No amount of transient transfection assays can substitute for analysing reporter gene expression in transgenic mouse models. Only then, can all the developmental constraints responsible for driving tissue and stage-specific expression be brought to bear upon the transgene. Even then, the reporter gene is integrated randomly into the chromatin and expression is subject to site-specific variegation from flanking chromatin.
Few transgenic mouse studies have been reported for GPCRs. Typically, appropriate spatiotemporal expression of a transgene is only achieved when large genomic fragments are used to drive reporter gene expression. This is partly a reflection of the need for distal regulatory elements and partly the inability of short genomic sequences to insulate themselves from the influences of the surrounding chromatin—the cause of site-specific variegation. When it comes to specificity—size matters. However, there are some exceptions. Studies of the A1 adenosine receptor showed that a 500 bp proximal promoter drove widespread expression and was sufficient to drive expression in CNS and atria (Rivkees et al. 1999). Interestingly, this proximal promoter sequence contains motifs for the cardiac transcription factors, GATA and Nkx2.5. Furthermore, both recombinant GATA and Nkx2.4 synergistically activated transcription from the A1 promoter, but the in vivo relevance of these sites was not tested in these transgenic experiments. If they recapitulate results obtained from transient reporter gene assay, then mutation of the Nkx2.4 and GATA sites should ablate cardiac expression but leave CNS expression unaffected.
When the native gene exhibits a widespread expression pattern like the A1 receptor, then it can be difficult to tell the difference between deregulated expression and correct expression. This distinction is clearer in the case of a gene that displays a very discrete pattern of expression such as the follicle-stimulating hormone (FSH) receptor that is limited to granulosa cells of the ovary and Sertoli cells of the testis. Here, over 5 kb of 5'flanking sequence was insufficient to drive reporter gene expression in Sertoli cells (Heckert etal. 2000). Transgenic mouse studies can also reveal differences between regulatory elements required to specify embryonic development and those required to maintain adult expression patterns. Hu et al. (Hu et al. 1999) showed that reporter genes containing 3 kb of 5'flanking sequence of the mouse ^ opioid receptor gene promoter was sufficient to recapitulate embryonic expression of the endogenous gene in the forebrain, midbrain, and hindbrain. However, this same construct was insufficient to drive peripheral expression in the adult. This distinction between regulatory regions required for CNS expression and peripheral expression is also seen with the neuropeptide Y (NPY) Y1 receptor gene. In this case, a reporter gene containing 1.3 kb of 5'flanking sequence was sufficient to drive appropriate spatiotemporal expression in the embryonic and adult CNS but was not capable of driving expression in heart, liver, and kidney (Oberto et al. 1998). Even over 9kb of the mGluR6 was not sufficient to repress ectopic transcription in the brain although developmentally regulated expression was seen in the retinal rod and ON-type cone bipolar cells, the major sites of expression of the endogenous mGluR1 gene (Ueda et al. 1997). In another study, 2.6 kb of 5'flanking sequence of the VIP1 receptor (VPAC1) was found to be sufficient to direct expression of a reporter gene to CNS and lung (Karacay etal. 2000). All of these studies have, to a greater or lesser extent, shown that discrete genomic fragments are capable of driving appropriate spatiotemporal expression of a reporter gene, with a fair degree of veracity. As yet, there is no report of a finer analysis that has identified specific regulatory domains that are required or sufficient to drive these expression patterns in vivo. Until this is achieved, the nature of the transcription factors that are responsible for these patterns of expression will remain elusive.
