Conservation of the deadenylase activity of proteins of the Caf1 family in human

Abstract

The yeast Pop2 protein, belonging to the eukaryotic Caf1 family, is required for mRNA deadenylation in vivo. It also catalyzes poly(A) degradation in vitro, even though this property has been questioned. Caf1 proteins are related to RNase D, a feature supported by the recently published structure of Pop2. Yeast Pop2 contains, however, a divergent active site while its human homologs harbor consensus catalytic residues. Given these differences, we tested whether its deadenylase activity is conserved in the human homologs Caf1 and Pop2. Our data demonstrate that both human factors degrade poly(A) tails indicating their involvement in mRNA metabolism.

FIGURE 1.

Alignment of the sequences of yeast Pop2, human Pop2, and human Caf1. Sequences were aligned with ClustalX (Thompson et al. 1997) and the resulting alignment manually edited. Sequence conservation is indicated below the aligned sequences with stars (*) indicating complete amino-acid conservation between the three sequences while semicolon (:) and dot (.) denote the presence of amino acids of high or low similarity, respectively. Conserved residues involved in magnesium binding and nuclease function in the RNase D subfamily of DEDD nucleases are indicated by arrows above the sequence. Amino acids at these first two positions were changed to Alanine by site-directed mutagenesis in the inactive human Caf1 mutant. The numbering, in the sequence, of the first and last aligned residues are indicated. Note that yeast Pop2 contains a large N-terminal extension of 148 residues compared to its human counterparts, and that it diverges in the first and last residues involved in nuclease function.

FIGURE 2.

The recombinant GST-Caf1 protein is a deadenylase. (A) Protein profile of the recombinant GST-hCaf1 fraction. A Coomassie stained gel of a typical purified GST-hCaf1 fraction is shown. The migration of a molecular weight size marker is indicated. (B) GST-hCaf1 is a RNAse. Gel-fractionated products of an RNA degradation assay using a 5′ end-labeled synthetic RNA oligonucleotide are shown. Various quantities of the recombinant protein were used in parallel time-course reactions; 60 fmol of substrate were used per reaction. Lane 1 shows the starting substrate. GST-hCaf1 pauses after degradation of the first seven A residues but resumes once the first U residue has been removed. (C) GST-hCaf1 specifically removes 3′ terminal A residues. Degradation of a 5′-labeled substrate ending by 5 A residues stops once the 3′ oligo(A) tail has been degraded.

FIGURE 3.

The deadenylase activity of hCaf1 does not result from the presence of contaminants. The deadenylase activity of recombinant GST-hCaf1 is specific. Time-course reaction was performed with the substrate ending with seven A residues and the indicated quantities of various proteins; 60 fmol of substrate were used per reaction. Lanes 1–16: GST preparations purified in parallel with GST-hCaf1 do not display deadenylase activity, ruling out the copurification of contaminating nucleases from E. coli. The deadenylase activity observed with GST-hCaf1 is not inhibited by RNasin (cf. lanes 17–20 and lanes 21–24) excluding contamination of the recombinant fraction by common RNAse. Finally, a recombinant catalytic site mutant is unable to degrade RNA (lanes 25–32). The free substrate is shown in lane 33. Overall, these results demonstrate unequivocally that GST-hCaf1 is responsible for the deadenylase activity detected.

FIGURE 4.

The Caf1 protein specifically degrades 3′ poly(A) tails. (A) Time-course reactions with the recombinant GST-hCaf1 factor and substrates ending with 20 A (lanes 1–4) or 10 A followed by 10 C (lanes 5–8). Protein quantities are indicated. This result demonstrates that GST-hCaf1 specifically degrades 3′ poly(A) tails. These data also demonstrate that GST-hCaf1 does not harbor oligo(A) specific endonuclease activity; 60 fmol of substrate were used per reaction. (B) To confirm the exonucleolytic activity of hCaf1, we also analyzed by TLC the product formed in a reaction containing hCaf1 and an RNA substrate, ending with poly(A), internally labeled with α³²P at A residues. This demonstrated the exclusive release of α³²P-AMP (AMP marker is show by arrow) (Sigma); 100 fmol of substrate were used per reaction.

FIGURE 5.

hPop2 is also a nuclease. hPop2 is a highly active RNase. (A) Purified recombinant hCaf1 and hPop2 (following purification tag removal with trombin and factor Xa, respectively) were incubated with (B) a synthetic RNA oligonucleotide ending by seven A residues. Time-course reactions were performed using the indicated quantities of the recombinant factor and the products detected after fractionation on denaturing gel. Note that, given the high activity of hPop2, time-course reactions lasted only 10 min. (C) hPop2 is not highly specific for 3′ oligo(A) tails. hPop2 degraded efficiently a substrate lacking a 3′ oligo(A) tail (lanes 9–16) while hCaf1 was unable to do so (lanes 1–8). The starting substrate was loaded in lane 17. (D) To confirm the activity of hPop2, we also analyzed by TLC the product formed in a reaction containing hPop2 and an RNA substrate, ending with a poly(G) internally labeled with α³²P at G residues. This result shows that hPop2 is able to release α³²P-GMP (GMP marker is show by an arrow) (Sigma); 100 fmol of substrate were used per reaction.