Evolution of a σ-(c-di-GMP)-anti-σ switch

Abstract

Filamentous actinobacteria of the genus Streptomyces have a complex lifecycle involving the differentiation of reproductive aerial hyphae into spores. We recently showed c-di-GMP controls this transition by arming a unique anti-σ, RsiG, to bind the sporulation-specific σ, WhiG. The Streptomyces venezuelae RsiG-(c-di-GMP)₂-WhiG structure revealed that a monomeric RsiG binds c-di-GMP via two E(X)₃S(X)₂R(X)₃Q(X)₃D repeat motifs, one on each helix of an antiparallel coiled-coil. Here we show that RsiG homologs are found scattered throughout the Actinobacteria. Strikingly, RsiGs from unicellular bacteria descending from the most basal branch of the Actinobacteria are small proteins containing only one c-di-GMP binding motif, yet still bind their WhiG partners. Our structure of a Rubrobacter radiotolerans (RsiG)₂-(c-di-GMP)₂-WhiG complex revealed that these single-motif RsiGs are able to form an antiparallel coiled-coil through homodimerization, thereby allowing them to bind c-di-GMP similar to the monomeric twin-motif RsiGs. Further data show that in the unicellular actinobacterium R. radiotolerans, the (RsiG)₂-(c-di-GMP)₂-WhiG regulatory switch controls type IV pilus expression. Phylogenetic analysis indicates the single-motif RsiGs likely represent the ancestral state and an internal gene-duplication event gave rise to the twin-motif RsiGs inherited elsewhere in the Actinobacteria. Thus, these studies show how the anti-σ RsiG has evolved through an intragenic duplication event from a small protein carrying a single c-di-GMP binding motif, which functions as a homodimer, to a larger protein carrying two c-di-GMP binding motifs, which functions as a monomer. Consistent with this, our structures reveal potential selective advantages of the monomeric twin-motif anti-σ factors.

Keywords: RsiG; Streptomyces; c-di-GMP signaling; protein evolution; second messenger.

Fig. 1.

Conservation and binding of c-di-GMP to single-motif RsiG proteins. (A) Schematic representation of twin-motif RsiG homologs, such as RsiG from S. venezuelae (RsiG_Sv), above sequence logos depicting amino acid sequence conservation in α1 and α5. Residues that form the c-di-GMP binding motifs are highlighted in green. An alignment including all 130 twin-motif RsiG homologs was used to generate the α1 and α5 logos using WebLogo (50). (B) Schematic representation of the single-motif RsiG homologs above an alignment of the sequences of the motif-containing helix from P. medicamentivorans (RsiG_Pm), T. album (RsiG_Ta), C. woesei (RsiG_Cw), R. radiotolerans (RsiG_Rr), and R. xylanophilus (RsiG_Rx). The sequence of the c-di-GMP binding α1 helix from RsiG_Sv is shown below for comparison. Residues that form the c-di-GMP binding motif are marked with asterisks. (C) Representative binding isotherm of RsiG_Cw binding to F-c-di-GMP. (D) Representative binding isotherm of RsiG_Rr binding to F–c-di-GMP. (E) Representative binding isotherm of RsiG_Cw+WhiG_Cw binding to F–c-di-GMP. (F) Representative binding isotherm of RsiG_Rr+WhiG_Rr binding to F-c-di-GMP. Three technical repeats were performed for each curve and the SEs from the three affinities were determined. For each panel, the x axis and y axis show protein concentration in micromolar (μM) and millipolarization units (mP), respectively.

Fig. 2.

Crystal structure of RsiG_Rr. (A) Overlay of RsiG_Rr apo structure homodimeric coiled-coil region (dark magenta and cyan) onto the twin-motif RsiG from S. venezuelae (green). Note disparate conformations of the N- and C-terminal regions. (B) Ribbon diagram of the apo RsiG_Rr (cyan and magenta) with the c-di-GMP binding residues shown as sticks and labeled. Below the structure is the c-di-GMP binding signature motif. (C) Shown in the same orientation as B is the Sv RsiG (green) with its c-di-GMP binding residues shown as sticks and labeled. Also shown as white sticks is the c-di-GMP dimer bound in the Sv RsiG–(c-di-GMP)₂–WhiG complex.

Fig. 3.

Crystal structure of the Rr (RsiG)₂–(c-di-GMP)₂–WhiG complex. (Left) A ribbon diagram of the Rr (RsiG)₂–(c-di-GMP)₂–WhiG complex; (Right) the Sv RsiG–(c-di-GMP)₂–WhiG complex shown in the same orientation. The WhiG molecules are colored green and their σ₂ and σ₄ domains and helices are labeled. The c-di-GMP dimers are shown as sticks. The monomeric twin-motif RsiG_Sv is colored magenta. For the homodimeric single-motif RsiG_Rr one subunit is colored magenta and the other yellow.

Fig. 4.

Contacts to c-di-GMP and WhiG_Rr from the homodimeric RsiG_Rr and flexibility of the WhiG_Rr binding regions of RsiG_Rr. (A) Close-up of the c-di-GMP binding region of the RsiG_Rr homodimer with one subunit colored yellow and the other magenta. WhiG_Rr is green. Residues that contact the c-di-GMP dimer are labeled. (B) Overlay of individual RsiG_Rr subunits from the WhiG-bound complex (magenta and yellow) and from apo RsiG_Rr (cyan), revealing that only the coiled-coil helices are similarly structured. In contrast, the N- and C-terminal extensions adopt distinct conformations upon binding the WhiG_Rr σ₂ and σ₄ domains. (C) Interactions between RsiG_Rr and the σ₂ domain of WhiG_Rr. (D) Interactions between RsiG_Rr and the σ₄ domain of WhiG_Rr.

Fig. 5.

Distribution of RsiG homologs in the phylum Actinobacteria. A maximum-likelihood phylogeny of 378 representative Actinobacterial species is shown, based on 37 concatenated housekeeping genes that were identified and aligned using PhyloSift (51). Sequences derived from five Chloroflexi genomes were used to root the tree (indicated in black). Genomes possessing an RsiG homolog are indicated by colored boxes, with red boxes signifying the presence of homodimeric RsiG homologs (with a single c-di-GMP binding motif), and blue boxes signifying the presence of monomeric RsiG homologs (with two c-di-GMP binding motifs). Major taxonomic groups with at least two representatives in which an RsiG homolog is found in >80% of genomes are indicated by the gray arcs. Tree scale is substitutions per site. A version of this tree that includes full taxonomic labeling and node support values can be found in SI Appendix, Fig. S8.

Fig. 6.

Predicted WhiG target promoters in R. radiotolerans and organization of the R. radiotolerans type IV pilus gene cluster, showing the positions of predicted WhiG target promoters. (A) Twenty matches to the well-established flagellar promoter consensus sequence (−35 TAAA; −10 GCCGATAA) (27) were identified bioinformatically in the intergenic regions of the R. radiotolerans genome, lying within 200 bp of a downstream start codon, and allowing for up to two base mismatches in total. Of these 20 sequences, the 12 shown were found to sit just upstream of appropriately positioned transcription start sites, demonstrating that they represent genuine promoters. Transcription start sites (TSS), shown in blue, were determined as part of a genome-wide 5′ triphosphate end-capture transcription start site mapping experiment. Putative −10 and −35 sequences are shown in red. Target genes with predicted functions in type IV pilus biosynthesis or c-di-GMP turnover are highlighted in purple. The logo based on the sequence alignment was created using Weblogo (50). (B) The genes in the type IV pilus gene cluster are shown as schematics with the predicted gene products listed below with the same color coding.