Studying GGDEF Domain in the Act: Minimize Conformational Frustration to Prevent Artefacts

life

Article Studying GGDEF Domain in the Act: Minimize Conformational Frustration to Prevent Artefacts

Federico Mantoni 1,†, Chiara Scribani Rossi 1,2,†, Alessandro Paiardini 1,2, Adele Di Matteo 3 , Loredana Cappellacci 4 , Riccardo Petrelli 4 , Massimo Ricciutelli 4, Alessio Paone 1,2, Francesca Cutruzzolà 1,2, Giorgio Giardina 1,2,* and Serena Rinaldo 1,2,*

1 Department of Biochemical Sciences “A. Rossi Fanelli”, Sapienza University of Rome, Piazzale Aldo Moro, 5, 00185 Rome, Italy; [email protected] (F.M.); [email protected] (C.S.R.); [email protected] (A.P.); [email protected] (A.P.); [email protected] (F.C.) 2 Laboratory Afﬁliated to Istituto Pasteur Italia—Fondazione Cenci Bolognetti, 00185 Rome, Italy 3 Institute of Molecular Biology and Pathology, CNR, Piazzale Aldo Moro, 5, 00185 Rome, Italy; [email protected] 4 Medicinal Chemistry Unit, School of Pharmacy, University of Camerino, Via S. Agostino 1, 62032 Camerino, Italy; [email protected] (L.C.); [email protected] (R.P.); [email protected] (M.R.) * Correspondence: [email protected] (G.G.); [email protected] (S.R.) † These authors contributed equally to the work.

Abstract: GGDEF-containing proteins respond to different environmental cues to ﬁnely modulate cyclic diguanylate (c-di-GMP) levels in time and space, making the allosteric control a distinctive trait of the corresponding proteins. The diguanylate cyclase mechanism is emblematic of this control:

two GGDEF domains, each binding one GTP molecule, must dimerize to enter catalysis and yield c-di-GMP. The need for dimerization makes the GGDEF domain an ideal conformational switch in

Citation: Mantoni, F.; Scribani Rossi, multidomain proteins. A re-evaluation of the kinetic profile of previously characterized GGDEF C.; Paiardini, A.; Di Matteo, A.; domains indicated that they are also able to convert GTP to GMP: this unexpected reactivity occurs Cappellacci, L.; Petrelli, R.; Ricciutelli, when conformational issues hamper the cyclase activity. These results create new questions regarding M.; Paone, A.; Cutruzzolà, F.; the characterization and engineering of these proteins for in solution or structural studies. Giardina, G.; et al. Studying GGDEF Domain in the Act: Minimize Keywords: enzyimatic assay; GGDEF; c-di-GMP; allostery; protein engineering Conformational Frustration to Prevent Artefacts. Life 2021, 11, 31. https://doi.org/10.3390/life11010031 1. Introduction Received: 1 December 2020 Accepted: 3 January 2021 Bacterial biofilm depends on the modulation of cyclic diguanylate (c-di-GMP) levels Published: 6 January 2021 in response to environmental cues. This second messenger is also involved in bacterial signal transduction, regulation, virulence and antibiotic resistance. It is synthesized Publisher’s Note: MDPI stays neu- by diguanylate cyclases (DGCs) harboring the GGDEF domain, starting from two GTP tral with regard to jurisdictional clai- molecules. Conversely, it is degraded by dedicated phosphodiesterases (PDEs), belonging 0 0 0 ms in published maps and institutio- to two families namely EAL and HD-GYP, to 5 -phosphoguanylyl-(3 -5 )-guanosine (pGpG). nal affiliations. Moreover, the HD-GYP family is also able to carry out a second hydrolytic cleavage to produce GMP through the so-called PDE-B activity, via the c-di-GMP → pGpG → GMP pathway (Scheme1)[ 1]. GGDEF, EAL and HD-GYP are named according to the sequence signature in the active site. These enzymatic domains are very often fused to a plethora Copyright: © 2021 by the authors. Li- of regulatory domains (e.g., REC; GAF; PAS; PAC; HAMP), and a large number of ho- censee MDPI, Basel, Switzerland. mologous but non-redundant multidomain enzymes per species exist, whose response to This article is an open access article environmental stimuli is either direct or mediated by protein-protein interaction with other distributed under the terms and con- sensors [1]. Many receptors able to sense the cellular levels of the dinucleotide have been ditions of the Creative Commons At- tribution (CC BY) license (https:// identified, including transcription factors, PilZ domains, degenerate DGCs or PDEs, and creativecommons.org/licenses/by/ also riboswitches [2]. 4.0/).

Life 2021, 11, 31. https://doi.org/10.3390/life11010031 https://www.mdpi.com/journal/life LifeLifeLife 20212021 2021,, 1111,,, 11x 31 FOR, x FOR PEER PEER REVIEW REVIEW 22 of of2 13 13of 13

SchemeSchemeScheme 1. 1.Synthesis Synthesis 1. Synthesis and and degradationand degradation degradation of of cyclic cyclicof cyclic diguanylate diguanylate diguanylate (c-di-GMP). (c- di(c-GMP).di-GMP).

TheTheThe variety variety variety of domainof domain architecturesarchitectures architectures inin c-di-GMP cin-di c--diGMP-GMP metabolizing metabolizing metabolizing enzymes enzymes enzymes and and theand the hetero- the het- het- erogeneitygeneityerogeneity of itsof receptors,ofits itsreceptors, receptors, allow allow bacteriaallow bacteria bacteria to react to toreact promptly react promptly promptly to many to tomany different many different different environmental environ- environ- mentalconditionsmental conditions conditions by modulating by bymodulating modulating their c-di-GMP their their c-di c based--diGMP-GMP adaptation based based adaptation adaptation strategy. strategy. As strategy. a consequence, As As a conse- a conse- it quence,hasquence, been it very hasit has been hard been very to very draw hard hard a uniqueto todraw draw mechanism a unique a unique mechanism for mechanism the c-di-GMP-mediated for for the the c-di c--diGMP-GMP regulation.-mediated-mediated regulation.Theregulation. only general The The only rule only general that general has rule emerged rule that that has to has emerged date, emerged is that to date,to the date, c-di-GMP is that is that the dependentthe c-di c--diGMP-GMP modulationdependent dependent modulationofmodulation protein function of proteinof protein occurs function function mainly occurs viaoccurs restriction mainly mainly via ofvia restriction the restriction possible of theof conformations the possible possible conformations conformations accessible to accessibletheseaccessible multi-domain to theseto these multi ormulti multi-subunit-domain-domain or ormulti proteinsmulti-subunit-subunit [3,4 ].proteins proteins [3,4]. [3,4]. TheTheThe GGDEF GGDEF GGDEF domain domain domain of of DGCsof DGCsDGCs is emblematic is emblematic of thisof of this thiscontrol; control; control; since since since only only onlyone one GTP one GTP mol- GTP mol- eculemoleculeecule can can bind can bind bindin thein inthe active the active active site, site, a site, dimerization a dimerization a dimerization step step is step mandatoryis mandatory is mandatory for for the forthe cyclization thecyclization cyclization reac- reac- tionreactiontion to to occur to occur occur and and and yield yield yield c- di c c-di-GMP.--diGMP.-GMP. Consequently, Consequently, Consequently, multidomain multidomain multidomain DGCs DGCs DGCs usually usually usually regulate regulate regulate theirtheirtheir activity activity activity by by byallowing allowing allowing the the the GGDEF GGDEF GGDEF domains domains domains to to dimerizeto dimerize dimerize (active (active (active dimer) dimer) dimer) or or orforc forcing forcinging them them them apartapartapart (inactive (inactive (inactive conformations), conformations), conformations), functioning functioning functioning as as asconformational conformational conformational switches switches switches (Scheme (Scheme (Scheme 22).). 2).

