Proc. Natl. Acad. Sci. USA
Vol. 75, No. 9, pp. 4102-4106, September 1978
Biochemistry

Nucleotide sequences related to the transforming gene of avian
sarcoma virus are present in DNA of uninfected vertebrates

(complementary DNA/RNA tumor viruses /oncogenesis)

DEBORAH H. SPECTOR, HAROLD E. VARMUS, AND J. MICHAEL BISHOP
Department of Microbiology, University of California, San Francisco, California 94143

Communicated by Howard M. Temin, May 22, 1978

ABSTRACT We have detected nucleotide sequences related
to the transforming gene of avian sarcoma virus (ASV) in the
DNA of uninfected vertebrates. Purified radioactive DNA
(cDNA,,,-) complementary to most or all of the gene (sre) re-
quired for transformation of fibroblasts by ASV was annealed
with DNA from a variety of normal species. Under conditions
that facilitate pairing of partially mal nucleotide sequences

1.5 M NaCl, 59°), cD formed duplexes with chicken,

uman, calf, mouse, and salmon DNA but not with DNA from
sea urchin, Drosophila, or Escherichia coli. The kinetics of
duplex formation indicated that cDNA,,,. was reacting with
nucleotide sequences present in a single copy or at most a few
copies per cell, In contrast to the preceding findings, nucleotide
sequences complementary to the remainder of the ASV genome
were observed only in chicken DNA. Thermal denaturation
studies of the duplexes formed with cDNA,arc indicated a high
degree of conservation of the nucleotide sequences related to
sre in vertebrate DNAs; the reductions in melting temperature
suggested about 3-4% mismatching of cDNAjgarc with chicken
DNA and 8-10% mismatching of cDNA,,,- with the other ver-
tebrate DNAs.

 

Nucleotide sequences encoding genomes of RNA tumor viruses
are present in the normal cellular DNA of a wide variety of
vertebrate species (see ref. 1 for review). The rate of evolution
of such sequences has been estimated from the final extent of
annealing and thermal stability of duplexes formed between
cellular DNA and virus-specific reagents. In general, virus-
specific DNA in cellular genomes appears to evolve at least as
rapidly as cellular unique-sequence DNA; this pattern has been
documented for viral DNA endogenous to primates (2), cats (3),
rodents (4), and birds (5-10). In some instances, the apparently
rapid evolution of virus-specific DNA could be explained by
its relatively recent introduction into the germ line of a species
by infection with a virus originating in another species (8).
In the cases cited, the hybridization reagents were not specific
for defined portions of the viral genome, and the endogenous
viruses under study were, in most cases, not oncogenic. Stehelin
et al. (11) isolated radioactive DNA (CDNAsare) complementary
to most or all of the viral gene(s) (sre) required for transfor-
mation of fibroblasts by avian sarcoma virus (ASV) (11-13).
cDNAg.rc annealed to DNA not only from normal chickens, the
presumed natural host for ASV, but also to DNA from several
other birds spanning 100 X 108 years of evolution (10). This
apparent conservation of nucleotide sequences related to src
(hereafter called “sarc sequences”) during avian speciation
contrasted sharply with the lack of conservation of sequences
encoding a nontransforming virus endogenous to chickens and
closely related to ASV. Moreover, the conservation of cellular
sare sequences suggested that they might have an important,
but as yet unknown, function, and that they might be present,
albeit in a diverged form, in vertebrates other than birds.

 

The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked “‘ad-
vertisement”™ in accordance with 18 U.S.C. §1784 solely to indicate
this fact.

4102

In the previous study (10), CDNAsgarc did not anneal signifi-
cantly to mammalian DNA (from calf and mouse) under rela-
tively stringent conditions of hybridization. In this study, we
have used conditions that would permit formation and detec-
tion of mismatched duplexes in order to test DNA from several
vertebrate and nonvertebrate species for cellular sare sequences.
Under these conditions, we find that CDNA,arc anneals with the
DNA of all tested vertebrates, but anneals only slightly to DNA
from sea urchin, and not at all to Drosophila and Escherichia
coli DNAs.

MATERIALS AND METHODS

Preparation of Cellular DNA. Unlabeled DNA was ex-
tracted from the following sources: 10- to 11-day-old chicken
embryos; liver and spleen of a GR mouse; human placental
tissue (provided by Y. W. Kan); sea urchin sperm and Dro-
sophila cells grown in culture (obtained from Brian McCarthy);
E. coli cells (provided by Herbert Boyer); and XC cells, derived
from a tumor produced in rats with the Prague C (Pr-C) strain
of ASV (14). Homogenized tissue or cells were suspended in
buffer containing 0.1 M NaCl, 10 mM Tris-HCl (pH 7.4), 1 mM
EDTA, 0.5% sodium dodecyl sulfate (NaDodSO,), and 500 pg
of Pronase per ml at 37° for 1-2 hr and extracted with phe-
nol/chloroform/isoamyl alcohol (25:24:1 vol/vol). The high
molecular weight DNA was then spooled out of solution after
addition of ethanol. High molecular weight salmon sperm and
calf thymus DNA (Worthington Biochemical Corp.) were
treated as above.

