulfur (about 20 per cent) that is precipitated by
antiphage serum (Table I) and either adsorbed by bacteria resistant to phage, or
not adsorbed by bacteria sensitive to phage (Table VII), indicates contamination
by membrane material. Contaminants of bacterial origin are probably negligible for
present purposes as indicated by the data given in Table I. For proof that our prin-
cipal findings reflect genuine properties of viable phage particles, we rely on some
experiments with inactivated phage cited at the conclusion of this paper.
The Chemical Morphology of Resting Phage Partides.--Anderson (1949)
found that bacteriophage T2 could be inactivated by suspending the particles
in high concentrations of sodium chloride, and rapidly diluting the suspension
with water. The inactivated phage was visible in electron micrographs as tad-
pole-shaped "ghosts." Since no inactivation occurred if the dilution was slow
TABLE I
Composition of Ghosts and Solution of Plasmoly'zed Pkage
Per cent of isotope]
Acid-soluble .................... : ..........
Acid-soluble after treatment with DNase ......
Adsorbed to sensitive bacteria ................
Precipitated by antiphage ...................
Whale phage labeled with ' Plasmolyzed
phage labeled
with
P~' ~~,~ P~ S'S
1 t 1 80 1
8s I 90 I 2 I 90
9o i ~ I s I 97
he attributed the inactivation to osmotic shock, and inferred that the particles
possessed an osmotic membrane. Herriott (1951) found that osmotic shock
released into solution the DNA (desoxypentose nucleic acid) of the phage
particle, and that the ghosts could adsorb to bacteria and lyse them. He pointed
out that this was a beginning toward the identification of viral functions with
viral substances.
We have plasmolyzed isotopically labeled T2 by suspending the phage
(1011 per ml.) in 3 M sodium chloride for 5 minutes at room temperature, and
rapidly pouring into the suspension 40 volumes of distilled water. The plas-
molyzed phage, containing not more than 2 per cent survivors, was then an-
alyzed for phosphorus and sulfur in the several ways shown in Table I. The
results confirm and extend previous findings as follows:--
I. Plasmolysis separates phage T2 into ghosts containing nearly all the
sulfur and a solution containing nearly all the DNA of the intact particles.
2. The ghosts contain the principal antigens of the phage particle detect-
able by our antiserum. The DNA is released as the free acid, or possibly linked
to sulfur-free, apparently non-antigenic substances.
42
VIRAL
PROTEIN AND NUCLEIC ACID
IN BACTERIOPHAGE
GROWTH
3. The ghosts are specifically adsorbed to phage-susceptible bacteria; the
DNA is not.
4. The ghosts represent protein coats that surround the DNA of the intact
particles, react with antiserum, protect the DNA from DNase (desoxyribo-
nuclease), and carry the organ of attachment to bacteria.
5. The effects noted are due to osmotic shock, because phage suspended in
salt and diluted slowly is not inactivated, and its DNA is not exposed to
DNase.
TABLE II
Sensitization of Phage DNA to DNase by Adsorption to Bacteria
Phage adsorbed to
Live bacteria ................................
Bacteria heated before infection ...............
Bacteria heated after infection ................
Heated unadsorbed phage: acid-
soluble t m
f
70 ......
80 ° .
900
t lO0 °.
Phage labeled
with
S $t
pn
835
p~
885
tm
p82
pn
pn
pn
Non-sedimentable isotope,"
per ce~a
After
DNase No
DNase
2 1
8 7
15 11
76 13
12 14
66 23
5
13
81
88
Phage adsorbed to bacteria for 5 minutes at 37°C. in adsorption medium, followed by
washing.
Bacteria heated for 10 minutes at 800C. in adsorption medium (before infection) or in
veronal buffer (after infection).
Unadsorbed phage heated in veronal buffer, treated with DNase, and precipitated with
trichloroacetic acid.
All samples fractionated by centrifuging 10 minutes at 1300 G.
Sensitization of Phage DNA to DNase by Adsorption to Bacteria.--The
struc-
ture of the resting phage particle described above suggests at once the possibil-
ity that multiplication of virus is preceded by the alteration or removal of the
protective coats of the particles. This change might be expected to show itself
as a sensitization of the phage
DNA
to DNase. The experiments described
in Table II show that this happens. The results may be summarized as fol-
lows:-
1. Phage DNA becomes largely sensitive to DNase after adsorption to
heat-killed bacteria.
2. The same is true of the DNA of phage adsorbed to live bacteria, and then
A. D. HERSHEY AND MARTHA CHASE
43
heated to 80°C. for 10 minutes, at which temperature unadsorb'ed phage is
not sensitized to DNase.
3. The DNA of phage adsorbed to unheated bacteria is resistant to DNase,
presumably because it is protected by cell structures impervious to the en-
zyme.
Graham and collaborators (personal communication) were the first to dis-
cover the sensitization of phage DNA to DNase by adsorption to heat-killed
bacteria.
The DNA in infected cells is also made accessible to DNase by alternate
freezing and thawing (followed by formaldehyde fixation to inactivate cellu-
lar enzymes), and to some extent by formaldehyde fixation alone, as illus-
trated by the following experiment.
Bacteria were grown in broth to 5 × 10 ~ cells per ml., centrifuged, resuspended in
adsorption medium, and infected with about two P=-labeled phage per bacterium.
After 5 minutes for adsorption, the suspension was diluted with water containing per
liter 1.0 mxs MgSO4, 0.1 m~ CaCI~, and 10 rag. gelatin, and recentrifuged. The cells
were resuspended in the fluid last mentioned at a concentration of 5 × 108 per ml.