Scheme 2. (A) In order to yield c-di-GMP, two GTP molecules must come in close proximity. Since SchemeScheme 2. 2.(A ()A In) In order order to to yield yield c-di-GMP, c-di-GMP, two two GTP GTP molecules molecules must must come come in in close close proximity. proximity. Since Since eacheach GGEDF GGEDF domain domain binds binds only only one one GTP GTP molecule, molecule, for forthe the condensatio condensation reactionn reaction to occurto occur two two each GGEDF domain binds only one GTP molecule, for the condensation reaction to occur two GGDEFGGDEF domain domain must must dimerize dimerize as shownas shown in panelin panel (B). ( BThe). The GGDEF GGDEF containing containing enzymes enzymes respond respond to to GGDEF domain must dimerize as shown in panel (B). The GGDEF containing enzymes respond to environmentalenvironmental stimuli stimuli by byswitching switching from from an anactive active conformation conformation (B) ( toB) anto aninactive inactive one one in whichin which theenvironmentalthe two two GGDEF GGDEF stimuli domain domain by are switching are not not able able fromto interactto aninteract active (panel ( conformationpanel (C )).(C )). (B) to an inactive one in which the two GGDEF domain are not able to interact (panel (C)). OnOn the the other other hand, hand, although although in inprinciple principle the the hydrolysis hydrolysis of ofc-di c--diGMP-GMP could could be beper- per- On the other hand, although in principle the hydrolysis of c-di-GMP could be per- formedformed by bya single a single EAL EAL domain, domain, EAL EAL containing containing proteins proteins have have been been also also shown shown to func-to func- formed by a single EAL domain, EAL containing proteins have been also shown to function tiontion as asdimers dimers and, and, as asthe the GGDEF GGDEF domains, domains, to functionto function as asconformational conformational switches switches [5– [58].– 8]. as dimers and, as the GGDEF domains, to function as conformational switches [5–8]. Fi- Finally,Finally, the the GGDEF GGDEF domain domain can can also also be befound found fused fused at theat the N- Nterminus-terminus of anof anEAL EAL domain, domain, nally, the GGDEF domain can also be found fused at the N-terminus of an EAL domain, in

Life 2021, 11, 31 3 of 13

the so-called hybrid proteins which potentially could display opposite catalytic activities. However, in these intriguing enzymes, usually only one of the two domains is active, with the inactive domain serving as allosteric regulator. In this scenario, it has been often reported that the GGDEF domain may control the catalytic properties of the downstream EAL domain of hybrid proteins [5,8–10]. The GGDEF-containing proteins represent a hot target to interfere with biofilm formation [11], including the rationale approach targeting the active or the allosteric sites [12–17]. However, to improve the target specificity, novel functional and structural data are required. From a biochemical point of view, working with such multi-domain proteins in vitro poses several issues which need to be taken into careful consideration in order to achieve a reliable characterization. Most importantly, given the many regulatory and/or transmembrane domains associated to DGCs, PDEs or hybrid proteins, these are often poorly soluble and shorter constructs that must be expressed to study them in vitro. Moreover, truncation of one or more domains may result in a modification of the allosteric regulation, whose characterization is the main goal of the study. Consequently, even when the design of the constructs is cautiously done, misleading results may be obtained. Here we report a re-evaluation of the kinetic profile of selected hybrid proteins as truncated catalytic constructs, showing that the GGDEF domain can slowly convert GTP to GMP as a side reaction. We also show that different GGDEF-containing proteins display this unexpected reactivity, and that this side reaction occurs when the physiological cyclase activity does not take place due to conformational or dimerization issues. We found that when using GTP as a substrate the GGDEF domain may catalyze opposite reactions (condensation or hydrolysis) depending on the presence of upstream domain(s) or on the assay conditions. This variability may bias the kinetic characterization of these proteins and raises some concerns on the design of soluble constructs of GGDEF-containing proteins. This work aims at highlighting these peculiar properties of the GGDEF domain and to propose some useful guidelines to minimize the structural and kinetics bias due to a possible disruption of the GGDEF conformational constraints in engineered constructs.

2. Materials and Methods 2.1. Protein Expression and Purification RmcA-DUAL construct and site-directed mutant was obtained and purified as previously published [5]. Synthetic genes PA0861 (Rbda, residues 373–818) and PA4601 (MorA, residues 970–1409) from P. aeruginosa harboring only the GGDEF and the C-terminal EAL domains were purchased from Life Technologies and subcloned in Pet28a vectors with an N- terminal His-tag between NdeI and XhoI/HindIII. YfinN GGDEF and PleD were obtained as reported in [18,19], respectively. Purification of each construct was done as reported below, and the corresponding buffers for each construct were reported in Table S1. All constructs were overexpressed in the Escherichia coli BL21 (DE3) strain. Freshly transformed colonies were grown at 37 ◦C in Luria-Bertani (LB) liquid medium supplemented with 30 µg/mL Kanamycin (Merck KGaA-Germany). At OD600 ~0.8, protein expression was induced by adding 0.5 mM IPTG (isopropyl B-d-thiogalactoside; Merck KGaA-Germany) and the temperature was reduced to 20 ◦C. Cells were harvested by centrifugation after 20 h and stored at −20 ◦C. All bacterial pellets were suspended in the lysis buffer and then they were lysed by sonication. The supernatant was separated on a Ni2+-IMAC HisTrap column (GE Healthcare, Chicago, IL, USA) equilibrated in Buffer A (Table S1), specific for each protein. Elution was carried out by increasing the imidazole concentration, with the proteins eluting at 300 mM imidazole. Fractions containing pure protein were pooled, imidazole was removed with PD-10 desalting columns (GE Healthcare, Chicago, IL, USA) and concentrated with Ultracel Amicon® concentrators (Merck KGaA-Germany). Monodisperse protein was obtained by FPLC, according to their molecular weights, and eluted with buffer A using an FPLC apparatus (AKTA system, GE Healthcare, Chicago, IL, USA) (Table S1). For PleD, multiple FPLC runs were done on diluted protein to allow c-di-GMP to dissociate on the column and then the collected dimer peak was concentrated. Life 2021, 11, 31 4 of 13

Purified proteins were flash-frozen in liquid nitrogen and stored at −20 ◦C and the protein content was evaluated with BCA assay (Merck KGaA-Germany) and spectroscopically. When affordable the Molar Extinction coefficient (ε) was calculated by BCA.

2.2. Kinetic Assay GTPase Activity The (α-β)-GTPase activity (hereinafter GTPase) of different RmcA derived constructs as well as of PleD, YfiN, RbdA and MorA was analyzed using Reverse Phase HPLC (RP- HPLC) as reported in [19]. Small variations in peak migration are ascribed to the difference in the room temperature during RP-HPLC runs; purified standards were run routinely to set the peak (data not shown). RmcA derived constructs and YfiN (10, 8 or 5 µM, as indicated) in 20 mM TrisHCl pH 8.0, 100 mM NaCl and 2.5 mM MnCl2 were incubated with 100 µM GTP and the reaction mixture was kept at 25 ◦C for different time intervals. No activity was observed when Zn2+ or Ca2+ were used as divalent cations (data not shown). The rates of hydrolase activity were also assayed in the presence of 30 or 60 µM c-di-GMP, if indicated. To test the putative GTPase activity of PleD, 5 µM of enzyme in 20 mM TrisHCl pH 8, 100 mM NaCl and 10 mM MgCl2 was incubated for 10 min at room temperature. To populate the active DGC form, if indicated, the enzyme was incubated at room temperature for 30 min with 10 mM NaF and 1 mM BeCl2. All the reactions, which were carried out with different enzymes, were stopped as described in [5]. All the experiments were done at least in triplicate; a sample experiment is reported in figures. Rates have been obtained as the linear fit of the initial time course and are the mean value ± SD. As a control, the reaction with the sole GTP was also carried out to monitor eventual auto hydrolysis of GTP; no degradation (and therefore GMP accumulation) was observed (data not shown).

2.3. Optimization of GMP Separation by RP-HPLC for Mass Spectrometry To assess the identity of the GTPase activity product, we optimized the RP-HPLC separation. We separated the unknown species using a 150 × 4.6 mm reverse phase column (Prevail C8, Grace Davison Discovery Science, particle size of 5 µm). Due to the incompatibility of 100 mM phosphate buffer with MS analysis, catalytic products were eluted with 0.1% Formic Acid/Methanol (98/2, v/v, 0.3 mL/min) a mobile phase, setting the UV detector at 254 nm. The product eluted was lyophilized and Mass Spectrometry analysis was done. HPLC/DAD/ESI-TRAP studies were performed using an Agilent 1100 series and an Ion Trap LC/MSD Trap SL G2445D from Agilent Technologies (Santa Clara, CA, USA) equipped with an ESI source operating in negative ionization mode.