The DNA was reduced to 6-7 S and residual RNA was di-
gested by boiling in 0.8 M NaOH for 20 min. The samples were
then neutralized with HCl, adjusted to 0.5% NaDodSQ,, treated
with 500 yg of Pronase per ml for 45 min at 37°, extracted twice
with phenol/chloroform/isoamy! alcohol (25:24:1), and pre-
cipitated with ethanol.

14C-Labeled unique-sequence DNA was prepared from
BALB/c 8T3 cells or chicken embryo fibroblasts grown in the
presence of {}4C\thymidine. The DNA was denatured and
reannealed to a Cot (concentration of DNA in moles of nu-
cleotides X time in sec) of 500 in 0.6 M NaCl at 68°; the DNA
that did not reassociate was isolated by its failure to adsorb to
a hydroxylapatite column in 0.14 M phosphate buffer (pH 6.8)
at 60°. Unlabeled unique-sequence calf thymus DNA was
prepared in the same way except that it was selected as DNA
that had not reannealed at a Cot of 1.5 X 105.

 

Abbreviations: ASV, avian sarcoma virus; src, gene required for
transformation of fibroblasts by ASV; sarc, nucleotide sequences de-
leted in transformation-defective deletion mutant of Prague C strain
of ASV (Pr-C td); cellular sarc, nucleotide sequences of cellular origin
related to sarc; cCDNAsarc, purified single-stranded DNA comple-
mentary to most or all of the gene (src); Cot, concentration of DNA in
moles of nucleotide X time in sec; NaDodSOg, sodium dodecyl sul-
ate.
Biochemistry: Spector et al.

Preparation and Characterization of cDNAs. Preparation
of both °2P-labeled single-strand DNA complementary to the
RNA genome of the B77 strain of ASV (cDNAgv77) and of
{SH|cDNAgarc have been described (11). For preparation of
cDNAgarc, the genome of Prague C strain of ASV was tran-
scribed into complementary DNA; hybridization was then used
to select for DNA specific for the region missing from the ge-
nome of a transformation-defective (td) deletion mutant of
Pr-C of ASV. We have examined the complexity and compo-
sition of this CDNAgare by two distinct procedures,

First, we hybridized cDNAgarc to 92P-labeled ASV 70S RNA
at various ratios of DNA:RNA. A maximum of 10-18% of the
viral RNA formed RNase-resistant duplexes with cCDNAgarc; this
value was achieved when the complementary sequences of
DNA and RNA were annealed at a ratio of 1-3 (Table 1). These
findings indicate that CDN Agare is a reasonably uniform rep-
resentation of 10-13% of the ASV genome (about 1300 nu-
cleotides), which represents about 60% of the deletion in the
strain of td virus used for the selection (15-18).

Second, we isolated the viral RNA that hybridized to
cDNA,,,; and analyzed the two-dimensional chromatogram
of oligonucleotides released from this RNA by hydrolysis with
T1 RNase (Sankyo) in low salt according to published proce-
dures (19, 20); homochromatography was carried out with
Homomix B (20). Autoradiography of the dried chromatograms
was carried out with Kodak X-omat film and Dupont Cronex
Lightening Plus screens. For comparison, a similar analysis was
performed with the entire genome of Pr-C ASV. Our chro-
matogram of the viral genome was sufficiently congruent with
that published by Wang et al. (13) to permit identification of
the individual larger oligonucleotides by their position in the
two-dimensional pattern (Fig. 1A). In particular, we could
identify the two oligonucleotides (nos. 8 and 10) that Wang et
al. have assigned to the region of the ASV genome that encodes
src (13). The chromatogram of the RNA that hybridized to
cDNAgare contained two large oligonucleotides whose positions
corresponded approximately to those of oligonucleotides 8 and
10 (Fig. 1B). However, we could not recover sufficient radio-
activity in these oligonucleotides to verify their identity by

Table 1. Genetic complexity of cCDNAsarc*

 

Ratio % RNA
Sample (DNA/RNA)*t _ hybridized

Total cCDNAsare 0.4/1 3

Vi 10.5
2.3/1 10
3.3/1 ul
7.3/1 13

16/1 11.3
cDNAwgar, hybridized 1/1 9
to calf thymus DNA! 3.5/1 10

cDNAsarc that did 2.2/1 10.2
not hybridize to 6.9/1 12
calf thymus DNA! 22/1 12

 

