J. gen. Virol. (I979), 4z, 1-26
Printed in Great Britain
REVIEW
ARTICLE
Structure, Replication, and Recombination of
Retrovirus Genomes: Some Unifying Hypotheses
By J O H N M. C O F F I N
Department of Molecular Biology and Microbiology and Cancer Research Center,
Tufts University School of Medicine, 136 Harrison Avenue,
Boston, Massachusetts o21I 1, U.S.A.
INTRODUCTION
The study of the molecular biology of retroviruses has revealed a number of features not
found in other groups of viruses. In recent years, a large number of excellent reviews of the
field have appeared (Table 1). In writing this article, it is not my intention to cover ground
already well treated in review form. Instead, I will consider only a limited number of interrelated subjects dealing with the organization of information in tumour virus genomes and
its transfer from parent to progeny. In general, I will ignore primary data and present only
the inferred structures and mechanisms and point out experimental approaches that might
fill in gaps in our knowledge.
Retroviruses form a large and diverse group, with isolates from many different vertebrates.
In most cases, our knowledge of the molecular biology of the viruses is derived from one
model system, the avian sarcoma-leukosis viruses (ASV). This bias is simply because these
are the longest known of the retroviruses and, in general, the easiest to deal with in the
laboratory. Except in a few cases where murine leukaemia viruses (MuLV) give cleaner
results, all of the examples I will use come from the ASV system. I believe that the general
picture that emerges will extend to all retroviruses, although details will differ in
different viruses.
By way of background, the following is a summary of some principal features of the
molecular biology of retroviruses.
Genes. The retrovirus genome has a coding capacity for about 3 × ~oS daltons of protein
(Beemon et al. I974; Billeter et al. 1974). All non-defective viruses have at least three genes (in
this case, a gene is defined as the region translated into a primary protein product): gag, which
codes for a precursor which is cleaved to yield four internal structural proteins of the virion;
pol, which codes for the virion RNA-directed D N A polymerase (reverse transcriptase); and
env, which codes for the virion envelope glycoproteins. These three genes may code for all
the information necessary for virus replication. A fourth region, e (for 'common'), is
found in a number of viruses (Wang et al. x975, I976). Although there is very tentative
evidence that it may have its own m R N A (Krzyzek et al. I978) and gene product (Purchio
et al. 1977), there is no genetic evidence for such an extra gene. A number of retrovirus
strains contain an additional gene which appears to have been added to the virus genome
by recombination with host cell information (Stehelin et al. I976). In general, this added
information appears to confer on the virus the ability to induce neoplastic disease with a
very short latent period, although genetic proof of this point exists only for the src gene of
ASV (Toyoshima & Vogt, I969; Martin, 197o). In non-defective avian sarcoma viruses, the
transforming gene is present in addition to gag, pol, env, and e. In other cases, it replaces a
I-:2
0O22-I317/79/0O0O-3487 $O2.0O © I979 S G M
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
2
J.M. COFFIN
Table 1. S o m e recent reviews on retroviruses*
Subject
Genetics
Genome structure
Biochemistry
Endogenous viruses
Transformation
References
Vogt 0977)
Vogt & Hu (I977)
Hunter (2978)
Beemon (2978)
Bishop et al. 0978)
Verma (I977)
Taylor (2977)
Eisenmann & Vogt 0978)
Weinberg (1977)
Robinson (I978)
Hanafusa (1977)
* This is not a complete list; only reviews with publication dates in I977 and I978 are included.
virus g e n e - e n v - in replication defective ASV; possibly parts of gag-pol in transforming
avian leukosis virus (such as MC29; P. Mellon, P. Duesberg & P. K. Vogt, personal communication) and murine sarcoma virus (Hu et al. 1977).
Replication. The general outline of retrovirus replication has been known for a long time
(Temin & Baltimore, 1972 ). Soon after entering a cell, the single-stranded virion R N A is
copied into double-stranded D N A by the polymerase carried in the virion. This D N A
provirus is transported to the nucleus and integrated into the cell genome. The integrated
provirus serves as a template for host R N A polymerase which synthesizes progeny virion
and messenger RNA. These R N A species are processed into forms resembling cell m R N A
and transported to the cytoplasm where virus protein synthesis and assembly of virions
Occur.
A prominent feature of this overall mechanism is that all steps in virus nucleic acid
replication seem to be carried out either by cellular systems or by virus coded enzyme(s) in the
virion itself and not by virus coded enzymes synthesized after infection. It should also be
pointed out that, in general, retrovirus replication does not lead to death o f the host cell.
Instead, the cell usually survives the infection, continues to divide and becomes permanently
virus producing. When a steady state is reached (one to two days after infection) approx. I
of cell R N A and protein synthesis, and lO -3 to IO-a ~ of D N A synthesis, is of viral species.
Diversity. Retroviruses appear to have a genetic plasticity unparalleled in other R N A
viruses. It is possible that no two isolates of exogenous virus from the same species are
identical to one another. Different laboratory strains of ASV, for example, differ in a
number of respects from each other, including point mutations, and additions and deletions
o f information. While the point mutation rate for ASV may not be significantly different
from other R N A viruses, the adaptability of virus populations to new selective conditions
seems to be quite high (J. Coffin & C. Barker, unpublished data). This phenomenon is likely
to be due in part to the extraordinarily high frequency of recombination between retrovirus
genomes, even as compared to D N A viruses. Additionally, deletion of non-essential
information (i.e. src) is an extremely frequent event (Vogt, I97Ia; Martin & Duesberg, 1972)
and it is quite difficult to maintain a population of ASV free of such deletion mutants. Frequent
deletions of essential information are also found (Shields et aL 1978). Addition o f cellular
to virus information does not appear to be as frequent, but is at least historically detectable,
as discussed above. It is likely that recombination of host genes and virus information happens
quite frequently, but lack of a suitable selective system prevents its detection.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
R e v i e w : Retrovirus g e n o m e s
3
In the sections that follow I will attempt to relate these various features of retrovirus
biology to one another. The underlying theme of the article will be the interrelationship of
structural features of the virus genome and the mechanism of the enzyme systems responsible
for its replication. In this way it is possible to develop an overall picture of retrovirus
biology which relates many diverse features in a reasonably simple way.
Structure and expression o f the viral genome
The genetic structure of ASV has been worked out primarily by oligonucleotide mapping
of deletion mutants and recombinants between different strains of viruses (Joho et al. 1975,
I976; Wang et al. I975; Coffin & Billeter, I976). The best available evidence yields the map
shown in Fig. I(A). This map is drawn to approximate scale. Although it is not yet known
exactly where the boundaries between the different coding regions lie, they are certainly in
the order 5'-gag-pol-env-src-c-3'. The order in which the four gag and two env polypeptides
are encoded can be deduced from pactamycin mapping experiments which locate them on
their precursor polypeptides (Vogt et al. I975; Diggelman & Klemenz, 1978; Shealey &
Rueckert, 1978). The coding portion of the genome must also contain control and recognition
sequences for m R N A processing, translation initiation and so forth, but these have not yet
been characterized and probably will not be until major portions of nucleotide sequence
have been determined.
The genome resembles typical eukaryotic mRNAs in its pattern of post-transcriptional
modifications. The 3' end is polyadenylated with approx. 2oo A residues (Lai & Duesberg,
r972), the 5' end is capped with an mTGS'pppS'Gm group (Furuichi et al. I975; Keith &
Fraenkel-Conrat, I975) and the RNA is methylated internally with approx. IO mnA residues
(Beemon & Keith, I976). The function of these modifications is not clearly understood.
In the virion, the native state of the genome is apparently a dimer of two such molecules,
linked by hydrogen bonds near their 5' ends (Kung et al. I976). Such a structure has not
been visualized for ASV, presumably because of the weakness of the dimer linkage and has
been inferred by analogy to other retroviruses. To obtain such a linkage between identical
molecules, there must either be a nucleic acid 'linker' base paired to both genomes, or a
palindromic sequence in each molecule at the linkage point. A linkage structure for ASV
has been proposed on the basis of the nucleotide sequence of the 5' end of the genome
(Haseltine et aL r977). This structure uses both a palindromic sequence and the tRNA '~p
primer as a linker. Such a structure is too near the 5' end to account for the structure seen
with other viruses and remains to be tested.
Since each molecule in the dimer contains the full genetic complement of the virus, the
virus is at least chemically diploid. This conclusion was first reached by dividing the weight
of high mol. wt. R N A present per virion by its sequence complexity as determined from
oligonucleotide fingerprinting (Beemon et al. I974; Billeter et al. I974). As I will discuss
laler, it remains to be established that the virus is genetically diploid - that is, that both
subunits make available all their genetic information to the progeny. It should be noted here
that, although there are on average two genome subunits per virion, there may exist in a
virus population substantial proportions of virions with three or more subunits as well as
virions with no genome RNA. This possibility is supported by the observation that, under
certain experimental conditions, more or less normal 'empty' virions can be prepared
(Levin et al. 1974; M. Linial, personal communication), suggesting that there is no stringent
requirement of precisely one dimer for virion assembly.
In addition to its genome, the virion contains numerous small R N A molecules derived
from the host cell (Sawyer & Dahlberg, I973). The vast majority of these are probably
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
4
J.M. COFFIN
Table 2. T e r m i n a l l y r e d u n d a n t s e q u e n c e s in A S V *
(I)
m7
(2) m 4
G p p p
G
Gm
p p p G~
5' terminus
t
t
t t
C C A U U U U A C C A U U C A C C A C A
C C A
U
U
C
U
A C C U
C
U
C A C C A C A
(3) m 7 G p p p G m C C A U U U U A C (C C C U C ) A C C A C A
(4) m ~ G p p p Gm C (C A U U U U ) G [C C A U U C A C C A C A]
(I)
(2)
(3)
(4)
3' terminus
G C C A U U U U A C C A U U C A C C A C A Azoo
G
C C A
U
U
C
U
A C C U
C
U
C A C C A
C A Azoo
G C C A U U U U A C (C C C U C) A C C A C A A~oo
[G C C A U U U U] G C C A U (U C)(A C C A C A) A2oo
* Sequences r and 2 have been determined exactly. Sequences 3 and 4 are tentative and derived by fitting
partial products to give minimum differences from the first two. Sequence I is found in Pr-A, B, and C,
B77 and RAV-o (Haseltine et aL ~977; Shine et aL t977; Coffin et aL I978a); sequence 2 in AMV (Stoll
et aL I977); sequence 3 in RAV-6 (Joho et al. I978); and sequence 4 in RAVq, RAV-2, and SR-D (Wang
et al. I977).
t Indicates the minimum points of sequence divergence. Nucleotides whose order is uncertain are in
parentheses and sequences not analysed but inferred from the other end are in square brackets.
fortuitous contaminants with no role in the life o f the virus, but one o f them, a molecule o f
t R N A ' % serves the important function o f primer for synthesis o f proviral D N A (Dahlberg
e t al. I974)- The observation by Taylor & Illmensee 0 9 7 5 ) that the t R N A trp is hydrogen
b o n d e d very near the 5' end o f the genome clarified the problem (which in fact occurs no
matter where the primer is located) that polymerization o f D N A by the virion polymerase
proceeds toward the 5' end o f the template and therefore must reach the end o f the molecule
before copying the entire genome. This finding led to a flurry o f model building (Junghaus
e t al. I975; Coffin, 1976; Collett & Faras, I976; Haseltine & Baltimore, I976; Stoll et al.
I977). To solve this problem these models postulated a terminally redundant sequence in
retroviral genomes. Therefore, by copying the 5' terminal region o f the virus genome, the
polymerase creates a new primer which is complementary to nucleotide sequences near the
3' end and can form a new template-primer pair allowing re-initiation and elongation o f the
nascent chain f r o m the 3' end. This prediction was quickly verified by nucleotide sequence
determination o f the terminal regions o f A S V and M u L V (Coffin & Haseltine, I 9 7 7 a ;
Haseltine e t al. I977; Schwartz e t al. 1977; Stoll et al. I977; W a n g et al. 1977).
The terminally redundant sequences o f a n u m b e r o f A S V strains are shown in Table 2.
A l t h o u g h some o f these sequences are tentative and based only on the composition o f
R N a s e A and Us digestion products o f the terminal oligonucleotides, it is clear that there is
similarity between different strains o f b o t h exogenous and endogenous viruses. There are
also, however, several points at which sequence variability occurs with no detectable
difference in function. It m a y be that this sequence serves no other function than to m a t c h
its opposite n u m b e r and that therefore almost any pair o f matching sequences will work.