2.4. Crystallography

YfiNGGDEF crystals were obtained by vapor diffusion methods (sitting drop) by mixing equal volumes of 2.5 mg/mL protein solution (buffer: 250 mM NaCl 10 mM Tris pH 8.0, 2% glycerol, 5 mM MgCl2) with precipitant solution (0.2 M ammonium acetate, 0.1 M Tri- sodium Citrate pH 5.5, 30% w/v PEG 4K). Before flash freezing in liquid nitrogen, crystals were soaked in a solution of mother liquor containing 20% glycerol as cryoprotectant. A 360◦ data-set (oscillation of 0.1◦) was collected at the BESSY synchrotron on ID-14.1 beamline, at 0.918 wavelength. The best crystal diffracted to 2.8 Å resolution. Data were integrated using XDS [20]. Scaling, conversion of intensities to amplitudes, systematic absence analysis and free flag (10%) assignment was performed using Aimless [21] within the CCP4 [22] software suite. Initial phases were obtained by molecular replacement using the PDB entry 4IOB [18] as search model in MOLREP [23] Cycles of model building and refinement were done using COOT [24] and Refmac-v5.8 [25] respectively. Data collection and refinement statistics are shown in Table1. Coordinates and structure factors were deposited in the protein data bank with accession code: 7A7E. Life 2021, 11, 31 5 of 13

Table 1. Data collection and reﬁnement statistics.

Data Collection PA1120-E232start a Beamline Bessy ID14.1

Space group P212121 79.64 85.10 126.97 Cell dimension; a b c (Å), α β γ (◦) 90.00 90.00 90.00 Resolution range (Å) 49.64–2.80 (2.95–2.80) 1 b Half-set correlation CC( 2 ) (%) 99.6 (68.2) 10.5 (1.6) Completeness (%) 100 (100) N. unique reflection (total) 21,951 (3175) Multiplicity 13.2 (13.5) Wilson B-factor (Å2) 56.1 Refinement Resolution range (Å) 49.0–2.80 N◦ of unique reflections 19,705

Rfree test set 2195 reﬂections (10.2%)

Rwork/Rfree (%) 21.96/22.36

Fo,Fc correlation 0.94 Total N. of atoms 3809 Average B-factor all atoms (Å2) 64.3 Bonds RMSD Length (Å) 0.007 Angle (◦) 1.520 Ramachandran plot (n. residues, %) Favoured 485 (94.4) Allowed 29 (5.6) a Values in parentheses refer to the highest-resolution shell; b Percentage of correlation between intensities from random half-data sets.

3. Results 3.1. The GGDEF Domain of RmcA Is Able to Hydrolyze GTP We recently characterized the catalytic moiety of RmcA from the P. aeruginosa strain PAO1, a transmembrane GGDEF-EAL containing phosphodiesterase able to recognize L- Arginine in the periplasmic space [26] and to degrade c-di-GMP [5]. We demonstrated that the GGDEF domain binds GTP and allosterically controls the phosphodiesterase activity by allowing the EAL domain to bind to c-di-GMP and dimerize [5]. Given that the GGDEF domain can bind GTP and metals as a genuine DGC, we decided to re-evaluate possible activities of GGDEF occurring in parallel to the characterized PDE one of the GGDEF-EAL (hereinafter DUAL) construct. Therefore, we analyzed the activity of DUAL in the presence of the sole GTP (excess, 100 µM) at protein concentration higher than that previously used (5–8 µM DUAL rather than 1 µM). Under these experimental conditions, where pseudo- ﬁrst order is still guaranteed, two different species accumulate: an unexpected early one (Figure1A), found to be GMP by mass spectrometry (Supplementary Figure S1) and a late one (appreciable after many hours) corresponding to pGpG (Figure1A and Supplementary Figure S1, panel D). The production of pGpG likely belongs to a residual DGC activity of the GGDEF domain, which slowly forms c-di-GMP, rapidly converted into pGpG by the EAL domain. Life 2021, 11, x FOR PEER REVIEW 6 of 13

3. Results 3.1. The GGDEF Domain of RmcA Is Able to Hydrolyze GTP We recently characterized the catalytic moiety of RmcA from the P. aeruginosa strain PAO1, a transmembrane GGDEF-EAL containing phosphodiesterase able to recognize L- Arginine in the periplasmic space [26] and to degrade c-di-GMP [5]. We demonstrated that the GGDEF domain binds GTP and allosterically controls the phosphodiesterase activity by allowing the EAL domain to bind to c-di-GMP and dimerize [5]. Given that the GGDEF domain can bind GTP and metals as a genuine DGC, we decided to re-evaluate possible activities of GGDEF occurring in parallel to the characterized PDE one of the GGDEF-EAL (hereinafter DUAL) construct. Therefore, we analyzed the activity of DUAL in the presence of the sole GTP (excess, 100 µM) at protein concentration higher than that previously used (5–8 μM DUAL rather than 1 μM). Under these experimental conditions, where pseudo-first order is still guaranteed, two different species accumulate: an unexpected early one (Figure 1A), found to be GMP by mass spectrometry (Supplementary Figure S1) and a late one (appreciable after many hours) corresponding to pGpG (Figure 1A and Supplementary Figure S1, panel D). The production of pGpG likely belongs to a residual DGC activity of the GGDEF domain, which slowly forms c-di-GMP, rapidly converted into pGpG by the EAL domain. Interestingly, when GTP and c-di-GMP are both present in the reaction mixture, we observed that GMP accumulates after a lag-phase of ~35 min (Figure 1B), contrary to the reaction with the sole GTP. During this lag-phase, c-di-GMP is actively consumed by the Life 2021, 11, 31 6 of 13 PDE domain (Figure 1C) and only when residual c-di-GMP reaches ~8 µM, GMP accumulation is observed (Figure 1C).

FigureFigure 1. GMP1. GMP production production by by DUAL. DUAL. ( A(A)) Chromatograms Chromatograms showing the the nucleotide nucleotide content content of of the the reaction reaction of 5 of µM 5 µ DUALM DUAL with 100 µM GTP: pGpG accumulates (black line) only after 18 h of incubation (kcat 0.035 min−1), thus suggesting that with 100 µM GTP: pGpG accumulates (black line) only after 18 h of incubation (k 0.035 min−1), thus suggesting that DUAL is unlikely to be a DGC under physiological conditions. Incubation times lowercat than two hours yields only a species DUALeluting is unlikely at 2.7 min, to bewhich a DGC is GMP under (grey physiological lines, see conditions.Figure S1 for Incubation details). (B times) Time lower-course than of twoGMP hours production yields onlyperformed a species elutingwith at8 μM 2.7 DUAL, min, which in the is presence GMP (grey of 60 lines, μM c see-di- FigureGMP or S1 in forits absence details). (as (B )indicated Time-course in the of legend), GMP productionunder experimental performed withconditions 8 µM DUAL, optimized in the for presence GTPase ofactivity. 60 µM (C c-di-GMP) Nucleotide or incontent its absence of the reaction (as indicated described in the in legend),(B) in the under presence experimental of c-di- conditionsGMP, as optimized a function forof time. GTPase Data activity. are the ( Cmean) Nucleotide of three contentexperiments of the ±SD. reaction Peaks described below the in GMP (B) in one the are presence due to ofGTP c-di-GMP, and as abuffe functionr components. of time. Data are the mean of three experiments ±SD. Peaks below the GMP one are due to GTP and buffer components. 3.2. (α-β)-GTPase Activity on Other Hybrid Proteins Interestingly,To investigate when whether GTP GTP and c-di-GMP hydrolysis are is botha general present feature in the of reaction the GGDEF mixture,-EAL we observedDUAL moiety, that GMP we qualitatively accumulates analyze after athe lag-phase reactivity of of ~35 the min motility (Figure regulator1B), contrary protein to theMorA reaction from P. with aeruginosa the sole, wh GTP.ose Duringkinetic characterization this lag-phase, has c-di-GMP been previously is actively carried consumed out by[27]. the Briefly, PDE domain Phippen (Figure and coworkers1C) and found only when that MorA residual is able c-di-GMP to convert reaches GTP to ~8 c-diµ-M,GMP GMP accumulationthrough its GGDEF is observed domain (Figure and to1C). hydrolyze c-di-GMP to pGpG by the EAL domain, in the presence of the sole GTP or with both GTP and c-di-GMP; on the other hand, MorA 3.2.phosphodiesterase (α-β)-GTPase Activity activity on is Other almost Hybrid absent Proteins in the absence of GTP. The same profile has beenTo observed investigate in this whether study GTPand no hydrolysis GMP accumulation is a general occurs feature (Supplementary of the GGDEF-EAL Figure DUAL S2 moiety,A–C). Nevertheless, we qualitatively experi analyzements carried the reactivity out under of the the conditions motility regulator optimized protein for RmcA, MorA from P. aeruginosa, whose kinetic characterization has been previously carried out [27]. Brieﬂy, Phippen and coworkers found that MorA is able to convert GTP to c-di-GMP through its GGDEF domain and to hydrolyze c-di-GMP to pGpG by the EAL domain, in the presence of the sole GTP or with both GTP and c-di-GMP; on the other hand, MorA Life 2021, 11, x FOR PEER REVIEW phosphodiesterase activity is almost absent in the absence of GTP. The same proﬁle7 of 13 has