* A constant amount of cDNAga,- (0.05 ng) was incubated with dif-
ferent amounts (0.02-1 ng) of 9?P-labeled 70S RNA (6 X 10? cpm/yg)
of Pr-C ASV at 68° in 5 wl of 0.8 M NaCl/10 mM Tris-HCI, pH 7.4/10
mM EDTA for 72 hr. Extent of hybridization of the RNA was
measured by hydrolysis with 50 yg of RNase A and 30 units of RNase
Tl per ml in 0.8 M NaCl/0.03 M sodium citrate at 37° for 45 min.

t Ratio of complementary nucleotide sequences.

1 CDNAsare was fractionated on the basis of its ability to hybridize with
calf thymus DNA. In this experiment, 35% of the cCDNAgare hybri-
dized with calf thymus DNA and 65% of the cDNAyar- did not hy-
bridize with the DNA.

Proc. Natl. Acad. Sci. USA 75 (1978) 4108

Homochromotography ~s

   

Electrophoresis — pe»

Fic. 1: Two-dimensional fractionation of RNase T1 digest of
Pr-C ASV RNA and of Pr-C ASV RNA complementary to cDNAgare.
32P-Labeled RNAs were digested with RNase T1, applied to a strip
of cellulose acetate, electrophoresed at pH 3.5, and transferred to a
thin-layer plate of DEAE-cellulose for homochromatography. (A)
32P_Labeled 70S Pr-C RNA. The numbers correspond to those used
by Wang et al. (13) for identification of RNase T1 oligonucleotides
of Pr-C ASV. (B) 32P-Labeled Pr-C RNA complementary to-cDNAgarc
was selected as follows: cCDNAgare (1.5 ng) was incubated with 10 ng
of 32P-labeled 70S RNA (6 x 104 cpm/ng) of PrC-ASV at 68° in 7.6
u] of buffer containing 0.9 M NaCl, 10 mM Tris-HCl (pH 7.4), 10 mM
EDTA, and 20 ug of yeast RNA for 16 hr. After hybridization the
mixture was adjusted to a final concentration of 0.3 M NaCl and 0.5
ig of RNase T'l per ml and incubated at 37° for 1 hr. The mixture was
then adjusted to 0.5% NaDodSO,, 1 mg of proteinase K per ml, and
0.2% diethylpyrocarbonate, incubated at 37° for 1 hr, and applied to
a 30-ml Bio-Gel A 0.5 column equilibrated with 0.6 M NaCl/10 mM
Tris-HCl, pH 7.4/10 mM EDTA/0.4% NaDodSO,. Yeast RNA (20 xg)
was added to the material in the void volume and the solution was
then extracted once with phenol/chloroform/isoamy) alcohol (25:24:1).
The aqueous phase was precipitated with 2 vol of isopropanol. The
precipitated RNA complementary to cDNAgar, was resuspended in
10 mM Tris-HCl, pH 7.4/10 mM EDTA, heated at 100° for 5 min,
quickly cooled, and reprecipitated with 2 vol of isopropanol before
redigestion with RNase T1.

chemical analysis. The remainder of the large oligonucleotides
isolated from the viral genome (Fig. 1A) allegedly map outside
the boundaries of src (13), and none of these oligonucleotides
were present in the chromatogram of the RNA that hybridized
to cCDNAgrc (Fig. 1B).

To prepare cDNA representing regions of the ASV genome
other than “sare,” we prepared a relatively uniform transcript
of the genome of Pr-C ASV as detailed elsewhere (21) and then
annealed it with RNA from td Pr-C virus at 4:1 ratio of
cDNA:RNA. The annealed product, cDNAtdrep, lacks se-
quences present in cDNAgre and is representative of other
portions of the ASV genome.

Annealing Conditions and Hydroxylapatite Fractionation
of DNA. Denatured DNAs were annealed with [SH]cDNAgare
or [SH]cDNAtd rep at 59° for up to 120 hr in 1.5 M NaCl/10 mM
Tris-HCl, pH 7.4/10 mM EDTA. Duplex formation was de-
tected by fractionation on columns of hydroxylapatite (Bio-
Gel-HTP, Bio-Rad} at 50°, The DNA was loaded onto columns
in 0.14 M phosphate buffer and eluted with 5 column volumes
of both 0.14 M phosphatebuffer (single-stranded DNA)and0.4
M phosphate buffer (double-stranded DNA). Fractions were
analyzed for acid-precipitable radioactivity.

