VIROLOGY - CHAPTER SIX 

 
ONCOGENIC VIRUSES  
Dr Richard C. Hunt

Cancers are the result of a disruption of the normal restraints on cellular proliferation. It is apparent that the number of ways in
which such disruption can occur is strictly limited and there may be as few as forty cellular genes in which mutation or some
other disruption of their expression leads to unrestrained cell growth.

There are two classes of these genes in which altered expression can lead to loss of growth control: 
(a) Those genes that are stimulatory for growth and which cause cancer when hyperactive. Mutations in these genes will be
dominant. These genes are called oncogenes.
(b) Those genes that inhibit cell growth and which cause cancer when they are turned off. Mutations in these genes will be
recessive. These are the anti-oncogenes or tumor-suppressor genes.

Viruses are involved in cancers because they can either carry a copy of one of these genes or can alter expression of the
cell's copy of one of these genes.

CLASSES OF TUMOR VIRUSES

There are two classes of tumor viruses:

 DNA tumor viruses

 RNA tumor viruses, the latter also being referred to as RETROVIRUSES.

We shall see that these two classes have very different ways of reproducing themselves but they often have one aspect of
their life cycle in common: the ability to integrate their own genome into that of the host cell. Such integration is not, however,
a pre-requisite for tumor formation.

TRANSFORMATION AND ONCOGENES

If a virus takes up residence in a cell and alters the properties of that cell, the cell is said to be transformed. Transformation by
a virus is the change in the biological properties of a cell that results from the regulation of the cell by viral genes and that
confer on the infected cells certain properties of neoplasia.

Transformation often includes loss of growth control, anchorage-independent growth, ability to invade extracellular matrix,
dedifferentiation and immortalization. In carcinomas, many epithelial cells undergo an epithelial-mesenchymal transformation.
Transformed cells often exhibit chromosomal aberrations and the changes seen in transformation often, but not always, result
from the integration of the viral genome into the host cell's chromosomes.

The region of the viral genome (DNA in DNA tumor-viruses or RNA in RNA-tumor viruses) that can cause a tumor is called an
oncogene. This foreign gene can be carried into a cell by the virus and cause the host cell to take on new properties.

The discovery of viral oncogenes in retroviruses led to the finding that they are not unique to viruses and homologous genes
(called proto-oncogenes) are found in all cells. Indeed, it is likely that the virus picked up a cellular gene during its evolution
and this gene has subsequently become altered. Normally, the cellular proto-oncogenes are not expressed in a quiescent cell
since they are involved in growth (which is not occurring in most cells of the body) and development; or they are expressed
under strict control by the cell. However, they may become aberrantly expressed when the cell is infected by tumor viruses
that do not themselves carry a viral oncogene.  We shall see later how this happens but it is clear that a virus may cause
cancer in two ways: It may carry an oncogene into a cell or it may activate a cellular proto-oncogene.
The discovery of cellular oncogenes opened the way to the elucidation of mechanisms by which non-virally induced cancers
may be caused. We shall investigate what the protein products of the viral and cellular oncogenes do in the infected cell and
in cells in which cellular proto-oncogenes are expressed. We shall see that their functions strongly suggest mechanisms by
which cells may be transformed to a neoplastic phenotype. The discovery of cellular oncogenes led to the discovery of
another class of cellular genes, the tumor repressor (suppressor) genes or anti-oncogenes.

Initially, the involvement of viral and cellular oncogenes in tumors caused by retroviruses was much more apparent than the
involvement of the DNA tumor virus oncogenes but the discovery of tumor repressor genes (as a result of our knowledge of
how retroviruses cause cancer) led to the elucidation of the mode of action of DNA virus oncogenes.

It should be noted that while retroviruses have been instrumental in elucidation of the mechanisms of oncogenesis, most
human cancers are probably not the result of a retroviral infection although retroviruses are important in cancers in some
animals. It is becoming much more apparent that many human tumors may result from infection by DNA tumor virsues.

DNA TUMOR VIRUSES

DNA tumor viruses have two life-styles:

In permissive cells, all parts of the viral genome are expressed. This leads to viral replication, cell lysis and cell death

In cells that are non-permissive for replication, viral DNA is usually, but not always, integrated into the cell chromosomes at
random sites. Only part of the viral genome is expressed. This is the early, control functions (e.g. T antigens) of the virus. Viral
structural proteins are not made and no progeny virus is released.

DNA TUMOR VIRUSES INVOLVED IN HUMAN CANCERS

The first DNA tumor viruses to be discovered were rabbit fibroma virus and Shope papilloma virus, both
discovered by Richard Shope in the 1930s. Papillomas are benign growths, such as warts, of epithelial cells. They
were discovered by making a filtered extract of a tumor from a wild rabbit and injecting the filtrate into another
rabbit in which a benign papilloma grew. However, when the filtrate was injected into a domestic rabbit, the
result was a carcinoma, that is a malignant growth. A seminal observation was that it was no longer possible to
isolate infections virus from the malignant growth. This was because the virus had become integrated into the
chromosomes of the malignant cells.

 SMALL DNA TUMOR VIRUSES

 FAMILY: PAPOVAVIRIDAE - PAPOVAVIRUSES

The papovaviridae are small non-enveloped icosahedral DNA viruses. The major capsid protein, VP1, is present
as 72 pentamers. Each pentamer is associated with one molecule of another minor capsid protein, either VP2 or
VP3. The DNA is complexed with histone proteins encoded by the host cell.

PAPILLOMAVIRUSES

Papilloma viruses have a genome size about 8 kilobases. They cause warts and also human and animal cancers. Warts are
usually benign but can convert to malignant carcinomas. This occurs in patients with epidermodysplasia verruciformis.