It seems unlikely that tissue- and stage-specific transcription of GPCRs, as a gene family, will be regulated in a co-ordinated manner. With a few notable exceptions (see below), there is no cis-linkage among GPCR genes and their tissue expression profiles and developmental activation are completely disparate. It seems equally likely, that, their expression will be controlled by disparate regulatory elements each operative in different tissues or at different times. This latter point appears to be borne out by the studies of Oberto et al. (Oberto et al. 1998) and Hu et al. (Hu et al. 1999) and probably reflects the fact that most GPCR expression patterns do not follow singular cell lineages. Genes expressed only in cells derived from a singular lineage might be expected to exhibit co-ordinate regulation since they are derived from the same progenitor population and consequently they have all been exposed to at least some of the same transcription factors during the cells' development. An example of such co-ordinate regulation can be seen in the noradrenergic cells of the PNS all of which are derived from the neural crest. Whereas, many of the pan-neuronal properties of sympathetic precursors are under the control of the MASH-1 bHLH protein (Sommer et al. 1995), expression of tyrosine hydroxylase and dopamine ^-hydroxylase are under the control (in part) of the Phox2a and Phox2b homeobox genes (Lo et al. 1997; Pattyn et al. 2000; Yang etal. 1998a). In this way, the transmitter phenotype (in this case noradrenaline) is intimately part of the sympathoadrenal lineage and it is not surprising to see that expression of its synthetic enzymes are co-ordinately regulated at least in part, by a singular transcription factor, Phox2a. If a GPCR shared this unique expression pattern then it too, might be co-ordinately controlled by Phox2a. However, a gene such as the M1 muscarinic receptor gene, which is also expressed on all sympathetic neurons (Brown et al. 1995), is additionally expressed by many telencephalic regions (Buckley et al. 1988) where Phox2a and Phox2b are not expressed. Accordingly, M1 expression is likely controlled by different mechanisms in these two neuronal areas. Even the example given here is a gross oversimplification since there is no known metazoan gene whose expression is exquisitely regulated by only one transcription factor. Nevertheless, it is probably true that genes which are expressed across many lineages, such as is the case for many GPCRs, will probably not necessarily be co-ordinately regulated. The predicted consequence of this is that transgenic mouse studies will not reveal individual cis elements that are required or necessary for expression in all regions where the endogenous gene is transcribed. This concept will only be tested when sufficient GPCR genes are characterised in transgenic mice. Exceptions would be the obverse—those GPCRs whose expression is highly restricted to a particular tissue. Probably the most spatially restricted GPCRs are represented by members of the opsin gene family, expression of which are restricted to the photoreceptors of the retina. Interestingly, as little as 1.1 kb of 5'flanking sequence of the human blue opsin promoter is sufficient to restrict expression to S-cone type photoreceptors and bipolar cells (Chen et al. 1994). Similarly, a discrete 2.2 kb of 5'flanking sequence of the rhodopsin promoter was sufficient to drive expression in photoreceptor cells but expression was also seen in rods and to a lesser extent in cones (Gouras et al. 1994). Similar fragments derived from the mouse gene also drive photoreceptor expression (Lem et al. 1991)—again, a certain 'leakiness' in expression was seen with cones expressing low amounts of transgene. Earlier studies had shown that shorter constructs containing as little as 222 bp of flanking sequence of the bovine rhodopsin promoter were sufficient to confine reporter gene expression to photoreceptor cells but gave aberrant expression patterns across the retina (Zack etal. 1991). Cross species comparison of the rhodopsin promoter revealed a 102 bp domain of conserved sequence (Zack etal. 1991) approximately 1.5-2 kb (dependent upon species) upstream from the T.I.S. but deletion of this region did not result in loss of correct spatial expression, although deletion did lead to lower levels of transgene expression (Nie et al. 1996). Neither did this region confer retina-specific expression upon a hetero-logous promoter. Nevertheless, the possible linkage between tightly controlled expression and relatively discrete regulatory elements is seductive. Although this retina enhancer region shows homology to the locus control region (LCR) of the cone opsins, the lack of correlation between copy number and transgene expression levels clearly rules out a role of this region as a LCR in the rhodopsin gene. Although putative proximal elements and cognate transcrip-tionfactors, suchasNrl (Kumar etal. 1996) and the Crxl homeobox (Chen etal. 1997),have been assessed in vitro, their in vivo relevance is unkown. The case for their potential in vivo relevance is increased by the observed synergy of their activation of the rhodopsin promoter and the demonstration of their physical interaction (Mitton et al. 2000). Importance of this region is further underlined by the demonstration Crx also interacts with the architectural transcription factor HMG I(Y) (Chau et al. 2000) and mutations in Crx have been associated with several retinal degenerative diseases including Leber congenital amaurosis and adult onset cone-rod dystrophy and retinitis pigmentosa (Sohocki et al. 1998).
Was this article helpful?