Fractionation of cDNA,,,_ on Basis of Hybridization to
Calf Thymus DNA. [3H|]cDNAsare was annealed with dena-
tured, unique-sequence calf thymus DNA to a Cot greater than
2X 104 and fractionated on columns of hydroxylapatite. The
0.14 M phosphate buffer washes (containing cDNAgarc that did
not hybridize to the calf DNA) and the 0.4 M phosphate buffer
4104 Biochemistry: Spector et al.

washes (containing the calf DNAcDNAgac hybrids) were
pooled separately. A portion of each fraction was analyzed for
acid-precipitable radioactivity. To remove phosphate, we ad-
justed the remainder of each fraction to 5 mM cetyltrimethy]-
ammonium bromide, kept it at 0° for 10 min, and centrifuged
it at 10,000 rpm for 20 min at 4°. The pelleted nucleic acids
were resuspended in 1 M NaCl and precipitated with 2 vol of
EtOH.

To separate CDNA arc from the calf DNA, we denatured the
pooled eluates from hydroxylapatite and annealed them with
a large excess of 70S RNA from Pr-C under conditions where
the unique sequence calf thymus DNA would not reanneal with
itself or with the CDNAsare (Cot of calf DNA was less than 102).
After fractionation on hydroxylapatite at 60°, the cDNAsare*
RNA duplexes were precipitated with 5 mM cetyltrimethy]l-
ammonium bromide, resuspended in 1 M NaCl, and adjusted
to 0.8 M NaOH to hydrolyze the RNA (87°, 2 hr). After pre-
cipitation with ethanol, the two populations of cCDNAgar. (one
that had and one that had not hybridized with calf DNA) were
again hybridized to Pr-C 70S RNA, recovered as described
above, dialyzed extensively against 10 mM Tris-HCl, pH 7.2/10
mM EDTA, and precipitated with ethanol.

Thermal Denaturation of DNA‘DNA Duplexes. The
thermal stability of DNA-cDNAsarc duplexes was examined by
adjusting the reaction mixture to 0.14 M PO,°~ and applying
it to a 6-cm column of hydroxylapatite at 50°. After three
washes (8 ml each) with 0.14 M phosphate buffer, 1500 cpm
of 82P-labeled cCDNApz77 annealed with 17 ug of XC cell DNA
(Cot greater than 10,000) was applied to the hydroxylapatite
as an internal standard. The column was washed with four
additional 8-ml volumes of 0.14 M phosphate buffer at 50°, and
the temperature of the water bath was raised in 4° increments;
at each temperature the column was allowed to equilibrate and
then washed with 8 ml of 0.14 M phosphate buffer. The frac-
tions were analyzed for acid-precipitable radioactivity.

RESULTS

Annealing of cDNAgarc to Vertebrate DNAs. In preliminary
studies, Stehelin et al. did not observe significant annealing of
cCDNAgarc to calf thymus or mouse DNA under relatively
stringent conditions of hybridization (0.6 M NaCl, 68°) and of
assay (resistance to S; nuclease and hydroxylapatite chroma-
tography at 60°) (table 2 of ref. 10). However, when cDNAsgarc
was annealed to uninfected human, mouse, calf, or salmon
sperm DNA in 1.5 M NaCl at 59°, 18-43% of the cDNA be-
haved like duplex DNA by hydroxylapatite chromatography
at 50° (Table 2). More extensive annealing was observed with
DNA from normal chicken embryos and with DNA from XC
cells. Chicken DNA has been shown to contain sequences closely
related to CDNAgerc; XC cell DNA serves as a control for the
formation of duplexes between homologous DNAs since XC
cells (rat cells transformed by Pr-C ASV) contain approxi-
mately 20 copies of ASV DNA (22). In contrast, CDNAgarc an-
nealed only slightly to the DNA from an echinoderm (sea ur-
chin), and not at all to the DNAs from bacteria (E. colf) and an
insect (Drosophila) (Table 2).

The various cellular DNAs were also tested for hybridization
with single-stranded DNA (cDNAid-rep) complementary to the
RNA genome of variants of ASV that contain deletions in sre.
A major fraction of CDNAtd-rep hybridized with DNA from
uninfected chicken cells and from the ASV-transformed XC
cells, but there was little or no reaction (<4%) with human, calf,
mouse, salmon, sea urchin, Drosophila, or E. colt DNAs (Table
2). These results conform to the previously observed pattern (10)
in that sequences related to cDNAgare are conserved in DNA

Proc. Natl. Acad. Sci. USA 75 (1978)

Table 2. Homology between normal DNAs and cDNAgarc,
cDNAta-rep, and chicken single-copy DNA

 