Epidermodysplasia verruciformis is also known as Lewandowsky-Lutz dysplasia or Lutz-Lewandowsky

epidermodysplasia verruciformis and is very rare. It is an autosomal recessive mutation that leads to abnormal,
uncontrolled papilloma virus replication. This results in the growth of scaly macules and papules on many parts of
the body but especially on the hands and feet. Epidermodysplasia verruciformis, which is associated with a high risk
of skin carcinoma, is typically associated with HPV types 5 and 8 (but other types may also be involved). These
infect most people (up to 80% of the population) and are usually asymptomatic.

Papilloma viruses are also found associated with human penile, uterine, cervical and anal carcinomas and are very likely to be
their cause; moreover, genital warts can convert to carcinomas.

Squamous cell carcinomas of larynx, esophagus and lung  appear very like cervical carcinoma histologically and these may
also involve papilloma viruses. Recently, a strong causal link between certain oral-pharyngeal cancers and HPV16 has
been demontsrated.

There are more than 100 types of human papilloma viruses but, clearly, not all are associated with cancers; however,
papillomas may cause 16% of female cancers worldwide and 10% of all cancers.

Vulvar, penile and cervical cancers associated with type 16 and type 18 papilloma viruses (and others) but the most common
genital human papilloma viruses (HPV) are types 6 and 11. As might be expected if they are indeed the causes of certain
cancers, types 16 and 18 cause transformation of human keratinocytes. In a German study, it was shown that 1 in 30 HPV
type16-infected women will develop malignant disease while 1 in 500 infected people develop penile or vulvar cancer. Since
not all infected persons develop a cancer, there are probably co-factors in stimulating the disease. Such co-factors have been
identified in alimentary tract carcinomas in cattle where a diet containing bracken fern is associated with the disease. People
with HIV infection or AIDS are at increased risk of HPV-associated cancers as are patients with other forms of
immunosuppression.

The fact that a virus is usually found in association with a disease (often, in the case of tumors, the presence of a
copy of the viral genome in the neoplastic cells) does not prove that the virus caused the cancer. The association
could be casual rather that causal. Nevertheless, in many instances the epidemiological data are very strong and, in the
case of human cervical cancer, the efficacy of the anti-HPV vaccine makes the contention that HPV does cause cervical
cancer very compelling.

POLYOMA VIRUSES

These are small viruses with genomes of about 5 kilobases

Murine Polyoma virus

Polyoma virus was so named because it causes a wide range of tumors in a number of animal species at many different sites.
It was originally isolated from AK mice and is fully permissive for replication in mouse cells. It  causes leukemias in mice and
hamsters.

Simian virus 40
SV40 virus was initially discovered in the rhesus monkey kidney cells that were used to make inactivated Salk polio vaccine
virus. It was found that when the inactivated polio virus made in these cells was added to African Green Money Kidney cells,
the vaccine gave a cytopathic effect indicative of the presence of a live virus that had not been killed by the formalin used
to inactivate the vaccine virus. SV40 replicates in rhesus monkey kidney cells but has no cytopathic effect on them. Many
early recipients of the Salk polio vaccine received contaminating SV40 since anti-SV40 antibodies (against a protein called the
large tumor antigen (T-antigen)) could be detected in their blood. No elevated incidence of cancer has been found in these
people.

Although SV40 is a monkey virus that has no apparent effect on the host animal, it causes sarcomas when injected into
juvenile hamsters. The hamster tumor cells produce no infective virus.

Human polyoma viruses

The first two human polyoma isolates, known as BK and JC were discovered in 1971. Neither came from a tumor. BK came
from the urine of a kidney transplant patient and JC came from the brain of a Hodgkin's lymphoma patient who progressed to
progressive multifocal leukoencephalopathy (PML); however, they cause tumors when injected into animals. 70 to 80% of the
human population is seropositive for JC. This virus is known to be the cause of PML (see slow viral diseases), a disease
associated with immunosuppression. In 1979, the rate of occurrence of this disease was 1.5 per 10 million population. It has
become much more common because of AIDS and is seen in 5% of AIDS patients. BK virus is an important cause of
nephropathy and graft failure in immuno-suppressed renal transplant recipients and almost everyone in western countries has
anti-BK virus antibodies by the age of 10. Recently, BK viral DNA has been associated with human prostate cancer.

Three other human polyoma viruses have recently been described: KI, WU and  Merkel cell polyoma virus. The
latter virus causes a rare skin cancer (Merkel cell carcinoma, see box below)

Polyoma viruses are usually lytic and when transformation occurs, it is because the transforming virus is defective. After
integration into host DNA, only early functions are transcribed into mRNA and expressed as a protein product. These are the
tumor antigens. Because the expression of the genes for tumor antigens is essential for transformation of the cells, they may
be classified as oncogenes.

DEFINITION OF AN ONCOGENE: AN ONCOGENE IS A GENE THAT CODES FOR A PROTEIN THAT POTENTIALLY CAN
TRANSFORM A NORMAL CELL INTO A MALIGNANT CELL. IT MAY BE TRANSMITTED BY A VIRUS IN WHICH CASE WE
REFER TO IT AS A VIRAL ONCOGENE.

FAMILY: ADENOVIRIDAE

ADENOVIRUSES

These viruses are somewhat larger than polyoma and papilloma viruses with a genome size of about 35 kilobases. They were
originally isolated from human tonsils and adenoids, are highly oncogenic in animals and only a portion of the virus is
integrated into the  host genome. This portion codes several T antigens that carry out early functions. Tumor-bearing animals
make antibodies against the T antigens.

No humans cancers have been unequivocally associated with adenoviruses.

  TUMOR ANTIGENS ARE ONCOGENES

Tumors caused by papilloma virus, adenovirus or polyoma virus contain viral DNA but do not produce infectious
virus. The presence of the virus, however, elicits the formation of antibodies against the tumor antigens. In the
case of adenoviruses, only part of the viral genome is found in the host cell chromosomes whereas SV40 may
integrate part or all of its genome. Whether or not the whole SV40 genome is integrated, only a part of the
genome is transcribed into mRNA and protein and this is the region that encodes the early functions of the virus
replication cycle.

