It has long been suggested that the human serpin genes are evolved from the same ancestor through intra- and interchromosomal duplications to develop into different clusters [31-33]. Interestingly, when the human genome sequence was unveiled a few years ago, several non-homologous genes were found to be mapped into the serpin gene clusters. One of these genes, PDCD10, not only forms a head-to-head organization with SERPINI1 at chromosome 3q26 but also shares a tiny promoter with it. The organization of this divergent gene pair is well preserved in many species throughout the evolution, indicating that PDCD10 and SERPINI1 need to be tightly together to maintain the gene co-regulation or their functional association. Further sequence analysis on the intergenic region revealed a 400-bp fragment near the vicinity of PDCD10 is highly similar among human, mouse and rat, a result strongly indicating that the transcriptional regulation mechanisms of these two genes are evolutionary conserved among different species. Indeed, by performing a series of luciferase assays, a 175-bp fragment from nt 1 to 175 in the vicinity of PDCD10 was determined to function as a minimal promoter of both genes (Fig. 5). The same 175-bp minimal promoter was also observed (data not shown) when we utilized the human intergenic fragment to perform the promoter analysis in rat adrenal gland PC12 cells [30]. In addition, we detected the existence of an important regulatory fragment from nt 176 to 473 in the vicinity of PDCD10 that possesses a repressive activity for SERPINI1 (Fig. 6). These findings are different from a previous study on mouse Serpini1 [30] claiming that a 200-bp segment near the transcription initiation site of Serpini1 was responsible for promoter activity. There are at least two possible reasons for the discrepancy observed in human and mouse SERPINI1 promoter. First, the constructs used for promoter analysis of mouse Serpini1 were all derived from 5' deletions of the full-length promoter region. These constructs might have contained the repressive element and/or lost the minimal promoter, and thus have very low promoter activities which could not be differentiated from each other. Secondly, different cell lines chosen might result in different conclusions; particularly the rat adrenal gland cell line PC12 has very low expression of Serpini1.
Trinklein and colleagues have pointed out that divergently transcribed gene pairs whose transcriptional start sites are separated by less than 1 kb represent more than 10% of genes in the whole human genome [26]. Such a high proportion indicates that there is a biological importance for those adjacent gene pairs even through a long evolution. Two feasible explanations for the existence of bidirectional promoters in the mammalian genome were recently given by Zhang and colleagues [34]. The first explanation is that they are the ancestral sequences survived in evolution. A bidirectional promoter, like the one for PDCD10 and SERPINI1, with the characteristics of TATA-less, GC-rich and multiple Sp1 sites was thought to be the preserved ancestral sequences. The second explanation is that the genes sharing the bidirectional promoter are functionally related. From the sequence alignment data, we observed that this highly conserved promoter sequence is maintained throughout the evolution. Besides, there are evidences indicating that PDCD10 and SERPINI1 are both related to brain diseases. Therefore, we suggest that the compact adjacent arrangement for PDCD10 and SERPINI1 might be due to the possibility that they are involved in the same biological pathway.
Apart from the PDCD10-SERPINI1 gene pair at chromosome 3q26, there was another human head-to-head gene pair found to be associated with brain diseases [35]. This gene pair, consisting of the parkinsonism-related parkin gene (PARK2) and parkin co-regulated gene (PACRG) at chromosome 6q25.2-27, has many features in common with the PDCD10-SERPINI1 pair. Like SERPINI1 and PDCD10, PARK2 and PACRG are also non-homologous. PARK2, like SERPINI1, has its tissue specificity and is mainly expressed in brain, heart, skeletal muscle and kidney. PACRG is expressed in all tissues like PDCD10. The bidirectional promoter of PACRG-PARK2 is also TATA-less, GC-rich, and with multiple Sp1 binding sites [35,36]. Furthermore, the genetic mutations of SERPINI1 and PARK2 are able to accelerate the formation of cellular inclusion bodies and to cause the dementia symptom [18,21,37]. Therefore, we suggest that brain, being the most sophisticated organ which controls almost all responses of the whole body, may use bidirectional promoters to achieve gene expressions more efficiently, especially in the pathogenesis of its diseases.
Gene expression and tissue specificity are often regulated by transcriptional activators or repressors [38-40]. In the case of the PDCD10-SERPINI1 gene pair, it is apparent that these two genes are not co-expressed genes (Fig. 1). This phenomenon can be best explained by the existence of a critical fragment, from nt 176 to 473 of the intergenic sequence, which contains an enhancer and a repressive element to coordinately regulate the transcriptions of both genes (Fig. 5 & 6). The intergenic enhancer is orientation-dependent exerting its effect on the promoter of PDCD10 in the native context (Fig. 5). It also works in an orientation-dependent manner with heterologous promoters (Fig. 6A & 6B). Previously, a similar orientation-dependent intergenic enhancer had been reported in a mouse gene pair: the divergent Cryab (αB-crystallin) and Hspb2 (heat shock 27 kDa protein 2) pair [41]. However, this is the first time that such an enhancer is identified in the human genome. Moreover, the neuron-specificity of SERPINI1 and its down-regulation in brain tumorigenesis may be accounted for by a repressive element located in the same critical regulatory fragment (Fig. 5 & 6). Together these results suggest that the fine coordination between the orientation-dependent cis-control elements may play an important role in the asymmetrical expressions of PDCD10 and SERPINI1 in different tissues. In addition to the cis-control elements, the possible involvement of certain trans-acting factors in this bidirectional promoter region need to be further elucidated.
Answers to all your Biology Questions