% % chicken
cDNAtarep single-copy
% cCDNAsare in DNA in
DNA in duplexes duplexes duplexes
XC 63 + 4.6 (57-65) 67 —_
Chicken 58 + 3.6 (55-52) 66 90
Human 28 + 4.4 (22-34) 2.7 —
Calf 30 + 7.1 (21-43) 3.8 5
Mouse 26 + 6.3 (18-33) 17 —
Salmon 24 + 6.8 (18-34) 18 5
Sea urchin 6.3 + 2.3 (2.8-9) <1 —_
Drosophila 2.6 + 2.5 (0-5) <1 _—
E. coli 0.9 + 1.2 (0-2.4) 2.1 <1

 

Reaction mixtures contained denatured cellular DNA (XC, 1-3.8
mg/ml; chicken, 2.54.7 mg/m]; human, 3.3-5 mg/ml; calf, 3.5-4.4
mg/ml; mouse, 1.9-4.6 mg/ml; salmon, 4.3-7.6 mg/ml; sea urchin,
3.1-6.2 mg/ml; Drosophila, 1.9-2.5 mg/ml; E. coli, 1.9-3.8 mg/ml) and
[PH] cDNAgare (1.25 ng/ml, 25,000 epm/ml) or [SH}¢DNAta-rep (0.94
ng/ral, 19,000 cpm/ml), or 4C-labeled unique-sequence chicken DNA
(550 ng/ml, 15,000 cpm/ml) in a final volume of 0.08-0.16 ml con-
taining 1.5 M NaCl/10 mM Tris-HCl, pH.7.4/10 mM EDTA, and were
incubated for 96-120 hr. Duplex formation was measured by frac-
tionation on hydroxylapatite. Values have been corrected for back-
grounds of 2% for CHIcDNAsas, hybridization, 4.2% for the [#H]-
cDNAtd-rep hybridizations, and 5% for the unique sequence chicken
(!4C]DNA. hybridizations. Numbers in parentheses represent the
range of values obtained in several independent experiments; SD is
given.

from divergent species, whereas sequences related to other
portions of the ASV genome are found only in the DNA from
normal chickens, ring-necked pheasants (9), and ASV-infected
cells. There was also little or no annealing (<5%) of labeled
unique-sequence chicken DNA with DNAs from calf, salmon,
or E. coli. These results also document the specificity of duplex
formation between cDNAgre and the various cellular DNAs
and exclude the possibility that an artifact of the hydroxyla-
patite assay can explain our results.

The kinetics of the formation of duplexes between cCDNAgerc
and DNAs from chicken, mouse, human, and calf are shown _
in Fig 2. For comparison, the rates of reassociation of chicken
and mouse unique-sequence DNA were also assayed (Fig. 2A
and B). Since the rate of association between cDNAgy- and the
DNA was similar to the kinetics for the reassociation of unique
nucleotide sequences, we conclude that the sequences related
to cCDNAgar¢ are present as single or at most a few copies in each
haploid complement of the DNAs tested.

Most of the reactions between cDNAgar- and the DNAs from
vertebrates were incomplete; 18-43% of cDNAgr- formed
duplexes with mammalian and fish DNAs and 55-62% with
chicken DNA. Either some of the nucleotide sequences repre-
sented by cDNA, were absent from the vertebrate DNAs, or
the amounts of DNA complementary to CDNAgic in the hy-
bridization reactions were insufficient to drive the reactions to
completion. We used two experimental approaches to distin-
guish between these possibilities.

(¢) We isolated cCDNAgare that failed to reassociate with calf
thymus DNA and tested its capacity to form duplexes with
additional calf thymus DNA during a second period of incu-
bation. The reassociations were carried out as described for
Table 2 (final value of Cot 2 X 104), with at least 2.4 mg of calf
thymus DNA per ng of CDNAgarc. After the first reaction, 65%
of cDNAgarc eluted from hydroxylapatite as single-stranded
DNA, 35% as duplex. The single-stranded cDNAsarc was Tre-
covered from the column eluate, incubated with fresh dena-
Biochemistry: Spector et al.