Many DNA viruses have early and late functions. Early functions are the result of the expression of proteins that
prime the cell for virus production and are involved in viral DNA replication. These proteins are expressed before
genome replication and do not usually end up in the mature virus particle. Late functions are the results of the
expression of viral structural proteins that combine to form the mature virus. They are expressed during and after
the process of DNA replication. Since early functions are involved in the replication of the viral genome, it is not
surprising that they can also alter the replication of host cell DNA.

SV40 expresses two such proteins, the T antigens (large T and small T antigen). The large T antigen acts as a cis-
regulatory element at the level of viral DNA replication by binding to the origin of replication and stimulating
transcription. It can also bind to and modulate the activity of host cell DNA polymerase alpha.

As we shall see later, DNA replication in the cell is controlled by suppressor proteins (the best studied of which
are the retinoblastoma (Rb) and p53 suppressor proteins). SV40 large T antigen can bind directly to these proteins
and inactivate them, thereby inducing the cell to go from Go to S phase. Because polyoma viruses have a small
genome, they rely on many cell functions for DNA replication and it is important that the virus causes the cell to
enter S phase because it creates a suitable environment for viral DNA replication.

Thus SV40 Large T antigen: 

 is necessary for transformation of a cell to the cancerous state

 stimulates the host cell to replicate its DNA
 is found mostly in the nucleus (to which it is directed by its nuclear localization signal) but a
small amount goes to the cell surface (where it is a tumor-specific transplantation antigen)
 binds to cellular DNA
 binds to p53 protein (see below)

A second T antigen (small T antigen) interacts with a family of cellular phosphatases (called pp2A) which results
in the failure of certain cellular proteins to be phosphorylated, thereby relieving cell cycle arrest.

In mouse polyoma virus, there is also a middle T antigen which can act as an oncogene.

Similarly, in adenovirus-induced tumors, only a part of the viral genome becomes integrated and again it is the
early region genes. This region codes for the E1A and E1B proteins. In papilloma virus-induced tumor, again, two
early genes, E6 and E7, are expressed.

Thus, papilloma, polyoma and adenoviruses seem to cause cell transformation in a similar manner: the integration of early
function genes into the chromosome and the expression of these DNA synthesis-controlling genes without the production of
viral structural proteins.  As we shall see later, all three virus types induce cell proliferation by interacting with tumor
suppressor genes.

Two important points that should be emphasized about T antigens of DNA tumor viruses as oncogenes:

 They are true viral genes. There are no cellular homologues in the uninfected cell
 They are necessary in lytic infections because they participate in the control of viral and cellular DNA
transcription

These properties should be contrasted  with retroviral oncogenes to be discussed later

COMPLEX TUMOR VIRUSES

FAMILY: HERPESVIRIDAE

HERPESVIRUSES

Herpesviruses are much larger than the DNA viruses described above and have a genome size of 100 to 200
kilobases. Because of their large size, a lot remains to be discovered concerning the way in which these viruses
transform cells.

There is considerable circumstantial evidence that implicates these large enveloped viruses in human cancers and they are
highly tumorigenic in animals. The herpes virus genome integrates into the host cell at specific sites and may cause
chromosomal breakage or other damage (see below). Herpesviruses are often co-carcinogens. They may have a hit and run
mechanism of oncogenesis, perhaps by expressing proteins early in infection that lead to chromosomal breakage or other
damage.
Herpesviruses have over 100 genes. When these viruses infect cells which are non-permissive for virus
production but which are transformed, only a subset (about 9) of viral genes are expressed. These genes code of
nuclear antigens or membrane proteins. Not all nine transformation-associated genes are expressed in all herpes-
transformed cells.
 

Epstein-Barr virus (Human herpes virus 4)

EBV is the herpes virus that is most strongly associated with cancer.  It infects primarily lymphocytes and epithelial cells. In
lymphocytes, the infection is usually non-productive, while virus is shed (productive infection) from infected epithelial cells.

EBV is causally associated with:     

 Burkitt's lymphoma in the tropics, where it is more common in malaria-endemic regions

 Nasopharyngeal cancer, particularly in China and SE Asia, where certain diets may act as co-
carcinogens
 B cell lymphomas in immune suppressed individuals (such as in organ transplantation or HIV)
 Hodgkin's lymphoma in which it has been detected in a high percentage of cases (about 40% of
affected patients)
 X-linked lymphoproliferative Disease (Duncan's syndrome)

EBV can cause lymphoma in Marmosets and transform human B lymphocytes i n vitro. 

EBV also causes infectious mononucleosis, otherwise known as glandular fever. This is a self-resolving infection of B-
lymphocytes which proliferate benignly. Often infection goes unnoticed (it is sub-clinical) and about half of the population in
western countries has been infected by the time they reach 20 years of age. Why this virus causes a benign disease in some
populations but malignant disease in others is unknown.

Human Herpes Virus 8 (HHV-8, Kaposi's Sarcoma Herpes Virus)

HHV-8 infects lymphocytes and epithelial/endothelial cells and is the causative agent of Kaposi's sarcoma. It has
also been associated with hematologic malignancies, including primary effusion lymphoma, multicentric
Castleman's (also Castelman's) disease (MCD), MCD-related immunoblastic/plasmablastic lymphoma and
various atypical lymphoproliferative disorders.

EBV and HHV-8 have been found to be associated with oral lesions and neoplasms in HIV-infected patients.
Among these diseases is oral hairy leukoplakia (OHL) which is benign and causes white thickenings on the
tongue epithelium in which these viruses proliferate.  