been observed in this study and no GMP accumulation occurs (Supplementary Figure S2A–C). Nevertheless, experiments carried out under the conditions optimized for RmcA, i.e., ~i.e.,4 folds ~4 foldslower lower protein protein concentration concentration and pseudo and pseudo-ﬁrst-first order order conditions, conditions, only onlyconfirm conﬁrm that MorAthat MorA is a GTP is a- GTP-dependentdependent PDE; PDE;no DGC no DGCactivity activity is detectable is detectable when when GTP GTPis the is sole the sole substratesubstrate (Figure (Figure 2A) while2A) while GMP GMP accumulates accumulates when when GTP GTPis present. is present. Therefore, Therefore, reactivity reactivity of MorAof MorA with withGTP depends GTP depends on the on experimental the experimental conditions. conditions.

Figure 2. Kinetic properties of other hybrid proteins. (A) RP-HPLC chromatograms of 1 h of reaction Figure 2. Kinetic properties of other hybrid proteins. (A) RP-HPLC chromatograms of 1 h of reac- µ µ µ tion ofof 10 10 µMM MorA MorA with with 100 100 µM MGTP GTP or 30 or µM 30 cM-di- c-di-GMPGMP or with or withboth bothnucleotides. nucleotides. (B) The (B same) The same experimentsexperiments reported reported in (A) inwere (A )run were with run 10 with µM RbdA. 10 µM A RbdA. sample A experiment sample experiment of a triplicate of a triplicateis is shown;shown; the peaks the peaksbelow below the GMP the GMPone are one due are to due GTP to and GTP buffer and buffer components. components.

We then analyzed the reactivity of another well-characterized hybrid protein: the positive regulator of biofilm dispersal RbdA from P. aeruginosa [8]. Under our experimental conditions, the DUAL portion of RbdA displays PDE activity and lacks the DGC one; moreover, in the presence of both GTP and c-di-GMP, GMP also accumulates (Figure 2B). The reactivity of this construct is different from that observed by Liu and coworkers, who found that the protein is both a DGC and a GTP-enhanced PDE, with no GMP accumulation [8]. Nevertheless, they worked on a RbdA construct harboring PAS-PAC- GGDEF-EAL domain, which is different from ours and harbors only the DUAL moiety; the PAS-PAC portion upstream the GGDEF domain likely allows GGDEF to productively dimerize and to populate a catalytically competent DGC able to enter catalysis.

3.3. GMP Production Is Not Due to an Unexpected PDE-B Activity of the EAL Domain Given that GMP production is a common trait of the GGDEF-EAL tandem, a series of experiments were done to exclude a side (and unusual) PDE-B activity. As mentioned above, pGpG may be further hydrolyzed into 2 molecules of GMP (the so-called PDE-B activity), as observed in the HD-GYP subtype of PDEs [27]. To probe or exclude this, we investigated the reactivity of RmcA DUAL with the sole pGpG. Kinetics data indicate that no pGpG depletion occurs even in the presence of GTP, while GMP accumulation is observed only when GTP is present (Figure 3A). These findings are consistent with what previously reported on the isolated RmcA EAL domain, which, although efficient as PDE, is not able to populate GMP from c-di-GMP or pGpG [5]. Moreover, we tested the reactivity of the double mutant of DUAL in the EAL domain D1136N/D1137N (hereinafter DUAL-DD), where the replacement of the two aspartic acid residues hampers the holo active site construction and finally PDE turnover [5]. As shown in Figure 3B, GMP still accumulates in the presence of GTP. To a lower extent also c-di- GMP is formed, belonging to a spurious DGC; c-di-GMP is not further hydrolyzed into pGpG since the PDE activity is functionally abolished by mutation. In the wildtype protein, on the contrary, the little amount of c-di-GMP, due to the spurious DGC, is rapidly hydrolyzed into pGpG, as commented above. Taken together, all these data indicate that conversion of GTP into GMP does not belong to a PDE-B activity.

Life 2021, 11, 31 7 of 13

We then analyzed the reactivity of another well-characterized hybrid protein: the positive regulator of bioﬁlm dispersal RbdA from P. aeruginosa [8]. Under our experimental conditions, the DUAL portion of RbdA displays PDE activity and lacks the DGC one; moreover, in the presence of both GTP and c-di-GMP, GMP also accumulates (Figure2B). The reactivity of this construct is different from that observed by Liu and coworkers, who found that the protein is both a DGC and a GTP-enhanced PDE, with no GMP accumulation [8]. Nevertheless, they worked on a RbdA construct harboring PAS-PAC- GGDEF-EAL domain, which is different from ours and harbors only the DUAL moiety; the PAS-PAC portion upstream the GGDEF domain likely allows GGDEF to productively dimerize and to populate a catalytically competent DGC able to enter catalysis.

FigureFigure 3.3. GMP does not not belong belong to to a a PDE PDE-B-B activity. activity. (A (A) )Chromatograms Chromatograms showing showing the the nucleotide nucleotide contentcontent of the the reaction reaction of of DUAL DUAL (5 (5 µM)µM) with with 30 30µMµ MpGpG pGpG with with or without or without 100 100µM µGTPM GTP(grey (grey and andblack black traces, traces, respectively) respectively) at 1 h ator 118 h h or (dashed 18 h (dashed and continuous and continuous lines, respectively). lines, respectively). GMP accumu- GMP accumulateslates only in onlythe presence in the presence of GTP. of (B GTP.) Chromatograms (B) Chromatograms showing showing the nucleotide the nucleotide content content of the ofreac- the reactiontion of 5 ofµM 5 µDUALM DUAL or DUAL or DUAL-DD-DD (dotted (dotted or continuous or continuous line, line,respectively), respectively), with with100 µM 100 GTPµM GTPafter after18 h of 18 incubation. h of incubation. GMP accumulates GMP accumulates in both in samples both samples while pGpG while is pGpG produced is produced only in the only DUAL in the sample. A sample experiment of a duplicate is shown; peaks below the GMP one are due to GTP DUAL sample. A sample experiment of a duplicate is shown; peaks below the GMP one are due to and buffer components. GTP and buffer components.

3.4. GGDEFMoreover, Incompetent we tested Dim theerization reactivity Promotes of the double GMP Production mutant of DUAL in the EAL domain D1136N/D1137NIf this background (hereinafter reactivity DUAL-DD), is a general where feature the replacementof the GGDEF of domains, the two aspartic it could acidbias residueskinetic studies hampers describing the holo GGDEF active site-containing construction proteins; and ﬁnally therefore, PDE we turnover probed [ 5the]. As reactivity shown inof FigureGGDEF3B, domains GMP still with accumulates GTP in inprotein the presence with dom of GTP.ain architectures To a lower extent lacking also the c-di-GMP DUAL ismotif. formed, We belonginganalyzed the to areactivity spurious of DGC; PleD c-di-GMP from C. crescentus is not further, since hydrolyzed it represents into the pGpG refer- sinceence system the PDE for activity studying is functionally DGCs: this abolishedprotein populates by mutation. a stable In theinhibited wildtype state, protein, which on is thereleased contrary, by phosphorylation the little amount ofto c-di-GMP,allow the GGDEF due to the moieties spurious to interface DGC, is rapidly and enter hydrolyzed catalysis intoupon pGpG, GTP binding as commented [28]. As above. shown Taken in Figure together, 4A, reaction all these of data inhibited indicate PleD that (“as conversion purified”) of GTPwith intoGTP GMP yields does GMP, not while belong c to-di a-GMP PDE-B synthesis activity. was negligible. As expected, the activated PleD (treated with beryllium fluoride to mimic phosphorylation) recovers the DGC 3.4. GGDEF Incompetent Dimerization Promotes GMP Production activity yielding c-di-GMP, while no GMP is observed (Figure 4A). Therefore, GMP accu- mulatesIf this from background GTP when reactivity the productive is a general dimerization feature of of the the GGDEF GGDEF domains, domains, it couldnecessary bias kineticfor the studiesDGC activity describing of PleD GGDEF-containing [29], is conformationally proteins; hampered. therefore,we probed the reactivity Since the dimer formation is a pre-requisite to enter DGC catalysis, we investigated the possible GMP production of the isolated GGDEF of YfiN from P. aeruginosa, an inactive truncated version, previously shown to be a monomer able to bind GTP with high affinity [18]. According to our hypothesis, the kinetic data clearly shows that the GGDEF domain of YfiN displayed the side activity accumulating GMP, while c-di-GMP production absent (Figure 4B).