Consistent with this idea, we have f o u n d variants o f Pr-RSV-B, which arise during passaging
o f the virus, to display greater variability in terminally redundant than in coding sequences
(S. Voynow, C. Barker & J. Coffin, unpublished data). In M o l o n e y murine leukaemia virus
( M o - M u L V ) the redundant sequence is m u c h longer - approx. 60 nucleotides as c o m p a r e d
to a b o u t 2i for A S V - and has no obvious sequence or structural relationship to any o f the
A S V sequences (Coffin et al. I978b).
As Table 2 shows, each strain o f virus has exactly the same sequence at the 5' and 3'
termini o f its genome a l t h o u g h there m a y be some heterogeneity in the length o f the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
Review: Retrovirus genomes
CAP
Rs I
Us
I PB
~/gag
(A)
pol
~ 1 9 I 1,~
-,~,~7 I 15 I
env
I
110,70
85
(B)
(C)
(l)p
[]
[]
[]
(De)
[]
prl80
pr76
-4
I
t
pr80
[ 35 I
60
[]
p60
]
D~
[]"
I
ISl,~
4- t
t
pr76
t R,..y.ol.y.!.A?
Us
[]'~
or
prl80
[]'~
4t
I
I = 1000 nucleotides
Fig. I. Structure of the ASV genome and mRNAs. (A) The genome of non-defective ASV drawn
to approximate scale (bar represents Iooo nucleotides). The boxes indicate terminal sequences
which are shown above the genome enlarged 5o-fold. Coding regions are labelled above the genome
and the approximate weights ( x io -3) of the protein products are shown below. (B) to (D) Probable
m R N A species. The mol. wt. ( x m -a) of the primary product is written above each molecule. The
translation terminator is shown by t and the box at the 5' end indicates the region from the 5' end
of the genome added by splicing. (B) src m R N A ; (C) env m R N A ; (D1) and (D2) two possibilities
for gag-pol mRNAs.
redundant sequence in a virus population (Stoll et al. I977). This identity (or 'concordance')
of terminal sequences is also found in recombinants between strains with differing termini
(Wang et al. I977; Joho et al. I978) and in spontaneous mutants arising in cloned virus
populations (S. Voynow, C. Barker & J. Coffin, unpublished data). As we shall see, these
results are almost certainly a direct consequence of the mechanism of replication and do not
provide arty information on either the mechanism of recombination or any selective disadvantage of molecules with slightly mismatching terminal sequences.
Fig. i shows, on a greatly expanded scale, the terminal regions of the genome of probable
importance for replication. These include, from left to right, the 5' copy of the redundant
sequence (Rs, I6 to 2I nucleotides); a sequence unique to the 5' end (Us, 8o nucleotides);
the primer binding site (PB, I6 nucleotides) complementary to the 3' terminal 16 nucleotides
of tRNA trp (Cordell et al. 1976; Eiden et al. I976); the unique sequence near the 3' end
(Us, approx. 2o0 nucleotides); and the 3' copy of the redundant sequence (Rs, I6 to 2I
nucleotides). No other sequences are known to be required for replication and it would
be interesting to see if any RNA flanked by these nucleotide sequences could be replicated
as a 'virus' when complemented with virus gene products.
As mentioned above, the U5 sequence, at least in ASV, may contain the dimer linkage
structure (Haseltine et al. I977). This sequence is almost exactly the same length in ASV and
MuLV (Haseltine & Coffin, I978) and may also serve as a 'spacer' to alleviate steric
problems that could arise when the nascent DNA chain is transferred from one end to the
other during DNA synthesis. Since the U5 sequence is spliced intact on to m R N A molecules
(Mellon & Duesberg, I977; Weiss et al. I977; M. Champion & J. Coffin, unpublished data),
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
6
J.M. COFFIN
it may have a role in ribosome binding, although adjacent sequences do not agree with the
amino acid sequence of the N-terminal peptide of pI9 [which is usually blocked (Herman
et aL I975) and therefore must be the first peptide synthesized]. Protein synthesis
must therefore be initiated to the right of this sequence (R. Eisenmann, personal communication).
The U3 sequence has recently attracted attention because of its presence near both ends
of the integrated provirus (Hughes et al. I978; J. Taylor, personal communication). As
discussed in a later section, it is likely that this sequence contains some signals for initiation
a n d / o r termination of RNA synthesis. At least portions of the Us sequence are identical in
all strains of exogenous ASV examined to date. In particular, all such viruses contain a
specific oligonucleotide (oligonucleotide C; Wang et aL I975, I976) which is immediately
adjacent to the R 3 sequence (D. Schwartz & W. Gilbert, personal communication) and
contains the A A U A A A sequence found near the poly(A) of many eukaryotic mRNAs
(Proudfoot & Brownlee, I976). [This sequence appears within the R~ sequence of Mo-MuLV
R N A (Coffin et aL I978b).] It is interesting that this region of the genome of RAV-o and
related endogenous avian oncoviruses does not contain oligonucleotide C and these viruses
have a Us region which contains no detectable sequence relationship to that of exogenous
viruses (Coffin et aL I978a ). This different sequence apparently confers a selective disadvantage to RAV-o relative to exogenous viruses since all recombinants between the two
(about 30 examined to date) invariably have this part of the genome from the exogenous
virus (P. Tsichlis & J. Coffin, unpublished data). It is tempting to speculate that the
difference in the Us sequence is related to the lower rate of virus production observed in
cells infected with or spontaneously producing RAV-o as compared with exogenous virus
(Linial & Neiman, i976; Robinson, 1976).
Virus m R N A s
Several species of virus RNA, tentatively identified as mRNAs, can be found in retrovirus
infected cells. These are shown in Fig. I(B-D). The mRNAs for env and probably src are
portions of the total genome and appear to consist of the translated portion and all information to the 3' end of the genome (Hayward, I977; Weiss et al. I977). Both of these
species have a segment which is derived from the 5' end of the genome which includes U5
and probably the PB segment (Weiss et al. I977; Mellon & Duesberg, I977).
The mRNAs for the pr76 gag precursor and the prI 8o p o l precursor are inseparable from
each other and from genome R N A in sucrose gradients (Opperman et al. I977 ) and are
therefore similar or identical in size to the total genome. In infected cells and in vitro the
two protein precursors are made in a ratio of approx. Io: I and the prI8o seems to contain
pr76 as its N-terminal portion. These results suggest that there is a weak terminator at the
junction between the gag and p o l regions, which, in normal ceils is suppressed some of the
time to allow read-through of pol. Such a mechanism is known to operate in the R N A
bacteriophage Qfl (Weiner & Weber, I973) and serves to regulate the relative amounts of
coat protein molecules. The same mechanism may also be used by Mo-MuLV, since
addition of amber suppressor t R N A to an in vitro translation system increases the proportion of prx8o to pr76 (Philipson et al. I978). This result suggests the presence of an amber
terminator at the MuLV gag-pol boundary. In the case of ASV, suppressor tRNAs do not
increase translation o f p r I 8o although addition of amber suppressor t R N A leads to synthesis
of a molecule slightly lower than pr76, suggesting that pr76 terminates naturally at a U A G
codon (Bishop et al. 1978 ). This result must mean that there are other 'anti-termination
factors' in eukaryotic cells, or that the mRNAs for gag and p o l are, in fact, different species.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
Review: Retrovirus genomes
7
The latter possibility is much more likely and two alternatives for their structure are shown
in Fig. I (D 1 and D~). In each case the two molecules differ by a splice near the gag-pol
boundary. In the first case (D1) the m R N A for prI8o more closely resembles the genome
R N A and is devoid of a terminator between gag and pol. A splice at the boundary would
create a termination codon to make the m R N A for pr76. In the second case, splicing removes
a terminator in the genome at the end of gag or shifts the reading frame to remove terminator(s). Note that MuLV and ASV may be the same in this respect, differing only in that in
ASV, more than one termination codon is removed and in MuLV splicing to create the
p o l m R N A does not induce a frame shift and removes only an amber terminator. It should
be noted that neither gag or pol m R N A may be identical to the genome RNA. One reason
that the mRNAs might be distinct from genome RNA would be to surmount the problem
that seems to occur soon after infection if the genome RNA is incorporated into polysomes.
Since RNA is translated 5' to 3' and copied 3' to 5', it is difficult to see how the genome
could be copied into DNA with the ribosomes and polymerase moving in opposite
directions. This apparent problem could be eliminated if an additional splicing event
created the gag and p o l i n i t i a t i o n (or ribosome binding) sequence, or, alternatively, removed
a blocking sequence. Other solutions involving virion proteins are also possible.
The considerations discussed above raise the possibility that the high tool. wt. R N A
isolated from virions may be a rather complex mixture of different species, differing only
slightly in size, including the authentic genome and spliced mRNAs. It is not known whether
the virus has some mechanism for excluding m R N A species from virions. The presence in
the 7oS complex of occasional molecules of env (Stacey & Hanafusa, I978) and probably
src (Purchio et al. I978) m R N A has been established. Since these molecules do not appear
as peaks in gels or gradients of virion RNA and do not contribute detectable amounts of
capped oligonucleotide to fingerprints (Coffin & Billeter, I976) they must be in substantially
lower concentration relative to high mol. wt. RNA than they are in infected cells (Hayward,
I977; Weiss et al. I977). The mRNAs in the virion are detectable by in vitro translation
experiments and at least the env m R N A can apparently be stably expressed following
infection of cells with virions containing this species (D. Stacey, personal communication).
This result is probably a reflection of the fact that splicing of the 5' terminal sequence of
the genome on to smaller RNAs could create a ' mini genome' containing all the structures
necessary to be replicated by virus systems. Heterodimers containing such minigenomes
may be important intermediates in formation of recombinants between virus genomes and
partial virus information expressed in certain cell types (Coffin et al. I978 a).
Most or all synthesis and processing of virus RNA takes place in the nucleus of infected
cells (Parsons et al. I974; Rymo et al. I974), presumably using the integrated DNA as
template. The major enzymatic activities are almost certainly from the host cell- for
example, the sensitivity of synthesis to a-amanitin implicates RNA polymerase II (Jacquet
et al. I974; Rymo et al. I974). It has been suggested that ASV RNA is synthesized as a
precursor slightly larger than genome size (Bishop et al. I976 ), but this question needs to
be re-examined in light of the recent progress in understanding cell m R N A processing.
The presence of both spliced (mRNA) and apparently unspliced (genome) virus RNA
species as finished products raises the problem of how the splicing is regulated. It seems
likely that the genome is the precursor to spliced mRNAs, since micro-injection of 35 S
ASV virion RNA into chicken cell nuclei yields active env m R N A (Stacey & Hanafusa,
I978). In other systems where splicing has been observed, such as globin (Tilghman et al.
I978), the precursor seems to be quite short lived. If this is a general phenomenon, the
problem arises of preserving a portion of the tumour virus genome molecules against
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
8
J. M. C O F F I N
Rs Us X PB G
P
E
S
XU3 R3
D,
(A)
-.°
(B)
rs us pb
(C)
XI-13
rs us X pb
U3 Rs Us X
9
(D)
XU3 Rs Us PB
-I
(E)
(F)
G
'.3
e
s
X tl 3 r 5 U 5 X
P
E
S
U3 Rs Us X
E
S
U~ Rs Us
+
U3 RsUsPB
~6) • - -
p
G
P
t'
t'
i
t
....
Fig. 2. Structures of proviral DNA. (A) The genome of ASV (not to scale) G, P, E, S are the coding
regions shown to indicate polarity. X is a possible additional repeat postulated to aid in circle
formation (S. Hughes, personal communication); the arrow-head indicates the 3' end of all
molecules. Other symbols are as in Fig. I. (B) and (C) In vitro ( - ) strand forms. The dotted line is
the t R N A ~rp primer. Sequences complementary to the genome are shown in lower case letters.
(B) Strong stop D N A ; (C) heterogeneous elongation products of strong stop DNA. (D) and
(E) In vivo cytoplasmic forms: (D) a partial ( - ) strand with a small fragment of ( + ) strand;
(E) complete ( - ) strand with complete ( + ) strand in pieces. (F) and (G) Nuclear forms: (F) two
covalently closed circles differing only in the number of times the U3-Rs-Un sequence appears;
((3) integrated provirus. The dashed lines indicate cell DNA. The arrows show initiation (i) and
termination (t) sites of transcription if the genome is not transcribed as a larger precursor.
splicing. It has been suggested that one of the gag proteins (pI9) regulates the extent of
splicing by binding to possible splice points in the virus RNA (Leis et al. 1978). This model
would require that a portion of intracellular pT9 be found in the nucleus of infected cells;
however, such a finding has not been reported.
