Systematic Nomenclature for GGDEF and EAL Domain-Containing Cyclic Di-GMP Turnover Proteins of Escherichia coli

Abstract

In recent years, Escherichia coli has served as one of a few model bacterial species for studying cyclic di-GMP (c-di-GMP) signaling. The widely used E. coli K-12 laboratory strains possess 29 genes encoding proteins with GGDEF and/or EAL domains, which include 12 diguanylate cyclases (DGC), 13 c-di-GMP-specific phosphodiesterases (PDE), and 4 “degenerate” enzymatically inactive proteins. In addition, six new GGDEF and EAL (GGDEF/EAL) domain-encoding genes, which encode two DGCs and four PDEs, have recently been found in genomic analyses of commensal and pathogenic E. coli strains. As a group of researchers who have been studying the molecular mechanisms and the genomic basis of c-di-GMP signaling in E. coli, we now propose a general and systematic dgc and pde nomenclature for the enzymatically active GGDEF/EAL domain-encoding genes of this model species. This nomenclature is intuitive and easy to memorize, and it can also be applied to additional genes and proteins that might be discovered in various strains of E. coli in future studies.

TEXT

More than 10 years ago, it was demonstrated that GGDEF domains can produce and EAL domains can degrade the bacterial second messenger cyclic di-GMP (c-di-GMP) (1–4). With these assignments, it also became clear that bacterial genomes—in particular, those of gammaproteobacteria—usually contain multiple genes encoding these diguanylate cyclases (DGC) and c-di-GMP phosphodiesterases (PDE) (5, 6). Crystal structures of GGDEF and EAL domains have been elucidated, and studies of structure-function relationships have identified the key amino acid residues required for substrate and cation binding and catalysis (7). This also allowed identification of a subset of GGDEF and EAL (GGDEF/EAL) domain proteins, in which these key amino acids are not conserved, as “degenerate” and enzymatically inactive. In a few cases, it could be demonstrated that these degenerate GGDEF/EAL domain proteins have alternative functions based on direct interactions with other macromolecules (8–11). A subset of proteins combine GGDEF and EAL domains in a single polypeptide, where one domain is usually enzymatically active and the other is degenerate and plays a regulatory role in these “composite” proteins (3). Most GGDEF/EAL domain proteins also contain N-terminal sensory input domains that control their output activities, and a majority are localized or anchored in the cytoplasmic membrane via their membrane-intrinsic or periplasmic sensory domains (12).

In studies of the molecular principles and physiological functions of c-di-GMP signaling, Escherichia coli has served as one of a few model species (13, 14). The commonly used E. coli K-12 laboratory strain has a total of 29 proteins with GGDEF and/or EAL domains, including 12 and 10 proteins featuring the GGDEF and EAL domains alone, respectively, and 7 composite proteins carrying both domains. Based on direct measurements of purified proteins and/or the presence of key conserved amino acids and their elimination by point mutations, DGC and PDE activities can be assigned to 12 and 13 proteins, respectively, whereas 4 of the 29 proteins can be classified as degenerate GGDEF/EAL proteins (Table 1). Genes involved in c-di-GMP signaling have also been studied in pathogenic E. coli, i.e., in uropathogenic E. coli (UPEC) (42), and two additional genes encoding PDEs have been detected in an enterohemorrhagic E. coli (EHEC) strain (38) and a meningitis-associated E. coli strain (39). Together with a recent analysis of genome sequences of 61 E. coli strains (29), which included commensal as well as pathogenic strains of the major pathotypes and phylogroups, a total of two additional GGDEF domain proteins and four more EAL domain proteins have been identified that are not found in E. coli K-12. On the basis of the presence of the key residues involved in enzymatic activities, these proteins should be active DGCs and PDEs (see an accompanying paper [29] in this issue).

TABLE 1.