Figure 4. The GGDEF domain of other DGCs is able to produce GMP. (A) RP-HPLC chromatograms of PleD DGC reaction. Catalytic reaction was carried out either using PleD enzyme activated as DGC by BeF3 (continuous lines) or with the inactive form of the enzyme, as purified (dotted lines). (B) RP-HPLC chromatograms of YfiN DGC reaction. The little amount of pGpG which

Life 2021, 11, x FOR PEER REVIEW 8 of 13

Figure 3. GMP does not belong to a PDE-B activity. (A) Chromatograms showing the nucleotide content of the reaction of DUAL (5 µM) with 30 µM pGpG with or without 100 µM GTP (grey and black traces, respectively) at 1 h or 18 h (dashed and continuous lines, respectively). GMP accumulates only in the presence of GTP. (B) Chromatograms showing the nucleotide content of the reaction of 5 µM DUAL or DUAL-DD (dotted or continuous line, respectively), with 100 µM GTP after 18 h of incubation. GMP accumulates in both samples while pGpG is produced only in the DUAL sample. A sample experiment of a duplicate is shown; peaks below the GMP one are due to GTP and buffer components.

3.4. GGDEF Incompetent Dimerization Promotes GMP Production If this background reactivity is a general feature of the GGDEF domains, it could bias kinetic studies describing GGDEF-containing proteins; therefore, we probed the reactivity of GGDEF domains with GTP in protein with domain architectures lacking the DUAL motif. We analyzed the reactivity of PleD from C. crescentus, since it represents the reference system for studying DGCs: this protein populates a stable inhibited state, which is Life 2021, 11, 31 released by phosphorylation to allow the GGDEF moieties to interface and enter catalysis8 of 13 upon GTP binding [28]. As shown in Figure 4A, reaction of inhibited PleD (“as purified”) with GTP yields GMP, while c-di-GMP synthesis was negligible. As expected, the activated PleD (treated with beryllium fluoride to mimic phosphorylation) recovers the DGC activityof GGDEF yielding domains c-di-GMP, with while GTP inno protein GMP is with observed domain (Figure architectures 4A). Therefore, lacking GMP the DUALaccu- mulatesmotif. Wefrom analyzed GTP when the reactivity the productive of PleD dimerization from C. crescentus of the, sinceGGDEF it represents domains, the necessary reference forsystem the DGC for studyingactivity of DGCs: PleD this[29], protein is conformationally populates a stable hampered. inhibited state, which is released by phosphorylationSince the dimer formation to allow theis a GGDEFpre-requisite moieties to enter to interface DGC catalysis, and enter we catalysis investigated upon theGTP possible binding GMP [28 ].production As shown of in the Figure isolated4A, reaction GGDEF of of inhibited YfiN from PleD P. aeruginosa (“as puriﬁed”), an inac- with tiveGTP truncated yields GMP, version, while previously c-di-GMP shown synthesis to be was a monomer negligible. able As to expected, bind GTP the with activated high affinityPleD (treated [18]. According with beryllium to our ﬂuoridehypothesis, to mimic the kinetic phosphorylation) data clearly recoversshows that the the DGC GGDEF activity domainyielding of c-di-GMP,YfiN displayed while the no GMPside activity is observed accumulating (Figure4A). GMP, Therefore, while c GMP-di-GMP accumulates produc- tionfrom absent GTP (Figure when the 4B). productive dimerization of the GGDEF domains, necessary for the DGC activity of PleD [29], is conformationally hampered.

Figure 4. The GGDEF domain of other DGCs is able to produce GMP. (A) RP-HPLC chromatograms Figure 4. The GGDEF domain of other DGCs is able to produce GMP. (A) RP-HPLC chromato- gramsof PleD of PleD DGC DGC reaction. reaction. Catalytic Catalytic reaction reaction was was carried carried out out either either using using PleD PleD enzyme enzyme activated acti- as vatedDGC as by DGC BeF3 by (continuous BeF3 (continuous lines) or lines) with or the with inactive the inactive form of form the enzyme,of the enzyme, as purified as purified (dotted (dot- lines). ted(B lines).) RP-HPLC (B) RP chromatograms-HPLC chromatograms of YfiN DGC of YfiN reaction. DGC reaction. The little The amount little of amount pGpG of which pGpG accumulates which after 2 h could be ascribed to the α-ß hydrolysis of the DGC intermediate pppGpG, thus further confirming the aberrant behavior of a truncated GGDEF. A sample experiment of a triplicate is shown; peaks below the GMP one are due to GTP and buffer components.

Since the dimer formation is a pre-requisite to enter DGC catalysis, we investigated the possible GMP production of the isolated GGDEF of YfiN from P. aeruginosa, an inactive truncated version, previously shown to be a monomer able to bind GTP with high affinity [18]. According to our hypothesis, the kinetic data clearly shows that the GGDEF domain of YfiN displayed the side activity accumulating GMP, while c-di-GMP production absent (Figure4B).

3.5. Structure of Inactive YfiNGGDEF The versatile reactivity of the GGDEF domain seems to be related to a conformational issue rather than to an altered active site (as compared to the catalytically competent site required for DGC reaction): this may explain the alternative reactivity of the same protein towards GTP, depending on the allosteric status of the protein. This hypothesis has been 2+ confirmed by solving the 3D structure of the YfiNGGDEF protein in complex with Mg by X-ray crystallography at 2.8 Å resolution. The crystal contained 3 monomers in the asymmetric unit (Figure5A) and protein interface surface analysis performed using the PISA server (https://www.ebi.ac.uk/msd-srv/prot_int/cgi-bin/piserver) showed that all protein-protein interfaces of the crystal lattice would not be stable in solution, confirming that YfiNGGDEF is monomeric (as previously demonstrated in solution [18]). A magnesium ion is bound in the GTP binding site, coordinated by the oxygen atoms of D287, S288(CO), D330 (belonging to the GGDEF signature), and two water molecules (Figure5B). Superpo- sition with the recently solved structure of RmcA in complex with GTP and calcium [5], shows that the metal ions occupy the same position and that the two water molecules replace the oxygen provided by the β and γ phosphate groups of GTP (Figure5C). The GTP substrate docks to the GGDEF domain in an extended conformation and a side reaction leading to the formation of c-GMP can be excluded given the position of the 30-OH, Life 2021, 11, x FOR PEER REVIEW 9 of 13

accumulates after 2 h could be ascribed to the α-ß hydrolysis of the DGC intermediate pppGpG, thus further confirming the aberrant behavior of a truncated GGDEF. A sample experiment of a triplicate is shown; peaks below the GMP one are due to GTP and buffer components.