Proviral D N A synthesis
Analysis of intermediate structures in proviral DNA synthesis is proceeding rapidly in
many laboratories, but a complete picture of the mechanism of virus DNA synthesis is not
yet available. One major problem is that retroviruses use a system different from all other
known groups of viruses to ensure that all portions of the genome are copied into double
stranded DNA. This system combines direct terminal repeats in the genome and the ability
of the polymerase to move, along with the nascent chain, from one template to another. As
we shall see, all the consequences of this surprising mechanism are probably not yet
appreciated.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
Review: Retrovirus genomes
9
Fig. 2 shows the genome of ASV (A; with the terminal regions expanded for clarity) and
the forms of DNA observed in vitro and in vivo which are currently considered to be intermediates. All the forms are shown aligned with the portion of the genome from which they
are believed to be copied and molecules of the same sense as the virus gene (+ strand) are
labelled with capital letters; complementary molecules (-strand) with small letters. The
3' end of each strand (and therefore the direction of its synthesis) is shown by an arrowhead.
The first two DNA forms shown (B and C) have been detected only in in vitro endogenous
or reconstructed polymerase reactions. When the total product of such a reaction is analysed
by polyacrylamide gel electrophoresis, various homogeneous DNA species can be isolated,
each of which is an elongation product of smaller species (Haseltine et al. I976). The most
prominent product (B), variously called' strong stop D N A ' , ' short stop DNA', 'iDNA', and
'cDNAs,', is a copy of the U5 and R 5 region of the genome terminating opposite the Gm
(Shine et al. I977; Haseltine et al. 1977; Coffin & Haseltine, I977a). In the case ofPr-C ASV
it is Ioi nucleotides long, in Mo-MuLV its length is I35 to 145 nucleotides. Its prominence
among the reaction products is presumably because initiation and elongation of DNA
chains are rapid events relative to the transfer from end to end. In the case of MuLV it has
been shown that DNA molecules longer than strong stop DNA (C) contain a single copy
of the redundant sequence, flanked by the U~ sequence at their 5' end and sequences copied
from the 3' end of the genome at their 3' end (Haseltine & Coffin, 1978). The presence of
only one copy of the repeated sequence shows that this sequence must be used as a
bridge to create a template-primer pair near the 3' end of the genome, as predicted by the
models. In the case of ASV, the same type of experiment has not yielded the same result.
Instead, DNA species only slightly longer than strong stop DNA have been proven to
contain sequences derived from many parts of the genome, particularly a region near the
middle (Cashion et aL I976; J. Coffin & W. Haseltine, unpublished data). Although this
result was first attributed to additional initiation sites in the virus genome, it is more likely
to be a consequence of'improper' elongation, since we have found that DNA in each single
band from polyacrylamide gels hybridized to both 5' terminal and internal sequences
(J. Coffin & W. Haseltine, unpublished data). It is possible that the difference between
MuLV and ASV in in vitro DNA species is a consequence of difference in length of the
terminally redundant sequences. The stability of the hybrid of strong stop DNA and the
6o-nucleotide MuLV R 3 sequence is much higher than that of the corresponding 2Inucleotide ASV hybrid (Coffin & Haseltine, I977a; Coffin et aL 1978b). It is probable,
under in vitro reaction conditions, that the specificity of hybrid formation between strong
stop DNA and the ASV R 8 sequence may be low relative to the affinity of the polymerase
for any sequence with perhaps a few nucleotides of homology to the 3' end of the strong
stop DNA. This difference probably also has some historical interest, since it has traditionally been much more difficult to prepare cDNA probes complementary to the whole
genome of MuLV as compared to ASV using only the endogenous virion polymerase and
primer. The ASV polymerase does not seem to have such a problem in vivo, since the
majority of the DNA found (at sufficiently late times after infection) is in full-genome
length forms (Varmus et al. I976) although some smaller species can be detected which
might be analogous to the 'improper' elongation products seen in vitro (Guntaka et al.
I976).
In the case of Mo-MuLV, an infectious product can be synthesized by virions in vitro
(Rothenberg et al. I977), but this species is only a small minority of the total reaction
product. This species is not made if actinomycin D is present in the reaction, suggesting
that a portion of the infectious molecule might be synthesized using a DNA template. The
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
IO
J.M. COFFIN
longest DNA molecule synthesized in such reactions contains a complete copy of the
genome including the terminal regions (Rothenberg et al. 1978), and it has been suggested
that it is somewhat longer than full length (Shank et al. 1978) and corresponds to the
minus strand of the form shown in Fig. 2(E).
The remaining forms of DNA shown in Fig. 2 have been detected and characterized in
cells shortly after infection, usually by hybridization of labelled virus RNA or cDNA with
cell DNA fractions often using the 'blotting' procedure of Southern (I975). Unfortunately,
only relatively long-lived forms can be observed in this way and many key intermediates
have not been found. Two of the DNAs are found in the cytoplasm o to 24 h after infection
(Varmus et al. 1978). The first of these (Fig. 2D) is a partial ( - ) strand of heterogeneous
length whose 5' end is the strong stop sequence. A short (250 to 30o nucleotides) piece of
( + ) strand DNA is found before full length ( - ) strands appear (Varmus et al. 1978). This
piece apparently is synthesized from a unique starting point and contains both U3 and U5
sequences. It must therefore be synthesized using the 5' terminal region of the ( - ) strand
DNA as a template before completion of ( - ) strand synthesis. At a somewhat later time
(3 to 4 h in ASV) double stranded DNA with a sedimentation coefficient similar to that
expected for a copy of the virus genome appears (E). This species consists of an intact ( - )
strand hydrogen bonded to short fragments of ( + ) strand DNA (Varmus et al. 1976). Very
recent evidence (Shank et al. I978; J. Taylor, personal communication) suggests that this
DNA has a surprising structure. As expected, its right end is a copy of the Us-Rs-U3
combination as found in structure (C). Its left end contains an additional copy of the Uz
sequence adjacent to Rs-Us. This juxtaposition of sequences can be created only by a
second 5' to 3' transfer of polymerase plus growing chain.
The presence of the remaining forms (Fig. 2 F) exclusively in the nucleus of infected cells
suggests that some cell enzymes which are exclusively nuclear must be involved in their
formation. Such cell enzymes must include a DNA ligase, since the ( + ) strand is now in a
continuous chain (Varmus et al. t976). The nuclear forms are covalently closed circles of
two different sizes (not including defective forms). The larger circle is similar to a closed form
of the linear molecule and has the sequence -C-U3-Rs-Us-Ua-Rs-Us-PB-. The smaller circle
contains only a single copy of the U3-Rs-U5 repeat. One or both of these forms is presumed
to be the form which is subsequently integrated into the host DNA, although there is as yet
no direct evidence on this point.
Possible mechanisms o f proviral D N A synthesis
The mechanisms which relate one proviral DNA structure with another are not yet fully
understood. The principal problems are: the mechanism of the transfers of the growing
chain from end to end; the nature of the primers for ( + ) strand synthesis; and the mechanism
of circle formation. From the in vitro results discussed above, the first end-to-end transfer
of the nascent DNA chain must utilize base pairing between sequences at the 3' end of the
DNA and the 3' end of the genome, i.e. rs-R~. It seems most probable that the DNA becomes
available for base pairing as a result of hydrolysis of the copied RNA by the polymeraseassociated RNase H (Moiling et al. I97I) and recent in vitro evidence (Collett et aL I978)
supports this view. For this purpose, the RNase H has two useful properties: it degrades
only as an exonuclease (Leis et al. I973) and therefore will not attack the RNA prior to
completion of the DNA strand; and it hydrolyses only the RNA strand of DNA-RNA
hybrids (Mrlling et al. I971) and therefore should not separate the tRNA trp primer from
the PB region of the genome. This specificity would ensure that end-to-end transfer of the
growing DNA chain is a unimolecular reaction. Loss of the Us-R5 sequence by RNase H
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
Review: Retrovirus genomes
II
(a)
D1
r5 us pb
g
p
e
s
u3 r5 u5
D2
~
-q
/,I
rs us
D3
~
LI3 F5
ll 5
//
U3 1"5 Ll5 1/U3
F5 ll5
1) 4
It 3 rs
pb
tl 5
g
p
e
s
113
/5
ll5
(b)
D~
pb
D2
Rs Us PB
u3
rs us pb
g
p
e
s
g
p
e
s
u 3 r s u~
g
p
e
s
u3 r5 us
U3 Rs Us P B ~
pb
D3
U3
~
U3 Rs Us P B ¢ I I~
pb
Fig. 3. Possible intermediates in synthesis of complete linear proviruses. (a) According to a
suggestion by J. Taylor e t al. (personal communication). The strong stop sequence is copied from
one genome (italic letters) and transferred to a second to synthesize a complete ( - ) strand (D1)
which is then transferred to the 3' end of (probably) the first genome (D~) to add the Ua sequence
(Da), yielding the complete structure shown in D4. A n experimental prediction of this model is the
distribution of sequences from each parent in the progeny. (b) According to a suggestion by
Shank e t a L (1978), before completion of the ( - ) strand, a s h o r t ( + ) strand is synthesized near its
5' e n d (D1). The sticky end (13) created by copying the primer allows a template-primer pair to be
formed (D2) and elongation of the ( - ) strand on the short ( + ) strand template leads to the
structure shown (D3).
hydrolysis would appear to create a major problem for synthesis of the structure shown in Fig.
2(E), since the information which must be copied again has been lost. Two mechanisms have
been proposed to eliminate this problem (Fig. 3)- In the first (J. Taylor, personal communication), shown in Fig. 3 (a), D N A synthesis is initiated on only one of the two genomes in
the complex and is subsequently transferred to the 3' end of the other which retains its 5'
terminal region. D 1 shows the structure resulting from elongation to the end of the genome.
In this molecule the left-hand r:Us sequence and all the coding regions are derived from
one parent genome and the right hand r:u5 sequence from the other (shown by italics).
The U3 sequence would then be added to the D N A by using the identical mechanism of
transfer as before (D~) and terminating specifically at the end of the Ua sequence (D3). This
process yields the structure shown in D 4. The second model, Fig. 3 (b), proposed by Shank
et al. 0978), need use only one genome as template. Initiation of synthesis, transfer of the
growing chain and elongation to the (now truncated) 5' end lead to the ( - ) strand shown
in D 1. If at the same time the short ( + ) strand initiated near the 5' end is elongated to copy
the tRNAtrv primer, a new redundancy (the PB sequence) is created. Removal of the primer
by RNase H creates another sticky end and allows the ( + ) strand fragment to be used as
template (D~) to arrive at the structure shown in D 3.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
I2
J.M.
COFFIN
At the moment there is too little information to decide between these models. The first
(Fig. 3a) has two unattractive features: the sacrificing of a complete genome for the sake
of IOO nucleotides of sequence and the problem of why the polymerase should terminate at
a specific point the second time it is copied, but not the first. In the second model (Fig. 3b),
the specific initiation site for ( + ) strand synthesis is somewhat difficult to explain, but such a
piece of D N A has been found. The sensitivity to actinomycin D of in vitro synthesis of infectious MuLV DNA, alluded to above, lends some support to model (b) since DNA-directed
synthesis would be required for a complete ( - ) strand. A possible experimental test is
suggested by the drawings in Fig. 3- In the case of model (a), the R3 sequence of progeny
virus will always be derived from one parent and the R~ from another. Thus, examination
of progeny of cells infected with heterozygotes between viruses with distinguishable
redundant sequences should allow us to distinguish between the two models.
The most reasonable guess for the mechanism of initiation of ( + ) strand synthesis is that
the genome RNA, after being copied into an R N A - R N A hybrid is nicked by a cellular
(endo) RNase H (or is broken before D N A synthesis, see later;Keller & Crouch, ~972). The
virion enzyme could then use the 3' ends of the fragments to initiate synthesis and simultaneously remove the remaining genome RNA and elongate the new ( + ) chains in a sort
of 'nick translation' synthesis (thus providing the only example I know of in nature of
wholesale destructive replication o f genetic information). A less likely possibility for initiation of ( + ) strand synthesis would be association of the ( - ) D N A with small cell RNAs
after removal of the genome. These could act as primers in a manner analogous to the
tRNA trP. A particular problem is posed by the small piece of ( + ) strand D N A found at the
left hand end of the replicating molecule (Varmus et al. 1978) (Fig. 2 D), since it would seem
that a very specific initiation site would have to be used. Either a site-specific cleavage,
or attachment of some very specific primer must be required for this step.