Novel designations for GGDEF/EAL domain-encoding genes of E. coli

Gene name	b no.	UniProt entry	New gene name	Other gene named	Domain architecturea	Comment and/or reference(s)e
Genes encoding diguanylate cyclases (intact GGDEF domains)
yaiC	b0385	P0AAP1	dgcC	adrA*	MASE2b-GGDEF	12, 15–17
ycdT	b1025	P75908	dgcT		MASE4b-GGDEF	(18)
ydaM	b1341	P77302	dgcM		PAS-PAS-GGDEF	19–21
yddV	b1490	P0AA89	dgcO	dosC	Globin sensor GGDEF	22
ydeH	b1535	P31129	dgcZ		CZB-GGDEF	23
yeaJ	b1786	P76237	dgcJ		GAPES1c-GGDEF	21
yeaP	b1794	P76245	dgcP		GAF-GGDEF
yedQ	b1956	P76330	dgcQ		CHASE7c-xCache-GGDEF	21, 24, 25
yegE	b2067	P38097	dgcE		MASE1b-PAS-PAS-PAS-GGDEF-xEAL	12, 21, 26
yfiN	b2604	P46139	dgcN	tpbB*	CHASE8c-HAMP-GGDEF	26–28
yliF	b0834	P75801	dgcI		GAPES2c-GGDEF
yneF	b1522	P76147	dgcF		xMASE1-GGDEF	Promoter and first 4 TM segments deleted in E. coli K-12 (18)
EC55989_0813		B7LBD9	dgcX		MASE4b-GGDEF	Extra DGC in EAEC (18)
EcSMS35_1716		B1LFF9	dgcY		MASE5b-GGDEF	Extra DGC in E. coli SMS35 and NMEC 07:K1 strain CE10 (29)
Genes encoding c-di-GMP phosphodiesterases (intact EAL domains)
rtn	b2176	P76446	pdeN		CSSc-EAL
yahA	b0315	P21514	pdeL		LuxR-EAL	30, 31
ycgG	b1168	P75995	pdeG		CSSc-EAL
yciR	b1285	P77334	pdeR	gmr	PAS-GGDEF-EAL	19, 20
yddU	b1489	P76129	pdeO	dosP	PAS-PAS-xGAF-xGGDEF-EAL	22, 30, 32, 33
yfeA	b2395	P23842	pdeA		MASE1b-xGGDEF-EAL
yfgF	b2503	P77172	pdeF		MASE1b-xGGDEF-EAL	34, 35
yhjH	b3525	P37646	pdeH		EAL	21, 26, 36
yhjK	b3529	P37649	pdeK	hmsP*	GAPES3c-HAMP-xGGDEF-EAL
yjcC	b4061	P32701	pdeC		CSSc-EAL
ylaB	b0457	P77473	pdeB		CSSc-EAL
yliE	b0833	P75800	pdeI		CHASE9c-xCache-HAMP-xGGDEF-EAL
yoaD	b1815	P76261	pdeD	adrB	CSSc-EAL	37
Z1528		Q8XAQ9	pdeT	vmpA	CSSc-EAL	Extra PDE in EHEC O157:H7 (18, 38)
EcE24377A_E0053		A7ZH68	pdeW		EAL	Extra PDE in ETEC E24377A (29)
ECP_2965		Q707K1	pdeX		EAL	Extra PDE in UPEC 536 (29)
UTI89_C1116		Q1RDG4	pdeY	sfaY	EAL	Extra PDE in several ExPEC strains (29, 39)
Genes encoding proteins with degenerate GGDEF and EAL domains
ycgF	b1163	P75990	bluF		BLUF-xEAL	9, 40
yeaI	b1785	P76236	cdgI		MASE4b-xGGDEF
ydiV	b1707	P76204	rflP		xEAL	Regulator of FlhDC proteolysis (10, 11)
yhdA	b3252	P13518	csrD		GAPES4c-xGGDEF-xEAL	8, 41

