In vitro and in vivo studies of the RNA conformational switch in Alfalfa mosaic virus

Abstract

The 3' termini of Alfalfa mosaic virus (AMV) RNAs adopt two mutually exclusive conformations, a coat protein binding (CPB) and a tRNA-like (TL) conformer, which consist of a linear array of stem-loop structures and a pseudoknot structure, respectively. Previously, switching between CPB and TL conformers has been proposed as a mechanism to regulate the competing processes of translation and replication of the viral RNA (R. C. L. Olsthoorn et al., EMBO J. 18:4856-4864, 1999). In the present study, the switch between CPB and TL conformers was further investigated. First, we showed that recognition of the AMV 3' untranslated region (UTR) by a tRNA-specific enzyme (CCA-adding enzyme) in vitro is more efficient when the distribution is shifted toward the TL conformation. Second, the recognition of the 3' UTR by the viral replicase was similarly dependent on the ratio of CBP and TL conformers. Furthermore, the addition of CP, which is expected to shift the distribution toward the CPB conformer, inhibited recognition by the CCA-adding enzyme and the replicase. Finally, we monitored how the binding affinity to CP is affected by this conformational switch in the yeast three-hybrid system. Here, disruption of the pseudoknot enhanced the binding affinity to CP by shifting the balance in favor of the CPB conformer, whereas stabilizing the pseudoknot did the reverse. Together, the in vitro and in vivo data clearly demonstrate the existence of the conformational switch in the 3' UTR of AMV RNAs.

FIG. 1.

The CPB and the TL conformations of the AMV RNA3 3′ terminus. The two conformers of AMV RNA3 3′ 145 nt are shown. (A) CPB conformer. The two major CP binding sites are indicated by brackets. Base pairing between loop D and stem A promotes TL conformation. (B) Secondary structure of the TL conformer.

FIG. 2.

Adenylation and replication of pseudoknot and hairpin mutants (A) Structure of five different pseudoknot variants and four different hairpin variants are depicted. Note that pk1, pk2, and pk3 can only be tested in combination with hairpins hp1 and hp2 (left panel), whereas pk4, pk5, and pk6 can only be tested in combination with hp3 and hp4 (right panel). Values between brackets are the calculated stacking energies (31) for the “top” four or five base pairs in the pseudoknots or all base pairs in the hairpins. An alternative conformation of hp2 is indicated by the dashed box. (B) Denaturing gel showing products of the adenylation reactions with the indicated constructs (C) Products of minus-strand assays using the indicated transcripts as templates with AMV polymerase.

FIG. 3.

CP blocks adenylation and minus-strand synthesis In vitro. (A) Adenylation of the indicated transcripts by CCA-adding enzyme was done in the absence (−) or presence (+ and ++) of AMV CP. RNA/CP ratios were 1 to 0.25 (+) and 1 to 1.25 (++). (B) Minus-strand assay in the presence of CP. RNA/CP ratios were 1 to 0.5 (+) and 1 to 5 (++).

FIG. 4.

Conformational switch investigated in the yeast three-hybrid system. (A) Outline of the various constructs used in Y3H. Wt, 3′ 112 nt of wild-type RNA3; Wt/1-7, wt pseudoknot (pk) but disruption of hpA; A4/1-7, disruption of pk and hpA; A4/Ts-3 and A4/Ts-4, disruption of pk and mutation of the CP binding site; Wt/Ts-1, strong pk, weak hpA (see also pk5/hp4 Fig. 2A); A4/Ts-1, pk disrupted, weak hpA. (B) Growth of yeast cells transformed with the various constructs. Two independent colonies are grown for each construct. The negative control (−/−) lacked the 3′ terminus of the AMV RNA3. The concentration of 3-AT in the growth medium is indicated, as well as the presence (+) or absence (−) of histidine.