The last detectable intermediate prior to integration is believed to be one of the two closed
circular forms (Guntaka et al. 1975; Fig. 2F). The smaller circle could be created by recombination of the identical sequences at the ends of form E as:
-
-
-
R5 U5
U 8
x
U3RsU 5 - - - -+ - - - U~RsU 5 - - -
+ U~RsU ~ (lost)
For the larger circle there are at least three possible mechanisms: (I) direct Iigation of the
ends of the linear molecule; (2) illegitimate recombination of non identical sequences near
each end of the molecule; or (3) recombination through a short sequence repeated at each
end (S. Hughes, personal communication). The latter possibility is illustrated by sequence
' X ' in Fig. 2. If this short sequence is present in the genome near the 5' end of the Us
sequence and repeated at the 3' end of the U5 sequence, then the linear form (Fig. 2E)
would have it near each end. Recombination to close the circle would then proceed:
- - - XU~ R5 UsX
×
XU3RsUsX---
~ -- - XU3RsUsXU3R~UsX---
+
X (lost)
Note that the X sequence is still present more than enough times for synthesis of progeny
containing this sequence twice. The complete sequence of the 3' terminal region of ASV
should soon be available for comparison with the U5 sequence to validate this hypothesis. In
principle it could be tested by hybridization; however, the presence of sequences in the
genome other than near the 5' terminus hybridizable to strong stop D N A is in dispute.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
Review: Retrovirus genomes
13
Using a method which would have detected homologous regions of 30 nucleotides or more,
we could find no such sequence (Coffin & Haseltine, I977b). Other hybridization protocols,
which should not have seen the R3-r~ hybridization, have sometimes detected sequences
complementary to strong stop DNA near the 3' end of the genome (Friedrich et al. I977;
W. S. Hayward, personal communication) and sometimes not (Weiss et al. I977; Krzyzek
et aL I978).
From the previous discussion, it should be clear that the pathway from parental virus
single stranded RNA to closed circular DNA is still a fruitful area for speculation. While
I have made no effort to hide my bias in this matter, it is unlikely that any of the mechanisms
proposed will be completely correct. Nevertheless, such model building has in the past
proved useful for designing experiments and as a pastime for tumour virologists. The
reader is invited to join in by following a few simple rules:
(I) Use virion systems (polymerase, RNase H, tRNA trp) as much as possible, bearing
in mind the in vitro biochemistry of the enzymes.
(2) Use cell systems (additional enzymes and primers) as little as possible.
(3) Rely on base pairing rather than protein molecules to join and hold the nucleic-acids
together.
(4) Favour intramolecular events over intermolecular events.
(5) Do not be afraid to invoke directly repeated sequences in the genome when it seems
necessary or useful.
(6) Be careful to ensure that all parts of the genome are passed on to the progeny.
Integration
Recently much has been learned about the structure of the integrated DNA (Fig. 2 G)
of ASV.
(I) It is, to within a few base pairs, identical to the double stranded linear form (Fig. 2E)
and therefore contains a complete copy of the genome plus repeats of sequences
derived from the ends (Hughes et al. I978 ; J. Taylor, personal communication).
(2) There is only one or a small number of copies per infected cell (Varmus et al. I973).
(3) Different clones of ASV-infected mammalian and quail cells have the provirus
integrated at different sites in the host genome - it has been calculated that there are
at least 40 such sites (Hughes et aL I978).
(4) In ASV-infected chicken cells, the integration site is apparently not unique even
within a clone of cells (J. Taylor, personal communication; H. Varmus, personal
communication).
The simplest mechanism to imagine for integration of the provirus is recombination
between the large circular form and cell DNA. In such a case, the integration system must
recognize a specific sequence in the DNA at the point of closure of the circle (the Uz-U 5
junction). Since both the U5 and U3 sequences appear twice in the circle, such specificity
would require that features of both sequences be recognized. Additionally, the system may
have no specificity for any sequence in the cell DNA of more than a few base pairs in length.
It is therefore most unlikely that integration is a result of a general recombination system
which recognizes and joins any homologous sequences. This latter conclusion is supported
by the failure to find recombinants among the virus progeny from super-infected cells
pre-infected with a different virus strain or harbouring a non-expressed endogenous virus
genome (Wyke et al. I974; Weiss et al. 1973).
The system for integration most likely pre-exists in the host cell, although some elements
from the virion may be retained in a complex with the DNA. It should be possible to infer
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
I4
J.M. COFFIN
the role of virion proteins in integration by transfection experiments with pure proviral
DNA. Although chicken cells transfected with any form of ASV proviral D N A synthesize
infectious virus (Fritsch & Temin, I977), it has been impossible to prove in this system that
the transfecting DNA is integrated into the cell genome (G. Cooper, personal communication). This result suggests that a virion protein may be required for some phase of the
integration process. However, transfection of ASV or murine sarcoma virus DNA into
mammalian cells seems to lead to integration of the transfecting DNA (Smotkin et aL
I976; G. Cooper, personal communication), although the low efficiency of the transfection
process and the fact that partial MSV DNA molecules lacking one end are infectious
(R. Weinberg, personal communication) leaves open the possibility that the usual mechanism
of integration is not involved in transfection.
The failure to find a specific integration site for the ASV genome in clones of infected
chicken cells raises the interesting possibility that the integrated genome can be excised and
re-integrated at reasonably high frequency. Recombination in the directly repeated sequence
(U3-Rs-Us) at each end of the integrated genome would lead to excision of the provirus
from the cell DNA. The excised provirus would be the same as the smaller circle in Fig. 2F.
It is difficult to see how this form could be re-integrated without loss of one of the repeated
sequences.
The absence of specific integration sites and the presence of additional information at
each end (U3 at the left and U5 at the right) of the integrated provirus strongly suggests that
control sequences for initiation (and possibly termination) of virus RNA synthesis are
encoded in the virus genome. If the progeny genome is not cleaved from a longer precursor,
these sites are as shown in Fig. 2 G (i and t).
As suggested by J. Taylor (personal communication), the promoter for virus RNA
synthesis is probably in the U3 portion of the provirus, and the U5 region may contain a
termination sequence. These considerations make the U3 sequence of particular interest,
since ASV RNA is synthesized at a very high rate - about I ~ of total cell RNA or some
Io ~ of total m R N A synthesis (Coffin et al. I974; Jacquet et al. I974; Rymo et al. I974)
and reaches a steady state level of some Ioooo copies per cell (Coffin & Temin, ~972), as
much as the most abundant cell m R N A species in specialized cells. As noted above, the
major difference in sequences between endogenous and exogenous ASV strains is in the
3' proximal sequence. This system should therefore provide a very useful model for a
genetic test of the ideas outlined above and for study of structure and function of sequences
regulating RNA synthesis in eukaryotic cells.
If the hypothesis of initiation and termination controlled by proviral sequences is correct,
then an interesting problem arises, since the RNA polymerase must cross the Rs-U5 termination sequence twice. The first time the polymerase encounters this, sequence transcription
must not be efficiently terminated or there would be no synthesis of the bulk of the genome
RNA. An analogous, although not so serious, problem seems to occur at the 3' end in that
the Us sequence might initiate RNA synthesis to the right of the provirus. Similarly, if the
final genome size is the result of cleavage at the indicated points, the system responsible
would have to recognize the Uz-Rs-U5 sequence twice and cleave at U3-R5 on the left side
of the genome and at Rs-U 5 on the right side.
A final point in the overall replication scheme concerns the fate of the redundant
sequences in the course of replication. From the scheme in Fig. 2 and from in vitro evidence
(Haseltine & Coffin, ~978), it should be clear that the Rz sequence possesses no genetic
continuity, since it is not copied during replication. Thus, although a provirus with slightly
different terminal sequences could be formed as a result of recombination, point mutation
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
Review: Retrovirus genomes
15
or transfer of the enzyme and D N A from one template genome to another, the termini of
any homozygous progeny formed would again be identical after one further round of
replication. This phenomenon almost certainly accounts for the 'concordance' of terminally
redundant sequence in recombinants (Wang et al. I977; Joho et al. I978) and point mutants
(S. Voynow, J. Coffin & C. Barker, unpublished data) which have been re-cloned to
homogeneity.
Recombination
The excellent review on recombination by Hunter (I978) is still current, so I will refer the
reader there for a complete discussion of all the models proposed to date and experimental
evidence for and against each. In place of such a complete review I will cover a few major
points and propose yet another model, one that seems to mesh well with the mechanisms
discussed in the previous section.
Although recombination has been demonstrated between MuLV strains (Elder et al.
I977; Failer & Hopkins, I978), the phenomenon has been studied extensively only in ASV,
due to the large selection of genetic markers available with the avian viruses. A typical
experiment (Blair, I977) is to infect a culture of cells with two genetically different viruses
and harvest and examine the progeny with some selective system after a few replication
cycles. Although determination of precise recombination frequencies is impossible in such
a system, as much as 4o ~ of the progeny of the original cross can be found as recombinants, even when closely spaced markers [for example, two ts mutants in src (Balduzzi et al.
I978)] are used.
Characteristics o f recombinant virus g e n o m e s
Fig. 4 shows genomes derived from two typical recombination experiments. In both cases,
recombinant progeny were selected to received src from one parent and env from another.
The figure shows simplified oligonucleotide maps in which regions of the genome derived
from the src + parent are shaded; regions from the env donor are open; and regions of
undetermined origin are shown as lines. The hatched and open boxes above the maps show
the regions selected from each parent. Examination of the figure reveals several points
related to the mechanism of recombination.
( 0 As previously demonstrated (Beemon et al. I974; Joho et aL I975), the genomes are
the products of intermolecular recombination events, since all contain nucleotide sequences
derived from each parent and also lack sequences derived from each parent.
(z) Unselected crossovers are very frequent. Of the 2I molecules shown in Fig. 4, only
two have no crossover points except between env and src and there is an average of more
than two additional detected crossovers per molecule. As a result, there is no detectable
physical linkage. This is most noticeable in the genome region immediately to the left of the
selected env marker. Within about Io ~ of the genome (or Iooo nucleotides), the recombinants contain sequences derived about equally from each parent. Thus the probability of
recombination between neighbouring nucleotides in such an experiment can be very
roughly estimated to be of the order of Io -3 (perhaps plus or minus an order of magnitude).
Even if a high negative interference is considered, this is an extraordinary figure. In T-even
bacteriophage, considered to have an extremely high frequency of recombination, the
corresponding value is about z × io -5 (Luria et al. I978). The result of this high frequency
can be seen to randomize the parentage of all unselected regions in the 5' half of the genome.