^aThe domain names indicate the following Pfam entries: BLUF, PF04940; Cache, PF02743; CHASE7, PF17151; CHASE8, PF17152; CHASE9, PF17153; CSS, PF12792; CZB, PF13682; EAL, PF00563; GAF, PF01590 or PF13492; GAPES1, PF17155; GAPES2, PF17156; GAPES3, PF17154; GAPES4, PF17157; GGDEF, PF00990; globin sensor, PF11563; HAMP, PF00672; LuxR, PF00196; MASE1, PF05231; MASE2, PF05230; MASE4, PF17158; MASE5, PF17178; PAS, PF08448 or PF13426. An “x” in front of a domain name indicates an enzymatically inactive or highly divergent domain. The Pfam entries for new sensor domains are to appear in the 29th release of the Pfam database (48).

^bAn integral membrane domain.

^cPredicted periplasmic domain.

^dAsterisks indicate designations used for homologous genes in other genera or species as follows: adrA, Salmonella (occasionally, adrA has also been used for E. coli); tpbB, Pseudomonas aeruginosa; hmsP, Yersinia.

^eETEC, enterotoxigenic E. coli; ExPEC, extraintestinal pathogenic E. coli; NMEC, neonatal meningitis E. coli; EAEC, enteraggregative E. coli; TM, transmembrane.

Being aware that a systematic nomenclature of the many E. coli genes encoding DGCs and PDEs might eventually be useful, most researchers have refrained from renaming single genes and proteins involved in c-di-GMP signaling in E. coli and have used the preliminary y designations instead, even though these were difficult to memorize and certainly not popular in oral scientific presentations. However, on the basis of the finding that the DGC YdeH is regulated by zinc, it was recently renamed “DgcZ” (23). Also, the newly identified genes encoding DGCs and PDEs in non-K-12 E. coli strains had to be given names, and it seemed obvious to use a dgc and pde nomenclature (29).

Therefore, as a group of researchers who in recent years have worked on the molecular mechanisms and/or the genomic basis of c-di-GMP signaling in E. coli, we now propose a general and systematic dgc and pde nomenclature for the enzymatically active GGDEF/EAL domain-encoding genes of E. coli (Table 1). By using these self-explanatory designations, we also reflect a trend for similar (though not yet systematically used) names for GGDEF/EAL domain proteins in some other species, including Caulobacter, Pseudomonas, Listeria, and Bdellovibrio spp. We are fully aware that nomenclature is a convention and sometimes has to include oversimplifications (for instance, for proteins with multiple functions), but its main function is to allow researchers to remember things and communicate more easily. In detail, the proposed nomenclature is based on the following considerations.

• Following the principle that genes should be named according to the molecular function of the gene product and not according to a mutant phenotype that may be due to a very indirect connection and may represent a functional side effect, “dgc” and “pde” designations are based either on experimentally determined DGC and PDE activities or on the presence of the conserved key amino acids required for these enzymatic functions. These conserved amino acids include the (G/A/S)G(D/E)EF motif in DGCs (43, 44) and the presence of the catalytic glutamyl residue and the main amino acids involved in c-di-GMP and cation binding in PDEs (45–47). The latter criterion reflects pragmatic reasons of feasibility—while isolated and usually soluble EAL domains alone often show PDE activity in vitro, isolated GGDEF domains are usually inactive, which makes measuring DGC activity of membrane-associated DGCs rather challenging. Not only should sensory input domains be integrated into an appropriately reconstituted lipid environment to allow dimerization of the GGDEF domains as a prerequisite for enzymatic activity, but these sensory domains might also need to bind as-yet-unknown ligands in order to promote enzymatic activity.