Human cytomegalovirus (Human Herpes Virus 5)

This herpes virus is frequently associated with Kaposi's sarcoma but  this disease is now thought probably to be caused by
human herpes virus 8.

For more on herpes viruses and the diseases that they cause, go to Virology Chapter 11 Herpes Viruses

FAMILY: HEPADNAVIRIDAE

HEPATITIS B VIRUS 

Hepatitis B virus is very different from the other DNA tumor viruses. Indeed, even though it is a DNA virus, it is much more
similar to the oncornaviruses (RNA tumor viruses) in its mode of replication. The DNA is transcribed into RNA not only for the
manufacture of viral proteins but for genome replication. Genomic RNA is transcribed back into genomic DNA. This is called
reverse transcription. The latter is not typical of most DNA tumor viruses but reverse transcription is a very important factor in
the life cycles of RNA-tumor viruses. See below.

For more information on the molecular biology of hepatitis B virus and the diseases it causes, go to chapter 18 and chapter
19, part 2.

Hepatitis B is a vast public health problem and  hepatocellular carcinoma (HCC),  which is one of world's most common
cancers, may well be caused by HBV. There is a very strong correlation between HBsAg (hepatitis B virus surface antigen)
chronic carriers and the  incidence of HCC. In  Taiwan, it has been shown that HBsAg carriers have  a risk of HCC that is 217
times that of a non-carrier. 51% of deaths of HBsAg carriers are caused by liver cirrhosis or HCC compared to 2% of the
general population.

RNA TUMOR VIRUSES (RETROVIRUSES)

Retroviruses are different from DNA tumor viruses in that their genome is RNA but they are similar to many DNA tumor
viruses in that the genome is integrated into host genome.

Since RNA makes up the genome of the mature virus particle, it must be copied to DNA prior to integration into the host cell
chromosome. This life style goes against the central dogma of molecular biology in which that DNA is copied into RNA.

 Retrovirus structure

The outer envelope comes from the host cell plasma membrane

Coat proteins (surface antigens) are encoded by env (envelope) gene and are glycosylated. One primary gene product is
made but this is cleaved so that there are more than one surface glycoprotein in the mature virus (cleavage is by host enzyme
in the Golgi apparatus). The primary protein (before cleavage) is made on ribosomes attached to the endoplasmic reticulum
and is a transmembrane (type 1) protein.

Inside the  membrane is an icosahedral capsid containing proteins encoded by the gag gene (Group- specific AntiGen). Gag-
encoded proteins also coat the genomic RNA. Again there is one primary gene product. This is cleaved by a virally-encoded
protease (from the pol gene)

There are two molecules of genomic RNA per virus particle with a 5' cap and a 3' poly A sequence. Thus, the virus is diploid.
The RNA is plus sense (same sense as mRNA).

About 10 copies of reverse transcriptase are present within the mature virus, these are encoded  by the pol gene.

Pol gene codes for several functions (again, as with gag and env, a polyprotein is made that is then cut up)

The pol gene products are:

a) Reverse transcriptase (a polymerase that copies RNA to DNA)

b) Integrase (integrates the viral genome into the host genome)

c) RNase H  (cleaves the RNA as the DNA is transcribed so that reverse transcriptase can make the second complementary
strand of DNA)

d) Protease (cleaves the polyproteins translated from mRNAs from the gag gene and the pol gene itself). This virally encoded
protease is the target of a new generation of anti-viral drugs.
GROUPS OF RETROVIRUSES

ONCOVIRINAE

These are the  tumor viruses and those with similar morphology. The first member of this group to be discovered was Rous
Sarcoma Virus (RSV) - which causes a slow neoplasm in chickens.

Viruses in this group that cause tumors in humans are: 

HTLV-1 (human T-cell lymphotropic virus-1) which causes Adult T-cell Leukemia (Sezary T-cell Leukemia). This disease is
found in some Japanese islands, the Caribbean, Latin America  and  Africa.  HTLV-1 is sexually transmitted 

HTLV-2 (human T-cell lymphotropic virus-2) which causes Hairy Cell Leukemia. The virus is endemic to very specific regions of the
Americas, particularly in native American populations.

LENTIVIRINAE

These have a long latent period of infection before disease occurs; they were mainly associated with diseases of ungulates
(e.g. visna virus) but HIV (formerly HTLV-III) which causes AIDS belongs to this group. It is much more closely related to
some Lentivirinae than it is to HTLV-I and HTLV-II which are Oncovirinae.

SPUMAVIRINAE

There is no evidence of pathological effects of these viruses. They establish persistent infections in many animal
species. They have been isolated from primates (including humans), cattle, cats, hamsters, and sea lions. Cells
infected by spumaviruses have a foamy appearance (because of numerous vacuoles) and often form syncytia of
giant multinucleate cells. Chimpanzee (simian) foamy virus is the type species. Human foamy virus is a variant of
simian foamy virus and is usually acquired from a monkey bite.

NFECTION AND TRANSFORMATION OF A CELL BY A RETROVIRUS

The following stages occur in the infection process:

1) Binding to a specific cell surface receptor

2) Uptake by endocytosis or by direct fusion to the plasma membrane. The virus may require entry into a low pH endosome 
before fusion can occur, although some (e.g. HIV) can fuse directly with the plasma membrane 

3) RNA (plus sense) is copied by reverse transcriptase to minus sense DNA. Here, the polymerase is acting as an RNA-
dependent DNA polymerase. Since reverse transcriptase is a DNA polymerase, it needs a primer. This is a tRNA that is
incorporated into the virus particle from the previous host cell.

4) RNA is displaced and degraded by a virus-encoded RNase H activity. Reverse transcriptase now acts as a DNA-dependent
DNA polymerase and copies the new DNA into a double strand DNA. This DNA form of the virus is known as a the provirus.

