In this paper, studies were done to show that an in vivo electroporation is an effective way to study promoters in the rodent epididymis, particularly in the initial segment, which is unique in its dependence on LTFs for normal gene expression and function. Control experiments using a plasmid encoding EGFP under the control of the CMV promoter showed expression targeted to the epithelial cells in the initial segment (Fig. 2B). These control experiments also showed that GGT promoter IV could target protein expression to the epithelial cells where this promoter is normally expressed regardless of the method of DNA injection (intraluminal vs. interstitial; Fig. 2, C and D). The in vivo electroporation method was then used to evaluate two previously characterized initial segment specific gene promoters: the cres promoter and GGT promoter IV.

Analysis of the cres promoter had been carried out previously in LTp2 gonadotroph cells, which normally express cres. The transcriptional activity of the 135-bp minimal promoter was dependent on two C/EBP DNA-binding sites found within this sequence. The same promoter was used in this study. Although the promoter is derived from the mouse sequence, the rat promoter is similar to the mouse, and the expression of cres is restricted to the rat initial segment as it is in the mouse (unpublished observations). In this study, only the 5′ C/EBP element in the cres promoter was necessary for promoter activity in vivo in the initial segment, suggesting that only the 5′ site may be necessary for the activity of this promoter in vivo (Fig. 3A). In addition, there may be differences in the transcriptional control of the cres promoter in the anterior pituitary compared to the initial segment that are revealed by these in vivo studies.

Substantial differences were observed when GGT promoter IV was analyzed in vivo. In primary cell cultures of initial segment epithelial cells, cotransfection of PEA3 with the various GGT promoter IV constructs was able to activate a promoter truncated at —135 bp. In vivo, the 135-bp construct had some transcriptional activity, which was equivalent to expression from the longer pGL3-b250 construct. However, the activity of both pGL3-b135 and pGL3-b250 was approximately 10% of that observed when the promoter was extended to —530 bp (Fig. 4), indicating the presence of a site or sites between —250 and —530 bp that result in increased transcriptional activity. Three consensus PEA3 sites are found within this region (Fig. 1B). When the core GGAA was mutated within any one of these sites, promoter activity decreased to approximately 10% of that seen with pGL3-b530, suggesting that all three of these sites are necessary for the increased activity seen with pGL3-b530 (Fig. 5). A control mutation made in a similar site within this region had no effect on promoter activity. These studies are in agreement with previous data from electrophoretic mobility shift assays that showed a protein from initial segment nuclear extracts bound an oligonucleotide containing the PEA3 site at —399 to —394 bp. Binding of this protein was blocked when a PEA3 monoclonal antibody was added to the extracts. Of particular interest is the fact that all three PEA3 sites appear to be necessary for pGL3-b530LUC activity, suggesting a coordinated regulation of this construct through these sites. This pattern of PEA3 DNA-binding sites appears to be unique to GGT promoter IV. PEA3 DNA-binding sites have been demonstrated to play roles in the regulation of urokinase plasminogen activator, matrix metalloproteinase 1, and ge-latinase B promoters. However, in these promoters, PEA3 acts in conjunction with adjacent AP-1 sites. Interestingly, there is an AP-1 site in the b135 construct, which, when mutated, has no effect on transcriptional activity (data not shown). Future studies are aimed at identifying which of the PEA3 family members binds to these sites and the stoichiometry of binding. Variable binding to these sites may play a role in the regulation of the level of GGT mRNA IV expression.

When GGT promoter IV was extended to —681 and —903 bp, a 79% and 95% decrease in promoter activity was observed, respectively, suggesting the presence of one or more negative cis-regulatory elements within this region (Fig. 4). However, when the sequence between —530 and —681 bp was cloned in front of the SV40 promoter of pGL3-control, no change in transcriptional activity of the SV40 promoter was observed (data not shown), suggesting that this sequence does not act as a general repressor. The action of this repressor sequence may be context dependent. For example, the PEA3 sites between —250 and —530 bp may be necessary for repressor activity. Analysis of the sequence between —530 and —903 bp reveals a highly AT-rich region. Some homeodomain proteins, which bind to AT-rich sequences, act as repressors of transcription. The homeodomain protein, Nkx3.1, represses the activity of the prostate-derived Ets factor (PDEF), another Ets family member, on the prostate-specific antigen promoter in prostate cells. Several other potential transcription factor binding sites are found within this region, including two GATA-1 sites. GATA-1 can interact with Ets family member, PU.1, and interfere with its ability to transactivate myeloid target genes. Alternatively, these experiments raise the possibility that perhaps structural hindrance from the increasing size of the promoter rather than a repressor may be the cause of the decreased transcriptional activity. This idea is reinforced by the results obtained with pGL3-b6500 (Fig. 6) and raises the question about what happens to this region in vivo when this sequence is found in the context of chromosomal DNA, which is affected by the regulation of chromatin condensation and unwinding. Future studies are aimed at elucidating the mechanism for the decreased transcriptional activity observed with the sequences upstream of —530 bp.

When the 2-kb promoter sequence was used to assay for promoter activity in the in vivo electroporations, minimal expression was observed compared to pGL3-b530, which suggested that perhaps the entire promoter sequence had not been identified (Fig. 4). In an attempt to analyze more 5′ promoter sequence, an additional construct was made that contained 4.5 kb of sequence 5′ to the 2 kb previously cloned. When this construct was used for in vivo electroporations, no additional promoter activity was observed (Fig. 6). One explanation for these results may be that pGL3-b6500 does not represent the entire promoter sequence. Approximately 10 kb of genomic DNA sequence exist between GGT 5′ UTR exon V and the transcriptional start site for GGT mRNA IV. These experiments examined the effects of only 6.5 kb of this sequence. Another explanation for these results may be the existence of additional cis-acting regulatory elements upstream of the GGT gene locus, such as an enhancer or locus control region (LCR). These elements would not be identified by these in vivo electroporation studies. LCRs are functionally characterized as DNA sequences that allow for copy number-dependent expression of transgenes in mice regardless of the chromosomal integration site. LCRs have been described for groups of genes that are coordinately regulated, such as the p-globin gene locus, the human growth hormone (hGH) locus, and the red and green cone pigment gene locus. Indeed, analysis of the other GGT promoters revealed similar results to those depicted in Figure 3, suggesting that additional sequences may be necessary for correct expression from the GGT locus. Possibly, these upstream sequences control the general expression from the GGT locus, whereas cell type-specific regulation comes from cis-regulatory elements present within the promoters themselves. Finally, structural hindrance from the increased sequence length as mentioned in the previous paragraph may also be contributing to the decreased expression seen with pGL3-b6500.

One of the most exciting benefits of the in vivo electroporation method for studying initial segment gene expression is the ability to assess the effects of LTFs on initial segment promoter activity by comparing results from experiments done either in the presence or absence of LTFs (±EDL). This type of experiment would allow one to determine if there were additional sites in the promoter sequence that are sensitive to the effects of LTFs. Studies are currently being performed to address these issues.

In conclusion, this report highlights the benefits of using in vivo electroporation as a method to analyze promoters in the rodent epididymis, in particular the initial segment, which is highly dependent on LTFs for the expression of multiple genes. The results of the experiments presented here demonstrate substantial differences compared with results obtained by in vitro cell culture methods, emphasizing the importance of evaluating these genes in their native environment. However, this method could be adapted for use in any tissue that can be easily manipulated, and it offers the advantage of testing promoters in the intact tissue rather than cell culture.