(3) There are no obvious 'hot spots' for recombination in which crossovers occur more
2
VIR
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
42
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
0q u~3 sd~m o~g~Iu!I ao 3!louo~, p a ~ p u ~ s , l~ql olq~qoaduI ! £Iq~.tq s! ~! 'suos~oa osoql ao d
•(q 'v8L6I 7v ~o q~i!~qs) sosna!A lugu!quaoooa snoaomnu u! su!o~oad g v g ,i~uaaouq~, j o
oou~a~odd~ ~uonboaj oq~ j o osn~o oq~ s! uouotuouoqd a~i!m!s ~ ~-eq~ ~IO~I!I~JOA S! ~I "(uop~3
-mntutuoo i~uosaod 's!Iqo!sL "d) oas j o uotssoadxo oql qllA~ oIqp~dmoou ! s! 13npoad ouo~
0SOqA~ o-AV~I JO uo!~oa g v g oql ut aouonbos ~ j o ~lnsoa oq~ oq o~ 1.~ OAO!loq 0A~ '~A0~AOJs.Iql Jo
odo3s oql puo£oq s! u o u o m o u o q d s!ql j o uo!l~u~ldxo pol!~lop ~ ~l!qAA "A~Oaa~oql £q UA~OqS
'luoa~d +oas oql tuoaJ uo!~oa poloops I~UO!l!pp~ ouo lseo I 1~ oA~q ii ~ 'JOAOA~Oq '(q)l~ "~!A Ut
SlU~U!qtuo30a ~q~L "(vSL6I 'l v lo u ~ o D ~9L6I 'aolOIl!8 ~g u~JoD) sd~tu I~3!sgqd poqs!lqnd
~isno!Aoad j o s!s~q oql uo ,uoa~d q3~o moaj poloOlOS uooq 0A~q O~ poloodxo suo~Boa oq3
a~oqs (q) pu~ (o)k "$M j o d o l oql ~ sa~q oqJ~ "suoi~oa polooIOS poloodxoun oa~ oaoq& (10
•somt~ 8 pu~ V 'L '~ '~ '£ 0 @ ! a o l ~3o[ tuoa3)
posn oau s u o p ! s o d JOAOSSOJO0Iq~,0010 p X!S 0q~'fl U.i somouog oql Jo j[~q 1jo[ oq] u I "fii,uonboaj
•o~s qi!a polzOlOS-Oa£IiUaaedde gog u! uo#aa laoqs ~ staoqs A~O~at~oq.L "(elvp poqs!lqndun 'U~OD "f
s!Iqo!s.L "d) s!Iqo!sz "d £q poz!aolaeaeqo pue paleios! OaaAx(+aasaaua) SlU~u~.quJo00"¢I"(_aaSaaUa
-+lod+~o~) O-AV~I x (+as~4ua+lod+go.~) ~t-ad (q) "(elep poqs!iqndun 'aolunH 3 ~ uo!dmvqD "I~I
'trio D "f) £Joleaoqv I alto u! pas£1eue pue (9L61) lBoA ~g aalun H £q po)e[os! O.~0A~£uo~oad +o.~so~tu~t
"G~assaaUa+lods~gog) ~E~sl x (+a~sa,tua+lod+go.~) Wad (o) "mol,oq lE u~oqs s! suo!~ox ~u!poo jo
uo!l~3o[ o~m!xoadd~ oqL "pou!ua.to~oplou S~A~o~eluoaed oSOqA~SUO!~Oa~oqs S0U!"I"aoqlo oql moaj
suo!~o.t uodo quoaed ouo moaj pOA!aop oa~ suol~o.t popeqs "om~s oql oi~3s oq~ o~I~m ol ~l~q~!Is
po~.tols!p '(~8L6I "]~ ]a uyjo~) '.9L6I 'aolo[I!~t~, u!~IoD) sosna!A [~luo~d oq~ jo sd~m op!~oopnu
-o~![o uo pos~q o.m stu~a~!p oq£ 'su.mals ASV U0OA~i0q slu~u!quao3oa jo sd~m [~3!s~qd "1~"~!~I
==================================================================
~
:::::::::::::::::::::: [[
====================================================================
f~:~:~:~:~:~:~:~:::;I.~-6
~
t::~:~:~:~:::~:~:~1.~0[
IZ~
===================================================================
===================================================================
==================================================================
====================================================================================
===============================================================
I:: ::::::::::-::~
[
1~
[:~:~:~:i3:ii:i:i:;'----1
?
I
I-10--AV~t ×
I
[]
~1 a([
(q)
~
6
E~-~E:]--E:2]- ~
I
~i:i:]|
!
I,.~::~:]
i
~
9
:::::::::::::::::::::::::::
=====================
(::::!
i
I-'1
"b£E .vi X
~
fl
(v)
-l(I
"I7
Review: Retrovirus genomes
G'
F'
E'
D'
C'
B'
A'
G
F
E
D
C
B
A
G'
F'
E'
D'
C'
B'
A'
(1)
b
G
G'
F
E
D
F'
E'
D'
L..L
(3)
G
F
E
D
G'
F'
E'
D'
¢,
(4)
G
(5)
~ g
C'
B'
A'
C'
b~
a
C
d ~'c'
F
E
D
f'
e
d
A
C
A
bN
C
c'
b
a
Fig. 5. Recombination by breakage-repair. Two genomes in a dimer are shown. The dimer linkage
and other terminal structures are omitted for clarity. In the course of D N A synthesis, the polymerase encounters a break in the template (I); RNase H removes the copied portion (2); and the
nascent chain is transferred by base-pairing to the other genome where synthesis continues (3).
On encountering another break, the process is repeated, leading to a crossover back to the first
genome (4)- The final ( - ) strand is thus the recombinant molecule shown in (5). The primes denote
sequences from different genomes. As usual, ( + ) strand sequences are shown by capital letters,
( - ) strand sequences by lower case letters.
developed for R N A tumour viruses, except, possibly, within very short distances (for
example between ts mutants in src; Balduzzi et al. 1978). While very rough linkage maps
have seemed to yield the correct gene order (i.e. gag-pol-env-src; Hunter, r978) I consider
this a fortuitous result of selective effects either of the sort described above, or some more
subtle effect, for example a higher specific infectivity of virions of one subgroup than those
of another or of wild type than temperature sensitive virus, even at permissive temperature.
A possible mechanism of recombination
The simple model for recombination I wish to propose is outlined in Fig. 5. It rests on
the following points:
0 ) The virion of a tumour virus protects its genome very poorly from breakage. Years of
experience with ASV R N A have shown that, unless extraordinary measures are taken (such
as harvesting of virus at very short intervals; Cheung et al. ~972), it is almost impossible
to obtain virions with a significant proportion of unbroken R N A (unpublished observations of nearly everyone in the field). Nevertheless, preparations of virus with no detectable
intact genomes contain substantial amounts of infectious virus. This is an unusual situation
for an R N A virus, since with most other R N A viruses the genomes are relatively easily
2-:2
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
I8
J.M. COFFIN
isolated as intact, homogeneous, species. This difference could also reflect on the technical
skills of R N A tumour virologists, but I do not favour this view, for obvious reasons.
(2) An old observation with ASV is that it is much more resistant to inactivation by u.v.
light than other R N A viruses with comparable genome size (Rubin & Temin, I959; Bister
et al. I977). Furthermore, such inactivation seems to be the result of cross-linking of R N A
and virion protein rather than single strand breaks (Owada et al. I976).
(3) The frequency of recombination is such that it is probable that one or more crossovers occur at each infection cycle.
(4) Recombination between different strains of virus seems to require infection with a
heterozygote virus containing a genome of each parent in the same virion (Wyke et al. I974).
This point has been contested recently (Alevy & Vogt, i978), but I consider the evidence
in favour of the heterozygote model to be much stronger than that against it. Note that the
requirement for heterozygotes suggests only that recombination is an event preceding
integration. At this time in the infection cycle the probability of two genomes (or proviruses)
becoming juxtaposed into whatever structure is necessary to mediate recombination is much
higher for a heterozygote virion than for two virions entering the cell at different points.
(5) The virion polymerase is capable of being transferred via redundant sequences from
one template to another.
(6) Synthesis of ( - ) strand proviral D N A is very slow compared to other polymerizing
systems (Varmus et al. I978).
Observations (I) and (2) suggest that the virus has a mechanism for surviving breaks in
its genome RNA. The other observations suggest what the mechanism might be. A model is
shown in Fig. 5. (The terminal sequences are omitted from the model for clarity.) The key
assumption is that when a polymerase molecule encounters a break in the template it
behaves the same as when it reaches the end of an R N A template. Thus, in Fig. 5 the nick
between B and C causes the polymerase to halt 0 ) , and the template to be subsequently
degraded by RNase H (2). As at the 5' end of the molecule, this frees the nascent chain for
base pairing, but this time with the corresponding sequence on the other molecule in the
dimer (3). If, in the course of subsequent synthesis, another break is encountered, the
growing chain will be transferred back to the first template (4) and so forth, leading finally to
the recombinant shown (5).
The proposed mechanism is a variant both of copy-choice recombination, which has been
proposed for numerous systems but never demonstrated, and of nick repair, which occurs
in at least some bacteriophage systems (Broker & Doerman, I975). As in the latter case, the
recombination rate should be a function of the number of breaks in the RNA. This is
probably the strongest prediction capable of being easily tested. It is expected that this process
would be quite slow and if numerous breaks are encountered could lead to a net rate o f
proviral synthesis much less than that expected from the nucleotide addition rate.
If the breaks which lead to crossing over are random effects o f ' a g i n g ' of virions, then the
expected frequencies of single and multiple crossovers can be calculated in certain cases,
since both RNA molecules in a dimer must be of the same age. For example, in virions
with an average of one break per genome, or two breaks per dimer, then by Poisson distribution 37 % of the molecules will be unbroken and transcribed in one piece. These will not
give rise to recombinants. Of the remaining 63 %, 37 % (or 23 % of the total) will transfer
to an unbroken molecule and lead to a single crossover. Of the 40 % that jump to a broken
molecule, slightly more than half (approx. 6o %) will transfer to the right of a break in the
second genome and lead to a second crossover. This will lead to a final distribution of 37 %
parental, 39 % single recombination and 24 % multiple recombinant. Clearly, as the number
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
Review:
Retrovirus
genomes
I9
of breaks per genome increases, the proportion of single recombinants will decrease very
rapidly. The derivation of both genomes from pools with the same breakage history would
thus lead to a rather peculiar effect. Recombinants selected for a crossover between markers
relatively near one another will come more frequently from more broken and (probably)
older molecules in a virus population. Therefore, the closer the selected markers are, the
more unselected crossovers should be found in the progeny.
Under the mechanism proposed, a dimeric genome leads to only a single provirus. Thus
heterodimers would not be genetically detectable as heterozygotes. Genetic heterozygotes,
however, do persist for long periods of time in some mixed virus populations, but not in
others (Weiss et al. I973; Vogt, I977; P. C. Balduzzi, personal communication; and our
unpublished data). As mentioned in the first section of this review, it is possible that these
genetic heterozygotes are a consequence of a small proportion of virions with two or more
dimers which contain complementing defects not easily repaired by recombination. Particles
suggesting multigenome virions have been observed in MuLV (Yuen & Wong, I977). Two
types of non-recombining structures could be imagined to lead to persistent complementing
heterozygotes. The first would be a deletion in each genome with no common sequence
between the deletions. The second could be a type of recombinant suggested by the genomes
in Fig. 4- Note that the env region always consists of a large block of information from one
parent, with no detectable internal recombination. This result, combined with the observation of substantial sequence homology in this region between different ASV subgroups
(Junghans e t al. I975; T a l e t al. I977; Coffin et al. I978 a) suggests that such recombinants
should occur but not be viable. In a cross between, for example, ASV strains of envAsrc and env~src + genotypes, a substantial proportion (probably a majority) of the recombinant
progeny would be e n v ~ s r c + where env A~ codes for non-functional glycoprotein. Thus the
majority of viable progeny detectable by selection of transforming virus of subgroup A
could easily come from a minority of multigenome virions that are genetic heterozygotes
between the two parents or between an envABsrc + recombinant and the envAsrc - parent.
Note that a corollary of this line of reasoning is that a substantial proportion of cells
infected with virions containing heterodimeric genomes coding for two subgroups of env
gIycoproteins should yield onIy non-infectious progeny as observed by Vogt 097Ib),
particularly when the virus is highly at risk for recombination (i.e. the genomes have large
numbers of breaks). This effect could lead to an apparent higher recombination frequency
between closely spaced point markers (for example, ts mutants in the same gene) than
between mote distant markers and env.
A similar phenomenon would occur in a cell infected with a heterozygote of a full length
genome and a deletion mutant. If a break is encountered by the polymerase in the full-length
genome in a region not found in the deleted genome, then synthesis will most likely terminate and the infection will be aborted. This effect will not occur during copying of the
deleted genome. Therefore, the majority of the cells infected with such heterozygotes will
receive a provirus containing the same deletion, although other parts of the genome may be
recombinant. The size of the majority depends on the physical state of the infecting RNA.
This would lead to an apparent 'dominance' of deletion mutants under appropriate experimental conditions as in the previously rather puzzling observation of Kawai & Hanafusa
0976) who performed a cross of SR-RSV ts68 (envAsrc t~) and SR-RSV NY8 ( e n v - s r c + ) .
They found among the progeny only e n v - s r c t~ and no envAsrc + recombinants. The same
sort of dominance could lead to a rapid selection of deletions in mixed virus populations
especially those including env, a result we have observed (J. Coffin & C. Barker, in preparation).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
20
J.M. COFFIN
An argument that could be made against this proposal is that a structure with numerous
breaks would tend to fall apart during replication. This is unlikely for three reasons:
(I) The 7oS structure isolated from 'aged' virions retains the identical sedimentation
coefficient even though the RNA after denaturation can be shown to consist of pieces only
a few hundred nucleotides in length (again, an observation made by almost all workers).
