Translational repression by PUF proteins in vitro

Abstract

PUF (Pumilio and FBF) proteins provide a paradigm for mRNA regulatory proteins. They interact with specific sequences in the 3' untranslated regions (UTRs) of target mRNAs and cause changes in RNA stability or translational activity. Here we describe an in vitro translation assay that reconstitutes the translational repression activity of canonical PUF proteins. In this system, recombinant PUF proteins were added to yeast cell lysates to repress reporter mRNAs bearing the 3'UTRs of specific target mRNAs. PUF proteins from Saccharomyces cerevisiae and Caenorhabditis elegans were active in the assay and were specific by multiple criteria. Puf5p, a yeast PUF protein, repressed translation of four target RNAs. Repression mediated by the HO 3'UTR was particularly efficient, due to a specific sequence in that 3'UTR. The sequence lies downstream from the PUF binding site and does not affect PUF protein binding. PUF-mediated repression was sensitive to the distance between the ORF and the regulatory elements in the 3'UTR: excessive distance decreased repression activity. Our data demonstrate that PUF proteins function in vitro across species, that different mRNA targets are regulated differentially, and that specific ancillary sequences distinguish one yeast mRNA target from another. We suggest a model in which PUF proteins can control translation termination or elongation.

FIGURE 1.

Puf5p represses translation in vitro. (A) The in vitro translation assay. Yeast cytoplasmic extracts and two luciferase reporter RNAs were combined with reaction buffers and allowed to translate for 1 h at 30°C. Each reaction contained two luciferase reporters: The HO reporter consisted of a short 5′UTR, the firefly luciferase ORF, and the HO 3′UTR. The boxes in the diagram labeled 4 and 5 represent the binding sites of Puf4p and Puf5p, respectively. The control reporter was composed of the 5′UTR of CAN1, the Renilla luciferase ORF, and the 3′UTR from PGK1. All reporters carried a m7GpppG cap structure and a 50-nt poly(A) tail. After incubation, levels of firefly and Renilla luciferase were determined using a luminescence-based detection system. (B) Puf5p specifically represses the HO reporter. The first graph shows the firefly (FF) luciferase response, the second graph shows the control Renilla (Ren) luciferase response, and the last graph shows normalized responses (FF/Ren). Normalization consists of dividing the firefly by the Renilla response and setting a control reaction (containing only buffer) to 1.0 for each reporter RNA. (C) Titration of Puf5p. Increasing concentrations of Puf5p were added to translation reactions containing either the HO or HO mut reporter.

FIGURE 2.

Puf5p represses continuously. (A) Kinetics of translation and repression. Firefly and Renilla luciferase levels were measured every 10 min during a 1-h incubation at 30°C. (B) Kinetics as a proportion of total luciferase activity. The data in panel A are expressed as the level of firefly luciferase from the HO reporter over time in the presence or absence of Puf5p. (C) RNA stability. Total RNA was re-isolated from translation reactions at various time points during a translation reaction, and firefly luciferase mRNA was detected via Northern blot. The control (cont.) RNA is a 500-nt fragment of firefly luciferase added after the translation reaction ends to control for variation in RNA re-isolation. To verify that Puf5p was active in this experiment, luciferase levels were measured at the 60-min time point of the same incubations (D). (Right) The graph quantifies the mRNA levels. To determine the percent RNA remaining, background readings were subtracted from the HO signal and then divided by the signal of the control RNA, where the RNA detected at 1 min is assumed to be equal to 100%. (D) Puf5p was active in the assays analyzing kinetics (see panel C).

FIGURE 3.

Puf4p represses the HO reporter. Puf4p was added to translation reactions with either the HO or HO mut reporter. The final concentrations of Puf4p were 0, 0.5, or 2.5 μM.

FIGURE 4.

C. elegans FBF-2 is active in yeast extracts. (A) Repression by FBF-2 protein. Purified FBF-2 or an FBF-2 mutant was added to translation reactions with reporters containing 0–3 FBF binding elements (FBEs). FBF-2 specifically repressed only those reporters containing FBEs, and repression was directly correlated with the number of FBEs. The FBF-2 mutant protein contains alanine substitutions in the RNA-contacting residues of repeat 6 that greatly reduce its affinity for RNA. (B) Titration of FBF-2. Translation reactions containing the 3 FBE reporter were incubated with increasing concentrations of FBF-2.

FIGURE 5.

Multiple target 3′UTRs mediate repression, but the HO 3′UTR is particularly efficient. Five reporters containing the firefly luciferase ORF and the 3′UTR from the mRNA indicated were tested in vitro. The PGK1 3′UTR served as a negative control (it is not a Puf5p target and contains no Puf5p binding elements). For detailed description of 3′UTRs, see Materials and Methods.

FIGURE 6.

The Puf4p site is not required for repression. The HO 3′UTR was mutated to determine the contribution of each PUF binding element. For each site, the core “UGU” sequence was mutated to “ACA,” which eliminates PUF binding (Hook et al. 2007). The crossed-out boxes indicate the site that was mutated in each reporter.

FIGURE 7.

Segments of HO enhance repression of CIN8. (A) CIN8 versus HO binding sites. In the HO CIN8 bs reporter, the Puf5p binding site in the HO 3′UTR (UGUAUGUAAU) was replaced with the Puf5p binding site from the CIN8 3′UTR (UGUAAUAUU). (B) Chimeric 3′UTRs. The HO 3′UTR was divided into four segments that were then inserted into the 3′UTR of the CIN8 reporter while maintaining the same orientation and distance relative to the Puf5p site, as shown.

FIGURE 8.

A sequence in the HO 3′UTR enhances Puf5p repression. A 9-nt mutation in the HO 3′ UTR, mut E, reduces repression. The mut E element was further subdivided into three mutants 3 nt long, E1–E3. Each mutant contained a transversion mutation in the bracketed sequence. Another variation of the HO reporter (HO long 3′) included an extra 25 nt of vector sequence that is present in all HO mutant reporters for technical reasons (see Materials and Methods). The sequences highlighted in gray indicate the PUF binding sites. The highlighted sequence constitutes the smallest region in mut E that is required for optimal repression of the HO 3′UTR.

FIGURE 9.

Mutation of segment E does not affect Puf5p binding. (A) Complete sequences of the two RNA oligos used in EMSA analysis. The Puf5p site is underlined, and the mut E mutation is highlighted. (B,C) Representative EMSA autoradiograms. The lower band in each gel is radiolabeled unbound RNA and the upper band is RNA bound to Puf5p. The K_d shown was calculated from four independent experiments. (D) Quantitation. The fraction of RNA bound by Puf5p was plotted versus protein concentration.

FIGURE 10.

3′-UTR length affects repression. (A) The HO 3′UTR was expanded by the inclusion of either 25 or 70 extra bases of downstream vector sequence. (B) The 3′UTR was expanded by introducing random sequence between the stop codon and the HO 3′-UTR sequence. The number indicates the length of the additional sequence. The HO 5′-spacer 185 mut E1 contains both a spacer sequence and mutation E1 from Figure 5.