This suspension was frozen at -15°C. and thawed with a minimum of warming,
three times in succession. Immediately after the third thawing, the cells were fixed by
the addition of 0.5 per cent
(v/v)
of formalin (35 per cent HCHO). After 30 minutes
at room temperature, the suspension was dialyzed free from formaldehyde and cen-
trifuged at 2200 G for 15 minutes. Samples of P~-labeled phage, frozen-thawedl fixed,
and dialyzed, and of infected cells fixed only and dialyzed, were carried along as
controls.
The analysis of these materials, given in Table III, shows that the effect
of freezing and thawing is to make the intracellular DNA labile to DNase,
without, however, causing much of it to leach out of the ceils. Freezing and
thawing and formaldehyde fixation have a negligible effect on unadsorbed
phage, and formaldehyde fixation alone has only a mild effect on infected cells.
Both sensitization of the intracellular p3~- to DNase, and its failure to leach
out of the cells, are constant features of experiments of this type, independently
of visible lysis. In the experiment just described, the frozen suspension cleared
during the period of dialysis. Phase-contrast microscopy showed that the ceils
consisted largely of empty membranes, many apparently broken. In another
experiment, samples of infected bacteria from a culture in salt-poor broth
were repeatedly frozen and thawed at various times during the latent period
of phage growth, fixed with formaldehyde, and then washed in the centrifuge.
Clearing and microscopic lysis occurred only in suspensions frozen during the
second half of the latent period, and occurred during the first or second thaw-
ing. In this case the lysed cells consisted wholly of intact cell membranes,
appearing empty except for a few small, rather characteristic refractile bodies
apparently attached to the cell walls. The behavior of intracellular p32 toward
DNase, in either the lysed or unlysed cells, was not significantly different from
44 VIRAL PROTEIN AND NUCLE]C ACID IN BACTERIOPHAGE GROWTH
that shown in Table III, and the content of Pa was only slightly less after
lysis. The phage liberated during freezing and thawing was also fitrated in
this experiment. The lysis occurred without appreciable liberation of phage
in suspensions frozen up to and including the 16th minute, and the 20 min-
ute sample yielded only five per bacterium. Another sample of the culture
formalinized at 30 minutes, and centrifuged without freezing, contained
66
per cent of the pa in non-sedimentable form. The yield of extracellular phage
at 30 minutes was 108 per bacterium, and the sedimented material consisted
largely of formless debris but contained also many apparently intact cell
membranes.
TABLE III
Sensitization of Intracdlular Phage to DNase by Freezing, Thawing, and Fixation
with Formaldehyde
J Unsdsorbed Infected cells I
phage frcaen, frozen,
thawed, Infected cells
thawed, fixed fixed fixed
only
Low speed sediment fraction
Total P~ .................................... -- 71 86
Acid-soluble ................................. -- 0 0.5
Acid-so.luble after DNase ..................... -- 59 28
Low
speed supematant fraction
Total P= .................................... -- 29 14
Acid-soluble ................................. 1 0.8 0.4
Acid-soluble after DNase ..................... 11 21 5.5
The figures express per cent of total P'~ in the original phage, or its adsorbed fraction.
We draw the following conclusions from the experiments in which cells
infected with PS!labeled phage are subjected to freezing and thawing.
1. Phage DNA becomes sensitive to DNAse after adsorption to bacteria in
buffer under conditions in which to known growth process occurs (Benzer,
1952; Dulbecco, 1952).
2. The cell membrane can be made permeable to DNase under conditions
that do not permit the escape of either the intracellular P= or the bulk of the
cell contents.
3. Even if the cells lyse as a result of freezing and thawing, permitting escape
of other cell constituents, most of the p3~ derived from phage remains inside
the cell membranes, as do the mature phage progeny.
4. The intracellular p3~ derived from phage is largely freed during spon-
taneous lysis accompanied by phage liberation.
A. D. HERSHEY AND MARTHA CHASE
45
We interpret these facts to mean that intracellular DNA derived from phage
is not merely DNA in solution, but is part of an organized structure at all times
during the latent period.
Liberation of DNA from Phage Particles by Adsorption to Bacterial Frag-
ments.--The
sensitization of phage DNA to specific depolymerase by adsorp-
tion to bacteria might mean that adsorption is followed by the ejection of
the phage DNA from its protective coat. The following experiment shows that
this is in fact what happens when phage attaches to fragmented bacterial
cells.
TABLE IV
Rdease of DNA from Phage Adsorbed to Bacterial Debris
Phage
labeled with
Sss ] pal
Sediment fraction
Surviving phage .................................... 16 22
Total isotope ...................................... 87 55
Acid-soluble isotope ............................... :. 0 2
Acid-soluble after DNase ............................ 2 29
Supematant fraction
Surviving
phage .................................... 5 5
Total isotope ...................................... 13 45
Acid-soluble isotope ................................. 0.8 0.5
Acid-soluble after DNase ............................ 0.8 39
S 35-
and P~-labeled T2 were mixed with identical samples of bacterial debris in adsorption
medium and warmed for 30 minutes at 37°C. The mixtures were then centrifuged for 15
minutes at 2200 G, and the sediment and supernatant fractions were analyzed separately.