Furthermore, such badly degraded structures still have all portions of the genome equally
present (Billeter et aL 1974).
(2) During the course of synthesis, the nascent chain could remain attached to the
structure via the interaction of the primer with the PB sequence (although, for simplicity,
Fig. 2 is not drawn this way).
(3) The strongest dimer linkage point is near the 5' end of the genome (Kung et aL 1976).
This would contribute to maintaining the dimer structure until ( - ) strand synthesis was
very nearly complete. Note that the result described in (I) shows that other, weaker linkage
points must also exist. Structures suggesting multiply linked subunits can be visualized by
electron microscopy (Mangel et al. 1974; Weissman et al. 1974).
Three other possible consequences of this model can be derived by considering a possibility suggested by in vitro experiments with ASV, described in a previous section. If the
polymerase plus growing chain are able to transfer to new templates with only limited
homology to the nascent chain, then such a mechanism could easily result in a reasonable
frequency of deletions or re-duplications. The former would occur by a jump on either
template to the left hand side of the region already copied (e.g. from C to E) and the latter
by a jump to the right on the other genome (e.g. from C to B'). Deletions of non-essential
information (for example, src) have been known for a long time to occur with high frequency
(Vogt, I97Ia). More recently, it has been recognized that other types of deletions also occur
with high frequency (Shields et al. 1978; J. Coffin & C. Barker, unpublished data). Reduplications of sequences have not been reported, although the heteroduplex maps of
partial td viruses found by Lai et al. (I977) are as consistent with re-duplication as with the
authors' interpretation of insertion of unrelated information.
If the mechanism ofproviral synthesis allows such jumps to relatively unrelated sequences,
then, as a very rare event, such a jump might take place between a virus genome and
unrelated cell m R N A which happened to be nearby. This kind of mechanism could lead
from an endogenous virus to a sarcoma virus, since the src gene is apparently closely related
to sequences expressed as m R N A in uninfected cells (Spector et al. 1978). As compared to
recombination between viruses, such an event would be extremely uncommon, but the
selection would be powerful. One virion with an src gene out of perhaps lOTM or more in a
viraemic animal would manifest itself as a visible tumour for the curious scientist to select
for laboratory propagation. Such a crossover would be much more f r e q u e n t - although
still r a r e - if there were a partial homology between viral and cell sequence. This would
account for the apparently frequent reappearance of sarcoma virus after injection of virus
with partial td deletions into chickens (Hanafusa et al. 1977). Note that this mechanism
would place no constraint on where such an insert occurred in the genome and that the
final location would be determined probably by chance small homologies in sequence. Also,
the most frequent event would involve replacement of virus with cell information, leading
to a defective transforming virus - the usual case.
Although copy-choice recombination is not a new proposal for R N A tumour viruses
(Vogt, I973), I believe the ability to use broken genomes to recombine and even to replicate
efficiently is new. This model has some experimental tests to distinguish it from other
possibilities. For one, unlike the model proposed by Hunter (1978), it would be possible,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
R e v i e w : Retrovirus genomes
2I
under certain experimental conditions, to repair large deletions, although deletions with
no intervening common sequence could not be repaired. Second, the progeny of cells
infected with heterozygote viruses should yield either parental or (probably more frequently)
recombination types, never heterozygous or partially heterozygous progeny, so long as
multigenome heterozygotes are excluded. The third prediction is the relationship between
intactness of the virion RNA and recombination frequency mentioned above.
The overall scheme of proviral DNA synthesis that would emerge from this model
appears quite complex - the growing DNA chain is transferred not only twice from end to
end of the genome, but also repeatedly from one genome to another before a complete
( - ) strand is synthesized. In spite of this complexity, however, such a mechanism would
allow the virus to survive both physical damage to its genome RNA and to undergo rather
rapid evolution by recombination. This would be an advantageous situation since it would
combine in retroviruses some of the best features of both RNA and DNA viruses. The
proposed schemes of replication and recombination also seem to tie together in a fairly
neat package many of the apparently diverse features of retroviruses which distinguish
them from all other groups of viruses.
I thank numerous workers, cited in the text, for the communication of unpublished material
without which this review would be hopelessly out of date. I also want to thank P. Balduzzi,
K. Conklin, P. Duesberg, M. Kotler, B. Mermer, S. Voynow, and P. Tsichlis for constructive
discussion and suggestions to improve the manuscript. Work in my laboratory is supported
by Public Health Research Grant CA I7659 from the National Cancer Institute and Grant
NPz46B from the American Cancer Society. I am the recipient of a Faculty Research
Grant from the American Cancer Society.
Note added in proof. After submission of this article, it came to my attention that a
similar mechanism for retrovirus recombination had previously been proposed (Haseltine
and Baltimore, I976).
REFERENCES
ALEVV, M. C. & VO6T, P. K. (I978). T s pol m u t a n t s o f avian s a r c o m a viruses: m a p p i n g a n d d e m o n s t r a t i o n o f
single cycle r e c o m b i n a n t s . Virology 87, 21-33.
BALDUZZI, P., BEAMAND, J. A., CHRISTENSEN, J. R., PEARSON, Y. M. & WYKE, J. A. 0978). Provisional m a p p i n g
o f t r a n s f o r m a t i o n defective t e m p e r a t u r e sensitive m u t a n t s o f r o u s s a r c o m a virus. I n Avian RNA Tumor
Viruses, pp. 112-I 2I. Edited by S. Barlati a n d C. d e G i u l i - M o r g h e n . P a d u a : Piccin Medical Books.
BEEMON, K. 0978)- Oligonucleotide fingerprinting with R N A t u m o r virus R N A . Current Topics in Microbiology and Immunology (in the press).
BEEMON, K. L. & KEITI.I, S. m. 0976). Structure of R o u s s a r c o m a virus R N A : I) Localization o f Nn-methyladenosine; 2) T h e sequence o f 23 nucleotides following the 5' capped t e r m i n u s mVGpppGmp. I n Animal
Virology, pp. 97-IO5. Edited by D. Baltimore, A. S. H u a n g a n d C. F. Fox. N e w York: A c a d e m i c Press.
BEEMON, K., DUESBERG, V. H. & VOGT, P. X. (I974). Evidence for crossing-over between avian t u m o r viruses
b a s e d o n analysis o f viral R N A s . Proceedings of the National Academy of Sciences of the United States
of America 7 x, 4254-4258.
BILLETER, M. A., PARSONS, S. T. & COFEIN, J. M. (I974). T h e nucleotide sequence complexity of a v i a n t u m o r
virus R N A . Proceedings of the National Academy of Sciences of the United States of America 7 I,
356o-3564.
BISHOP, J. M. 0 9 7 8 ) . Retroviruses. Annual Review of Biochemistry 47, 35-88.
BISHOP, J. M., WEISS, S. R., OPPERMAN,H., HACKETT, P., QUINTRELL, N., CHEN, L. S., LEVINTOW, L. & VARMUS,I-I. E.
(I978). T h e strategy of retrovirus gene expression. In Avian RNA Tumor Viruses, pp. I 8 I - I 8 9 . Edited by
S. Barlati a n d C. d e G i u l i - M o r g h e n . P a d u a : Piccin Medical Books.
BISHOP, J. M., DENG, C. T., MAHY, B. W. J., QUINTRELL, N., STAVNEZER, F. C. & VARMUS, H. E. 0976). Synthesis
o f viral R N A in cells infected by avian s a r c o m a viruses. I n Animal Virology, pp. I-2o. Edited by D.
Baltimore, A. S. H u a n g a n d C. F. Fox. N e w Y o r k : A c a d e m i c Press.
BISTER, K., VARMUS,H. E., STAVNEZER, E., HUNTER, E. & VO6T, P. K. (1977). Biological a n d biochemical studies
o n the inactivation o f avian oncoviruses by ultraviolet irradiation. Virology 77, 689-7o4.
BLAIR, D.G. 0977). Genetic r e c o m b i n a t i o n between avian leukosis a n d s a r c o m a viruses. Experimental
variables a n d the frequencies o f recombination. Virology 77, 534-544.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
22
ft. M. C O F F I N
BROKER, T. R. & DOERMANN, A. H. (I975)- Molecular a n d genetic r e c o m b i n a t i o n of bacteriophage T4. Annual
Review of Genetics 9, zI 3-244.
CASHION, L. M., JOHO, R. H., PLANITZ, M. A., BILLETER, M. A. & WEISSMANN, C. (1976). Initiation sites o f R o u s
s a r c o m a virus R N A - d i r e c t e d D N A synthesis in vitro. Nature, London 262, 186-I9O.
CHEUNG, K. S., SMITH, R. E., STONE, M. P. & JOKLIK, W. K. 0972). C o m p a r i s o n o f i m m a t u r e (rapid harvest)
a n d m a t u r e R o u s s a r c o m a virus particles. Virology 50, 851-864.
COFFIN, J. M. (1976). G e n e s responsible for t r a n s f o r m a t i o n by avian R N A t u m o r viruses. Cancer Research
36, 4282-4288.
COFFIN, J. M. & BILLETER,M. A. (1976). A physical m a p of the R o u s s a r c o m a virus genome. JournalofMolecular
Biology xoo, 293-318.
COFFIN, J. M. & HASELTINE, W. A. (I977a). T e r m i n a l r e d u n d a n c y a n d the origin of replication o f R o u s s a r c o m a
virus R N A . Proceedings of the National Academy of Sciences of the United States of America 74,
19o8-1912.
COEFIN, J. M. & HASELTINE, W. A. (1977b)- Nucleotide sequence o f R o u s s a r c o m a virus R N A at the initiation
site o f D N A synthesis. Journal of Molecular Biology xx7, 8o5-814.
COFFIN, J. M. & TEMIN, H. M. 0972). Hybridization o f R o u s s a r c o m a virus deoxyribonucleic acid polymerase
p r o d u c t a n d ribonucleic acids f r o m chicken a n d rat cells infected with R o u s s a r c o m a virus. Journal of
Virology 9, 766-775.
COFFIN, J. M., PARSONS, J. T., RYMO, L., HAROZ, R. K. & WEISSMANN, C. (1974). A n e w a p p r o a c h to the isolation
o f R N A - D N A hybrids a n d its application to the quantitative d e t e r m i n a t i o n o f labelled t u m o r virus
R N A . Journal of Molecular Biology 86, 373-396.
COFFIN, J., CHAMPION, M. A. & CHABOT, F. (I978a). G e n o m e structure o f avian R N A t u m o r viruses: relationships between e x o g e n o u s a n d e n d o g e n o u s viruses. I n Avian RNA Tumor Viruses, pp. 68-87. Edited by
S. Barlati a n d C. d e G i u l i - M o r g h e n . P a d u a : Piccin Medical Books.
COFFIN, J. M., HAGEMAN, T. C., MAXAM,A. M. & HASELTINE, W. A. (I978b). Structure of the g e n o m e o f M o l o n e y
m u r i n e leukemia virus: a terminally r e d u n d a n t sequence. Cell 13, 761-773.
COLLETT, M. S. & EARAS, A. J. (I976). Evidence for circularization of the avian o n c o r n a v i r u s R N A g e n o m e
d u r i n g proviral D N A synthesis f r o m studies o f reverse transcription in vitro. Proceedings of the National
Academy of Sciences of the United States of America 73, I329-I332COLLETT, M. S., DIERKS, P., PARSONS, J. T. & FARAS, A. J. (I978)- R N a s e H hydrolysis of the 5' t e r m i n u s o f
avian s a r c o m a virus g e n o m e d u r i n g reverse transcription. Nature, London 272, I81-I83.
CORDELL, B., STAVNEZER, E., FRIEDRICH, R., BISHOP, J. M. & GOODMAN, H. M. (1976). Nucleotide sequence that
binds primer for D N A synthesis to the avian s a r c o m a virus genome. Journal of Virology I9, 548-558.
DAHLBERG, J. E., SAWYER, R. C., TAYLOR, J. M., FARAS, A. J., LEVINSON, W. E., GOODMAN~H. M. & BISHOP, J. M.
(I974). Transcription o f D N A f r o m the 7oS R N A of R o u s s a r c o m a virus. I. Identification o f a specific
4S R N A which serves as primer. Journal of Virology x3, I 126-1133DIGGELMAN, H. & KLEMENZ, R. (I978). Characterization o f the c o m m o n precursor polypepfide to the two
envelope glycoproteins. I n Avian RNA Tumor Viruses, pp. 191-2o8. Edited by S. Barlati a n d C. deGuiliM o r g h e n . P a d u a : Piccin Medical B o o k s
EIDEN, J. J., QUADE, K. & NICHOLS, J. W. (I976). Interaction o f t r y p t o p h a n transfer R N A with R o u s s a r c o m a
virus 35 S R N A . Nature, London 259, 245-247.