5) Double strand DNA is circularized and integrated into host cell DNA (see below) using a virally encoded integrase enzyme.
This DNA is copied every time cellular DNA is copied. Thus, at this stage the provirus is just like a normal cellular gene.

6) Full length, genomic RNA (plus sense) is copied from the integrated DNA by host RNA polymerase II which normally copies
a gene to mRNA. The genomic RNA is capped and poly adenylated, just as an mRNA would be.
Since the full length genomic RNA is the same sense as message, it also acts as the mRNA for GAG and POL polyproteins.

The genomic RNA is spliced by host nuclear enzymes to give mRNA for other proteins such as ENV. The RNA of some more
complex retroviruses such as HTLV-1 and HIV undergoes multiple splicing (see chapter 7, HIV).

Note that mRNA comes from splicing genomic RNA or is the genomic RNA. As a result, both mRNA and genomic RNA must
be the same sense - since mRNA must be plus sense, the genomic RNA of all retroviruses must also be plus sense.

An advantage of this mode of replication is that it allows growth in terminally differentiated cells since the only host cell
polymerase usurped by the virus is RNA polymerase II which is present in all cells.

MECHANISM OF VIRAL GENOME REPLICATION

If host RNA polymerase II is used to copy the DNA back to RNA, there are major problems with having a DNA provirus form
but an RNA genome in the mature virus particle

These problems include: 

1) RNA polymerase II does not copy the upstream and down stream control sequences of genes. It only copies the
information necessary to make a protein 

2) The lack of proof reading by RNA polymerase II

 Failure of RNA polymerase II to copy the entire gene

The problem is that, when transcribing genes, RNA polymerase II needs control and recognition sites upstream from the
transcription initiation site. The upstream site at which the polymerase molecule binds is called the PROMOTOR. Promotors
are not themselves copied into mRNA since they have no function in the translation of protein. After binding to the promotor,
the polymerase begins transcription at a downstream site, the RNA initiation site. The polymerase continues to transcribe
DNA into RNA until it reaches a termination/polyadenylation signal, part of which is not copied since,agaun, it also has no
function in the making of the protein. Furthermore, both up and downstream from the transcribed region are control
sequences that modulate the transcription of the gene. These are called ENHANCERS. These are essential parts of any gene
and must be present for RNA polymerase II to work but they are not copied to RNA. This is because RNA polymerase II in the
host cell has the function of making messenger RNA which is dispensed with after translation. To make a protein, the actual
mRNA molecule does not need the control sequences of the original gene. Thus, the use of host RNA polymerase II means
that the control sequences in the original genome should not get into the RNA genome of progeny virions.

This means that either the DNA copy of the viral RNA genome virus must integrate into host DNA downstream from a host
promotor and upstream from host termination sites (a tall order indeed!) or it must find a way of providing its own control
sequences (which, as we said, are not copied into progeny genome). It does the latter in a most complex manner.
A   The structure of the RNA genome of the mature retrovirus
How can a retrovirus provide its own control promotors and enhancers if they are not
transcribed when the DNA provirus is copied to the genomic RNA form?

Here is a brief (and very incomplete) summary of how a retrovirus does it:

1) The viral RNA is composed of three regions. At each end are repeats (called, not surprisingly, terminal repeats). The repeat
sequences (R) (shown in green in figure 15) do not code for proteins. In between the two repeats, there is a unique (not
repeated) region that contains the viral genes that code for the proteins (GAG, POL and ENV) plus other unique sequences at
either end that do not code for protein. Near the 5' end of the RNA genome is the U5 region and near the 3' end is the U3
region. PBS (in figure 15) is the primer binding site. The tRNA binds here when reverse transcriptase starts copying the RNA.
PPT is a polypurine tract.
 
2) In the integrated form (when transcribed into DNA and inserted into the host cell chromosome), the provirus is more
complicated. We find that part of the 3' unique region (called U3) of the RNA genome has been copied and transposed to the
opposite end of the genome. Conversely, part of the 5' end of the unique region (called U5) has been copied and transposed
to the other end. This gives the integrated DNA the structure shown in figure 15B. For convenience, only one strand of the
DNA is shown.

Now, of course, there are larger terminal repeats since the U3 and U5 regions are also repeated. The U3-R-U5 regions are
known as long terminal repeats or LTRs. The U3 region contains all of the promotor information that is necessary to start RNA
transcription at the beginning of the R (repeat) region while the U5 region contains all of the information necessary to
terminate after the other R repeat. In addition, the LTRs contain information that enhances the degree of transcription of the
three retroviral genes (enhancer regions). These enhancers can be up or downstream from the protein-encoding part of  the 
genes.

Host RNA polymerase II copies the provirus DNA to genomic RNA which can be also spliced to mRNAs. Since the
polymerase starts after the promotor (in U3), at the transcription initiation site, it begins exactly at the beginning of the R
region (figure 16). Thus we get a faithful (almost - see below) copy of the RNA that entered the cell. The termination
sequences and poly A signal are in U5 which is also not copied.

Because of this mechanism, there can be only one promotor site (from U3) for all three viral genes so they must be all
transcribed together. Splicing enzymes, from the host cell nuclear splicing machinery, cut the primary transcript to form the
individual mRNAs where necessary. (See chapter 7, HIV in which this has been well elucidated ). Unlike the situation with
DNA tumor viruses, there is no distinction between early/late functions.

You may ask why, if U5 contains the termination and polyadenylation sites, does the transcript not just terminate at the end of
the first R region of the LTR (figure 15b) and never get into the structural genes. The termination site in the first U5 is
suppressed, often by complex secondary structure mechanism. In some retroviruses there is a sequence in the gag gene that
provides the context to suppress the termination activity of the first U5. Clearly the second U5 does not have a gag gene
following it.

