Evaluation and characterization of the predicted diguanylate cyclase-encoding genes in Pseudomonas aeruginosa.

Opportunistic pathogen Pseudomonas aeruginosa can cause acute and chronic infections in humans. It is notorious for its resistance to antibiotics due to the formation of biofilms. Cyclic-di-GMP is a bacterial second messenger that plays important roles during biofilm development. There are 40 genes in P. aeruginosa predicted to participate in c-di-GMP biosynthesis or degradation. It is time-consuming for the functional characterization of these genes. Here, we cloned 16 genes from P. aeruginosa PAO1 that are predicted to encode diguanylate cyclases (DGCs, responsible for c-di-GMP biosynthesis) and constructed their corresponding in-frame deletion mutants. We evaluated the methods to measure the intracellular c-di-GMP concentration by using deletion mutants and PAO1 strains containing a plasmid expressing one of the 16 genes, respectively. Functional outputs of all PAO1-derived stains were also detected and evaluated, including biofilm formation, production of exopolysaccharide, swimming and swarming motilities. Our data showed that measuring the c-di-GMP level only characterized a few DGC by using either pCdrA::gfp as a reporter or LC/MS/MS. Functional output results indicated that overexpression of a DGC gave more pronounced phenotypes than the corresponding deletion mutant and suggested that the swimming motility assay could be a quick way to briefly estimate a predicted DGC for further studies. The overall evaluation suggested 15 out of 16 predicted DGCs were functional DGCs, wherein six were characterized to encode DGCs previously. Altogether, we have provided not only a cloning library of 16 DGC-encoding genes and their corresponding in-frame deletion mutants but also paved ways to briefly characterize a predicted DGC.

INTRODUCTION

Pseudomonas aeruginosa is an opportunistic pathogen that can infect humans, animals, and plants (Plotnikova, Rahme, & Ausubel, 2000; Stover et al., 2000). It can cause nosocomial infections such as pneumonia and cystic fibrosis (CF) that are refractory to eradicate due to its capability to form biofilms (Lebeaux, Ghigo, & Beloin, 2014; Lyczak, Cannon, & Pier, 2000; Mulcahy, Isabella, & Lewis, 2014). The surface‐associated microbial communities called biofilms are surrounded by a matrix of extracellular polymeric substances including exopolysaccharides, extracellular DNA, and proteins (Decho & Gutierrez, 2017; Fong & Yildiz, 2015; Koo, Allan, Howlin, Stoodley, & Hall‐Stoodley, 2017; Wei & Ma, 2013). The bis‐(3′‐5′)‐cyclic guanosine monophosphate (c‐di‐GMP) is an important second messenger that controls the switch from a planktonic to biofilm lifestyle (Boyd & O'Toole, 2012; Hengge, 2009; Romling, Galperin, & Gomelsky, 2013) and regulates a wide range of biological activities (Fu et al., 2018; Ryan, Fouhy, Lucey, & Dow, 2006). For example, c‐di‐GMP regulates the expression of biofilm matrix polysaccharide Psl and Pel in P. aeruginosa (Ha & O'Toole, 2015; Klausen, Aaes‐Jorgensen, Molin, & Tolker‐Nielsen, 2003; Liang, 2015) and regulates the virulence in a murine model of acute infection (Kulasakara et al., 2006). The c‐di‐GMP has also been involved in the regulation of bacterial motility such as flagella‐mediated swimming and swarming motilities. It was demonstrated that c‐di‐GMP could fine‐tune swimming speed by binding to a molecular brake, YcgR (Boehm et al., 2010).