EISENMANN, R. N. & "VOGT, V. M. (1978)- T h e biosynthesis o f o n c o v i r u s proteins. Biochimica et Biophysica
Acta 473, 187-239.
ELDER, J. n., GAUTSCH, J. W., JENSEN, F. C., LERNER, R. A., HARTLEY, J. A. & ROWE, W. P. (I977). Biochemical
evidence t h a t M C F m u r i n e leukemia viruses are envelope (env) gene r e c o m b i n a n t s . Proceedings of the
National Academy of Sciences of the United States of America 74, 4676--468o.
FALLER, D. V. & HOPKINS, N. (~978). I"1 oligonucleotides that segregate with t r o p i s m a n d with properties o f
gp7o in r e c o m b i n a n t s between N - a n d B-tropic m u r i n e leukemia viruses. Journal of Virology 26, 153-I 58.
FRIEDRICH, R., KUNG, H. J., BAKER, B., VARMUS,H. E., GOODMAN, H. M. & BISHOP, J. M. (I977). Characterization
o f D N A c o m p l e m e n t a r y to nucleotide sequences at the 5'-terminus of the avian s a r c o m a virus g e n o m e .
Virology 79, 198-215.
FRITSCH, E. & TEMIN, H. M. (1977). F o r m a t i o n a n d structure of infectious D N A o f spleen necrosis virus.
Journal of Virology 2x, 119-13o.
FURUICHI, Y., SHATKIN, A. J., STAVNEZER, E. & BISHOP, J. M. (I975). Blocked, m e t h y l a t e d 5'-terminal sequence
in a v i a n s a r c o m a virus R N A . Nature, London 257, 618-62o.
GUNTAKA, R. V., MAHY, B. W. J., BISHOP, J. M. & VARMUS,H. E. (I975). E t h i d i u m b r o m i d e inhibits a p p e a r a n c e
o f closed circular viral D N A a n d integration o f virus specific D N A in duck cells infected by a v i a n
s a r c o m a virus. Nature, London 253, 5o7-5I I.
GUNTAKA, R. V., RICHARDS, O. C., SHANK, P. R., KUNG, H. J., DAVIDSON, N., FRITSCH, E., BISHOP, J. M. & VARMUS,
H. E. (I976). Covalently closed circular D N A of avian s a r c o m a virus: purification f r o m nuclei of infected quail t u m o r cells a n d m e a s u r e m e n t by electron microscopy a n d gel electrophoresis. Journal of
Molecular Biology xo6, 337-357.
HANAFUSA, H. (1977). Cell t r a n s f o r m a t i o n by R N A t u m o r viruses. I n Comprehensive Virology, vol. 9,
pp. 4Ol-483. Edited by H. F r a e n k e l - C o n r a t a n d R. R. W a g n e r . N e w Y o r k : P l e n u m Press.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
Review: Retrovirus genomes
23
HANAFUSA, H., HALPERN, C. C., BUCHHAGEN, D. L. & KAWAI, S. (I977). Recovery o f avian s a r c o m a virus f r o m
t u m o r s induced by transforming-defective m u t a n t s . Journal of Experimental Medicine x46 , 1735-I747.
HASELTINE, W. A. & BALTIMORE, D. (I976). In vitro replication of R N A t u m o r viruses. In Animal Virology,
PP. I75-213. Edited by D. Baltimore, A. S. H u a n g a n d C. F. Fox. N e w Y o r k : A c a d e m i c Press.
HASELTINE, W. A. & COFFIN, J. M. (I978). In vitro studies o f the replication o f M o l o n e y m u r i n e leukemia virus.
Cold Spring Harbor Symposium on Quantitative Biology 43 (in t h e press).
HASELTINE, W. A., KLEID, D. G., PANET, A., ROTHENBERG, E. & BALTIMORE, D. (I976). Ordered transcription o f
R N A t u m o r virus genomes. Journal of Molecular Biology xo6, IO9-I3I.
HASELTINE, W. A., MAXAM, A. M. & GILBERT, W. (I977). R o u s s a r c o m a virus g e n o m e is terminally r e d u n d a n t :
the 5' sequence. Proceedings of the National Academy of Sciences of the United States of America 74,
989-993.
HAYWARD, W. S. (I977). Size a n d genetic c o n t e n t of viral R N A s in a v i a n oncovirus infected cells. Journal of
Virology 24, 47-63.
HERMAN, A. C., GREEN, R. W., BOLOGNESI, D. P. & VANAMAN,T. C. (1975). C o m p a r a t i v e chemical properties o f
a v i a n o n c o n a v i r u s polypeptides. Virology 64, 339-348.
HU, S., DAVIDSON, N. & VERMA, I. M. (I977). A heteroduplex s t u d y o f the sequence relationships between the
R N A s o f M - M S V a n d M - M L V . Cell to, 469-477.
HUGHES, S., VOGT, P. R., SHANK, P. R., SPECTOR, P. H., KUNG, H. J., BREITMAN, M. L., BISHOP, J. M. & VARMUS,
H.E. (I978)- Proviruses o f avian s a r c o m a viruses are terminally r e d u n d a n t , coextensive with unintegrated linear D N A , a n d integrated at m a n y sites in rat cell D N A . Cell (in the press).
HUNTER, E. (I978)- T h e m e c h a n i s m for genetic r e c o m b i n a t i o n in the avian retroviruses. Current Topics in
Microbiology and Immunology 79, 295-309.
HUNTER, E. & VOGT, P. K. (I976). Temperature-sensitive m u t a n t s of a v i a n s a r c o m a viruses. Genetic recombin a t i o n with wild type s a r c o m a virus a n d physiological analysis o f multiple m u t a n t s . Virology 69, 23-34.
JACQUET, M., GRONER, Y., MONROY, G. & HURWITZ, J. (I974). T h e in vitro synthesis o f avian m y d o b l a s t o s i s
viral R N A sequences. Proceedings of the National Academy of Sciences of the United States of America
71, 3045-3049.
JOHO, a. H., BILLETER, M. A. & WEISSMANN, C. (1975). M a p p i n g o f biological f u n c t i o n o n R N A o f a v i a n t u m o r
viruses: location of regions required for t r a n s f o r m a t i o n a n d determination o f h o s t range. Proceedings
of the National Academy of Sciences of the United States of America 72, 4772-4776.
JOHO, R. H., STOLL, E., FRITS, R. R., BILLETER, M. A. & WEISSMANN, C. (I976). A partial genetic m a p o f R o u s
s a r c o m a virus R N A : location o f polymerase, envelope a n d t r a n s f o r m a t i o n markers. In Animal Virology,
pp. 127-145. Edited by D. Baltimore, A. S. H u a n g a n d C. F. Fox. N e w Y o r k : A c a d e m i c Press, Inc.
JUNO, R. H., BILLETER, M. A. & WEISSMANN, C. (1978). C o n c o r d a n c e o f the R N A termini of r e c o m b i n a n t s f r o m
crosses between avian retroviruses with different termini. Virology 85, 364-377.
JUNGHANS, R. P., HU, S., KNIGHT, C. g. & DAVIDSON, N. (1975). Heteroduplex analysis o f avian R N A t u m o r
viruses. Proceedings of the National Academy of Sciences of the United States of America 74, 477-48 I.
KAWAI, S. & HANAFUSA, H. (I976). R e c o m b i n a t i o n between a t e m p e r a t u r e sensitive m u t a n t a n d a deletion
m u t a n t o f R o u s s a r c o m a virus. Journal of Virology x9, 389-397.
KEITH, J. & FRAENKEL-CONRAT,H. (I975). Identification o f the 5' e n d o f R o u s s a r c o m a virus R N A . Proceedings of the National Academy of Sciences of the United States of America 72, 3347-3350.
KELLER, W. & CROUCH, a. (I972). D e g r a d a t i o n o f D N A - R N A hybrids by ribonuclease H a n d D N A polym e r a s e of cellular a n d viral origin. Proceedings of the National Academy of Sciences of the United States
t~fAmerica 69, 336o-3364.
KRZYZEK, R. A., COLLETT, M. S., LAW, A. F., PERDUE, M. L., LEIS, J. P. & FARAS, g. J. (I978). Evidence for splicing
o f avian s a r c o m a virus 5'-terminal g e n o m i c sequences o n t o viral-specific R N A in infected cells. Proceedings of the National Academy of Sciences of the United States of America 75, I284 - I 2 8 8 .
KUNG, H.-J., HU, S., BENDER, W., BAILEY, J. M. & DAVIDSON, N. 0976). R D - I I4, b a b o o n , a n d woolly m o n k e y
viral R N A s c o m p a r e d in size a n d structure. Cell 7, 6o9.
LAI, M. M. C. & DUESBERG, P. H. (I972). Adenylic acid-rich sequence in R N A s o f R o u s s a r c o m a virus a n d
R a u s c h e r m o u s e leukemia virus. Nature, London 235 , 383-386.
LAI, M. M. C., HU, S. & VOGT, P. K. (r977)- Occurrence o f partial deletion a n d substitution o f the src gene in
the R N A g e n o m e o f a v i a n s a r c o m a virus. Proceedings of the National Academy of Sciences of the
United States of America 74, 4781-4785.
LEIS, J., McGINNIS, J. & GREEN, R. W. (1978). R o u s s a r c o m a virus p I 9 binds to specific d o u b l e - s t r a n d e d regions
o f viral R N A : effects o f PI9 o n cleavage of viral R N A by R N a s e III. Virology 84, 87-98.
LEIS, J., BERKOWER, I. & HURWITZ, J. (I973). R N A - d e p e n d e n t D N A polymerase activity o f R N A t u m o r
viruses. I n DNA Synthesis In vitro, pp. 287-308. Edited by R. D. Wells a n d R. B. I n m a n . Baltimore:
University P a r k Press.
LEVIN, J. G., GRIMLEY, P. M., RAMSEUR, J. N. & BEREZESKEY, I.K. (1974)- Deficiency o f 6o to 7oS R N A in
m u r i n e leukemia virus particles a s s e m b l e d in cells treated with a c t i n o m y c i n D. Journal of Virology x4,
I52-161.
LINIAL, M. & NEIMAN, P. E. (I976). Infection o f chick cells by s u b g r o u p E viruses. Virology 73, 5o8-52o.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
24
J.M. COFFIN
LURIA, S. E., DARNELL, J. E., BALTIMORE, D. & CAMPBELL, A. ( I 9 7 8 ) . I n
General Virology,
3rd Edition, p. I9o.
New York: John Wiley and Sons.
MANGLE, W. F., DELIUS, H. & DUESBURG, P. H. (I974). Structure and molecular weight of the 6o-7oS R N A
and the 3o-4oS R N A of the Rous sarcoma virus. Proceedings of the National Academy of Sciences
of the United States of America 7 x, 4541-4545.
MARTIN, G. S. (197o). Rous sarcoma virus: a function required for the maintenance of the transformed state.
Nature, London 227, IO2I--IO23.
MARTIN, G. S. & DUESBERG,P. H. (I972). The a subunit in the R N A of transforming arian tumor viruses.
I. Occurence in different virus strains. II. Spontaneous loss resulting in non-transforming variants.
Virology 47, 494-497MELLON, P. & DUESBERG, P.H. 0977). Subgenomic, cellular Rous sarcoma virus RNAs contain oligonucleotides from the 3' half and the 5' terminus of virion RNA. Nature, London 270, 631-634.
M()LLING, K., BOLOGNESI, D. P., BAUER, H., BUSES, W., PLASSMAN, H. W. & HAUSEN, P. (1971). Association of
viral reverse transcriptase with an enzyme degrading the R N A moiety of R N A - D N A hybrids. Nature
New Biology 234, 24o-243.
OPPERMANN,H., BISHOP,J. M., VARMUS,H. E. & LEVINTOW,L. (I977). A joint product of the genes gag and pol
of avian sarcoma virus: a possible precursor of reverse transcriptase. Cell x2, 993-Ioo5.
OWADA, m., IHARA,S., TOVASHIMA,K., I,:OZAL,Y. & SUCINO,Y. (I976). Ultraviolet inactivation of avian sarcoma
viruses: biological and biochemical analysis. Virology 69, 710-718.