The results axe expressed as per cent of input phage or isotope,
Bacterial debris was prepared by infecting cells in adsorption medium with
four particles of T2 per bacterium, and transferring the cells to salt-poor
broth at 37°C. The culture was aerated for 60 minutes, ~r/50 HCN was added,
and incubation continued for 30 minutes longer. At this time the yield of ex-
tracellular phage was 400 particles per bacterium, which remained unadsorbed
because of the low concentration of electrolytes. The debris from the lysed
cells was washed by centrifugation at 1700 G, and resuspended in adsorption
medium at a concentration equivalent to 3 X 10 ~ lysed cells per ml. It consisted
largely of collapsed and fragmented cell membranes. The adsorption of radio-
active phage to this material is described in Table IV. The following facts
should be noted.
46
VIRAL PROTEIN" AND NUCLEIC ACID llg BACTERIOPHAGE GROWTH
1. The unadsorbed fraction contained only 5 per cent of the original phage
particles in infective form, and only 13 per cent of the total sulfur. (Much
of this sulfur must be the material that is not adsorbable to whole bacteria.)
2. About 80 per cent of the phage was inactivated. Most of the sulfur of
this phage, as well as most of the surviving phage, was found in the sediment
fraction.
3. The supernatant fraction contained 40 per cent of the total phage D~qA
(in a form labile to DNase) in addition to the DNA of the unadsorbed surviving
phage. The labile DNA amounted to about half of, the DNA of the inactivated
phage particles, whose sulfur sedimented with the bacterial debris.
4. Most of the sedimentable DNA could be accounted for either as surviving
phage, or as DNA labile to DNase, the latter amounting to about half the DNA
of the inactivated particles.
Experiments of this kind are unsatisfactory in one respect: one cannot tell
whether the liberated DNA represents all the DNA of some of the inactivated
particles, or only part of it.
Similar results were obtained when bacteria (strain B) were lysed by large
amounts of UV-killed phage T2 or T4 and then tested with P~-labeled T2
and T4. The chief point of interest in this experiment is that bacterial debris
saturated with UV-killed T2 adsorbs T4 better than T2, and debris saturated
with T4 adsorbs T2 better than T4. As in the preceding experiment, some of
the adsorbed phage was not inactivated and some of the DNA of the inacti-
vated phage was not released from the debris.
These experiments show that some of the cell receptors for T2 are different
from some of the cell receptors for T4, and that phage attaching to these spe-
cific receptors is inactivated by the same mechanism as phage attaching to
unselected receptors. This mechanism is evidently an active one, and not
merely the blocking of sites of attachment to bacteria.
Removal of Phage Coats from Infected Bavteria.--Anderson (1951) has ob-
tained electron micrograph s indicating that phage T2 attaches to bacteria
by its tail. If this precarious attachment is preserved during the progress of
the infection, and if the conclusions reached above are correct, it ought to
be a simple matter to break the empty phage membranes off the infected
bacteria, leaving the phage DNA inside the cells.
The following experiments show that this is readily accomplished by strong
shearing forces applied to suspensions of infected cells, and further that in-
fected cells from which 80 per cent of the sulfur of the parent virus has been
removed remain capable of yielding phage progeny.
Broth-grown bacteria were infected with S 3s- or P3*-labeled phage in ad-
sorption medium, the unadsorbed material was removed by centrifugation,
and the cells were resuspended in water containing per liter 1 m~ MgSO,,
0.1 rn~ CaCl~, and 0.1 gin. gelatin. This suspension was spun in a Waring
A. D. HERSHEY AND MARTHA CHASE
47
blendor (semimicro size) at 10,000 g.P.~. The suspension was cooled briefly
in ice water at the end of each 60 second running period. Samples were removed
at intervals, titrated (through antiphage serum) to measure the number of
bacteria capable of yielding phage, and centrifuged to measure the proportion
of isotope released from the cells.
The results of one experiment with each isotope are shown in Fig. 1. The
data for S as and survival of infected bacteria come from the same experiment,
in which the ratio of added phage to bacteria was 0.28, and the concentrations
JO0<
8o
+~
60
~ 40
2O
®
'®
Znfected baczez~ia
® ---------.
- ® ~.xt~acellulaz~ ;5 35
-- &
~ExZ~acellula~ 1 ~ 32
I I I I I I I
YIin.
0 i 2 3 ~ 8
:Runnin 9 %/me in blendo~
FIG. I. Removal of $a5 and 1 ~ from bacteria infected with radioactive phage, and
survival of the infected bacteria, during agitation in a Waring blendor.
of bacteria were 2.5 X 10 s per mh infected, and 9.7 X 108 per ml. total, by
direct titration. The experiment with P3~-labeled phage was very similar.
In connection with these results, it should be recalled that Anderson (1949)
found that adsorption of phage to bacteria could be prevented by rapid stir-
ring of the suspension.
At higher ratios of infection, considerable amounts of phage sulfur elute
from the cells spontaneously under the conditions of these experiments, though
the elution of P~ and the survival of infected cells are not affected by multi-
plicity of infection (Table V). This shows that there is a cooperative action
among phage particles in producing alterations of the bacterial membrane
which weaken the attachment of the phage. The cellular changes detected in
48
VIRAL PROTEIN AND NUCLEIC ACID IN BACTERIOPHAGE GROWTH
this way may be related to those responsible for the release of bacterial com-
ponents from infected bacteria (Prater, 1951; Price, 1952).
