Hepatic glutamine synthetase (GS) shows a unique expression pattern limited to a few hepatocytes surrounding the terminal hepatic veins. Starting from the genomic clone of the rat GS gene, λ GS1 [Van de Zande, L. P. G. W., Labruyère, W. T., Arnberg, A. C., Wilson, R. H., Van den Bogaert, A. J. W., Das, A. T., Frijters, C., Charles, R., Moorman, A. F. M. & Lamers, W. H. (1990) Gene (Amst.) 87, 225–232] additional genomic clones containing up to 9 kb of 5′flanking region were isolated in order to characterize cis-acting elements involved in the regulation of GS expression.

Sequence analysis of the 5′flanking region up to −2520 bp revealed a putative AP2-binding site at −223 bp and a second GC box at −2343 bp in addition to the canonical TATA, CCAAT and GC boxes found proximal to the transcription-start site. A possible negative glucocorticoid-responsive element (GRE) and regions with very weak similarity to a GRE and to a known silencer element were noted at −506 bp, −406 bp and at −798 bp, respectively.

Within the sequenced part of the 5′flanking region no known regulatory elements associated with liver-specific gene expression were found except for a putative HNF3-binding site at −896 bp. Functional analysis by transient transfection assays using constructs with the pSSCAT or the pXP1 vector revealed that the elements present within the first 153 bp and particularly the first 368 bp of upstream sequence consititute an active promoter the activity of which is decreased by additional sequences up to −2148 bp. The presence of dexamethasone led to a 2–4-fold increase in the promoter activity of all these constructs.

Using the heterologous truncated thymidine-kinase-gene promoter of the plasmid pT81-luc a strong enhancer element was located between −2520 bp and −2148 bp. Its activity was not affected by dexamethasone but was negatively influenced by flanking sequences in both directions. This enhancer was also effective with the homologous GS promoter (−153 to +59 bp) and the heterologous full thymidine-kinase-gene promoter (pT109luc). No further enhancers were found up to −6200 bp.

Using the same approach, a second enhancer was found between +259 bp and +950 bp within the first intron.

Deoxyribonuclease-I hypersensitivity studies confirmed the presence of a hypersensitive site between +350 bp and +550 bp and suggested a second site between +850 bp and + 1200 bp.

The ultimate GS promoter (−153 to +59 bp) as well as the two strong enhancer regions identified drove luciferase expression most efficiently or almost exclusively in HepG2 cells but not mouse embryo fibroblasts indicating that these regions might be involved in the cell-type specificity of GS expression.

In accord with known data on GS activity and localization these findings suggest that the regulation of the GS gene in the liver is different from that of liver-specific enzymes and proteins. Although our results do not yet indicate whether the regulatory regions identified play a role in the positional regulation of GS gene expression, they provide a good basis for further studying this complex phenomenon.