PARSONS, J. Y., COFFIN, J. M., HAROZ, R. K., BROMLEY, P. A. & WEISSMANN, C. (t974). Quantitative determination
and location of newly synthesized virus-specific ribonucleic acid in chicken cells infected with Rous
sarcoma virus. Journal of Virology IX, 761-774.
PHILIPSOS, L, ANDERSSON,P., OLSnEVSKV,U., WEINBERG, R. & BALTIMORE, D. (I978). Translation of MuLV
and MSV RNAs in nuclease treated reticulocyte extracts: enhancement of the gag-pol polypeptide
with yeast suppressor tRNA. Cell x3, I89-2oo.
PROODFOOT, n. J. & BROWNLEE,G. 6. (t976). 3' non-coding region sequences in eukaryotic messenger RNA.
Nature, London 263, 2I 1-214.
PURCHIO, A. F., ERICKSON,E. & ERICKSON,R. L. (1977). Translation of 35S and of subgenomic regions of avian
sarcoma virus RNA. Proceedings of the National Academy of Sciences of the United States of America
74, 4661-4665.
PURCHIO, A. F., ERICKSON, E., BRUGGE, J. S. & ERICKSON, R. L. 0978). Identification of a polypeptide encoded
by the avian sarcoma virus src gene. Proceedings of the National Academy of Sciences of the United
States of America 75, I567-157 I.
ROTHENBERG, E., SMOTKIN, D., BALTIMORE, D. & WEINBERG, R. D. 0977). In vitro synthesis of infectious D N A
of murine leukemia virus. Nature, London 269, 122-I 26.
ROTHENBERG, E., DONOGHUE, D.J. & BALTIMORE, D. (1978). Analysis of a 5' leader sequence on murine
leukemia virus R N A : heteroduplex mapping with long reverse transcriptase products. Cell x3, 435-45 I.
ROBINSON, H. (I976). Intracellular restriction on the growth of induced subgroup E avian type C viruses in
chicken cells. Journal of Virology xM, 856-866.
ROBINSON, H. (I978). Inheritance and expression of chicken genes which are related to avian leukosissarcoma viruses. Current Topics in Microbiology and Immunology (in the press).
RUBIN, rI. & TEMIN,H. M. 0959)" A radiological study of cell-virus interactions in the Rous sarcoma. Virology
7, 75-9I.
RYMO, L., PARSONS,J. T., COFFIN, J. M. & WEISSMANN, C. (I974). In vitro synthesis of Rous sarcoma virusspecific R N A is catalyzed by a DNA-dependent R N A polymerase. Proceedings of the National Academy
of Sciences of the United States of America 7x, 2781-2786.
SAWYER, R. C. & I)AHLBERG, J.E. 0973). Small RNAs of Rous sarcoma virus: characterization by twodimensional polyacrylamide gel electrophoresis and fingerprint analysis. Journal of Virology x2,
1226-I237.
SCHWARTZ, D. E., ZAMECNIK, P. C. & WEITH, U. L. ( I 9 7 7 ) . R o u s s a r c o m a v i r u s g e n o m e is t e r m i n a l l y r e d u n d a n t :
the 3' sequence. Proceedings of the National Academy of Sciences of the United States of America 74,
994-998.
SHAIKH, R., L1NIAL, M., COFFIN, J. M. & EISENMAN, R. W. ( I 9 7 8 a ) . Recombinant avian oncoviruses. I. Alterations
in precursor to the internal structural proteins. Virology 87, 326-338.
SHAIKH, R., LINIAL, M., BROWN, S., SEN, A. & EISENMANN, R. 0978b). Recombinant avian oncoviruses. II.
Alterations in the gag proteins and evidence for intragenic recombination. Virology, (in the press).
SHANK, P. R., HUGHES, S., KING, H. J., MAJORS, J. E., QUINTRELL~ N., GUNTAKA, R. V., BISHOP, J. M. & VARMUS,
H. E. (I978). Mapping unintegrated forms of avian sarcoma virus DNA: both termini of linear D N A
bear a 3oo nucleotide sequence present once or twice in two species of circular DNA. Cell (in the
press).
SHEALY, D. J. & REUCKERT,R. R. (I978). Proteins of Rous-associated virus 6I, an avian retrovirus: common
precursor for glycoproteins gp85 and gP35 and use of pactamycin to map translational order of proteins
in the gag, pol, and env genes. Journal of Virology 26, 38o-388.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
Review: Retrovirus genomes
SHIELDS, A., WITTE, O. N., ROTHENBERG, E. & BALTIMORE, D. (I978).
25
High frequency of aberrant expression of
moloney murine leukemia virus in clonal infections. Cell x4, 6oi-6IO.
SHINE, J., CZERNILOFSKY, A. P., FRIEDRICH, a . , BISHOP, J. M. & GOODMAN, H. M. (I977). Nucleotide sequence at
the 5" terminus of the avian sarcoma virus genome. Proceedings of the National Academy of Sciences of
the United States of America 74, I473-I477.
SMOTKIN, D., GIANNI, A. M., YOSHIMURA, F. & WEINBERG, R. A. (I976). Transfection of Moloney MuLV DNA.
In Animal Virology, pp. 237-242. Edited by D. Baltimore, A. S. Huang, and C. F. Fox. New York:
Academic Press, Inc.
SPECTOR, D. H., SMITH, K., PADGETT, T., McCOMBE, P., ROULLAND-DUSSOIX, D., MOSCOVIC1, C., VARMUS, H. E. &
BISt-IOP, J. M. (I978). Uninfected avian cells contain R N A related to the transforming gene of avian
sarcoma viruses. Cell x3, 371-379.
SOUTHERN, E. M. (1975). Detection of specific sequences among D N A fragments separated by gel electrophoresis. Journal of Molecular Biology 98, 5o3-517.
STACEY, D. & HANAFUSA,H. (I978). Nuclear conversion of microinjected avian leukosis virion R N A into an
envelope-glycoprotein messenger. Nature, London 273, 779-782.
STEHELIN, D., VAR~aUS,H. E., BISHOP, J. M. & VOGT, P. K. 0976)- D N A related to the transforming gene(s) of
avian sarcoma virus is present in normal avian DNA. Nature, London 26o, I7O-I72.
STOLL, E., BILLETER, M.A., PALMENBERG, A. & WEISSMANN, C. (I977). Avian myeloblastosis virus R N A is
terminally redundant: implications for the mechanism of retrovirus replication. Cell x2, 57-72.
TAL, J., KUNG, H. J., VARMUS, H. E. & BISHOP, J. M. (I977). Purification of D N A complementary to the env gene
of avian sarcoma virus and analysis of relationships among the env genes of avian leukosis-sarcoma
virus. Journal of Virology 2x, 497-505.
TAYLOR, J. M. (I977). An analysis of the role of t R N A species as primers for the transcription into D N A of
R N A tumor virus genomes. Biochimica et Biophysica Acta 473, 57-7I.
TAYLOR, J. M. & ILLMENSEE,R. (1975). Site on the R N A of an avian sarcoma virus at which primer is bound.
Journal of Virology x6, 553-558.
TEMIN, H. M. & BALTIMORE, D. (I97Z)- RNA-directed D N A synthesis and R N A tumor viruses. Advances in
Virus Research x7, I29-185.
TILGHMAN, S. M., CURTIS, P. J., TIEMEIER, D. C., LEDER, P. & WEISSMANN, C. (I978). The intervening sequence of
a mouse fl-globin gene is transcribed within the I5S fl-globin m R N A precursor. Proceedings of the
National Academy of Sciences of the United States of America 75, I3O9-I313.
TOYOSHIMA, K. & VOGT, P. K. (I969). Temperature sensitive mutants of an avian sarcoma virus. Virology 39,
93o-93I.
VAR~V.;S, H. E., VOGT, P. K. & BISHOP, J.M. 0973)- Integration of deoxyribonucleic acid specific for Rous
sarcoma virus after infection of permissive and nonpermissive hosts. Proceedings of the National
Academy of Sciences of the United States of America 7o, 3o67-3o71.
VARMUS, H. E., HENSLEY, S., LINN, J. & WHEELER, K. (I976). Use of alkaline sucrose gradients in a zonal rotor
to detect integrated and unintegrated avian sarcoma virus-specific D N A in cells. Journal of Virology x8,
574-585.
VARMUS, H. E., HENSLEY, S., KUNG, H. J., OPPERMANN, H., SMITH, V. C., BISHOP, J. M. & SHANK, P. K. (I978).
Kinetics of synthesis, structure, and purification of avian sarcoma virus-specific D N A made in the
cytoplasm of acutely infected ceils. Journal of Molecular Biology I2o, 55-82.
VERMA, I. M. 0977). The reverse transcriptase. Biochimica et Biophysica Acta 473, 1-38.
VOGT, P.K. (I97I a). Spontaneous segregation of nontransforming viruses from cloned sarcoma viruses.
Virology 46, 939-946.
VOGT, e. K. (I97I b). Genetically stable reassortment of markers during mixed infection with avian tumor
viruses. Virology 46, 947-952.
VOGT, P. K. (I973). The genome of avian R N A tumor viruses: a discussion of four models. In Possible
Episomes in Eukaryotes, pp. 35-4I. Edited by L. Sylvestri. Amsterdam: North-Holland.
VOGX, P. K. (I977). The genetics of R N A tumor viruses. In Comprehensive Virology, vol. 9, PP. 341-455.
Edited by H. Fraenkel-Conrat and R. R. Wagner. New York: Plenum Press.
VOGT, V. M., EISENMANN, R. & DIGGELMAN, H. (1975). Generation of avian myeloblastosis virus structural
proteins by proteolytic cleavage of a precursor polypeptide. Journal of Molecular Biology 96,
471-493.
VOGT, P. K. & l~tr, S. (I977). The genetic structure of R N A tumor viruses. Annual Review of Genetics H ,
2o3-238.
WANG, L.-H., DUESBERG, P. N., BEEMON, K. & VOGT, P. K. (1975). Mapping RNase Tl-resistant oligonucleotides
of avian tumor virus RNAs: sarcoma-specific oligonucleotides are near the poly(A) end and oligonucleotides common to sarcoma and transformation-defective viruses are at the poly(A) end. Journal of
Virology I6, IO5I-IO7O.
WANG, L.-H., DUESBERG, P. H., MELLON, P. & VOGT, P. K. (I 976). Distribution of envelope-specificand sarcomaspecific nucleotide sequences from different parents in the RNAs of avian tumor virus recombinants.
Proceedings of the National Academy of Sciences of the United States of America 73, 447-45 t.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39
26
J . M . COFFIN
WANG, L.-H., DUESBERG, P. H., ROBBINS, T. YOKOTA, H. & VOGT, P. K. (1977). The terminal oligonucleotide o f
avian tumor virus RNAs are genetically linked. Virology 82, 472-493.
WEINBERG, R.A. (I977). Structure of the intermediates leading to the integrated provirus. Biochimica et
Biophysica Acta 473, 39-55.
WEINER, A. i . & WEBER, K. (I973). A single U G A codon functions as a natural termination signal in the
coliphage Qfl coat protein cistron. Journal of Molecular Biology 8o, 837-855.
WEISS, R. A., MASON, W. S. & VOGT, P. K. (1973). Genetic recombinants and heterozygotes derived from endogenous and exogenous avian R N A tumor viruses. Virology 52, 535-55z.
WEISS, S. R., VARMUS,H. E. & BISHOP,J. M. (I977)- The size and genetic composition of virus-specific RNAs in
the cytoplasm of cells producing avian sarcoma-leukosis viruses. Cell x2, 983-992.
WEISSMANN, C., PARSONS, J. T., COFFIN, J. M., RYMO, L., BILLETER, M. A. & HOFSTETTER, H. (1974). Studies on
the structure and synthesis of Rous sarcoma virus RNA. CoM Spring Harbor Symposium on Quantitative Biology 34, IO43-1o56.
WYKE, J. A., BELL, J. G. & BEAMAND, J. A. (1974). Genetic recombination among temperature-sensitive mutants
of Rous sarcoma virus. Cold Spring Harbor Symposium on Quantitative Biology 39, 897-9o5.
YUEN, P. H. & WONG, 1". K. V. (I977). A morphological study on the ultrastructure and assembly of murine
leukemia virus using a temperature-sensitive mutant restricted in assembly. Virology 8o, 260-274.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Fri, 16 Jun 2017 07:58:39