Introduction

When a pool of various genes is generated by random mutations introduced into a gene, and then the preferred genes are selected out of the pool, it can be a simulation of molecular evolution. This process is referred to as directed evolution (1,2). We used this process to generate and improve an Escherichia coli promoter out of a DNA fragment that had no promoter activity (3). We were presumably successful in doing this because we were dealing with a prokaryote that contains a simple promoter structure and whose transcription mechanism is not very complex. For eukaryotes whose promoter structures are more complex, it is probably much more difficult to construct and improve such a promoter with a similar technique. Remans and colleagues (4) attempted to apply a combination of error-prone PCR and DNA shuffling to analyze a plant virus promoter. When they measured the activities of the mutated promoters, no promoter demonstrated any increase in its activity. Instead, most of the mutants were found to have a decreased activity, thus suggesting difficulties in improving eukaryotic promoter activities by error-prone PCR (4).

Eukaryotic transcription is controlled by transcription proteins that can be roughly classified as basic transcription proteins and transcription activating proteins. The basic transcription proteins, with the TATA box binding protein as the center unit, form a gigantic protein complex that can bind to the TATA box sequences to control the basic transcriptional activity and its start site. Alternatively, transcription activating proteins alone, or a smaller complex with other protein molecules, bind to a cis-acting element composed of a unique nucleotide sequence to control the transcriptional timing and/or amount. The basic transcriptional complex does not significantly induce transcription in vivo, and for activation of transcription, a transcription activating protein is required to bind to its corresponding cis-acting element, thus playing a key role in the control of gene expression.

A promoter may be constructed by combining cis-acting elements of DNA fragments in addition to the TATA box sequence. In fact, similar methods have been applied for constructing synthetic promoters with tissue or cell type specificity utilizing cis-acting elements derived from tissue or cell type specific promoters (5,6,7).

In the present study, we attempted to construct a strong promoter by randomly combining cis-acting elements known to be active in a variety of tissues. We herein show this method to be very efficient for creating strong eukaryotic promoters.

Materials and Methods

Cells and Bacteria

PC-3 and DU145 from human prostate carcinoma, SBC-5 from human small cell lung carcinoma, A549 from human lung cancer, HEK 293 from human embryonic kidney, Saos-2 from human osteosarcoma, and PC-12 from rat adrenal carcinoma were used in this study. The other cells used that were not of cancer origin included hFOB1. 19 of osteoblasts transformed with the large T antigen gene of simian virus 40 (SV40), human umbilical vein endothelial cells (HUVEC), and human dermal fibroblasts (HDF). PC-3, DU145, and SBC-5 cells were purchased from the Health Science Research Resources Bank (Tokyo, Japan). HUVEC and HDF were purchased from Toyobo Co. Ltd. (Osaka, Japan), while the rest of the cell lines were purchased from ATCC (Manassas, VA, USA). All of the cancer cells were grown and maintained in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS) and appropriate antibiotics at 37°C in a 5% CO₂ atmosphere. HUVEC and HDF were cultured in special media provided by the manufacturer, and hFOB1.19 cells were cultured in Dulbecco's modified Eagle's medium (DMEM):nutrient mixture F-12 Ham (1:1; Invitrogen, Carlsbad, CA, USA) supplemented with 10% FCS and penicillin and streptomycin. These cultures were also maintained at 37°C in a 5% CO₂ atmosphere.

Two E. coli strains, the DH5α strain purchased from Toyobo Co. Ltd. and the SURE® strain purchased from Stratagene (La Jolla, CA, USA), were used for the experiments in DNA manipulation. The E. coli cells were grown in LB medium at 37°C. All the medium compositions were purchased from BD Diagnostics (Difco™; Sparks, MD, USA). DNA manipulation experiments with E. coli were performed following the methods described by Sambrook and Russell (8).

Plasmid Constructions

A DNA fragment containing the TATA box sequence of the heme oxygenase 1 (HO-1) gene (9) was PCR-amplified with a pair of primers, 5′-ATGGTACCTTCCGCCTGGCCCACGTGAC-3′ and 5′-ATGAGCTCGATTCTGCAGGGCAGGGGCTCC-3′, using genomic DNA extracted from DU145 cells as a template. The amplified fragment was digested with SacI and KpnI and inserted into the SacI and KpnI sites of pGL3-Basic plasmid (Promega, Madison, WI, USA) to construct pGL3-TATA.