Showing papers by "Daniel H. Haft published in 2023"

Myxosortases process MYXO-CTERM and other bacterial C-terminal protein-sorting signals that have invariant Cys residues

Abstract: The LPXTG protein-sorting signal, found in surface proteins of various Gram-positive pathogens, was the founding member of a growing panel of prokaryotic small C-terminal sorting domains. Sortase A (SrtA) cleaves LPXTG, exosortases (XrtA and XrtB) cleave the PEP-CTERM sorting signal, archaeosortase A (ArtA) cleaves PGF-CTERM, and rhombosortase (RrtA) cleaves GlyGly-CTERM domains. Three sorting signal domains without previously known processing proteases are the MYXO-CTERM, JDVT-CTERM, and SYNERG-CTERM domains. These exhibit the standard tripartite architecture of short signature motif, then a hydrophobic transmembrane segment, then an Arg-rich cluster. Each has an invariant cysteine in its signature motif. Here, we show computational evidence that these three Cys-containing sorting signals are processed by corresponding subfamilies of glutamic-type intramembrane proteases, related to type II CAAX-processing proteases found in eukaryotes. We name these sorting enzymes generally as myxosortases, and identify MXAN_2755 from Myxococcus xanthus as MrtX (myxosortase X). Additional myxosortases families MrtC and MrtP have radically different N-terminal domains, suggesting most myxosortases act as bifunctional enzymes. Myxosortase-like processing enzymes are identified also for the JDVT-CTERM (MrtJ) and SYNERG-CTERM (MrtS). This work establishes a major new family of protein-sorting housekeeping enzymes for the surface attachment of proteins on bacterial outer membranes.

Eight Unexpected Selenoprotein Families in Organometallic Biochemistry in Clostridium difficile, in ABC Transport, and in Methylmercury Biosynthesis

Abstract: This work adds 8 novel prokaryotic selenoproteins to the 80 or so families previously known. It describes the SAO (selenocysteine-assisted organometallic) locus, with the most selenoproteins of any known system. ABSTRACT The bioinformatics of a nine-gene locus, designated selenocysteine-assisted organometallic (SAO), was investigated after identifying six new selenoprotein families and constructing hidden Markov models (HMMs) that find and annotate members of those families. Four are selenoproteins in most SAO loci, including Clostridium difficile. They include two ABC transporter subunits, namely, permease SaoP, with selenocysteine (U) at the channel-gating position, and substrate-binding subunit SaoB. Cytosolic selenoproteins include SaoL, homologous to MerB organomercurial lyases from mercury resistance loci, and SaoT, related to thioredoxins. SaoL, SaoB, and surface protein SaoC (an occasional selenoprotein) share an unusual CU dipeptide motif, which is something rare in selenoproteins but found in selenoprotein variants of mercury resistance transporter subunit MerT. A nonselenoprotein, SaoE, shares homology with Cu/Zn efflux and arsenical efflux pumps. The organization of the SAO system suggests substrate interaction with surface-exposed selenoproteins, followed by import, metabolism that may cleave a carbon-to-heavy metal bond, and finally metal efflux. A novel type of mercury resistance is possible, but SAO instead may support fermentative metabolism, with selenocysteine-mediated formation of organometallic intermediates, followed by import, degradation, and metal efflux. Phylogenetic profiling shows SOA loci consistently co-occur with Stickland fermentation markers but even more consistently with 8Fe-9S cofactor-type double-cubane proteins. Hypothesizing that the SAO system forms organometallic intermediates, we investigated the known methylmercury formation protein families HgcA and HgcB. Both families contained overlooked selenoproteins. Most HgcAs have a CU motif N terminal to their previously accepted start sites. Seeking additional rare and overlooked selenoproteins may help reveal more cryptic aspects of microbial biochemistry. IMPORTANCE This work adds 8 novel prokaryotic selenoproteins to the 80 or so families previously known. It describes the SAO (selenocysteine-assisted organometallic) locus, with the most selenoproteins of any known system. The rare CU motif recurs throughout, suggesting the formation and degradation of organometallic compounds. That suggestion triggered a reexamination of HgcA and HcgB, which are methylmercury formation proteins that can adversely impact food safety. Both are selenoproteins, once corrected, with HgcA again showing a CU motif. The SAO system is plausibly a mercury resistance locus for selenium-dependent anaerobes. But instead, it may exploit heavy metals as cofactors in organometallic intermediate-forming pathways that circumvent high activation energies and facilitate the breakdown of otherwise poorly accessible nutrients. SAO could provide an edge that helps Clostridium difficile, an important pathogen, establish disease.

In silico discovery of the myxosortases that process MYXO-CTERM and three novel prokaryotic C-terminal protein-sorting signals that share invariant Cys residues

Abstract: The LPXTG protein-sorting signal, found in surface proteins of various Gram-positive pathogens, was the founding member of a growing panel of prokaryotic small C-terminal sorting domains. Sortase A (SrtA) cleaves LPXTG, exosortases (XrtA and XrtB) cleave the PEP-CTERM sorting signal, archaeosortase A (ArtA) cleaves PGF-CTERM, and rhombosortase (RrtA) cleaves GlyGly-CTERM domains. Four sorting signal domains without previously known processing proteases are the MYXO-CTERM, JDVT-CTERM, Synerg-CTERM, and CGP-CTERM domains. These exhibit the standard tripartite architecture of short signature motif, then a hydrophobic transmembrane segment, then an Arg-rich cluster. Each has an invariant cysteine in its signature motif. Computational evidence strongly suggests that each of these four Cys-containing sorting signals is processed, at least in part, by a cognate family of glutamic-type intramembrane endopeptidases, related to eukaryotic type II CAAX-processing protease Rce1. For the MYXO-CTERM sorting signals of different lineages, their sorting enzymes, called myxosortases, include MrtX (MXAN_2755 in Myxococcus xanthus), MrtC, and MrtP, all with radically different N-terminal domains but with a conserved core. Predicted cognate sorting enzymes were identified also for JDVT-CTERM (MrtJ), Synerg-CTERM (MrtS), and CGP-CTERM (MrtA). This work establishes a major new family of protein-sorting housekeeping endopeptidases contributing to surface attachment of proteins in prokaryotes. Importance Homologs of the eukaryotic type II CAAX-box protease Rce1, a membrane-embedded endopeptidase found in yeast and human ER and involved in sorting proteins to their proper cellular locations, are abundant in prokaryotes but are not well understood there. This bioinformatics paper identifies several subgroups of the family as cognate endopeptidases for four protein-sorting signals processed by previously unknown machinery. Sorting signals with newly identified processing enzymes include three novel ones, but also MYXO-CTERM, which had been the focus of previous experimental work in the model fruiting and gliding bacterium Myxococcus xanthus. The new findings will substantially improve our understanding of Cys-containing C-terminal protein-sorting signals and of protein trafficking generally in bacteria and archaea.