 

 

 

 

 

 

 

 

 

 

= A gz /¢ 30
e 3 e
| 80 480 € 20 |
& s =

4 |
ao “08 1g
c eo? 2
5 2 | 30 §
27) 8 $0) ° 2
20 so § 20
5 3 é z
6 10 140.8 109
x 3

zx Le *

107 10° 10% 10? 10° 10%

Cyt, mol-sec/titer

Fic. 2. Annealing of cDNA,,;, to normal cellular DNAs. Dena-
tured DNAs (0.25-0.365 mg) were incubated with [SH]cDNAwgorc (0.1
ng, 2000 cpm) and !4C-labeled unique-sequence DNA in 0.08 ml of
1.5 M NaCl/10 mM Tris-HCl, pH 7.4/10 mM EDTA at 59? for various
amounts of time (6 min-120 hr). Duplex formation was measured by
fractionation on hydroxylapatite. (A) Chicken DNA (3.1 mg/ml) with
PH] cDNAgarc (1.25 ng/ml, 25,000 cpm/ml) (@) and unique-sequence
chicken [!4C]DNA (550 ng/ml, 15,000 cpm/ml) (0). (B) GR mouse
DNA (4.6 mg/ml) with BH]cDNAsgare (1.25 ng/ml, 25,000 cpm/ml) (@)
and unique-sequence mouse [!4C]DNA (500 ng/ml, 6900 cpm/ml) (0).
(C) Calf DNA (3.6 mg/ml) with [8H}cDNAsare (1.25 ng/ml, 25,000
cpm/ml). (D) Human DNA (3.3 mg/ml) with [SH]cDNAgarc (1.25
ng/ml, 25,000 cpm/ml).

tured calf thymus DNA, and fractionated on hydroxylapatite;
33% of the cDNA eluted as duplex (data not shown). These
results indicate that stoichiometric limitations impede the re-
action between CDNAgg;- and calf thymus DNA, and that the
extent of this reaction does not fully reflect the extent of
homology between the cDNA and nucleotide sequences in calf
thymus DNA.

(i) We used molecular hybridization to determine the ge-
netic complexity of the nucleotide sequences in CDNA sarc that
formed duplexes with calf thymus DNA. After reassociation
of CDNAsare with calf thymus DNA and fractionation on hy-
droxylapatite, both the single-stranded cDNAsarc and the
cDNAsare in duplexes were separated from the calf thymus
DNA. Both of the selected populations of cCDNAsarc were hy-
bridized with °2P-labeled 70S RNA of Pr-C ASV at various
ratios of DNA:RNA as described (Table 1). At saturation,
10-12% of the viral RNA formed RNase-resistant duplexes with
either sample of cCDNA,gr<; saturation occurred when comple-
mentary sequences of DNA and RNA were annealed at ratios
of 1-2. These results conform to those obtained with unfrac-
tionated CDNA sarc (Table 1). As expected, neither fractionated
nor unfractionated cDNAgare hybridized to 32P-labeled 70S
RNA of Pr-C tdASV (data not shown).

We conclude that the CDNAgarc that hybridized to calf thy-
mus DNA represents the same proportion of the ASV genome
as the unselected cDNAgarc. The failure of CDNAga,, to hybri-
dize completely with the DNA from vertebrate cells is probably
due to stoichiometric limitations imposed on the reactions by
mismatched base pairs (see Table 3).

Thermal Stability of Duplexes Formed between cDNAgarc
and Vertebrate DNAs. In order to analyze the degree of mis-
matching in duplexes formed between cDNA gare and the DNA
from various species, we determined the stability of the du-
plexes by elution from hydroxylapatite columns with a thermal
gradient (Fig. 3 and Table 3). The thermal elutions were stan-
dardized internally by adding duplexes formed in a separate
annealing reaction between the ASV sequences in XC cell DNA
and the homologous [82P}cDNAgz77 (tm = 79° + 1°); since the
virus used to prepare cCDNAgr7 consisted of at least 90% td

Proc. Natl. Acad. Sct. USA 75 (1978) 4105

 

100} r r

BEES
_

J a 4 1

 

=

% radioactivity eluted
a ® @
oo 6 8
T
T

20; r

 

 

 

 

 

L 1. 1 1 1 jo
60 70 80 90 60 70 80° 90

Temperature, °C -

Ld
50 60 70 80 90

Fic. 3. Thermal denaturation of duplexes formed between
cDNAsare and cellular DNAs. Denatured DNAs (0.4-1 mg) were in-
cubated with [3H]cDNAgere (0.2 ng, 4000 cpm) in 0.16 ml of 1.5 M
NaCl/10 mM Tris-HCl, pH 7.4/10 mM EDTA at 59° for 96 hr (final
Cot > 25,000). The samples were applied to hydroxylapatite and de-
natured with a thermal gradient. Denaturation was standardized
internally by adding duplexes formed in a separate annealing reaction
between XC cell DNA and [°2P]cDNAgz77. @, Duplexes with cCDNAgarc}
O, duplexes between cCDNAg7; and XC DNA (internal standard). (A)
XC DNA (0.45 mg); (B) salmon DNA (1 mg); (C) chicken DNA (0.4
mg); (D) human DNA (0.53 mg); (E) calf DNA (0.58 mg); (F) mouse
DNA (0.73 mg).