• The seven composite proteins with both GGDEF and EAL domains can be unequivocally assigned DGC or PDE functions and therefore the corresponding gene designations. In six of the seven cases, one of the two domains is clearly degenerate (Table 1). Only YciR features intact GGDEF and EAL domains, but it has been demonstrated that, in this case, the purified protein shows strong PDE activity in vitro (19, 20), whereas its GGDEF domain binds GTP but has only very minor and, in fact, cryptic DGC activity (20). We therefore propose pdeR as a new gene designation for yciR, with the “R” also seeming appropriate because PdeR is the core component of a regulatory switch that controls the expression of the CsgD major biofilm regulator in E. coli. Thus, PdeR is a multifunctional “trigger protein,” whose ability to bind and degrade c-di-GMP plays a regulatory role as it modulates the direct inhibitory interactions of PdeR with the transcription factor complex that controls csgD expression (20).

• In order to make the transition to the systematic nomenclature easier for people who have been working with these genes and proteins of E. coli, we propose to retain the capital letter currently found in the y designations (e.g., ydaM becoming dgcM, etc.) in as many cases as possible. There is only a single case of overlap—we suggest that yfeA should be renamed pdeA but that yahA should be renamed pdeL (referring to its N-terminal LuxR domain). In the case of ydeH, dgcZ, which alludes to the zinc binding of the sensory domain of the gene product, has already been introduced (23).

• In the few cases where genes are in operons, we propose to use the same capital letter; i.e., we propose that yliF and yliE should become dgcI and pdeI and that ycdT and a pde gene that follows ycdT in certain EHEC strains should become dgcT and pdeT.

• For a few already renamed genes (e.g., genes dosC and dosP and genes vmpA and sfaY), we suggest retaining these names as alternative designations but also reserving systematic names (e.g., genes dgcO and pdeO and genes pdeT and pdeY, respectively) and leaving it to the researchers working with these genes to determine which designation they want to use (for clarity, we suggest also mentioning the other designations in future publications). In addition, Table 1 also includes a few established designations for corresponding homologs in other bacterial species.

• Among the four genes encoding proteins with degenerate GGDEF/EAL domains only, yeaI is the only one that encodes a protein that binds c-di-GMP (F. Skopp and R. Hengge, unpublished data), indicating that this protein serves as a c-di-GMP-binding effector. We therefore propose cdgI as a new designation. For the other three genes, which encode proteins that do not bind c-di-GMP, we suggest that the previously assigned designations that reflect the functions of the encoded proteins should be retained. Thus, ycgF was already renamed bluF (alluding to its blue-light sensing BLUF domain) (9, 40) and yhdA was renamed csrD (as it controls the CsrA/CsrB/CsrC system) (8). Degenerate EAL-only protein YdiV was shown to directly inhibit and promote proteolysis of the flagellar master regulator FlhDC (10, 11), and we therefore propose rflP (regulator of FlhDC proteolysis) as a new gene name.

Besides assigning new systematic names to the genes and proteins involved in c-di-GMP signaling in E. coli, we also use this opportunity to introduce systematic designations for several N-terminal sensory input domains present in some of these proteins that have not been described before (Table 1). In particular, these are (i) two novel MASE (membrane-associated sensor) domains, i.e., MASE4, an eight-transmembrane helix domain found in DgcX and DgcT (YcdT) and the degenerate c-di-GMP-binding protein CdgI (YeaI) (18, 29), and MASE5, a six-transmembrane helix domain present in DgcY (29); (ii) four distinct “GAPES” domains (referring to gammaproteobacterial periplasmic sensory domains), which occur in DgcJ (YeaJ), DgcI (YliF), PdeK (YhjK), and CsrD; and (iii) three novel CHASE (cyclases/histidine kinase-associated sensory) domains present in DgcQ (YedQ), DgcN (YfiN), and PdeI (YliE). In contrast to CHASE domains, GAPES domains seem to be restricted to GGDEF/EAL domain proteins. The molecular functions in c-di-GMP signaling of all of these sensory input domains have yet to be elucidated.

The views expressed in this Commentary do not necessarily reflect the views of the journal or of ASM.

Funding Statement

Research in the laboratory of the corresponding author has been funded by the European Research Council under the European Union's Seventh Framework Programme (ERC-AdG 249780 to R.H.).