3.5. Structure of Inactive YfiNGGDEF The versatile reactivity of the GGDEF domain seems to be related to a conformational issue rather than to an altered active site (as compared to the catalytically competent site required for DGC reaction): this may explain the alternative reactivity of the same protein towards GTP, depending on the allosteric status of the protein. This hypothesis has been confirmed by solving the 3D structure of the YfiNGGDEF protein in complex with Mg2+ by X-ray crystallography at 2.8 Å resolution. The crystal contained 3 monomers in the asymmetric unit (Figure 5A) and protein interface surface analysis performed using the PISA server (https://www.ebi.ac.uk/msd-srv/prot_int/cgi-bin/piserver) showed that all protein- protein interfaces of the crystal lattice would not be stable in solution, confirming that YfiNGGDEF is monomeric (as previously demonstrated in solution [18]). A magnesium ion is bound in the GTP binding site, coordinated by the oxygen atoms of D287, S288(CO), D330 (belonging to the GGDEF signature), and two water molecules (Figure 5B). Super- position with the recently solved structure of RmcA in complex with GTP and calcium [5], Life 2021, 11, 31 shows that the metal ions occupy the same position and that the two water molecules9 of 13 replace the oxygen provided by the β and γ phosphate groups of GTP (Figure 5C). The GTP substrate docks to the GGDEF domain in an extended conformation and a side reaction leading to the formation of c-GMP can be excluded given the position of the 3′-OH, α whichwhich isis tootoo far far from from the the -phosphateα-phosphate to to perform perform a nucleophilica nucleophilic attack. attack. A closer A closer look look at the at position of the phosphate groups with respect to the substrate-binding pocket indicates that the position of the phosphate groups with respect to the substrate-binding pocket indi- the α-phosphate is particularly exposed to the solvent and, if the catalytically competent cates that the α-phosphate is particularly exposed to the solvent and, if the catalytically dimer does not form, may undergo a nucleophilic attack by a hydroxyl ion, thus producing competent dimer does not form, may undergo a nucleophilic attack by a hydroxyl ion, GMP (Figure5C). thus producing GMP (Figure 5C).

Figure 5. Structure of a monomeric GGDEF (YfiNGGDEF). (A) Content of the asymmetric unit showing 3 molecules. Each Figure 5. Structure of a monomeric GGDEF (YfiNGGDEF). (A) Content of the asymmetric unit showing 3 molecules. Each chainchain consistsconsists ofof aa singlesingle GGDEFGGDEF domain.domain. ((BB)) Close-upClose-up viewview ofof thethe GTPGTP bindingbinding sitesite (chain(chain C).C). TheThe residuesresidues belongingbelonging toto the GGDEF signature are colored in white. The magnesium ion (dark green) is coordinated by the main-chain carboxyl the GGDEF signature are colored in white. The magnesium ion (dark green) is coordinated by the main-chain carboxyl group of S288, the carbonyl groups of D287 and D330 (GGDEF) and two water molecules. (C) Structural superposition of group of S288, the carbonyl groups of D287 and D330 (GGDEF) and two water molecules. (C) Structural superposition of GTP (grey) and calcium (light green) as observed in complex with chain B of RmcA (PDB id: 5m3c [5])—and YfiN GGDEF GTP(blue). (grey) The andstructure calcium of YfiN (light GGDEF green) asis observedshown as inspheres complex to highlight with chain the B steric of RmcA hindrance (PDB id: surrounding 5m3c [5])—and the GTP YfiN molecule. GGDEF (blue).The direction The structure of a possible of YfiN nucleophilic GGDEF is shown attack as at spheres the α-phosphate to highlight is theindicated steric hindrance with a black surrounding arrow, the the other GTP phosphate molecule. Thegroups direction appear of much a possible less accessible nucleophilic due attackto the presence at the α-phosphate of the surrounding is indicated protein with residues. a black arrow, the other phosphate groups appear much less accessible due to the presence of the surrounding protein residues. 4. Discussion 4. DiscussionProteins involved in controlling c-di-GMP levels (and c-di-GMP itself) are characterized Proteinsby an extraordinary involved in plasticity controlling and c-di-GMP their function levels is deeply (and c-di-GMP governed itself) by allosteric are charac- con- terizedtrol of other by an signals extraordinary or domains. plasticity The structural and their functionanalysis, iswhere deeply possible, governed report bys allosteric on mul- controltiple conformation of other signals and oroligomeric domains. state The governing structural analysis,protein function where possible, [3,29]. This reports feature on multiplemakes the conformation rigorous biochemical and oligomeric characterization state governing of these protein proteins function difficult, [3,29]. Thisparticularly feature makesfrom a thequantitative rigorous biochemical point of view, characterization slowing down of our these understanding proteins difficult, of such particularly potential from a quantitative point of view, slowing down our understanding of such potential therapeutic targets. To further complicate the scenario, the results reported in this study indicate that the GGDEF domain can exhibit a novel and unusual GTPase activity, related to its conformational frustration (see for a definition [30,31]). We observed this side-reaction both on GGDEF-EAL hybrid constructs and on a stand-alone GGDEF domain. In the first case, GTP-dependent GMP production is negatively affected by c-di-GMP, which in turn sustains the pGpG production. Thus, the apparent GTP → GMP conversion is likely to be an alternative to the PDE activity (see Figure6 for a scheme). It is not the first time that a crosstalk between the EAL and the GGDEF domain takes place: according to mutagenesis studies on RmcA and structure data on both RmcA and RbdA, binding of GTP to GGDEF leads to a dramatic re-arrangement on the EAL domain, which, in turn, fasten the overall structure of the GGDEF-EAL tandem once c-di-GMP is bound to the EAL active site [5,8]. Once c-di-GMP levels drops (close to the KM value measured for the EAL domain, [5]), the EAL active dimer is not fully saturated and likely dissociates (yielding an OFF conformation for the EAL); under these conditions, GTP-bound GGDEF is likely less conformationally constrained and may hydrolyse, as a monomeric entity, GTP into GMP. Structural data [5] suggest that under these conditions the GGDEF is constrained from interfacing with its other unit to produce c-di-GMP as for a DGC enzyme. The same fate probably is observed in the inactive PleD: if the GTP-bound active site is conformationally trapped as an isolated moiety, the (α-β)-GTPase activity takes place. Life 2021, 11, x FOR PEER REVIEW 11 of 13

Life 2021, 11, 31 10 of 13 that characterization of the reactivity of GGDEF-containing proteins requires special caution, particularly in the protein engineering design stage.

E E G G GTP E = EAL GTP G = GGDEF c-di-GMP

Slow Rapid pGpG c-di-GMP

E E E E E E E E G G G G G G G G GTP GTP GTP GTP GTP GMP c-di-GMP

Figure 6. 6. SchemeScheme of of the the reactivity reactivity of of DUAL DUAL RmcA RmcA with with GTP, GTP, with with or orwithout without c-di c-di-GMP.-GMP. E, EAL E, EAL domain, G, G, GGDEF GGDEF domain. domain. In In green green the the domains domains performing performing turnover, turnover, in red in redthe thedomain domain in a in a restingresting state, in orange non non-activated-activated domain. It is not excluded that the red andand thethe orangeorange statestate of ofthe the EAL EAL domain domain could could show show the the same same conformation. conformation.