variants, the complementary DNA synthesized with this virus
was deficient in sequences homologous to cDNAwgarc- Duplexes
between cDNAgr. and DNA from XC cells had the highest tm
(79°; Table 3). Although rat cells presumably contain endog-
enous cellular sarc sequences, CDNAgarc Should react principally
with the 20 copies of ASV provirus contained in XC cells. As
shown previously (10), the duplexes between cDNAsgarc and
chicken DNA were slightly less stable; the 4° depression of tn
represents about 3% mismatching of bases (23). Although the
duplexes between cDNAsere and mammalian or fish DNA were
even less stable (tm = 65-66.5°; Table 2), the reductions in t,
suggest only about 8-10% mismatching of bases (23). We con-
clude that vertebrate DNAs contain a highly conserved set of
nucleotide sequences that are related to at least a portion of the
transforming gene(s) of ASV.

Table 3. Thermal stabilities of duplexes between cDNAsere

 

 

and cellular DNAs
% DNAgare
DNA in duplexes tin Atm

XC 65 19 0
Chicken 55 74.5 4.5 -
Human 29 66.5 12.5
Calf 24 65 14
Mouse 28 65.5 —13.5
Salmon 19 66.5 ~12.5

 

Denatured DNAs were annealed with [7H]cDNAgar- and adsorbed
to hydroxylapatite as described in legends to Fig. 2 and Table 2. The
values of tm have been normalized to at, of 79° for XC DNA hybri-
dized to [#2P]cDNAgz, included as an internal standard in each
analysis.
4106 Biochemistry: Spector et al.

DISCUSSION

Stehelin et al. (10) have reported that although nucleotide se-
quences related to src of ASV were present and highly con-
served in the DNA of various avian species ranging from
chicken to emu, similar sequences were not detected in the
DNA of two mammalian species (mouse and calf) when hy-
bridizations were performed under relatively stringent con-
ditions. In this report we have shown that by reducing the
stringency of conditions for the formation and measurement
of duplexes, at least partial homology with src can be detected
in the DNA of all vertebrate species tested but not in the DNA
from invertebrates such as sea urchins and Drosophila or from
E. coli.

It is difficult, however, to determine the exact degree of
homology between the cellular DNA and src. Thermal dena-
turation studies of duplexes formed between cCDNAgarc and the
various DNAs permit an estimate of base mismatching between
cDNAgye and cellular DNAs (3-4% for chicken DNA, 6-8% for
other avian DNAs, and 8-10% for the fish and mammalian
vertebrate DNAs). However, these studies were limited by the/
fact that only a portion of cCDNAgarc reacted with vertebrate
DNAs; 18-43% formed duplexes with mammalian and fish
DNAs and 55-62% with chicken DNA. Since the reaction be-
tween mismatched complementary nucleotide sequences is
slower than that between well-matched sequences, the reaction
of cellular DNAs with cCDNAgre may not reach completion
before the homologous cellular DNA strands have completely
reassociated. Thus, our techniques provide minimum estimates
of the extent of annealing, and maximum estimates of the fi-
delity of base pairing, since the most well-matched duplexes
are most likely to form. However, we have demonstrated (Table
1) that the population of cDNAsgarc that hybridized to a mam-
malian DNA (calf) has the same genetic complexity as the un-
annealed cDNAgarc- We conclude that at least a major fraction
of the nucleotide sequences in CDNAgarc have homologues in
mammalian DNA. In quail cells, we have shown that up to 90%
of CDNAgare can hybridize to RNA present in uninfected cells
(24); this indicates that most of the sequences in cCDNAgarc are
represented in avian DNA, although the annealing of cCDNAsarc
to cel] DNA is incomplete.

The kinetics of the hybridization of CDNAsare with DNAs
from chicken, mouse, human, and calf are similar to the kinetics
of reassociation of cellular DNA present in a single copy per
haploid cellular genome (Fig. 2). Although mismatched base
pairs can slow the rate of annealing between complementary
nucleotide sequences (25, 26), this effect should be no greater
than 5-fold for any of the reactions between CDNAgarc and the
cell DNAs we have studied (3-10% mismatching). Thus, a
maximum of 5-10 copies of DNA related to cCDNAsarec can be
present in each cell.

According to the fossil record, the separation of birds,
mamunals, and teleosts occurred 400 million years ago (27-29);
throughout this period at least a portion of the avian cellular
sarc sequences has been conserved. Moreover, in a large number
of avian tissues and cells, we have recently found cellular sarc
sequences to be universally transcribed into RNA, albeit at low
concentrations (24). Although the conservation and ubiquity
of expression of the cellular sarc sequences suggest that they
might perform some critical function, the nature of that func-
tion remains obscure, and no protein product of the cellular sarc
sequence has been identified.