A variant of the preceding experiments was designed to test bacteria at a
later stage in the growth of phage. For this purpose infected cells were aerated
in broth for5 or 15 minutes, fixed by the addition of 0.5 per cent
(v/v)
com-
mercial formalin, centrifuged, resuspended in 0.1 per cent formalin in water,
and subsequently handled as described above. The results were very similar
to those already presented, except that the release of PS~ from the cells was
slightly less, and titrations of infected cells could not be made.
The S86-1abeled material detached from infected cells in the manner de-
scribed possesses the following properties. It is sedimented at 12,000 G, though
less completely than intact phage particles. It is completely precipitated by
TABLE V
Effect of Multiplicity of Infection on F_,lution of Pkage Membranes from Infested Bacteria
Running
thne
in blendor
m/..
0
2.5
0
2.5
Multiplicity
of
infection
0.6
0.6
6.0
6.0
Plffi-labeled
phage
Isotope eluted
per
cePgt
10
21
1.3
24
iInfected
bacteria
surviving
per
cen~
120
82
89
86
Sss-labeled
phage
Isotope eluted
per cent
16
81
46
82
Infected bacteria
surviving
per
¢~1
101
78
9O
85
The infected bacteria were suspended at 109 cells per ml. in water containing per liter
1 mM MgSO4, 0.1 mM CaC12, and 0.1 gin. gelatin. Samples were withdrawn for assay of
extracellular isotope and infected bacteria before and after agitating the suspension. In
either case the cells spent about 15 minutes at room temperature in the eluting fluid.
antiphage serum in the presence of whole phage carrier. 40 to 50 per cent of
it readsorbs to sensitive bacteria, almost independently of bacterial concen-
tration between 2 )< 108 and 109 cells per ml., in 5 minutes at 37°C. The ad-
sorption is not very specific: 10 to 25 per cent adsorbs to phage-resistant bac-
teria under the same conditions. The adsorption requires salt, and for this
reason the efficient removal of S ~ from infected bacteria can be accomplished
only in a fluid poor in electrolytes.
The results of these experiments may be summarized as follows:--
1. 75 to 80 per cent of the phage sulfur can be stripped from infected cells
by violent agitation of the suspension. At high multiplicity of infection, nearly
50 per cent elutes spontaneously. The properties of the S3S-labeled material
show that it consists of more or less intact phage membranes, most of which
have lost the ability to attach specifically to bacteria.
2. The release of sulfur is accompanied by the release of only 21 to 35 per
A. D. HERSHEY AND MARTHA CHASE
49
cent of the phage phosphorus, half of which is given up without any mechan-
ical agitation.
3. The treatment does not cause any appreciable inactivation of intracellu-
lar phage.
4, These facts show that the bulk of the phage sulfur remains at the cell
surface during infection, and takes no part in the multiplication of intracellu-
lar phage. The bulk 0f the phage DNA, on the other hand, enters the cell soon
after adsorption of phage to bacteria.
Transfer of Sulfur and Phosphorus from Parental Phage to Progeny.--We
have concluded above that the bulk of the sulfur-containing protein of the
resting phage particle takes no part in the multiplication of phage, and in
fact does not enter the cell. It follows that little or no sulfur should be trans-
ferred from parental phage to progeny. The experiments described below show
that this expectation is correct, and that the maximal transfer is of the order
1 per cent
Bacteria were grown in glycerol-lactate medium overnight and subcultured
in the same medium for 2 hours at 37°C. with aeration, the size of seeding
being adjusted nephelometrically to yield 2 × 108 cells per ml. in the sub-
culture. These bacteria were sedimented, resuspended in adsorption medium
at a concentration of 109 cells per ml., and infected with S35-1abeled phage
T2. After 5 minutes at 37°C., the suspension was diluted with 2 volumes of
water and resedimented to remove unadsorbed phage (5 to 10 per cent by
titer) and S a5 (about 15 per cent). The cells were next suspended in glycerol-
lactate medium at a concentration of 2 X 1@ per ml. and aerated at 37°C.
Growth of phage was terminated at the desired time by adding in rapid suc-
cession 0.02 m~t HCN and 2 X 1011 UV-kflled phage per ml. of culture. The
cyanide stops the maturation of intracellular phage (Doermann, 1948), and
the UV-killed phage minimizes losses of phage progeny by adsorption to bac-
terial debris, and promotes the lysis of bacteria (Maal~e and Watson, 1951).
As mentioned in another connection, and also noted in these experiments,
the lysing phage must be closely related to the phage undergoing multiplica-
tion
(e.g.,
T2H, its h mutant, or T2L, but not T4 or T6, in this instance) in
order to prevent inactivation of progeny by adsorption to bacterial debris.
To obtain what we shall call the maximal yield of phage, the lysing phage
was added 25 minutes after placing the infected ceils in the culture medium,
and the cyanide was added at the end of the 2nd hour. Under these condi-
tions, lysis of infected ceils occurs rather slowly.
Aeration was interrupted when the cyanide was added, and the cultures
were left overnight at 37°C. The lysates were then fractionated by centrifuga-
tion into an initial low speed sediment (2500 G for 20 minutes), a high speed
supernatant (12,000 G for 30 minutes), a second low speed sediment obtained
by recentrffuging in adsorption medium the resuspended high speed sediment,
and the clarified high speed sediment.
50 VIRAL PROTEIN" AND N'UCLEIC ACID IN" BACTERIOPHAGE GROWTH
The distribution of S 8~ and phage among fractions obtained from three cul-
tures of this kind is shown in Table VI. The results are typical (except for the
excessively good recoveries of phage and S a6) of lysates in broth as well as
lysates in glycerol-lactate medium.