5. ConclusionsIt is worth mentioning that the GGDEF reactivity deeply changes depending on the experimentalDespite that setup: the in GGDEF the case domain of MorA, is evolutionarily the experimental conserved conditions and reported structurally in Phippen stable, itsand function coworkers is strictly includes dependent a very high on enzymethe neighboring concentration domain(s) (40 µ M)or sequences and littleGTP that, excess inde- pendently(100 µM), whichof their promotes nature, aare DGC requi turnover.red to ensure It is known the proper that DGC dimerization reaction is to dependent enter DGC on catalysis.protein concentration It is possible forto obtain both the recombinant active dimer stable formation and properly step [18 ,folded28] and GGDEF the cooperative domain fromsubstrate different binding proteins step [ (and32]. The obtain high crystals protein and concentration solve the 3- usedD structure, for MorA, as favoringreported above),stochastic both intermolecular in isolation or interactions in combination between with GGDEF other domains, domains, but may their randomly intrinsic populate reactiv- itythe is active difficult dimer to forbe theclearly c-di-GMP assessed. production, Their putative thus leading DGC and/or to an apparentallostericDGC switch activity. roles couldAccordingly, be significantly the DGC phenotypeaffected or iseven lost abolished when using by lower the boundaries protein concentration, chosen in the favoring engi- neeringthe GTPase design reaction. step. Last On but the not other least, hand, the the dependence results on on RbdA protein raises concentration some concerns should on bethe carefully characterization evaluated of to the characterize truncated GGDEF-EAL these proteins proteins biochemically and, possibly, and to onassign the GGDEF-to a cer- taincontaining GGDEF proteins; domain the its upstreamrole in a multidomain domains are likelyprotein. strategic to allow the GGDEF portion to enter DGC catalysis, as previously observed for the proteins containing the sole GGDEF, Supplementarysuch as YfiN [Materials:18]. For YfiN,The following the conserved are available active online site, at confirmed www.mdpi.com/xxx/s1, by structural Supple- data, mentaland the Methods; high affinity Figure forS1: Identification the substrate of are the notproduct sufficient of GTP to consumption promote DGC, by DUAL; given Figure that theS2: Nucleotidesconformational content activation of MorA isreaction(s). hampered by the upstream domain(s) truncation. Taken together, this analysis confirms that GGDEF-containing multi-domain proteins Author Contributions: Conceptualization, S.R., F.C. and G.G.; general methodology, F.M., A.D.M. show not only conformational plasticity but also versatility in their activities at least in vitro: and S.R.; structural biology G.G., A.D.M. and A.P. (Alessandro Paiardini); molecular biology A.P. (theyAlessio may Paone act) as and DGCs, C.S.R.; GTP-dependent validation, F.M. and conformational C.S.R.; formal switchesanalysis, F.M., or, to A.P a lower. (Alessandro extent, as PaiardiniGTPases), (see S.R.;Table mass spectrometry S2 for kcat comparison). and separation The optimization, role of additional M.R., R.P. domains and L.C.; flankingwriting— the reviewcatalytic and one editing, is not S.R., new A.D.M. in the and field G.G.; of GTP-dependentfunding acquisition, proteins. S.R. and In G.G. circularly All authors permutated have readGTPases and agreed (cpGTPases) to the published the presence version on of an the additional manuscript. domain located at the C-terminus with respect to the catalytic one is required [33] to prevent that an otherwise unfastened catalytic Funding: This research was funded by Sapienza University of Rome [RM11715C646D693E and domain fails to stabilize GTP [34,35]. RM11916B414C897E to S.R.; RP11916B407928AA to G.G.]. Up to now, we have no evidence that the GTPase activity has physiological relevance; Datamost Availability likely it is aStatement: spurious The readout data presented due to a in certain this study experimental are available set in up.the figures, According in sup- to plementary material and PDB databse. in vitro parameters, GTPase occurs when c-di-GMP levels are low (close to the KM of Acknowledgments:RmcA for c-di-GMP) Diffraction [5]. Looking data have at thebeen intracellular collected on BL14.1 c-di-GMP at the levels BESSY and II electron GTP cellular stor- ageconcentration ring operated even by the under Helmholtz starvation-Zentrum [9,36 Berlin,37], the[36]. DUAL We would domain particularly of RmcA like couldto be pre- acknowledgedominantly inthe the help GTP-bound and support state of local (considering contacts during the sub-the experiment.µM affinity for GTP). Therefore, Conflictsthe eventual of Interest: GTPase The activity authors in declare the cellular no conflict background of interest. should be modulated mainly by a drop in the c-di-GMP levels; if relevant, its role may just prevent c-di-GMP consumption under a threshold of intracellular (or local) c-di-GMP concentration. At this stage, any further consideration on this aspect is over speculative and out of the scope of this work. Regardless of the physiological relevance of such activity, our experiments clearly indi-

Life 2021, 11, 31 11 of 13

cate that characterization of the reactivity of GGDEF-containing proteins requires special caution, particularly in the protein engineering design stage.

5. Conclusions Despite that the GGDEF domain is evolutionarily conserved and structurally stable, its function is strictly dependent on the neighboring domain(s) or sequences that, inde- pendently of their nature, are required to ensure the proper dimerization to enter DGC catalysis. It is possible to obtain recombinant stable and properly folded GGDEF domain from different proteins (and obtain crystals and solve the 3-D structure, as reported above), both in isolation or in combination with other domains, but their intrinsic reactivity is difﬁcult to be clearly assessed. Their putative DGC and/or allosteric switch roles could be signiﬁcantly affected or even abolished by the boundaries chosen in the engineering design step. Last but not least, the dependence on protein concentration should be carefully evaluated to characterize these proteins biochemically and to assign to a certain GGDEF domain its role in a multidomain protein.

Supplementary Materials: The following are available online at https://www.mdpi.com/2075-172 9/11/1/31/s1, Supplemental Methods; Figure S1: Identification of the product of GTP consumption by DUAL; Figure S2: Nucleotides content of MorA reaction(s). Author Contributions: Conceptualization, S.R., F.C. and G.G.; general methodology, F.M., A.D.M. and S.R.; structural biology G.G., A.D.M. and A.P. (Alessandro Paiardini); molecular biology A.P. (Alessio Paone) and C.S.R.; validation, F.M. and C.S.R.; formal analysis, F.M., A.P. (Alessandro Paiardini), S.R.; mass spectrometry and separation optimization, M.R., R.P. and L.C.; writing—review and editing, S.R., A.D.M. and G.G.; funding acquisition, S.R. and G.G. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Sapienza University of Rome [RM11715C646D693E and RM11916B414C897E to S.R.; RP11916B407928AA to G.G.]. Data Availability Statement: The data presented in this study are available in the figures, in supplementary material and PDB databse. Acknowledgments: Diffraction data have been collected on BL14.1 at the BESSY II electron storage ring operated by the Helmholtz-Zentrum Berlin [36]. We would particularly like to acknowledge the help and support of local contacts during the experiment. Conflicts of Interest: The authors declare no conflict of interest.

References 1. Römling, U.; Liang, Z.-X.; Dow, J.M. Progress in Understanding the Molecular Basis Underlying Functional Diversification of Cyclic Dinucleotide Turnover Proteins. J. Bacteriol. 2016, 199, e00790-16. [CrossRef][PubMed] 2. Chou, S.-H.; Galperin, M.Y. Diversity of Cyclic Di-GMP-Binding Proteins and Mechanisms. J. Bacteriol. 2015, 198, 32–46. [CrossRef][PubMed] 3. Schirmer, T.; Jenal, U. Structural and mechanistic determinants of c-di-GMP signalling. Nat. Rev. Genet. 2009, 7, 724–735. [CrossRef][PubMed] 4. Römling, U.; Galperin, M.Y.; Gomelsky, M. Cyclic di-GMP: The First 25 Years of a Universal Bacterial Second Messenger. Microbiol. Mol. Biol. Rev. 2013, 77, 1–52. [CrossRef] 5. Mantoni, F.; Paiardini, A.; Brunotti, P.; D’Angelo, C.; Cervoni, L.; Paone, A.; Cappellacci, L.; Petrelli, R.; Ricciutelli, M.; Leoni, L.; et al. Insights into the GTP-dependent allosteric control of c-di-GMP hydrolysis from the crystal structure of PA0575 protein from Pseudomonas aeruginosa. FEBS J. 2018, 285, 3815–3834. [CrossRef][PubMed] 6. Bharati, B.K.; Mukherjee, R.; Chatterji, D. Substrate-induced domain movement in a bifunctional protein, DcpA, regulates cyclic di-GMP turnover: Functional implications of a highly conserved motif. J. Biol. Chem. 2018, 293, 14065–14079. [CrossRef] 7. Phippen, C.W.; Mikolajek, H.; Schlaefli, H.G.; Keevil, C.W.; Webb, J.S.; Tews, I. Formation and dimerization of the phosphodiesterase active site of the Pseudomonas aeruginosa MorA, a bi-functional c-di-GMP regulator. FEBS Lett. 2014, 588, 4631–4636. [CrossRef] 8. Liu, C.; Liew, C.W.; Wong, Y.H.; Tan, S.T.; Poh, W.H.; Manimekalai, M.S.S.; Rajan, S.; Xin, L.; Liang, Z.-X.; Grüber, G.; et al. Insights into Biofilm Dispersal Regulation from the Crystal Structure of the PAS-GGDEF-EAL Region of RbdA from Pseudomonas aeruginosa. J. Bacteriol. 2017, 200.[CrossRef] Life 2021, 11, 31 12 of 13

