staining. However, the mice from both lines also had clearly ectopic, extra-renal LacZ staining in tissues such as brain, cartilage and bone, possibly due to an integration site dependent influence over the weak podocyte specific elements contained in this fragment.
Whether or not suppressor elements exist in the region -2.1; -1.25 has not yet been determined. However, based on the results of the analysis of the nephB-∆700 transgenic mice, it could be concluded that this region was not important for podocyte expression of the transgene. The introduction of 700 bp deletion between -1.25 and -1.9 kb did not affect the expression of LacZ in podocytes and brain (Fig. 2C, Article II). Spinal cord and pancreas, on the other hand, lacked the expression of the transgene (Fig. 2C, Article II). Comparison of the results for nephB, nephC and nephB-∆700, showed that the region between -1.25 and -1.9 kb was required for expression in pancreas and spinal cord, while the sequence -4; -2.1 contained podocyte specific elements.
Very recently two independent studies showed that WT1 binding elements exist in close proximity to the basic promoter of the nephrin gene in both human and mouse (Guo et al., 2004; Wagner et al., 2004). Moreover, a 186 bp fragment of the human nephrin promoter, located 743 bp upstream of the ATG codon and later shown to contain WT1 binding sites, is capable of directing podocyte specific expression of LacZ marker gene when cloned in front of a minimal promoter (Guo et al., 2004). This short fragment is active when cloned in single or multiple copies and no neuronal expression is seen in any of the transgenic lines generated with it. It is worth mentioning that one of the two transgenic lines carrying a single copy of the 186 bp fragment showed mosaic LacZ expression - not all glomeruli were positively stained. In one of the two mouse lines generated using multiple copies of the fragment extra-renal staining in whisker follicles and epicardium was observed. This unspecific expression of the transgene probably results from integration site-dependent influence. The weaker enhancer activity in the first case and the extra-renal expression in the second case, are in agreement with our hypothesis that more than one element is needed for directing endogenous, nephrin-like expression. It is possible that the 186 bp WT1 responsive element needs to work in co-operation with the -4; -2.1 kb enhancer region to ensure the proper tissue specific expression of nephrin.
In an extensive, EMSA based screening, we managed to identify a short region (reg3.7) within the -4; -2.1 kb enhancer, which was recognized and bound by, as yet undetermined, protein factor(s) in nuclear extracts from a podocyte cell line. Similar results were observed for nuclear extracts from human embryonic kidney cells (HEK293), but not with nuclear proteins from macrophage-like cells. This indicated that HEK293 cells probably had part of
the transcription factors needed for initiation of transcription from the nephrin gene. The inability of the macrophage-like nuclear extracts to induce an EMSA shift indicated that the transcription factor in question was not widely expressed, or required additional cell specific factors.
The 20 bp reg3.7 was a part of a highly conserved stretch of DNA, located between exon 1B of nephrin and exon 1 of the newly identified gene – filtrin or kirrel2. A mutations screening of the region revealed the importance of six successive guanine nucleotides for the binding. An inhibition assay using chelating agents abolished the binding of the nuclear protein(s) to reg3.7, which indicated that the putative transcription factor(s) belong to the family of the zinc-coordinated DNA binding proteins. The zinc-finger family is the most abundant class of transcription factors - almost half of the predicted 2,000 human transcription factors are dependant on zinc ions for binding DNA (Urnov, 2002). As discussed earlier, many members of this family have been implicated in kidney development - WT1, RAR, Krueppel-like transcription factors, SP1, etc. Using different TRANSFAC® based prediction programs we tried to identify candidate factors that could be binding within or close by the six G stretch in reg3.7. As shown in Figure 7, among those were several zinc-finger proteins – SP1, WT1, retinoid X receptor α, vitamin D receptor, MZF1. Based on the data in the literature we considered two of these factors, namely WT1 and MZF1, good candidates. WT1 has been shown to play important role in podocyte development in general and in nephrin gene regulation in particular (Armstrong et al., 1993; Guo et al., 2002; Natoli et al., 2002a; Menke et al., 2003; Wagner et al., 2003; Guo et al., 2004; Wagner et al., 2004).
MZF1 is a zinc-finger protein previously shown to be involved in regulation of gene expression through alternative splicing (Nomoto et al., 1999). Using antibodies against the two candidate factors we showed that neither of them was likely to be binding reg3.7.
Since the prediction programs could not help us in identifying the putative transcription factor and because it is always possible that we were dealing with an unknown protein we tried to purify it by DNA affinity chromatography. However, our efforts have failed, so far.
The lack of success in these experiments is probably due to the low amounts of the transcription factor in nuclear extracts, which, together with the high background, makee the visualization of the protein on Coomassie stained gels very difficult.
It is interesting to note that reg3.7, which contained a podocyte specific transcription factor binding site, was not well conserved in the human sequence (Fig.1, Article III). This observation came as a surprise, since the work on region III was based on the homology between the two species. However, there are data showing that this is not always the case.
Recently Dermitzakis et al. have shown that between 32% and 40% of the human transcription binding sites are not functional in rodents due to the turnover of the promoter sequences in evolution (Dermitzakis and Clark, 2002). It is quite possible that the reg3.7 transcription factor binding site is located in another region in the human promoter. This human sequence remains to be identified.
An important question about the nature of the reg3.7 enhancer rises in connection with the recent discovery of kirrel2/filtrin. This novel gene is located upstream of nephrin and transcribed in opposite direction from a transcription start site located in close proximity to exon 1B. It codes for an immunoglobuline–like protein, with several splice isoforms and is expressed in kidney podocytes, pancreatic islet cells and nervous system (Ihalmo et al., 2003;
Sun et al., 2003). The similarity in the expression pattern of the two genes is striking. It is tempting to speculate that they both share some regulatory elements as it has been shown for other genes (Heikkila et al., 1993). It would be interesting to test if the reg3.7 enhancer can work in both directions and whether it is involved in the regulation of both genes.