The striking result of this experiment is that the distribution of S 35 among
the fractions is the same for early lysates that do not contain phage progeny,
and later ones that do. This suggests that little or no S 36 is contained in the
mature phage progeny. Further fractionation by adsorption to bacteria con-
firms this suggestion.
Adsorption mixtures prepared for this purpose contained about 5 X 10 ~
heat-killed bacteria (70°C. for 10 minutes) from 18 hour broth cultures, and
TABLE VI
Per Cent Distributions of Phage and S a~ among Centrifugally Separated Fractions of Lysates
after Infection with SS~-Labded 1"2
Fraction
1st low speed sediment .......................
2nd
" " "
High speed
" .......................
" "
supernatant .......................
Recovery ...................................
Lysis
at
tsar°
79
2.4
8.6
10
100
Lysls
st
t~lO
Sss
81
2.1
6.9
10
100
Maximal
yield
S s5 Phage
82 19
2.8 14
7.1 61
7.5 7.0
96 100
Infection with S3S-labeled T2, 0.8 particles per bacterium. Lysing phage UV-killed k
mutant of T2. Phage yields per infected bacterium: <0.1 after lysis at t -~ 0; 0.12 at t =
10; maximal yield 29. Recovery of S s5 means per cent of adsorbed input recovered in the
four fractions; recovery of phage means per cent of total phage yield (by plaque count before
fractionation) recovered by titration of fractions.
about 1011 phage (UV-killed lysing phage plus test phage), per ml. of adsorp-
tion medium. After warming to 37°C. for 5 minutes, the mixtures were diluted
with 2 volumes of water, and centrifuged. Assays were made from supernatants
and from unwashed resuspended sediments.
The results of tests of adsorption of S 35 and phage to bacteria (H) adsorbing
both T2 progeny and k-mutant lysing phage, to bacteria (B/2) adsorbing lysing
phage only, and to bacteria (B/2h) adsorbing neither, are shown in Table VII,
together with parallel tests of authentic SaS-labeled phage.
The adsorption tests show that the S 85 present in the seed phage is adsorbed
with the specificity of the phage, but that S 85 present in lysates of bacteria
infected with this phage shows a more complicated behavior. It is strongly
adsorbed to bacteria adsorbing both progeny and lysing phage. It is weakly
adsorbed to bacteria adsorbing neither. It is moderately well adsorbed to bac-
A, D. HERSHEY AND MARTHA CHASE
51
teria adsorbing lysing phage but not phage progeny. The latter test shows that
the S a5 is not contained in the phage progeny, and explains the fact that the
S aB in early lysates not containing progeny behaves in the same way.
The specificity of the adsorption of SSS-labeled material contaminating the
phage progeny is evidently due to the lysing phage, which is also adsorbed
much more strongly to strain H than to B/2, as shown both by the visible re-
duction in Tyndall scattering (due to the lysing phage) in the supematants
of the test mixtures, and by independent measurements. This conclusion is
further confirmed by the following facts.
TABLE VII
Adsorption Tests uqth Uniformly SS6-Labeled Pkage and with Products of Their Growth in
Non-Radioavtive Medium
Adsorbing bacteria
Sensitive (H) ......................
Resistant (B/2) ....................
Resistant (B/2k) ...................
Percent adsorbed
Uniformly labeled
~ss
phage
+ UV-__~h No UV-~ ;~Ju
S~ S*~
S '6
84 I 86
15 I 11
13 / 12
Products of
lysis at
t= 10
w --
(
79
46
29
Phage progeny
(Maximal yield)
Su Phage.
49 10
8
The uniformly labeled phage and the products of their growth are respectively the seed
phage and the high speed sediment fractions from the experiment shown in Table VI.
The uniformly labeled phage is tested at a low ratio of phage to bacteria: +UV-k means
with added UV-killed h mutant in equal concentration to that present in the other test
materials.
The adsorption of phage is measured by plaque counts of supematants, and also sedi-
ments in the case of the resistant bacteria, in the usual way.
1. If bacteria are infected with S 35 phage, and then lysed near the midpoint
of the latent period with cyanide alone (in salt-poor brothi to prevent read-
sorption of S 35 to bacterial debris), the high speed sediment fraction contains
S a~ that is adsorbed weakly and non-specifically to bacteria.
2. If the lysing phage and the S35-1abeled infecting phage are the same
(T2), or if the culture in salt-poor broth is allowed to lyse spontaneously (so
that the yield of progeny is large), the S 85 in the high speed sediment fraction
is adsorbed with the specificity of the phage progeny (except for a weak non-
specific adsorption). This is illustrated in Table VII by the adsorption to H
and
B/2h.
It should be noted that a phage progeny grown from S35-1abeled phage and
containing a larger or smaller amount of contaminating radioactivity could
not be distinguished by any known method from authentic S85-1abeled phage,
52
VmA5 PROTEIN AND NUCLEIC ACID llq BACTERIOPHAGE GROWTH
except that a small amount of the contaminant could be removed by adsorp-
tion to bacteria resistant to the phage. In addition to the properties already
mentioned, the contaminating S 8s is completely precipitated with the phage
by antiserum, and cannot be appreciably separated from the phage by further
fractional sedimentation, at either high or low concentrations of electrolyte.
On the other hand, the chemical contamination from this source would be
very small in favorable circumstances, because the progeny of a single phage
particle are numerous and the contaminant is evidently derived from the
parents.