9. Christen, M.; Christen, B.; Folcher, M.; Schauerte, A.; Jenal, U. Identification and Characterization of a Cyclic di-GMP-specific Phosphodiesterase and Its Allosteric Control by GTP. J. Biol. Chem. 2005, 280, 30829–30837. [CrossRef] 10. An, S.; Wu, J.; Zhang, L.-H. Modulation of Pseudomonas Aeruginosa Biofilm Dispersal by a Cyclic-Di-GMP Phosphodiesterase with a Putative Hypoxia-Sensing Domain. Appl. Environ. Microbiol. 2010, 76, 8160–8173. [CrossRef] 11. Cho, K.H.; Tryon, R.G.; Kim, J.-H. Screening for Diguanylate Cyclase (DGC) Inhibitors Mitigating Bacterial Biofilm Formation. Front. Chem. 2020, 8, 264. [CrossRef][PubMed] 12. Sambanthamoorthy, K.; Sloup, R.E.; Parashar, V.; Smith, J.M.; Kim, E.E.; Semmelhack, M.F.; Neiditch, M.B.; Waters, C.M. Identification of Small Molecules That Antagonize Diguanylate Cyclase Enzymes to Inhibit Biofilm Formation. Antimicrob. Agents Chemother. 2012, 56, 5202–5211. [CrossRef][PubMed] 13. Zhou, J.; Watt, S.; Wang, J.; Nakayama, S.; Sayre, D.A.; Lam, Y.-F.; Lee, V.T.; Sintim, H.O. Potent suppression of c-di-GMP synthesis via I-site allosteric inhibition of diguanylate cyclases with 20-F-c-di-GMP. Bioorg. Med. Chem. 2013, 21, 4396–4404. [CrossRef] [PubMed] 14. Fernicola, S.; Paiardini, A.; Giardina, G.; Rampioni, G.; Leoni, L.; Cutruzzolà, F.; Rinaldo, S. In SilicoDiscovery andIn VitroValida- tion of Catechol-Containing Sulfonohydrazide Compounds as Potent Inhibitors of the Diguanylate Cyclase PleD. J. Bacteriol. 2015, 198, 147–156. [CrossRef][PubMed] 15. Shanahan, C.A.; Gaffney, B.L.; Jones, R.A.; Strobel, S.A. Identification of c-di-GMP Derivatives Resistant to an EAL Domain Phosphodiesterase. Biochemistry 2013, 52, 365–377. [CrossRef] 16. Fernicola, S.; Torquati, I.; Paiardini, A.; Giardina, G.; Rampioni, G.; Messina, M.; Leoni, L.; Del Bello, F.; Petrelli, R.; Rinaldo, S.; et al. Synthesis of Triazole-Linked Analogues of c-di-GMP and Their Interactions with Diguanylate Cyclase. J. Med. Chem. 2015, 58, 8269–8284. [CrossRef] 17. Ching, S.M.; Tan, W.J.; Chua, K.L.; Lam, Y. Synthesis of cyclic di-nucleotidic acids as potential inhibitors targeting diguanylate cyclase. Bioorg. Med. Chem. 2010, 18, 6657–6665. [CrossRef] 18. Giardina, G.; Paiardini, A.; Fernicola, S.; Franceschini, S.; Rinaldo, S.; Stelitano, V.; Cutruzzolà, F. Investigating the Allosteric Regulation of YfiN from Pseudomonas aeruginosa: Clues from the Structure of the Catalytic Domain. PLoS ONE 2013, 8, e81324. [CrossRef] 19. Stelitano, V.; Brandt, A.; Fernicola, S.; Franceschini, S.; Giardina, G.; Pica, A.; Rinaldo, S.; Sica, F.; Cutruzzolà, F. Probing the activity of diguanylate cyclases and c-di-GMP phosphodiesterases in real-time by CD spectroscopy. Nucleic Acids Res. 2013, 41, e79. [CrossRef] 20. Kabsch, W. Software XDS for Image Rotation, Recognition and Crystal Symmetry Assignment. Acta Crystallogr. Sect. D Biol. Crystallogr. 2010, D66, 125–132. [CrossRef] 21. Evans, P.R.; Murshudov, G.N. How good are my data and what is the resolution? Acta Crystallogr. Sect. D Biol. Crystallogr. 2013, 69, 1204–1214. [CrossRef][PubMed] 22. Number 4 Collaborative Computational Project. The CCP4 suite: Programs for protein crystallography. Acta Crystallogr. Sect. D Biol. Crystallogr. 1994, 50, 760–763. [CrossRef][PubMed] 23. Vagin, A.A.; Teplyakov, A. Molecular replacement withMOLREP. Acta Crystallogr. Sect. D Biol. Crystallogr. 2009, 66, 22–25. [CrossRef][PubMed] 24. Emsley, P.; Cowtan, K. Coot: Model-building tools for molecular graphics. Acta Crystallogr. Sect. D Biol. Crystallogr. 2004, 60, 2126–2132. [CrossRef] 25. Murshudov, G.N.; Skubák, P.; Lebedev, A.A.; Pannu, N.S.; Steiner, R.A.; Nicholls, R.A.; Winn, M.D.; Long, F.; Vagin, A.A. REFMAC5 for the refinement of macromolecular crystal structures. Acta Crystallogr. Sect. D Biol. Crystallogr. 2011, 67, 355–367. [CrossRef][PubMed] 26. Paiardini, A.; Mantoni, F.; Giardina, G.; Paone, A.; Janson, G.; Leoni, L.; Rampioni, G.; Cutruzzolà, F.; Rinaldo, S. A novel bacteriall-arginine sensor controlling c-di-GMP levels inPseudomonas aeruginosa. Proteins Struct. Funct. Bioinform. 2018, 86, 1088–1096. [CrossRef] 27. Stelitano, V.; Giardina, G.; Paiardini, A.; Castiglione, N.; Cutruzzolà, F.; Rinaldo, S. C-di-GMP Hydrolysis by Pseudomonas aeruginosa HD-GYP Phosphodiesterases: Analysis of the Reaction Mechanism and Novel Roles for pGpG. PLoS ONE 2013, 8, e74920. [CrossRef] 28. Paul, R.; Abel, S.; Wassmann, P.; Beck, A.; Heerklotz, H.; Jenal, U. Activation of the Diguanylate Cyclase PleD by Phosphorylation- mediated Dimerization. J. Biol. Chem. 2007, 282, 29170–29177. [CrossRef] 29. Schirmer, T. C-di-GMP Synthesis: Structural Aspects of Evolution, Catalysis and Regulation. J. Mol. Biol. 2016, 428, 3683–3701. [CrossRef] 30. Toulouse, G. Theory of Frustration Effects in Spin-Glasses: I. Comm. Phys. 1977, 2, 115. 31. Gianni, S.; Camilloni, C.; Giri, R.; Toto, A.; Bonetti, D.; Morrone, A.; Sormanni, P.; Brunori, M.; Vendruscolo, M. Understanding the Frustration Arising from the Competition between Function, Misfolding, and Aggregation in a Globular Protein. Proc. Natl. Acad. Sci. USA 2014, 111, 14141–14146. [CrossRef][PubMed] 32. Oliveira, M.C.; Teixeira, R.D.; Andrade, M.O.; Pinheiro, G.M.; Ramos, C.H.; Farah, C.S. Cooperative Substrate Binding by a Diguanylate Cyclase. J. Mol. Biol. 2015, 427, 415–432. [CrossRef][PubMed] 33. Anand, B.; Verma, S.K.; Prakash, B. Structural stabilization of GTP-binding domains in circularly permuted GTPases: Implications for RNA binding. Nucleic Acids Res. 2006, 34, 2196–2205. [CrossRef] Life 2021, 11, 31 13 of 13

34. Rocchio, S.; Santorelli, D.; Rinaldo, S.; Franceschini, M.; Malatesta, F.; Imperi, F.; Federici, L.; Travaglini-Allocatelli, C.; Di Matteo, A. Structural and functional investigation of the Small Ribosomal Subunit Biogenesis GTP ase A (RsgA) from Pseudomonas aeruginosa. FEBS J. 2019, 286, 4245–4260. [CrossRef][PubMed] 35. Anand, B.; Surana, P.; Bhogaraju, S.; Pahari, S.; Prakash, B. Circularly permuted GTPase YqeH binds 30S ribosomal subunit: Implications for its role in ribosome assembly. Biochem. Biophys. Res. Commun. 2009, 386, 602–606. [CrossRef] 36. Simm, R.; Morr, M.; Kader, A.; Nimtz, M.; Römling, U. GGDEF and EAL domains inversely regulate cyclic di-GMP levels and transition from sessility to motility. Mol. Microbiol. 2004, 53, 1123–1134. [CrossRef] 37. Weinhouse, H.; Sapir, S.; Amikam, D.; Shilo, Y.; Volman, G.; Ohana, P.; Benziman, M. c-di-GMP-binding protein, a new factor regulating cellulose synthesis inAcetobacter xylinum. FEBS Lett. 1997, 416, 207–211. [CrossRef]