RNA-protein interactions in the yeast three-hybrid system: affinity, sensitivity, and enhanced library screening

Abstract

The yeast three-hybrid system has become a useful tool in analyzing RNA-protein interactions. An RNA sequence is tested in combination with an RNA-binding protein linked to a transcription activation domain (AD). A productive RNA-protein interaction activates a reporter gene in vivo. The system has been used to test candidate RNA-protein pairs, to isolate mutations in each interacting partner, and to identify proteins that bind a given RNA sequence. However, the relationship between reporter gene activation and in vitro affinity of an RNA-protein interaction has not been examined systematically. This limits interpretation of the data and complicates the development of new strategies. Here, we analyze several key parameters of the three-hybrid system, using as a model the interaction of a PUF protein, FBF-1, with a range of RNA targets. We compare activation of two reporter genes as a function of the in vitro affinity of the interaction. HIS3 and LacZ expression levels are directly related to affinity over a 10-fold range of Kd. Expression of the reporter genes also is directly related to the abundance of the activation domain fusion protein. We describe a new yeast strain, YBZ1, that simplifies screening of cDNA/AD libraries. This strain possesses a tandem, head-to-tail dimer of a high-affinity variant of MS2 coat protein, fused to a monomer of the LexA DNA-binding protein. We show that the use of this strain in cDNA library screens increases the number of genuine, sequence-specific positives detected, and at the same time reduces the background of false, RNA-independent positives.

FIGURE 1.

Three-hybrid system to detect and analyze RNA–protein interactions. Commonly used protein and RNA components are depicted. Both LacZ and HIS3 are present in strains L40coat and YBZ-1, and are placed under the control of LexA operators. The hybrid RNA contains two MS2-binding sites, and each of the reporters possesses multiple LexA operators.

FIGURE 2.

Comparison of HIS3 and LacZ expression. 3-AT resistance was quantified by plating cells on different concentrations of 3-AT. The highest concentration of 3-AT on which cells grew is plotted on the X-axis. The β-galactosidase activity was measured using the Beta-Glo assay (see Materials and Methods). The light emission, in arbitrary units, is depicted on the Y-axis. Each point derived from analysis of a different RNA as indicated in the figure.

FIGURE 3.

Comparison of in vitro binding affinities to reporter gene activation. (A) Electrophoretic mobility shift assay with purified FBF-1 and 100 fmol of 5′-end-labeled synthetic FBEa-wt RNA. Bound and free FBEa-wt RNA are labeled. The total concentration of FBF-1 present in each incubation is indicated above each lane. Analogous experiments were performed with each of the RNAs. (B) Kd versus 3-AT resistance. (C) Kd versus β-galactosidase activity.

FIGURE 4.

Effect of AD protein abundance on reporter expression. (A) Protein abundance versus LacZ activity. FBF-1/AD was produced at different levels in yeast using four different vectors. Y-axis: β-galactosidase activity. X-axis: relative protein concentration, determined by Western blotting, and normalized to that of using the pACTII vector. (B) Levels of FBF-1/AD determined by Western blotting using and anti-AD antibody. Anti-actin is used as a loading control.

FIGURE 5.

cDNA/AD library screen using YBZ1 and L40coat yeast strains. (A) In the yeast strain L40coat, LexA is fused to a monomer of MS2 coat protein. (B) In strain YBZ1, a head-to-tail dimer of LexA–MS2 coat protein fusion is introduced. MS2*, the MS2 coat protein used in YBZ1, possesses mutation N55K, which decreases K_d sevenfold (Lim et al. 1994). (C) Three-hybrid screen in YBZ1 versus L40coat. Transformants were plated on 0.5 mM 3-AT to select for activation of HIS3. The identity of the RNA-dependent positives were determined by sequencing with pACT-specific primers.

FIGURE 6.

Comparison of three-hybrid systems to detect RNA and small molecule ligands with protein partners. (A) The small-molecule-based three-hybrid system. Fusion proteins contain the LexA DNA-binding domain linked to dihyrdrofolate reductase, and protein Y with AD. A small molecule containing methotrexate linked to moiety X connects the two proteins and thereby leads to activation of the reporter gene. (B) RNA-based three-hybrid system. The RNA contains two MS2 stem loops fused to RNA sequence X. (C) Comparison of the relationship between Kd and reporter gene activity in the two systems. Data for the RNA-based system (Fig. 3C) are compared with the data of de Felipe et al. (2004).