A pineal regulatory element (PIRE) mediates transactivation by the pineal/retina-specific transcription factor CRX

Abstract

The circadian hormone melatonin is synthesized predominantly in the pineal gland by the actions of two pineal-specific enzymes: serotonin N-acetyltransferase (NAT) and hydroxyindole-O-methyltransferase (HIOMT). Pineal night-specific ATPase (PINA), another pineal- and night-specific protein we recently identified, is produced as a truncated form of the Wilson disease gene (Atp7b) product. To identify the regulatory elements required for pineal-specific gene expression, we isolated sequences upstream of the rat PINA gene and discovered a cis-acting element that is recognized by a novel pineal/retina-specific nuclear factor. This pineal regulatory element (PIRE) has a consensus of TAATC/T and is present in six copies in the 5' regulatory region of the PINA gene, at least three copies in the rat NAT promoter, and at least one copy in each of the putative HIOMT promoters A and B. A recently identified retina-specific protein, cone rod homeobox (CRX), binds to PIRE in vitro and transactivates PIRE-reporter constructs. These data suggest that Crx may play a crucial role in regulating pineal gene expression through interactions with PIRE.

Figure 1

Tissue specificity of PINA and its upstream intronic sequence. (A) Northern blot analysis of a panel of rat tissue total RNAs (10 μg each) at night (02:00) with PINA cDNA as probe. Equal loading and quality of RNAs were confirmed by glyceraldehyde-3-phosphate dehydrogenase (GAPDH) controls. (B) The PINA upstream intronic sequence. Schematic representation of the exon-intron organization of the PINA upstream region. Open boxes represent exons 8 and 9 of the rat Atp7b gene. The black box represents the part of the PINA exon 1 that belongs to intron 8 in Atp7b gene. The DNA sequence shown includes the PINA intronic promoter and small portions of flanking exons (bold italic). The arrow indicates the most upstream transcription initiation site as determined by primer extension and 5′-rapid amplification of cDNA ends (data not shown). The 28-bp sequence used as probe for EMSA (Fig. 2A) is underlined with the CRE in bold.

Figure 2

Identification of PIRE. (A) EMSA with a panel of rat tissue nuclear extracts (5 μg total protein each) and probe as indicated. The CRE site is underlined and the positions of shifted bands are indicated by arrows. (B) Result of EMSAs with pineal nuclear extract and a series of seven overlapping oligonucleotides. The sequence shared among the oligonucleotides used for EMSAs is in bold. (C) Mutation analysis of the 12 nucleotides sufficient for binding. All probes used have EcoRI sites at their 5′ termini and XbaI sites at their 3′ termini, plus a 12-nt sequence as indicated. The 12 nucleotides sufficient for binding were mutagenized one at a time, and each mutant was used as the core sequence of a probe. Arrows indicate the positions of shifted bands. The four nucleotides essential for binding are underlined.

Figure 3

Multiple PIRE sites in PINA, NAT, and HIOMT upstream regions. (A) Upstream sequences of rat PINA, NAT, and human HIOMT with putative PIREs underlined. Functional binding sites as determined by EMSA competition test (B) are further indicated in bold. Transcription initiation sites for PINA and NAT, as determined by primer extension and 5′-rapid amplification of cDNA ends (data not shown), and HIOMT (4) are indicated by arrows. (B) EMSA competition test with pineal nuclear extract and putative PIREs as competitors. The probe contains the 12-nt core sequence (CACTAATCTCCC) flanked by a 5′ EcoRI site and a 3′ XbaI site. The putative PIREs are numbered in order from 3′ to 5′. Sequences of double-stranded probes and competitors are indicated with the cores of PIREs in uppercase. Each competitor was in 100-fold excess of the concentration of probe. When no competitor was added, an equivalent amount of poly(dI-dC) was used as nonspecific competitor. ∗, PIREs as indicated are in reversed sequence.

Figure 4

CRX accounts for the pineal-retina specific binding activity. (A) Supershift assay with an anti-CRX antiserum. The same probe as in Fig. 3B was used. The antiserum was incorporated in EMSAs with a series of nuclear extracts indicated. The arrow indicates the position of supershifted band that is absent when no antibody or no nuclear extracts was incorporated. (B) DNase I footprint assay of PINA intronic promoter. A ³²P end-labeled PINA promoter fragment was incubated with or without CrxHD-glutathione S-transferase fusion protein (8). Protected regions are indicated by brackets with the nucleotide positions indicated. The positions of PIREs are indicated by solid bars.

Figure 5

Diurnal variations of Crx expression compared with that of NAT in pineal gland. Northern blot analysis of a panel of pineal total RNA isolated at different times over a 24-hr time course was quantified by PhosphorImaging (Molecular Dynamics). The Crx mRNA and NAT mRNA levels were normalized to those of glyceraldehyde-3-phosphate dehydrogenase mRNA and presented relative to their levels at 16:00.

Figure 6

Transactivation by CRX. (A) Reporter constructs for transactivation assays. The pGL3-promoter construct contains a simian virus 40 promoter upstream of the luciferase gene. The PIRE-GL3P construct contains three repeats of PIRE upstream of the simian virus 40 promoter whereas the PIRE-M-GL3P construct contains three repeats of PIRE mutant. (B) Luciferase assay results. Each reporter construct (25 ng) was cotransfected into HEK293 cells with 500 ng of either pcDNA 3.1/HisC expression vector alone (CRX −) or pcDNA3.1/HisC-Crx plasmid (CRX +) (8). Firefly luciferase activity was normalized by reference to renilla luciferase activity derived from the cotransfected internal control pRL-TK plasmid. These experiments were replicated four times.