Date Published: April 01, 2017
Publisher: International Union of Crystallography
Author(s): Kimberly A. Stanek, Jennifer Patterson-West, Peter S. Randolph, Cameron Mura.
The structure of an Hfq homolog from the deep-branching thermophilic bacterium Aquifex aeolicus, determined to 1.5 Å resolution both in the apo form and bound to a uridine-rich RNA, reveals a conserved, pre-organized RNA-binding pocket on the lateral rim of the Hfq hexamer.
The bacterial protein Hfq, initially identified as an Escherichia coli host factor required for the replication of RNA bacteriophage Qβ (Franze de Fernandez et al., 1968 ▸, 1972 ▸), is now known to play a central role in the post-transcriptional regulation of gene expression and mRNA metabolism (Vogel & Luisi, 2011 ▸; Sauer, 2013 ▸; Updegrove et al., 2016 ▸). Hfq has been linked to many RNA-regulated cellular pathways, including stress response (Sledjeski et al., 2001 ▸; Zhang et al., 2002 ▸; Fantappie et al., 2009 ▸), quorum sensing (Lenz et al., 2004 ▸) and biofilm formation (Mandin & Gottesman, 2010 ▸; Mika & Hengge, 2013 ▸). The diverse cellular functions of Hfq stem from its fairly generic role in binding small, noncoding RNAs (sRNAs) and facilitating base-pairing interactions between these regulatory sRNAs and target mRNAs. A given sRNA might either upregulate (Soper et al., 2010 ▸) or downregulate (Ikeda et al., 2011 ▸) one or more target mRNAs via distinct mechanisms. For example, the sRNA RhyB downregulates several Fur-responsive genes under iron-limiting conditions (Masse & Gottesman, 2002 ▸), whereas the DsrA, RprA and ArcZ sRNAs stimulate translation of rpoS mRNA, encoding the stationary-phase σs factor (Soper et al., 2010 ▸). In general, Hfq is required for cognate sRNA·mRNA pairings to be productive, and abolishing Hfq function typically yields pleiotropic phenotypes, including diminished viability (Fantappie et al., 2009 ▸; Vogel & Luisi, 2011 ▸).
The organism A. aeolicus belongs to the taxonomic order Aquificales, in the phylum Aquificae, within what may be the most phylogenetically ancient and deeply branching lineage of the Bacteria. Thus, this species offers a potentially informative context in which to examine the evolution of sRNA-based regulatory systems, such as those built upon Hfq. The Aae genome contains an open reading frame with detectable sequence similarity to characterized Hfq homologs (e.g. from E. coli and other proteobateria), and an RNomics/deep-sequencing study has shown that, upon heterologous expression in the γ-proteobacterium Salmonella enterica, this putative Hfq homolog can immunoprecipitate host sRNAs (Sittka et al., 2009 ▸). Sequence analysis confirms that this putative Hfq can be identified via database searches (Fig. 1 ▸), and that this homolog exhibits enhanced residue conservation at sequence positions that correspond to the three RNA-binding sites on the surface of Hfq, proximal, distal and lateral rim, denoted in the consensus line in Fig. 1 ▸. As the first step in our crystallographic studies, we cloned, expressed and purified recombinant Aae Hfq: in these initial experiments, Aae Hfq generally resembled hitherto characterized Hfq homologs in terms of biochemical properties (e.g. resistance to chemical and thermal denaturation, and hexamer formation).
The apo form of Aae Hfq, refined to 1.49 Å resolution in space group P1, reveals a dodecamer comprised of two hexamers in a head-to-tail orientation. The individual subunits of Aae Hfq are similar in structure, with a mean pairwise r.m.s.d. of less than ∼0.3 Å for all monomer backbone atoms. The largest differences among the 13 independently refined Hfq monomer structures (12 in P1, one in P6) occur in the N-terminal and L4 loop regions; notably, these are the two regions that mediate much of the interface between rings (distal⋯proximal face contacts in Fig. 5 ▸), as well as the intermolecular contacts between dodecamers across the lattice. The patterns of structural differences are also captured in the symmetric matrix of pairwise r.m.s.d.s between chains: hierarchical clustering on this distance matrix results in the monomers that comprise the PE (chains A–F) and DE (chains G–L) hexameric rings partitioning into two distinct groups (Fig. 6 ▸c, Supplementary Fig. S5c).