Host factor Hfq of Escherichia coli stimulates elongation of poly(A) tails by poly(A) polymerase I

Abstract

Current evidence suggests that the length of poly(A) tails of bacterial mRNAs result from a competition between poly(A) polymerase and exoribonucleases that attack the 3' ends of RNAs. Here, we show that host factor Hfq is also involved in poly(A) tail metabolism. Inactivation of the hfq gene reduces the length of poly(A) tails synthesized at the 3' end of the rpsO mRNA by poly(A) polymerase I in vivo. In vitro, Hfq stimulates synthesis of long tails by poly(A) polymerase I. The strong binding of Hfq to oligoadenylated RNA probably explains why it stimulates elongation of primers that already harbor tails of 20-35 A. Polyadenylation becomes processive in the presence of Hfq. The similar properties of Hfq and the PABPII poly(A) binding protein, which stimulates poly(A) tail elongation in mammals, indicates that similar mechanisms control poly(A) tail synthesis in prokaryotes and eukaryotes.

Figure 1

Hfq affects polyadenylation in vivo. The autoradiograph compares the kinetics of elongation of the poly(A) tail of rpsO mRNA in strains containing (+) or lacking (−) Hfq protein. Times after inhibition of transcription by rifampicin are indicated at the top. Relative amounts of rpsO mRNA were quantified with a PhosphorImager (Molecular Dynamics) and plotted as a function of time to estimate its stability.

Figure 2

Hfq affects PAP I activity in vitro. (a) The labeled 97RNA and 1 μg yeast RNA were incubated with 500 fmol PAP I. Purified Hfq was added as indicated. (b) and (c) 360 fmol of labeled 97RNA (b) or polyadenylated 97RNA(18–27A) (c) RNA primers were incubated with 20 fmol PAP I. A total of 4.5 pmol of purified Hfq was added. RNAs withdrawn at different times were analyzed on sequencing gels. Lanes 1 (a–c) show the nonelongated primers. Addition of Hfq, times of incubation, positions, and the nature of the RNA primers and markers are indicated. The star in b shows a shorter RNA produced by T7 RNA polymerase. The upper part of b was overexposed to show the smears of long molecules (lanes 14–18). Reactions of lanes 10 and 19 of b and c contained 10 times more PAP I than lanes 9 and 18.

Figure 3

Hfq strongly binds poly(A) RNA. Gel shift experiments were performed with RNA primers with or without a poly(A) extension indicated in a (Top) or b and c (Left). (a) Increasing amounts of Hfq (55 fmol in lanes 2, 6, 10, and 14; 275 fmol in lanes 3, 7, 11, and 15 and 1100 fmol in lanes 4, 8, 12, and 16) were added to the different primers. (b and c) The complexes formed by the RNA primers and 1100 fmol of Hfq were competed by poly(A≈300–6000) or poly(C) (Sigma) at concentrations indicated in pmols of nucleotides (Top). S and F in b show positions of the slow and fast-migrating complexes, respectively.

Figure 4

Functional and evolutionary relationships between polyadenylation machineries and telomerases. The thick rectangular frame encompasses the Hfq and PABP II polyadenylation stimulatory factors. Hfq and PABP II are linked to enzymes that they modulate by thin rectangular frames. Members of the nucleotidyltransferase family are surrounded by an oval, and telomerase-related enzymes that maintain the 5′ extremities of RNA tagged by a hairpin structure are circled. The functional relationships described here are shaded.