This strategy of virus replication in which viral RNA is first copied to DNA (by reverse transcriptase) which then gives rise to
mRNA and protein poses another problem for the virus. The initial step (RNA to DNA) is carried out by a viral enzyme which is
not normally in the cell. Yet this transcription step must occur before any mRNA transcription or protein translation can occur.
The problem is solved by the virus carrying about 10 copies of the reverse transcriptase protein into the cell with it. These
were packaged when the virus was assembled in the previous host cell. In theory, the viral genomic RNA coming into the cell
could act as an mRNA but it is too coated with protein to do so. Thus, new mRNA must be made which requires the synthesis
of reverse transcriptase.

ONCOGENES IN RETROVIRUSES 
  The structure shown in figure 15A and the upper part of figure 17 is that of a typical retrovirus with three structural genes
(gag, pol and env) but none of these is oncogenic. If the virus is to transform a cell, it must have sequences that alter cellular
DNA synthesis and provide the other functions that are typical of a transformed cell. These are in addition to the gag/pol/env
genome.

Thus we also find an ONCOGENE (onc) in the viral genome of many retroviruses that transform cells to neoplasia (figure 17).
It should be emphasized that the oncogene in RNA tumor viruses is not necessary for viral replication. It is an additional gene
that gives the virus its capacity to transform the host cell.

Definition of virally-induced transformation: The changes in the biologic function and antigenic specificity of a cell that result
from integration of viral genetic sequences into the cellular genome and that confer on the infected cell certain properties of
neoplasia. Note, however, that transformation can be induced by factors other than viruses e.g. carcinogens.
 
What are the oncogenic genes in retroviruses?

In retroviruses, these were first discovered as an extra gene in Rous sarcoma virus (RSV) (figure 17). This gene was called
src (for sarcoma). src is not needed for viral replication. It is an extra gene to those (gag/pol/env) necessary for the continued
reproduction of the virus. RSV has a complete gag/pol/env genome. Deletions/mutations in src abolish transformation and
tumor promotion but the virus is still capable of other functions. RSV is unusual in that it has managed to retain its whole
genome of gag/pol/env. 

  In sharp contrast to RSV, many retroviruses have lost part of their genome to accommodate an oncogene (figure 18). This
has two consequences:

1) The protein encoded by the oncogene is often part of a fusion protein with other virally-encoded amino acids attached

2) Virus is in trouble as it cannot make all of itself. To replicate and bud from the host cell needs products of another virus, that
is a helper virus.

About forty oncogenes have now been identified. Note that they are referred so by a three letter code (e.g. src, myc) often
reflecting the virus from which they were first isolated. Some viruses can have more than one oncogene (e.g. erbA, erbB).
Because they are viral oncogenes, we put a letter v in front of the three letter name. Here are a few of the most studied:

CELLS  HAVE PROTO-ONCOGENES

Once retroviral oncogenes had been discovered, a surprising observation was made: Unlike the situation with DNA virus
oncogenes which are true viral genes, there are homologs of all retrovirus oncogenes in cells that are not infected by a
retrovirus. These cellular homologs are often genes involved in growth control and development/differentiation (as might be
expected) and have important non-transforming functions in the cell; some can cause cancer under certain circumstances
and, presumably, those not shown to cause cancer have the ability to do so under the correct conditions. The cellular
homologs of viral oncogenes are called proto-oncogenes. To distinguish viral oncogenes from cellular proto-oncogenes, they
are often referred to as v-onc and c-onc respectively. Note that c-oncs are not identical to their corresponding v-oncs. It
appears that the virus has picked up a cellular growth controlling or differentiation  gene and, after the gene was acquired by
the virus, it has been subject to mutation.

Definition of a proto-oncogene: A host gene that is homologous to an oncogene that is found in a virus but which can induce
transformation only after being altered (such as mutation or a change of context such as coming under the control of a highly
active promotor). It usually encodes a protein that functions in DNA replication or growth control at some stage of the normal
development of the organism.
 

Characteristics of cellular proto-oncogenes

1) These are typical cellular genes with typical control sequences. As with most eucaryotic genes, most have introns (while
retroviral oncogenes - v-oncs - do not)

2) They show normal Mendelian inheritance because they are normal genes, essential to the functions of the cell.

3) As with all genes in the eukaryotic genome, they are always at same place in genome (cf. what would be expected of
endogenous retroviruses that had, over time, become incorporated into the cellular genome)

4) There are no LTR sequences (v-oncs always are in an LTR context)

5) Viral oncogenes are most like the c-onc of the animal from which the virus is thought to have acquired the gene. Thus, v-
src of RSV is more like chicken src than human src. Note that v-onc was long ago acquired accidentally by the virus from the
genome of a previous host cell

6) Cellular oncogenes are expressed by the cell at some period in the life of the cell, often when the cell is growing, replicating
and differentiating normally. They are usually proteins that are involved in growth control.

7) Cellular oncogenes are highly conserved

If v-onc and c-onc are so alike, why does the v-onc introduced by a virus cause havoc in the cell? This is due to differences in
the genes, mutations that have occurred in the gene once it was picked up by the virus. Such changes include:

 Amino acid substitutions or deletions which result in altered translation products

 Many v-onc proteins are fusion proteins translated from a v-onc that is a hybrid gene of a c-onc and a
viral gene.
 V-oncs are inserted into the host genome along with LTRs which contain promotors/enhancers. This is likely
to result in over expression of a gene that we know  is probably involved in control of DNA transcription and
replication!

Chronically transforming retroviruses do not have a v-onc

The observation that an acutely transforming virus such as RSV contains an extra gene, the oncogene, explains their high
neoplastic potential but, in contrast, chronically-transforming retroviruses only produce tumors slowly and they carry no gene
equivalent to a v-onc. At best, these viruses have just the three usual viral genes (gag/pol/env). An example is avian leukosis
virus (ALV) (figure 18).

How do chronically transforming viruses induce a tumor if they do not have an oncogene?