The properties of the SSS-labeled contaminant show that it consists of the
remains of the coats of the parental phage particles, presumably identical
with the material that can be removed from unlysed cells in the Waring blen-
dor. The fact that it undergoes little chemical change is not surprising since it
probably never enters the infected cell.
The properties described explain a mistaken preliminary report (Hershey
et al.,
1951) of the transfer of S 35 from parental to progeny phage.
It should be added that experiments identical to those shown in Tables VI
and VII, but starting from phage labeled with p3~, show that phosphorus is
transferred from parental to progeny phage to the extent of 30 per cent at
yields of about 30 phage per infected bacterium, and that the p3~ in prema-
turely lysed cultures is almost entirely non-sedimentable, becoming, in fact,
acid-soluble on aging.
Similar measures of the transfer of P~ have been published by Putnam and
Kozloff (1950) and others. Watson and Maal~e (1952) summarize this work,
and report equal transfer (nearly 50 per cent) of phosphorus and adenine.
A Progeny of S85-Labeled Pkage Nearly Free from tke Parental
Labd.--The
following experiment shows dearly that the obligatory transfer of parental
sulfur to offspring phage is less than 1 per cent, and probably considerably
less. In this experiment, the phage yield from infected bacteria from which
the S3S-labeled phage coats had been stripped in the Waring blendor was
assayed directly for S s~.
Sensitive bacteria grown in broth were infected with five particles of SaS-labeled
phage per bacterium, the high ratio of infection being necessary for purposes of as-
say. The infected bacteria were freed from unadsorbed phage and suspended in water
containing per liter 1 na~ MgSO4, 0.1 m~r CaC12, and 0.1 gin. gelatin. A sample of
this suspension was agitated for 2.5 minutes in the Waring blendor, and centrifuged
to remove the extraceUular S% A second sample not run in the blendor was centri-
fuged at the same time. The cells from both samples were resuspended in warm
salt-poor broth at a concentration of l0 s bacteria per ml., and aerated for 80 min-
utes. The cultures were then lysed by the addition of 0.02 mM HCN, 2 × 10 u UV-
killed T2, and 6 rag. NaCI per ml. of culture. The addition of salt at this point causes
S as that would otherwise be eluted (Hershey
et al.,
1951) to remain attached to the
A. D. HERSHEY AND MARTHA CHASE 53
bacterial debris, The lysates were fractionated and assayed as described previously,
with the results shown in Table VIII.
The data show that stripping reduces more or less proportionately the S aS-
content of all fractions. In particular, the SSS-content of the fraction contain-
ing most of the phage progeny is reduced from nearly 10 per cent to less than
1 per cent of the initially adsorbed isotope. This experiment shows that the
bulk of the S 36 appearing in all lysate fractions is derived from the remains of
the coats of the parental phage particles.
Properties of Phage Inactivated by Formaldehyde.--Phage
T2 warmed for
1 hour at 37°C. in adsorption medium containing 0.1 per cent
(v/v)
com-
mercial formalin (35 per cent HCHO), and then dialyzed free from formalde-
TABLE VIII
Lysates of Bacteria Infected wltk SSS-Zabeled 7"2 and Stripped in the Waving Blendor
Per cent of adsorbed S Is or of phage yield:
Eluted in blendor fluid
1st low-speed sediment..
2nd
" " "
High-speed "
" " supernatant.
Recovery .....
Ceils stripped
ss~
86
3.8
(0.2)
(0.7)
(2.0)
93
Phage
9.3
11
58
1.1
79
Ceils
not stripped
Sis
Phage
39
'31 13
2.7 11
9.4 89
0.7) 1.6
84 115
All the input bacteria were recovered in assays of infected cells made during the latent
period of both cultures. The phage yields were 270 (stripped cells) and 200 per bacterium,
assayed before fractionation. Figures in parentheses were obtained from counting rates
close to background.
hyde, shows a reduction in plaque titer by a factor 1000 or more. Inactivated
phage of this kind possesses the following properties.
1. It is adsorbed to sensitive bacteria (as measured by either S 36 or ps~
labels), to the extent of about 70 per cent.
2. The adsorbed phage kills bacteria with an efficiency of about 35 per
cent compared with the original phage stock.
3. The DNA of the inactive particles is resistant to DNase, but is made
sensitive by osmotic shock.
4. The DNA of the inactive particles is not sensitized to DNase by adsorp-
tion to heat-killed bacteria, nor is it released into solution by adsorption to
bacterial debris.
5. 70 per cent of the adsorbed phage DNA can be detached from infected
cells spun in the Waring blendor. The detached DNA is almost entirely re-
sistant to DNase.
54
VIRAL PROTEIN" AND
NUCLEIC ACID IN BACTEKIOPHAGE GROWTH
These properties show that T2 inactivated by formaldehyde is largely in-
capable of injecting its DNA into the cells to which it attaches. Its behavior in
the experiments outlined gives strong support to our interpretation of the cor-
responding experiments with active phage.
DISCUSSION"
We have shown that when a particle of bacteriophage T2 attaches to a
bacterial cell, most of the phage DNA enters the cell, and a residue contain-
ing at least 80 per cent of the sulfur-containing protein of the phage remains
at the cell surface. This residue consists of the material forming the protective
membrane of the resting phage particle, and it plays no further role in infec-
tion after the attachment of phage to bacterium.
