Introduction

The mammalian genome contains a great number of functional cis-regulatory elements, such as promoters, enhancers, silencers, insulators, etc. (1,2). Although the nucleotide sequences of the human and some other genomes are already known (3,4,5), the information about genomic location of these elements is rather scarce. This is currently one of the main obstacles for the understanding of complex gene regulation networks, where cis-elements serve as nodes that bind certain transcription factors and co-factors at a certain time, thus enabling proper reactions of a certain class of cells in response to a variety of extra- and intracellular signals. A complete functional annotation of sequenced eukaryotic genomes is supposed to include positions of all non-coding cis-regulatory nodes.

RNA polymerase II promoters are usually located closely upstream of genes, which facilitates their identification and mapping, provided that the location and transcription start sites of genes are known (6,7,8). In contrast to promoters, enhancers can function over long distances, and their positions relative to genes under their control vary over a wide range (9,10). Moreover, a single promoter can be regulated by several enhancers with different tissue specificities. These uncertainties greatly complicate identification and mapping of enhancers.

Up until now enhancers located within long genomic sequences have been identified by indirect methods (for example, by identification of evolutionarily conserved sequences (11)), by their ability to activate replication of plasmids in mouse cells (12), or in chromatin by methylation analysis of enhancer-associated histones H3 (13). There is also the direct enhancer identification technique based on antibodies to putative enhancer-binding proteins (ChIP-on-chip) combined with in silico analysis (14). However, the latter method is limited by our knowledge of enhancer-binding proteins and can be used only for identifying regulatory elements with at least one known specifically binding protein. Most enhancers are tissue- or cell-specific, and prediction of these elements by computational methods is difficult and not always reliable. In general, the problem of identification and mapping of enhancer elements remains unresolved, and the development of high-throughput experimental approaches to identifying and mapping genomic enhancers is highly desirable.

Here we propose a straightforward method for identification of sequences that are able to enhance the transcription of a reporter gene. Using this method, we have identified 15 potential enhancers from 1-Mb FXYD5-COX7A1 region of human chromosome 19.

Materials and Methods

Basic Protocols

Growth and transformation of Escherichia coli cells, preparation of plasmid DNA, agarose gel electrophoresis, blot-hybridization, and other standard manipulations were performed as described previously (15).

Plasmids

A pCMV-SGTN plasmid containing the GTN gene coding for EGFP-Herpes Simplex Virus thymidine kinase (HSV-tk)-neomycin phosphotransferase II (NPTII) fusion protein under control of a cytomegalovirus (CMV) minimal promoter and enhancer was constructed by E. Snezhkov based on pSGTN plasmid (16).

The pQCXIX self-inactivating retroviral vector (Clontech, Mountain View, CA, USA) was modified as follows: the vector was cut with EcoRV and XbaI, and a smaller fragment containing the CMV promoter, enhancer, internal ribosome entry site (IRES), and a multiple cloning site, was removed. To obtain pQCXIX-ENH vector, pSGTN plasmid was cut with NotI, the sticky ends filled in with Klenow enzyme, and then cut with AsuNHI restriction endonuclease. The resulting fragment containing a CMV minimal promoter and the GTN gene was ligated with a pQCXIX EcoRV-XbaI fragment. The plasmid obtained was cut with ClaI and a synthetic multiple cloning site CGATCTAGACTCGAGAATTC, which includes restriction sites for ClaI, XbaI, XhoI and EcoRI, was inserted. The structure of the resulting pQCXIX-ENH vector (Figure 1A) was verified by sequencing.

Figure 1. General scheme of the enhancer selection technique. (Click to enlarge)

Plasmids pQCXIX-ENH(+) (positive control containing the GTN gene under CMV promoter and enhancer control) and pQCXIX-ENH(−) (negative control containing a promoterless GTN gene) were obtained as above, except that ClaI-AsuNHI, and BamHI-AsuNHI fragments of pSGTN plasmid, respectively, were used for ligation, and no multiple cloning site was inserted.

Construction and Cloning of a Short Fragments Library

A library of short fragments representing the 1-Mb FXYD5-COX7A1 region of human chromosome 19 was prepared as described earlier (17,18) by exhaustive digestion with Sau3A and Csp6–1 of 30 overlapping cosmids that completely cover the region. In this work, the library primer was modified (ACTGAGGTCGACTATCCATGAACA) to include the SalI recognition site (underlined). The library was treated with SalI and ligated into the XhoI site of the pQCXIX-ENH vector. The resulting pool of plasmids will be referred to below as the pQCXIX-ENH(L) pool of plasmids. To preserve the representativeness of the library, the transformed E. coli cells were grown on agar plates. About 16,000 individual colonies for the first and 10,000 for the second selection round were collected, and plasmid DNA for transfection was isolated using a Wizard Plus Minipreps DNA Purification System (Promega, Madison, WI, USA)