A seminal observation was made: Just as any other retrovirus does, ALV can integrate into the cell genome at many different
sites but, in ALV-induced tumors, the virus is ALWAYS found in a similar position (very important!). This means that the
crucial transforming event must be rare and that the cells that form the tumor are a clone (cf. the acute transformers which are
found all over the place). In all cases of ALV-induced tumors, the viral genome is inserted near a cellular gene called c-myc.
This is the cellular proto-oncogene that, in an altered form (i.e. as a v-onc), is carried by some acutely transforming
retroviruses (e.g. avian myelocytoma virus which causes carcinoma, sarcomas and leukemias). In addition, the level of
translation of c-myc in the ALV-transformed cell is much greater than in uninfected cells. Thus, inserting the genome of ALV or
other chronically transforming retroviruses next to a c-onc has the same effect as carrying in a v-onc.

So, during integration, the virus comes to lie upstream from c-myc which then comes under the influence of the strong LTR
promotors of the virus which leads to over expression of c-myc. This is called oncogenesis by promotor insertion (Figure 19).

But in some tumors the virus is downstream from the c-myc gene. However, we saw that LTRs also have enhancers in
addition to promotors. We know that enhancer sequences can be upstream or downstream to have their effect. This is called
oncogenesis by enhancer insertion (Figure 20).

Why is insertion near c-myc important? The protein coded for by this gene is found in the nucleus of normal cells and is
involved in control of DNA synthesis. It can be shown that over-expression of c-myc leads to rapid DNA replication.

Can cellular ongogenes be involved in non-virally induced cancer?

Once it had been shown that viruses can either bring an oncogene into the cell or can take control of a cellular proto-
oncogene to give rise to a tumor, the question arose of whether cellular proto-oncogenes could give rise to tumors in the
absence of retroviral infection. The answer is yes! Other chromosomal rearrangements can bring a c-onc under the control of
the wrong promotor/enhancer (Figure 21). Alternatively, the c-onc might be mutated in a particular way so that it was over-
expressed or it might code for a mutant protein with an altered function.

Chromosomal mapping allows the precise localization of the site of a gene on a particular chromosome and many cancers are
associated with alterations in chromosomes, particularly translocations (the breakage of a chromosome so that the two parts
associate with two parts of another chromosome). 

Many break sites in tumor cells are very close to a known c-onc. This is highly suggestive and unlikely to have occurred by
chance!

 Disease C-onc translocation

Burkitt's lymphoma * myc 8 to 14
Acute myeloblastic leukemia mos 8 to 21
Chronic myelogenous leukemia abl 9 to 22
Acute promyelocytic leukemia fes 15 to 17
Acute lymphocytic leukemia myb 6 deletion
Ovarian cancer myb 6 to 14

* In Burkitt's lymphoma the c-myc on chromosome 8 is brought to a site on chromosome 14 close to the gene for
immunoglobulin heavy chains. It seems that the proto-oncogene may thus be brought under the control of the immunoglobulin
promotor, which is presumably very active in B lymphocytes . This explains why this tumor arises in B cells. In other
lymphomas, a c-onc is brought next to the immunoglobulin light chain promotor. These are also B cell lymphomas.

Epstein-Barr virus is probably the cause of Burkitt's lymphoma. This is a herpes virus and herpes viruses commonly cause
chromosomal breaks. If such a break causes an 8:14 translocation, the cell's myc gene will become adjacent to the cell's
immunoglobulin promotor and c-myc expression will rise in the cells in which this promotor is active.

Is there evidence that mutations in cellular oncogenes might also result in transformation?

The best evidence comes from the cellular oncogene that is the homologue of the viral oncogene found in the Harvey strain of
murine sarcoma virus (the v-onc is called HaRas). This c-onc was isolated from bladder carcinomas and compared to the
normal c-onc proto-oncogene. In many tumor cells only one change was found in the amino acid sequence of the protein;
glycine at amino acid position 12 was changed to valine. At position 12 only glycine and proline gave normal growth. All other
amino acids at this position gave a transformed cell. In a lung carcinoma, the transforming DNA also contained c-HaRas,
again it had a point mutation, this time at position 61.

What is the normal function of oncogenes?

As mentioned above, c-oncs are normal cellular genes that are expressed and function at some stage of the life of the cell.
We should expect them to be involved in DNA synthesis or perhaps the signaling pathways that lead to proliferation. More
than 40 oncogenes have been identified and there are probably a few undiscovered ones.

We can sub-divide the cellular oncogenes into those that encode nuclear proteins and those that encode extra-nuclear
proteins. The latter are mostly associated with the plasma membrane of the cell (Figure 22 and 23).

Products of oncogenes that are nuclear proteins: e.g. myc, myb. These are involved in control of gene expression (that is the
regulation of transcription - they are transcription factors) or the control of DNA replication. Neoplasia is associated with
elevated transcription of the oncogene but strong expression is not always necessary, rather there is a need to make the gene
constitutively active rather than under control of normal regulatory processes.

Products of oncogenes that are cytoplasmic or membrane-associated proteins: e.g. abl, src, ras. This type does not exhibit
altered expression but seems to convert from proto-oncogene to oncogene by mutation. Thus, in src-induced tumors, strong
over expression of the oncogene has no effect.

FUNCTION OF PROTO-ONCOGENE- ENCODED

EXAMPLE
PROTEINS
Control of DNA transcription (found in nucleus) myc
Signaling of hormone/growth factor binding such as a
src is a membrane-bound tyr kinase.
tyrosine kinase
GTP-binding proteins involved in signal transduction
ras
from a surface receptor to the nucleus
sis is an altered form of platelet-derived growth factor
Growth factors
B chain
erb-B is a homolog of the epidermal growth factor
receptor (it is also a tyrosine kinase). fms is a homolog
Growth factor receptors
of the macrophage colony-stimulating factor (M-CSF)
receptor

ANTI-ONCOGENES (TUMOR SUPPRESSOR GENES)

The way in which retroviruses cause tumor formation via oncogenes was established before anything was known about how
DNA tumor viruses cause tumors. Certainly, DNA tumor viruses carry oncogenes (e.g. SV40 T-antigen) but how do these
proteins, encoded in true viral genes with no cellular homologs, cause the formation of tumors?