These facts leave in question the possible function of the 20 per cent of sul-
fur-containing protein that may or may not enter the cell. We find that little
or none of it is incorporated into the progeny of the infecting particle, and that
at least part of it consists of additional material resembling the residue that
can be shown to remain extracellular. Phosphorus and adenine (Watson and
Maal~e, 1952) derived from the DNA of the infecting particle, on the other
hand, are transferred to the phage progeny to a considerable and equal ex-
tent. We infer that sulfur-containing protein has no function in phage multi-
plication, and that DNA has some function.
It must be recalled that the following questions remain unanswered. (1)
Does any sulfur-free phage material other than DNA enter the cell? (2) If
so, is it transferred to the phage progeny? (3) Is the transfer of phosphorus (or
hypothetical other substance) to progeny direct--that is, does it remain at
all times in a form specifically identifiable as phage substance or indirect?
Our experiments show clearly that a physical separation of the phage T2
into genetic and non-genetic parts is possible. A corresponding functional
separation is seen in the partial independence of phenotype and genotype in
the same phage (Novick and Szilard, 1951; Hershey et a/., 1951). The chemi-
cal identification of the genetic part must wait, however, until some of the
questions asked above have been answered.
Two facts of significance for the immunologic method of attack on problems
of viral growth should be emphasized here. First, the principal antigen of the
infecting particles of phage T2 persists unchanged in infected cells. Second, it
remains attached to the bacterial debris resulting from lysis of the cells. These
possibilities seem to have been overlooked in a study by Rountree (1951) of
viral antigens during the growth of phage T5.
S~ARY
1. Osmotic shock disrupts particles of phage T2 into material containing
nearly all the phage sulfur in a form precipitable by antiphage serum, and
capable of specific adsorption to bacteria. It releases into solution nearly all
A. D. HERSHEY AND MARl'HA CHASE
55
the phage DNA in a form not precipitable by antiserum and not adsorbable
to bacteria. The sulfur-containing protein of the phage particle evidently
makes up a membrane that protects the phage DNA from DNase, comprises
the sole or principal antigenic material, and is responsible for attachment of
the virus to bacteria.
2. Adsorption of T2 to heat-killed bacteria, and heating or alternate freezing
and thawing of infected cells, sensitize the DNA of the adsorbed phage to
DNase. These treatments have little or no sensitizing effect on unadsorbed
phage. Neither heating nor freezing and thawing releases the phage DNA
from infected cells, although other cell constituents can be extracted by these
methods. These facts suggest that the phage DNA forms part of an organized
intracellular structure throughout the period of phage growth.
3. Adsorption of phage T2 to bacterial debris causes part of the phage
DNA to appear in solution, leaving the phage sulfur attached to the debris.
Another part of the phage DNA, corresponding roughly to the remaining half
of the DNA of the inactivated phage, remains attached to the debris but can
be separated from it by DNase. Phage T4 behaves similarly, although the
two phages can be shown to attach to different combining sites. The inactiva-
tion of phage by bacterial debris is evidently accompanied by the rupture of
the viral membrane.
4. Suspensions of infected cells agitated in a Waring blendor release 75 per
cent of the phage sulfur and only 15 per cent of the phage phosphorus to the
solution as a result of the applied shearing force. The cells remain capable of
yielding phage progeny.
5. The facts stated show that most of the phage sulfur remains at the cell
surface and n~ost of the phage DNA enters the cell on infection. Whether
sulfur-free material other than DNA enters the cell has not been determined.
The properties of the sulfur-containing residue identify it as essentially un-
changed membranes of the phage particles. All types of evidence show that
the passage of phage DNA into the cell occurs in non-nutrient medium under
conditions in which other known steps in viral growth do not occur.
6. The phage progeny yielded by bacteria infected with phage labeled
with radioactive sulfur contain less than 1 per cent of the parental radioactiv-
ity. The progeny of phage particles labeled with radioactive phosphorus con-
tain 30 per cent or more of the parental phosphorus.
7. Phage inactivated by dilute formaldehyde is capable of adsorbing to
bacteria, but does not release its DNA to the cell. This shows that the inter-
action between phage and bacterium resulting in release of the phage DNA
from its protective membrane depends on labile components of the phage
particle. By contrast, the components of the bacterium essential to this inter-
action are remarkably stable. The nature of the interaction is otherwise un-
known.
8. The sulfur-containing protein of resting phage particles is confined to a
$6 VIRAL PROTEIN AND NUCLEIC ACID Ilq BACTERIOPHAGE GROWTH
protective coat that is responsible for the adsorption to bacteria, and functions
as an instrument for the injection of the phage DNA into the cell. This pro-
tein probably has no function in the growth of intracelIular phage. The DNA
has some function. Further chemical inferences should not be drawn from the
experiments presented.
REFERENCES
Anderson, T. F., 1949, The reactions of bacterial viruses with their host Cells, Bot.
Rev., 15, 464.
Anderson, T. F., 1951, Tr. New York Acad. So., 13, 130.
Anderson, T. F., and Doermann, A. H., 1952, J. Gen. Physiol., 35, 657.
Benzer, S., 1952, J. Bact., 63, 59.
Doermann, A. H., 1948, Carnegie Institution of Washington Yearbook, No. 47, 176.
l)oermann, A. H., and Dissosway, C., 1949, Carnegie Institution of Washingf.o~a Yexw-
book, No. 48, 170.
Dulbecco, R., 1952, J. Bact., 63, 209.
Herriott, R. M., 1951, J. Baa., 61, 752.
Hershey, A. D., 1946, Genaics, 31, 620.
Hershey, A. D., Roesel, C., Chase, M., and Forman, S., 1951, Carnegie I~titution of
Washington Yearbook, No. 50, 195.
