The data presented in this thesis argues that different MVV isolates show modified
mechanisms of transcriptional regulation. This conclusion is based on the observation that the
EV-1 LTRs lack a consensus AP-1 site and that the positions of transcription factor sites are
not fixed between the EV-1, 1514 or 1772 viruses. Both 1514, and its derivative 1772,
contain a consensus AP-1 site and in the case of 1514 this sequence has been demonstrated
to bind AP-1. The sequence within the EV-1 LTRs with closest homology to the consensus
AP-1 sequence did not bind AP-1 in vitro and did not appear to do so in vivo. It was not
possible to conclude for certain that this sequence was non-functional in vivo due to the
presence of a second transcription factor site in this region of the LTR. This factor, which binds
the same sequence in the 1514 and EV-1 viruses, shows a transactivating activity. This activity
was lower than that seen for the AP-1 complex present within these cells. There appear to be
two sites for this factor within the MVV LTR, one proximal to the TATA box the other at an
upstream site. This factor appears to be distinct from AP-1 both on the basis of site recognition
and the failure to observe binding of an anti-Jun antibody in gel supershift experiments. UV
crosslinking experiments identified a protein of approximately 34-40kD which interacts with
the recognition site. A full characterisation of this factor, and other cellular factors with which it
interacts, will allow further dissection of the mechanism by which MVV controls its transcription
rate.
Various questions need to be addressed regarding the nature of the VSS binding
protein. The simplest involves determining the tissue distribution of this factor; is there any
evidence for tissue specific expression? A second related question is whether this factor is a
component of a multi member family; do the family members show tissue specific expression?
In relation to MVV the key question is the expression of this factor during
monocyte/macrophage differentiation and activation. The transcription factor AP-1 is activated
during this process and it is important to determine if the factor which binds the VSS is under
similar control.
An analysis of the expression of transcription factors involved in the regulation of
MVV during monocyte/macrophage differentiation may also explain the restriction/latency of
MVV in the majority of infected cells in vivo. In vitro the MVV LTR has been shown to be a
strong promoter which is active in a variety of different cell types. This is in marked contrast to
102
the observations of in vivo infection where viral transcription appears under tight control. The
basis of this in vivo restriction is unclear. It is possible that the VSS binding protein is involved
in this process, however, nothing is known about its expression in vivo. In addition to these
two factors, AP-1 and the VSS binding protein, the nature of other factors interacting with the
LTR remains to be defined. It is possible that neither of these two factors is critical for the
regulation of MVV transcription and some additional factor is required.
One important question raised in this thesis which needs to be fully resolved is the
apparent variation between the EV-1 and 1514 viruses in relation to AP-1 binding. Sequence
comparison suggested that none of the non-consensus AP-1 sites within the EV-1 LTR would
be able to bind AP-1 but this still requires formal proof. The EV-1 TATA box proximal AP-1 site
was shown to be unable to bind AP-1 in vitro. From the data using the gel shift
oligonucleotides in a CAT reporter vector it was not possible to completely rule out AP-1
binding to the EV-1 sequence in vivo due to the presence of the VSS binding site. However,
this data does indicate that even if AP-1 is binding to the EV-1 sequence it is binding with a
much lower affinity than in 1514.
In relation to other factors which interact with the LTR the nature of the factor
interacting with the sequence at position 90 remains to be defined. From the data on the EV-1
LTR variants this factor appears to be playing a role in determining the activity of duplications of
c/s-regulatory sequences in the LTR. The effect of this sequence illustrates the
codependence of the factors driving transcription from the LTR. As has been discussed
crosstalk can involve direct protein-protein interactions or occur indirectly through effects on
DNA conformation and the factors which are recruited to the basal transcription complex
assembled on the TATA box. One factor which may be affecting LTR activity is integration
state. To date studies on the MVV LTR have all made use of transient transfection assays
where the DNA is present in a non-integrated form. In contrast during infection the viral
genome will be integrated into the hosts DNA. It is possible that this integrated state may affect
viral transcription and play a role in latency. Viral latency in vivo is likely to be the product of
several interacting regulatory pathways. A comparison of the factors interacting with the LTR
during latency and active transcription will be required.