The c‐di‐GMP is synthesized from two molecules of GTP by diguanylate cyclases (DGCs) and degraded into 5′‐phosphoguanylyl‐(3′‐5′)‐guanosine (pGpG) by specific phosphodiesterases (PDEs); later, pGpG was split into two GMP molecules by ribonuclease Orn (Orr et al., 2015; Stelitano et al., 2013). Bioinformatic analysis showed that DGC activity is associated with the presence of conserved GGDEF domain while PDE activity is associated with conserved EAL or HD‐GYP domains (Chan et al., 2004; Paul et al., 2004; Schmidt, Ryjenkov, & Gomelsky, 2005). Therefore, the intracellular c‐di‐GMP is depended on the active DGCs and PDEs. In P. aeruginosa, there are at least 40 predicted c‐di‐GMP metabolizing proteins (Ha, Richman, & O'Toole, 2014; Kulasakara et al., 2006). Comprehensive analyses of those c‐di‐GMP metabolizing genes were performed on strain PA14, which is a native Psl‐negative strain (Liberati et al., 2006). Afterward, the deletion library of c‐di‐GMP metabolizing genes was constructed and characterized briefly for their ability to form biofilms and motilities (Ha et al., 2014). Those data have provided means to demonstrate the potential use of this library. However, the phenotype of mutants was very complex and difficult to figure out a gene functioned as DGC or PDE. Thus, it is necessary to evaluate the current methods to find an effective way to characterize a DGC or PDE.

In this study, we selected 16 predicted DGC‐encoding genes from P. aeruginosa strain PAO1 that only contain conserved GGDEF domain to evaluate the methods for DGC characterization, which includes the measurement of intracellular c‐di‐GMP concentration and the functional outputs of all PAO1‐derived stains, including biofilm formation, the production of exopolysaccharide Psl, swimming and swarming motilities. Our work would help the researcher to go depth into their interested DGCs in bacteria.

MATERIALS AND METHODS

Bacterial strains and growth conditions

Strains, plasmids, and primers used in this study are listed in Tables A1 and A2. Pseudomonas aeruginosa strain PAO1 and Escherichia coli were used in this study. PAO1 and E. coli strains were routinely grown at 37°C in Luria‐Bertani (LB) medium. Pseudomonas aeruginosa strains were grown in LB medium without sodium chloride or in Jensen's, a chemically defined medium. Antibiotics were added at the following final concentrations when required: ampicillin (Ap), 100 µg/ml; carbenicillin (Carb), 300 µg/ml; gentamicin (Gm), 10 µg/ml; Irgasan (Irg) 25 µg/ml.