Lwoff, A., and Gutmann, A., 1950, Ann. Inst. Pasteur, 78, 711.
MaalCe, O., and Watson, J. D., 1951, Proc. Na. Acad. Sc., 37, 607.
Novick, A., and Szilard, L., 1951, Science, 113, 34.
Prater, C. D., 1951, Thesis, University of Pennsylvania.
Price, W. H., 1952, J. Gen. Physiol., 35, 409.
Putnam, F. W., and Kozloff, L.,1950, J. Biol. Chem., 182, 243.
Rountree, P. M., 1951, Brit. Y. Exp. Path., 39., 341.
Watson, J. D., and Maal~e, O,, 1952, Aaa path. a raicrobiol, stand., in press.

Discussion

Tap the paper where you want to comment
This seminal paper by Hershey and Chase confirmed that DNA, and not protein, is the hereditary molecule. Alfred Hershey would later share the Nobel Prize in 1969 with Salvador Luria and Max Delbrück for "their discoveries concerning the replication mechanism and the genetic structure of viruses." Bacteriophages are viruses that infect bacteria and they resemble Ring Pops with spider legs. The phage body is composed entirely of protein, with the viral DNA enclosed in a head structure. Landing on a target cell induces conformational shifts in the phage body proteins that allow the tail to act as a spring-loaded syringe, penetrating the host cell membrane and depositing its DNA. What happens to the DNA next depends on the type of phage: Lytic phage immediately hijack the host's metabolic machinery to produce new phage until the cell bursts. Lysogenic phage incorporate their genomes into that of the host and remain dormant (a prophage) until cellular stress causes them to become lytic and produce more phage particles. As such, lysogenic prophage are passed to daughter cells during cell division. Throughout the paper, Hershey and Chase frequently refer to sulfur and phosphorous because their experiment harnesses the presence or absence of either in DNA or protein to distinguish between the two molecules. More specifically, sulfur is present in certain amino acids, and thus protein, while phosphorous is present in the backbone of the DNA double helix. The "isotopically labeled" phage to which their refer is one in which radioactively-labeled sulfur and phosphorous are present in its protein and DNA, respectively. Among other things, these data show that the "sulfur" (synecdochic for protein) is adsorbed to sensitive bacteria. This means that protein, not DNA, makes contact with the bacterial surface. This is interesting because it posits a custodial relationship between the phage protein and DNA. In absence of all other information, it is clear that DNA is precious enough cargo that it must be protected from DNase, an enzyme that cleaves it. In order for the phage to multiply, it must shed its protein coat. Notice how carefully they word this passage; a more outwardly-assertive scientist would include the obvious conclusion: that DNA is the replicative molecule. Instead, they opt for more a cautious phrasing, stating that shedding the protein coat exposes the phage DNA, making it susceptible to DNase cleavage. This is sensible: phage DNA is no longer protected after it is injected into the cell. The experiment examining the range of temperatures at which the DNA of unadsorbed (free-floating) phage is sensitive to cleavage is interesting because it shows a precipitous drop in phage protein stability between 80-90ºC. The phage capsid, the "head" protecting its DNA, is composed mainly of a single protein that self-associates into connected hexamers (160 in the case of the T phages used in this study). Thus, once the temperature at which that protein denatures is reached, the whole capsid falls apart, exposing the phage DNA. "DNAse" is a common, tempting typo. Again, "sulfur" stands for "protein" for the purposes of this study. The protein coat sediments with bacterial debris because it associates with the membrane. However, because these are membrane fragments, the phage shoots its DNA into the medium itself. This experiment is kind of unsettling because they test whether or not phage will attach to pieces of dead bacteria and try to inject their DNA. Given that viral protein-bacterial membrane interactions mediate phage adsorption, their results are not surprising. These are the classic "Waring blender" experiments we learn about in high school. Based on how phage attach to cells, Hershey and Chase attempt to dislodge the attached protein coats from infected bacterial cells using a kitchen blender. In the first, they find that the majority of phage sulfur is extracellular and, as such, has no apparent role in viral reproduction. Based on this, they reason that sulfur should not be transmitted during phage multiplication. They see that the progeny of phage labeled with isotopic sulfur show almost no radioactive label. By contrast, very little isotopic phosphorous (DNA) is present extracellularly and it is indeed passed from parental phage to progeny. The picture that emerges is that the phage protein is structural and protective while its DNA serves the multiplicative role. Notice how exhaustively they eliminate a role for the radiolabeled phage protein beyond adsorption. At the time, protein was thought to have a hereditary role, so their data would need to be hermetic contravene that thinking. Even though they show isotopic phosphorous passes from parent phage to progeny, they are hesitant to propose anything more than DNA having "some function." In truth though, that is all they show aside from protein having no function. Despite this work being depicted in textbooks as determinate proof that DNA is the hereditary molecule, would their results look any different if DNA were somehow associated with, but not directly responsible for, viral multiplication? To be completely aboveboard, I wrote the preceding comment prior to reading these questions, which follow the same reasoning: DNA may be associated with phage reproduction, but because their study was biased to the radiolabels used, there could be another, unlabeled molecule involved that interacts with DNA. The point is, this study does not rule that out. They understand that their results imply that DNA is the hereditary molecule but want to remain cautious until a more thorough study of the molecule confirms it. This is the result with which we are all familiar: sulfur (protein) remains on the outside of cells while phosphorous (DNA) stays in and survives successive generations.