It has long been known that most tumors are the result of dominant mutations, i.e. a function is gained that makes the cell
grow when it should not  (Figure 24). For example, as noted above, if we have a receptor that sends a signal when it binds a
growth factor by switching on its tyrosine kinase activity and that receptor becomes mutated so that its tyrosine kinase activity
is permanently activated, the cell will get the aberrant growth signal even in the heterozygote. Thus the mutant allele is
dominant over the normal allele.

In each of these cases, the mutation is dominant. Thus, for example if one allele of erb-B (a homolog of the EGF receptor) is mutated so
that it is a constitutively switched on (i.e. does not need epidermal growth factor to bind to switch on the tyrosine kinase activity), then the
signal is on, regardless of the fact that the other allele is normal.

Retinoblastoma: A recessive tumor

There is a curious class of tumors that do not fit the usual characteristics in which the mutant oncogene is dominant over the
wild type.

In retinoblastoma, there appears to be a lesion that is recessive, that is the cancer causing mutation causes a loss of function 
(Figure 24). (This is recessive because, in a diploid organism, there are two genes. If one allele is mutated so that it does not
work, the other can still code for the normal protein and function is retained. In order to lose the function and have no protein
being made, both genes must be mutated, i.e. we have a recessive mutation) . Thus it appears that the protein that is encoded
in the retinoblastoma (Rb) gene is a growth suppressor. If a homozygous mutation occurs in the Rb gene, there will be no Rb
gene product at all and the cell will grow abnormally because the growth suppressor is no longer present. The product of the
Rb gene has been identified and shown to be a nucleus-located protein of 105 kDaltons.

A heterozygote at the Rb allele still has normal Rb and tumors can still be suppressed but homozygote has no functional Rb
and tumors cannot be suppressed

Above, we have noted that the adenovirus E1A (early function) protein is somehow involved in tumorigenesis. It has been found that E1A
protein in the transformed adenovirus-infected cell is complexed with a 105kD protein! This turns out to be the Rb gene product (Figure
25). Thus, it seems that adenovirus may cause a cell to grow abnormally by complexing (and thereby inactivating) a cellular protein
whose normal function is growth inhibition.

p53 and human cancer

Over the past two decades, since its discovery in 1979, a gene known as the p53 gene (after the size of its encoded protein)
has been linked to many cancers including many that are inherited. In these inherited cancers, it turns out that the p53 gene is
mutated. Alterations in this protein seem to be the basis (direct or indirect) of most human cancers. In total, 60% of human
cancers involve p53. 80% of colon cancers involve the p53 gene

 Human cancers that involve p53

cervix liver
breast lung
bladder skin
prostate colon

Initially, it was thought that the p53 gene product caused cancers but further investigation showed the opposite; p53 is, like
the retinoblastoma gene product, a tumor suppressor. p53 protein has been referred to as The Guardian of the Genome since
it regulates multiple components of the DNA damage control system.

How does p53 work in a functional cell? Normally, there are only a few of the suppressor p53 molecules in a healthy cell and
these are constantly turning over; but when the DNA becomes damaged (perhaps by radiation or chemical mutagens) and
DNA replication results, p53 turnover ceases. The rise in p53 stops DNA replication.

p53 is a transcription factor. When it builds up, p53 binds to a specific site(s) on the chromosomes and switches on other
genes and these, in turn, shut down mitosis. p53 can also act in another way: When it builds up it can set the cell on course to
apoptosis. Whether or not p53 causes reversible growth arrest or apoptosis depends on the state of cellular activation; for
example, extensive, unrepaired DNA damage can lead to sustained p53 production committing the cell to apoptosis. In
inherited cancers, there is a mutation in the p53 gene; often it is a single point mutation and the protein can no longer bind to
its correct site on the DNA and so cannot suppress DNA replication.

Like the Rb gene, product, you would expect the effects of p53 to be recessive since the second normal p53 allele should
make functional protein and should shut off DNA replication as usual; however, if you are heterozygous for the mutation, you
are, of course, only one mutation away from carcinogenesis. So why do cells that are heterozygous for the p53 mutation also
have problems? Unfortunately, p53 protein forms tetramers in a ribbon-like array and so if half of the p53 proteins are mutant,
there is a good chance that each tetramer will have one mutant p53 molecule and this inactivates the tetramer, a dominant-
negative effect.

Although we have learned a lot from families that inherit p53 mutations, it is clear that most p53 mutations come from non-
inherited environmental factors: carcinogens (benzopyrene in smoke, aflatoxin in molds on peanuts and corn, UV light) that
result in point mutations. There are also gain of function p53 mutations that lead to very aggressive tumors. These turn on
DNA replication genes.

What has this got to do with DNA tumor viruses? Just as with retinoblastoma gene product, the presence of a virus can mimic
mutation and take the tumor suppressor out of action by complexing it in an inactive form that cannot bind to the specific site
on DNA. This is what appears to happen in hepatitis C which causes hepatocellular carcinoma. In the case of a human
papilloma virus-infected cell, p53 is bound by the E6 protein and directed to a protease that recognizes a cleavage site in p53,
thereby destroying it (Figure 26). In addition, E7 protein binds and inactivates Rb protein.

Much research is now going on to see whether one can introduce healthy p53 genes into cells to shut down tumor growth.

Thus, our knowledge of how retroviruses cause cancer has led to an understanding of the formerly cryptic manner in which
DNA tumor viruses do the same thing.