Table A1

Primers used in this study

Primer name	Primer sequence (5′‐3′)	Restriction sites	Application
PA0169‐UPFw	GGGGTACCGACCTGCACATACCCGGCAC	KpnI	Generation of pEX18Gm‐PA0169
PA0169‐UPRv	GACGCCTGAGGAGGCCAGTTGCTCCAGCGA
PA0169‐DWFw	TCGCTGGAGCAACTGGCCTCCTCAGGCGTC
PA0169‐DWRv	GCTCTAGAGCGCAGGCGTTCATCGCCGC	XbaI
PA0290‐UPFw	GCTCTAGACATCGCTACCCTGGTCTTC	XbaI	Generation of pEX18Gm‐PA0290
PA0290‐UPRv	GATGCAGTCCTTGCCGGTTAGGTCGTCCATGGATC
PA0290‐DwFw	ATGGACGACCTAACCGGCAAGGACTGCATCGTCGT
PA0290‐DwRv	CCCAAGCTTGAACGGAAGCGTGTGCATG	Hind III
PA0338‐UPFw	GCTCTAGACTCTACGACTACCTGCATCC	XbaI	Generation of pEX18Gm‐PA0338
PA0338‐UPRv	GCCATCATTCAGCACCAGGCAGTCTATACGCTTGC
PA0338‐DwFw	CGTATAGACTGCCTGGTGCTGAATGATGGCGAGAC
PA0338‐DwRv	CCCAAGCTTGCTTCAACCTGCTCATTGGC	Hind III
PA3702‐UPFw	GCTCTAGACGGTTCGCTGGTGTACACC	XbaI	Generation of pEX18Gm‐PA3702
PA3702‐UPRv	GTCGGCCATCTCGATGAGCTCTCATGAGGGTTG
PA3702‐DwFw	CAACCCTCATGAGAGCTCATCGAGATGGCCGAC
PA3702‐DwRv	CCCAAGCTTGAACTCCAGCTTCTCGAGG	Hind III
PA3343‐UPFw	GCTCTAGACCTTCCTCGATGTCCTCAC	XbaI	Generation of pEX18Gm‐PA3343
PA3343‐UPRv	GGCATGCTCGTAGAGGCACACCTCTTCTCTTGGAG
PA3343‐DwFw	AGAGAAGAGGTGTGCCTCTACGAGCATGCCGATC
PA3343‐DwRv	CCCAAGCTTCTGGTAGAGGTAGACGAAGC	Hind III
PA3177‐UPFw	GCTCTAGAGCACAAGGCCTTCATCATCG	XbaI	Generation of pEX18Gm‐PA3177
PA3177‐UPRv	GTCCAGTTCCTCCACGGTATTGGACTGCTTGGGAG
PA3177‐DwFw	AAGCAGTCCAATACCGTGGAGGAACTGGACTACC
PA3177‐DwRv	CCCAAGCTTCACAGGATCCTGGCTATCG	Hind III
PA2870‐UPFw	GCTCTAGACGTCTTTCTCTGTTGCCTGG	XbaI	Generation of pEX18Gm‐PA2870
PA2870‐UPRv	GCTCACGGTCATTCGCAGCTTTGTCGGATCGAGC
PA2870‐DwFw	GATCCGACAAAGCTGCGAATGACCGTGAGCATCG
PA2870‐DwRv	CCCAAGCTTGGAGTTCACTCGCTACATCG	Hind III
PA1851‐UPFw	GCTCTAGATCACGTCCAGGCTCATCAC	XbaI	Generation of pEX18Gm‐PA1851
PA1851‐UPRv	CTTGGCCTCGTAGAGTGCAAGCATGCTGACTCCC
PA1851‐DwFw	GTCAGCATGCTTGCACTCTACGAGGCCAAGGATG
PA1851‐DwRv	CCCAAGCTTGAGACCGACTCCCATGAAC	Hind III
PA1120‐UPFw	GCTCTAGAGGACGTCGATCGAGCAAC	XbaI	Generation of pEX18Gm‐PA1120
PA1120‐UPRv	GCATGCTCGGGATAGAGGCGATGGATTGGAACC
PA1120‐DwFw	GGTTCCAATCCATCGCCTCTATCCCGAGCATGC
PA1120‐DwRw	CCCAAGCTTCGAGCGCAGGATTACTCG	Hind III
PA1107‐UPFw	GCTCTAGACGATTCCCAACCTGCTGG	XbaI	Generation of pEX18Gm‐PA1107
PA1107‐UPRv	GCTGTAGAGCTGCTCGTAGGACGAATTCACCTCG
PA1107‐DwFw	CGAGGTGAATTCGTCCTACGAGCAGCTCTACAGC
PA1107‐DwRv	CCCAAGCTTGAGGTTGGAGGTCCAGTTG	Hind III
PA0847‐UPFw	GCTCTAGACTGCAACTGAAGCAGTGGTC	XbaI	Generation of pEX18Gm‐PA0847
PA0847‐UPRv	GAAGTCGATGCGTTCGAAGTTGATGGCCTTCTCG
PA0847‐DwFw	AAGGCCATCAACTTCGAACGCATCGACTTCGTCAC
PA0847‐DwRv	CCCAAGCTTGATCATCGCCAGCAGTTCG	Hind III
PA4332‐UPFw	GGAATTCCGCCGACCCGTATGTCA	EcoRI	Zhu et al. (2016)
PA4332‐UPRv	GGGGTACCGAAGACCTGCTGCCGAACC
PA4332‐DwFw	GGGGTACCGCAAGCCACCACCAGGG
PA4332‐DwRv	CCCAAGCTTCACGGTCCTCTTCAGGCG	Hind III