Proc. Natl. Acad. Sci. USA 75 (1978)

We thank Y. W. Kan, Brian McCarthy, Herbert Boyer, Janet Stav-
nezer, and Don Fujita for the reagents provided and K. Smith for
technical assistance. This work was supported by grants from the U.S.
Public Health Service (CA 12705, CA 19287, 1T32, and CA 09043) and
the American Cancer Society (VC-70), and Contract NO1 CP 33293
within The Virus Cancer Program of the National Cancer Institute,
National Institutes of Health, PHS. D.H.S. was supported by the Helen
Hay Whitney Foundation. H.E.V. is the recipient of a Research Career
Development Award (CA 70193) from the National Cancer Insti-
tute.

1. Aaronson, S. A. & Stephenson, J. R. (1976) Biochim. Biophys. Acta
458, $23-354.
2. Benveniste, R. E. & Todaro, G. J. (1974) Proc. Natl. Acad. Sci.
USA 71, 4513-4518.

8. Benveniste, R. E. & Todaro, G. J. (1974) Nature (London) 252,

456-459.

Morris, V. L., Medeiros, E., Ringold, G., Bishop, J. M. & Varmus,

H. E. (1977) J. Mol. Biol. 114, 73-92.

Kang, C. Y. & Temin, H. M. (1974) J. Virol. 14, 1179-1188.

Neiman, P. E. (1973) Virology 53, 196-204.

Shoyab, M. & Baluda, M. A. (1975) J. Virol. 16, 1492-1502.

Varmus, H. E., Heasley, S. & Bishop, J. M. (1974) J. Virol. 14,

895-908.

Tereba, A., Skoog, L. & Vogt, P. K. (1975) Virology 65, 524-

534.

10. Stehelin, D., Varmus, H. E., Bishop, J. M. & Vogt, P. K. (1976)
Nature (London) 260, 170-178.

ll. Stehelin, D., Guntaka, R. V., Varmus, H. E. & Bishop, J. M. (1976)
J. Mol. Biol. 101, 349-365.

12. Bernstein, A., MacCormick, R. & Martin, G. S. (1976) Virology
70, 206-209.

13. Wang, L.-H., Duesberg, P., Beemon, K. & Vogt, P. K. (1975) J.
Virol. 16, 1051-1070.

14. Svoboda, J., Chyle, P. P., Simkovic, D. & Hilgert, I. (1963) Folia
Biol. (Prague) 9, 77-91.

15. Duesberg, P. H. & Vogt, P. K. (1970) Proc. Natl. Acad. Sct. USA
67, 1673-1680.

16. Lai, M.M.C., Duesberg, P. H., Horst, J. & Vogt, P. K. (1973) Proc.
Natl. Acad. Sci. USA 70, 2266-2270.

17. Duesberg, P. H. & Vogt, P. K. (1973) J. Virol. 12, 594-599.

18. Neiman, P. E., Wright, S. E., McMillin, C. & MacDonnell, D.
(1974) J. Virol. 13, 8837-846.

19. Friedrich, R., Morris, V. L., Goodman, H. M., Bishop, J. M. &
Varmus, H. E. (1976) Virology 72, 330-340.

20. Barrell, B. G. (1971) in Procedures in Nucleic Acid Research, eds.
Cantoni, A. L. & Davies,'D.! R: (Harper and Row, New York),
Vol. 2, pp. 751-795.

21. Deng, C. T., Stehelin, D., Bishop, J. M. & Varmus, H. E. (1976)
Virology 76, 313-330.

22. Varmus, H. E., Vogt, P. K. & Bishop, J. M. (1978) Proc. Natl.
Acad. Sci. USA 70, 3067-3071.

23. Ullman, J. S. & McCarthy, B. J. (1973) Biochim. Biophys. Acta
294, 405-424.

24. Spector, D. H., Smith, K., Padgett, T., McCombe, P., Roulland-
Dussoix, D., Moscovici, C., Varmus, H. E. & Bishop, J. M. (1978)
Cell 13, 371-379.

25. McCarthy, B. J. & Farquhar, M. N. (1972) Brookhaven Symp.
Biol, 23, 1-48.

26. Bonner, T. 1, Brenner, D. J., Neuteld, B. R. & Britten, R. J. (1973)
J. Mol. Biol. 81, 123-135.

27. Young, J. Z. (1962) The Life of Vertebrates (University of Chi-
cago Press, Chicago).

28. Romer, A. S. (1966) Vertebrate Paleontology (University of
Chicago Press, Chicago).

29. Bakker, R. (1975) Sct. Am. 232, 58-78.

>

of AIA MK