Identification of eukaryotic promoter regulatory elements using nonhomologous random recombination

Abstract

Understanding the regulatory logic of a eukaryotic promoter requires the elucidation of the regulatory elements within that promoter. Current experimental or computational methods to discover regulatory motifs within a promoter can be labor intensive and may miss redundant, unprecedented or weakly activating elements. We have developed an unbiased combinatorial approach to rapidly identify new upstream activating sequences (UASs) in a promoter. This approach couples nonhomologous random recombination with an in vivo screen to efficiently identify UASs and does not rely on preconceived hypotheses about promoter regulation or on similarity to known activating sequences. We validated this method using the unfolded protein response (UPR) in yeast and were able to identify both known and potentially novel UASs involved in the UPR. One of the new UASs discovered using this approach implicates Crz1 as a possible activator of Hac1, a transcription factor involved in the UPR. This method has several advantages over existing methods for UAS discovery including its speed, potential generality, sensitivity and lack of false positives and negatives.

Figure 1.

Generation of a nonhomologous random recombination library of a yeast promoter. The bottom figure represents the plasmid into which the library is cloned for blue/white screening.

Figure 2.

Composition of selected UAS elements from the KAR2 promoter. The selected fragments (colored arrows) are compared to the native promoter to reveal consensus regions. Numbering across the top represents the nucleotide position in the native promoter relative to the start of translation. Arrow colors indicate the order of fragment assembly (5′-red-green-blue-purple-3′) and the arrow direction indicates a sense (pointing right) or antisense (pointing left) promoter fragment. The shaded gray encompasses clones with the strongest phenotype. Four consensus regions (I, II, III and IV) are indicated with a vertical line.

Figure 3.

Partial sequence alignments of clones from the NRR analysis of the KAR2 promoter to reveal consensus regions. For regions I, II and III, all clones with the conserved region were used for the alignments. For region IV, clones with a strong phenotype that also lacked other conserved regions were used for the alignment. Consensus regions are shown in red. For region III, an extended consensus region is highlighted in blue. Each consensus region is compared to a known UPRE, with the putative transcription-factor binding site underlined.

Figure 4.

Quantitative β-galactosidase assays with the KAR2 consensus regions serving as UASs in the CYC1-lacZ construct. Multiple copies of regions I–IV, 3 ×, 2 ×, 3 × and 4 × respectively, were used in these constructs. Consensus region III and the extended region III (III +, Figure 3) were each analyzed. Assays were performed in wild-type (wt) and Δire1 strains in the presence and absence of tunicamycin. β-Galactosidase activity is shown in arbitrary units and error bars represent 1 SD.

Figure 5.

Composition of selected UAS elements from the SIL1 promoter. The selected fragments (colored arrows) are compared to the native promoter to reveal consensus regions. Numbering across the top represents the nucleotide position in the promoter relative to the start of translation. Arrow colors indicate the order of fragment assembly (5′-red-green-blue-3′) and the arrow direction indicates a sense (pointing right) or antisense (pointing left) promoter fragment. The shaded gray encompasses clones with the strongest phenotype. One consensus region (I) is indicated with a vertical line.

Figure 6.

Analysis of the consensus region for the SIL1 promoter. (Left) A partial sequence alignment of the clones with the strong phenotype reveals a consensus region that contains UPRE2 with a putative Gcn4-binding site (underlined). (Right) Quantitative β-galactosidase assays with one copy or three copies of the SIL1 consensus region serving as a UAS in the CYC1-lacZ construct. Assays were performed in wild-type (wt) and Δire1 strains in the presence and absence of tunicamycin. β-Galactosidase activity is shown in arbitrary units and error bars represent 1 SD.

Figure 7.

Composition of selected UAS elements from the HAC1 promoter tested in a wild-type strain. The selected fragments (colored arrows) are compared to the native promoter to reveal consensus regions. Numbering across the top represents the nucleotide position in the promoter relative to the start of translation. Arrow colors indicate the order of fragment assembly (5′-red-green-blue-purple-cyan-orange-3′) and the arrow direction indicates a sense (pointing right) or antisense (pointing left) promoter fragment. The shaded gray encompasses clones with the strong phenotype. Two consensus region (I and II) are indicated with a vertical lines.

Figure 8.

Partial sequence alignments of clones from the NRR analysis of the HAC1 promoter. (Top) Region I has a consensus sequence (red) common to all the clones and an extended consensus sequence (blue) shared by half of the clones. (Bottom) Wild-type region II contains the previously described UPRE1-like sequence and is shown with the KAR2 UPRE1 sequence for comparison. The putative transcription-factor binding site(s) in UPRE1 is underlined.

Figure 9.

Quantitative β-galactosidase assays with one copy of the HAC1 consensus regions I and II each serving as a UAS in the CYC1-lacZ construct. Assays were performed in wild-type (wt), Δire1 and Δcrz1 strains in the presence and absence of tunicamycin. β-Galactosidase activity is shown in arbitrary units and error bars represent 1 SD.