PA4396‐UPFw	GCTCTAGAGCTCCTTGATCAGACCAACG	XbaI	Generation of pEX18Gm‐PA4396
PA4396‐UPRv	CAGGCTGATGAAGCCGTTCCGTCTGTGAACCCTG
PA4396‐DwFw	GTTCACAGACGGAACGGCTTCATCAGCCTGAAAGC
PA4396‐DwRv	CCCAAGCTTCTGCAGCAACAGCAGTTCG	Hind III
PA4843‐UPFw	GCTCTAGAGTAGGTACGCACATCATCCAC	XbaI	Generation of pEX18Gm‐PA4843
PA4843‐UPRv	GTGGATCTCGGCGAAGTCATCGTGCTCGGTCATC
PA4843‐DwFw	ACCGAGCACGATGACTTCGCCGAGATCCACTACC
PA4843‐DwRv	CCCAAGCTTGGACCATTGTTCCAGGTTGC	Hind III
PA4929‐UPFw	GCTCTAGAGGAACGGACCTTCTTGTCC	XbaI	Generation of pEX18Gm‐PA4929
PA4929‐UPRv	GATTCAGGAGATACGCAGTAGCAGGAAGATGACG
PA4929‐DwFw	ATCTTCCTGCTACTGCGTATCTCCTGAATCAGTGG
PA4929‐DwRv	CCCAAGCTTACCTCAGTGCTGCCGAACTG	Hind III
PA5487‐UPFw	GCTCTAGAGCTGGAAGATCATCCGTTGC	XbaI	Generation of pEX18Gm‐PA5487
PA5487‐UPRv	GAAATGGAACGGACACTCCTGCTGCTCGATGTTC
PA5487‐DwFw	ATCGAGCAGCAGGAGTGTCCGTTCCATTTCAAGGG
PA5487‐DwRv	CCCAAGCTTTGAGGAAGAGCAGCAGCAG	Hind III
PA0169‐Fw	GGGGTACCGATGGACATCCTCGACCTGC	KpnI	Generation of pPA0169
PA0169‐Rv	GCTCTAGACGTCCTGACCCTGGAGTTG	XbaI
PA0290‐Fw	GGGGTACCCAGGGACACGAAGAACGACC	KpnI	Generation of pPA0290
PA0290‐Rv	GCTCTAGACCCAAAGAGGTATTCCGTGG	XbaI
PAO338‐Fw	GCTCTAGAGAGCATCTGCGGCGAGATG	Xbal	Generation of pPA0338
PA0338‐Rv	CCCAAGCTTCAGCGTCGTACGGGTACACG	Hind III
PA3702‐Fw	GGGGTACCCTCAAGCAGATGCGCGAG	KpnI	Generation of pPA3702
PA3702‐Rv	GCTCTAGACAGGGTTTCCACCACCTG	XbaI
PA3343‐Fw	GCTCTAGAGCAGCGAGAAGAACGTCTAC	Xbal	Generation of pPA3343
PA3343‐Rv	CCCAAGCTTGGATCAAGCGCTGCATGG	Hind III
PA3177‐Fw	GGGGTACCGCTACTGACGATCTTCGTGC	KpnI	Generation of pPA3177
PA3177‐Rv	GCTCTAGAGACGGCAAGGAAACCAGC	XbaI
PA2870‐Fw	GGGGTACCGAACGACTACTCGGCGATC	KpnI	Generation of pPA2870
PA2870‐Rv	GCTCTAGACCCATGGAGGTTTCCCAC	XbaI
PA1851‐Fw	GGGGTACCCACCAACAGGGTGTCCAG	KpnI	Generation of pPA1851
PA1851‐Rv	GCTCTAGAGTCCGTAGTCGATCGCAC	XbaI
PA1120‐Fw	GGGGTACCGACGTCGATCGAGCAACG	KpnI	Generation of pPA1120
PA1120‐Rv	GCTCTAGACAGCACAAGGCTGAACAGG	XbaI
PA1107‐Fw	GCTCTAGACGATTCCCAACCTGCTGG	Xbal	Generation of pPA1107
PA1107‐Rv	CCCAAGCTTGGTTCTCCCCTGGTAATGC	Hind III
PA0847‐Fw	GCTCTAGACCATCTCGTCCACCTCACG	Xbal	Generation of pPA0847
PA0847‐Rv	CCCAAGCTTGGTTCTCCGATGGGATGC	Hind III
PA4332‐Fw	GCTCTAGACAGTCGACGATCGAGTCG	Xbal	Generation of pPA0847
PA4332‐Rv	CCCAAGCTTCATCCAGCAGGTTGCTGC	Hind III
PA4396‐Fw	GCTCTAGAGCTTGCGCTCCTTGATCAG	Xbal	Generation of pPA4396
PA4396‐Rv	CCCAAGCTTGCTGACGATAGGGCGAAG	Hind III
PA4843‐Fw	GCTCTAGACGAAGTAGGTACGCAC	Xbal	Generation of pPA4843
PA4843‐Rv	CCCAAGCTTGAGTCTCGTCCGACTACG	Hind III
PA4929‐Fw	GCTCTAGAGAGGAACGGACCTTCTTGTCC	Xbal	Generation of pPA4929
PA4929‐Rv	CCCAAGCTTGCTTCTGAGCAAGGTCACG	Hind III
PA5487‐Fw	GGGGTACCCGACCCTGCTGGAAGATC	KpnI	Generation of pPA5487
PA5487‐Rv	GCTCTAGACGAGTCGTCGGAATGTGC	XbaI

