All DACs in a Row: Domain Architectures of Bacterial and Archaeal Diadenylate Cyclases

Abstract

Cyclic dimeric AMP (c-di-AMP) is a widespread second messenger that controls such key functions as osmotic homeostasis, peptidoglycan biosynthesis, and response to various stresses. C-di-AMP is synthesized by diadenylate cyclases that contain the DAC (DisA_N) domain, which was originally characterized as the N-terminal domain in the DNA integrity scanning protein DisA. In other experimentally studied diadenylate cyclases, DAC domain is typically located at the protein C termini and its enzymatic activity is controlled by one or more N-terminal domains. As in other bacterial signal transduction proteins, these N-terminal modules appear to sense environmental or intracellular signals through ligand binding and/or protein-protein interactions. Studies of bacterial and archaeal diadenylate cyclases also revealed numerous sequences with uncharacterized N-terminal regions. This work provides a comprehensive review of the N-terminal domains of bacterial and archaeal diadenylate cyclases, including the description of five previously undefined domains and three PK_C-related domains of the DacZ_N superfamily. These data are used to classify diadenylate cyclases into 22 families, based on their conserved domain architectures and the phylogeny of their DAC domains. Although the nature of the regulatory signals remains obscure, the association of certain dac genes with anti-phage defense CBASS systems and other phage-resistance genes suggests that c-di-AMP might also be involved in the signaling of phage infection.

Keywords: c-di-AMP; diadenylate cyclase; homeostasis; osmotic pressure; phage defense; pyruvate kinase allosteric regulation; sensor domain; signal transduction.

FIG 1

Domain organization and genomic neighborhoods of CdaG family diadenylate cyclases in four proteobacteria. (A) Domain organization of the Orf26 protein (GenBank accession AAP70311) from E. coli strain ECOR31 (89), the numbers indicate the domain boundaries. The C-terminal 72-aa fragment includes three putative α-helices (see the AlphaFold-predicted structure of Orf26 at https://alphafold.ebi.ac.uk/entry/Q6XGD6). (B to E) Genomic neighborhoods of four CdaG family diadenylate cyclases. Diadenylate cyclase genes are indicated by red arrows, genes for other oligonucleotide synthases are indicated by green arrows, phospholipase genes are indicated by blue arrows, and genes for predicted endonucleases are indicated by gray arrows. The genes are drawn approximately to size, and their functions are assigned based on the results of HHpred (98) searches. (B) Type II CBASS (cyclic oligonucleotide-based antiphage signaling system) operon from the high-pathogenicity island of the E. coli strain ECOR31 (89). This DNA fragment (GenBank accession AY233333.1, ORF28 to ORF22) encodes patatin-like cUMP-AMP-activated phospholipase CapE; cUMP-AMP synthase CdnE; CdaG-family diadenylate cyclase with DACNG and DACNH domains (DAC, red arrow); VC0178-like cGAMP-activated phospholipase CapV; VC0179-like cGMP-AMP synthase DncV (93); VC0180-like E2-E1 fusion protein Cap2, and VC0181-like putative deubiquitinase (JAB1/MPN domain) Cap3 (see references and for details). This operon is found in a variety of gammaproteobacteria, including more than 120 strains of Klebsiella pneumoniae. Association of DAC domains with SMODS, JAB, E1, and E2 domains has been noted in (99). (C) Genomic DNA fragment from Pseudogulbekiania sp. strain NH8B (GenBank accession AP012224.1, positions 2000384 to 2015929). The indicated ORFs NH8B_1909 to NH8B_1917 encode, respectively, a transcriptional regulator of the LysR family (NH8B_1909), restriction endonuclease of the EcoK Mrr family; SMODS-associated 2TM effector (NH8B_1911); nucleotidyltransferase domain of 2′,5′-oligoadenylate synthetase (OAS); CdaG-family diadenylate cyclase with DACNG and DACNH domains (DAC, red arrow); bifunctional protein with predicted endonuclease (EndoQ) and chromosome partition protein (SMC) domains (NH8B_1914); predicted endonuclease with HNH_5 domain (HNH); phospholipase D superfamily, FAM83 (PF07894) protein (PLDc), and phage defense pattern-recognition ATP-binding protein Avs4 (100). (D) Genomic DNA fragment from Shewanella sp. strain ANA-3 (GenBank accession CP000469.1, positions from 1463938 to 1475392). The indicated ORFs Shewana3_1258 to Shewana3_1250 encode, respectively, a transcriptional regulator of the HTH-XRE family (Shewana3_1258); patatin-like phospholipase A (PLA); CdnE-like nucleotidyltransferase domain; CdaG-family diadenylate cyclase with DACNG and DACNH domains (DAC, red arrow); a protein related to the pre-mRNA-splicing factor CWF11, contains DNA/RNA helicase domains; unknown protein Shewana3_1253; Cap3-like JAB1/MPN domain; IS21 element helper ATPase IstB, and IS21 family transposase IstA. (E) Genomic DNA fragment from Bordetella parapertussis Bpp5 (GenBank accession HE965803.1, positions 3944001 to 3953004). The indicated ORFs BN117_3672 to BN117_3666 encode, respectively, a transcriptional regulator of the HTH-XRE family (BN117_3672); patatin-like phospholipase A (PLA); CdnE-like nucleotidyltransferase domain; CdaG-family diadenylate cyclase with DACNG and DACNH domains (DAC, red arrow); unknown, putative RING-type E3 ubiquitin transferase (BN117_3668); AbiJ-like abortive infection protein, contains N-terminal AbiJ-NTD (PF18863) and HEPNAbiC (PF05168/PF14355) domains (101), and predicted ADP-ribosyltransferase (RES domain).

FIG 2

Predicted structures of the newly described N-terminal domains of diadenylate cyclases. AlphaFold (65) structural predictions for the newly described domains. (A) DACNG domain; (B) DACNH domain; (C) DACND domain; (D) DACNK domain; (E) DACNV domain; (F) DacZ_P domain; (G) DacZ_A domain; (H) DacZ_T domain. The structures are displayed in sequential colors (blue to red) starting from the N-terminal α-helix. The domain structures were extracted from the AlphaFold Protein Structure Database (69) using the protein entries and domain boundaries listed in Table 2.

FIG 3

Maximum-likelihood phylogenetic tree of the DAC domains from various diadenylate cyclase families. Sequences of DAC domains were extracted from the indicated UniProt entries using InterPro-defined boundaries and aligned with MUSCLE (102), generating an alignment with 178 informative sites. The maximum-likelihood phylogenetic tree was inferred using IQ-TREE 2 (103) with selection model LG+I+G4 (automatically chosen by the ModelFinder). Bootstrap values below 60% are not shown. Source organisms, GenBank accession numbers, and genomic locus tags of the selected proteins, as well as DAC domain coordinates, are listed in Table S1a in the supplemental material. DAC family assignments are from Table 3. The protein names are colored based on the taxonomy of the source organisms: Firmicutes, red; Cyanobacteria, cyan; Spirochaetes, orange; Tenericutes, purple; Planctomycetes, Verrucomicrobia, and Chlamydia, brown; Proteobacteria, blue (gammaproteobacteria with light blue shading, betaproteobacteria with gray shading, and deltaproteobacteria with no shading); other bacteria and archaea, black. Yellow shading indicates the CdaG family protein from Nitrosarchaeum koreense, the only thaumarchaeal protein in the current collection.