In Chapter 3 it was seen that alterations in the structure, and sequence, of the MVV
LTR modified its transcription rate. This data illustrated the interaction between various
sequences within the LTR and the variability of the LTR sequence within the EV-1 virus
103
population. Comparisons of the LTRs from the 3 viruses 1514, 1772 and EV-1 suggests that
alterations in LTR architecture are tolerated. These three viruses appear to have transcription
factors binding in distinct locations in the LTR. A question which remains to be answered is
whether such alterations in LTR structure alter the process of viral pathogenesis in vivo. The 3
strains do differ in the pathology they induce; EV-1 the British isolate causes predominatly
Maedi disease, in contrast the 1514 and 1772 viruses cause primarily Visna disease, with the
1772 virus being selected in vivo by serial passage for increased neuropathogenesis. Of
course, these viruses do not solely differ in their LTR sequences so the importance of this
variation to the distinct pathologies induced remains unclear. Comparing the 1514 and 1772
LTRs then it could be suggested that the separation of the VSS and the AP-1 site may result
in an elevated transcription, and replication rate, so accelerating the disease course. This
separation of these two binding sites also raises the question of synergistic interaction
between the two factors. It is still unclear whether both these factors interact simultaneoulsy at
the TATA box proximal site in 1514. This could be addressed using a A56 vector containing
the AP-1/VSS region but with mutations in the VSS region. The activity of this sequence
could then be compared with the 1514 and EV-1ex-v (AP-1 mutant) vectors. Such a
comparison should identify any interaction between AP-1 and the VSS binding protein which
modifies the transcription rate.
One method for determining the role of the LTR in determining the disease course
would be by constructing chimeric molecular clones differing only in the LTR sequence. Such
constructs could then be introduced directly into animals by DNA injection. The data on the
EV-1 LTR population within the infected animal and the comparison of these sequences with
those seen in the infecting population strongly suggests that the LTR is under selective
pressure in vivo. This selection appears to limit the number of LTR types capable of
establishing infection. Dissection of the interactions involved in the regulation of transcription
from these promoters will depend on targeted disruption of transcription factor binding sites
and the manipulation of sites by altering orientation and spacing. While these studies may be
performed using isolated LTRs and reporter gene assays, in order to shed light on the LTRs
role in disease, it will also be neccessary to perform such experiments using infectious
molecular clones. This will allow the effect of alterations of LTR structure on viral replication to
be monitored.
One aspect of the regulatory process not touched on in this work is the activity and
104
targets of the MVV Tat protein, it appears that the MVV Tat protein does possess a potent
transcriptional activation domain (Carruth et at., 1994). The absence of a TAR region in the
MVV viruses has led to the proposal that the Tat protein is interacting with cellular factors and is
activating transcription via this route. A second observation on the MVV Tat protein, that it only
weakly transactivates the MVV LTR, has been used to suggest that Tat may be primarily
targeting cellular genes so maintaining the cell in an activated state and disrupting normal
homeostasis. This mechanism of action is consistent with observations that MVV Tat is not
required for in vitro replication. It is possible that it is more important for the successful
maintenance of infection in vivo, and may be involved in the generation of pathology. This
remains to be tested by experimental infection with Tat deleted virus.
Due to the apparent requirement for integration in the lentiviral lifecycle these viruses
behave essentially as cellular genes. However they have the great advantage of containing all
their regulatory elements within a defined sequence of DNA which can be readily manipulated
in vitro. In contrast, cellular genes contain widely dispersed regulatory elements and these
elements cannot be easily manipulated in their normal context. Analysis of the molecular basis
of replication in MVV, and other lentiviruses, will shed light on the mechanisms by which these
pathogens maintain infection and evade the host immune response. Further, a fuller
understanding of the mechanisms involved in the control of lentiviral transcription is useful in
the dissection of cellular transcriptional control mechanisms.