Table A2

Strains used in this study

Strains	Relevant phenotype	References of source
PAO1	Prototroph	Holloway (1955)
PAO1∆0169	PAO1, In‐frame deletion of 0169 gene	This study
PAO1∆0290	PAO1, In‐frame deletion of 0290 gene	This study
PAO1∆0338	PAO1, In‐frame deletion of 0338 gene	This study
PAO1∆0847	PAO1, In‐frame deletion of 0847gene	This study
PAO1∆1107	PAO1, In‐frame deletion of 1107 gene	This study
PAO1∆1120	PAO1, In‐frame deletion of 1120 gene	This study
PAO1∆1851	PAO1, In‐frame deletion of 1851 gene	This study
PAO1∆2870	PAO1, In‐frame deletion of 2870 gene	This study
PAO1∆3177	PAO1, In‐frame deletion of 3177 gene	This study
PAO1∆3343	PAO1, In‐frame deletion of 3343 gene	This study
PAO1∆3702	PAO1, In‐frame deletion of 3702 gene	This study
PAO1∆4332	PAO1, In‐frame deletion of 4332 gene	Zhu et al. (2016)
PAO1∆4396	PAO1, In‐frame deletion of 4396 gene	This study
PAO1∆4843	PAO1, In‐frame deletion of 4843 gene	This study
PAO1∆4929	PAO1, In‐frame deletion of 4929 gene	This study
PAO1∆5487	PAO1, In‐frame deletion of 5487 gene	This study
PAO1/pPA0169	Overexpression of PAO169 gene in PAO1	This study
PAO1/pPA0290	Overexpression of PA0290 gene in PAO1	This study
PAO1/pPA0338	Overexpression of PA0338 gene in PAO1	This study
PAO1/pPA0847	Overexpression of PA0847gene in PAO1	This study
PAO1/pPA1107	Overexpression of PA1107 gene in PAO1	This study
PAO1/pPA1120	Overexpression of PA1120 gene in PAO1	This study
PAO1 pPA1851	Overexpression of PA1851 gene in PAO1	This study
PAO1/pPA2870	Overexpression of PA2870 gene in PAO1	This study
PAO1/pPA3177	Overexpression of PA3177 gene in PAO1	This study
PAO1/pPA3343	Overexpression of PA3343 gene in PAO1	This study
PAO1/pPA3702	Overexpression of PA3702 gene in PAO1	This study
PAO1/pPA4332	Overexpression of PA4332 gene in PAO1	This study
PAO1/pPA4396	Overexpression of PA4396 gene in PAO1	This study
PAO1/pPA4843	Overexpression of PA4843 gene in PAO1	This study
PAO1/pPA4929	Overexpression of PA4929 gene in PAO1	This study
PAO1/pPA5487	Overexpression of PA5487 gene in PAO1	This study

Molecular techniques

The pUCP20 vector used to overexpress the GGDEF genes by standard transformation technique and pEX18Gm vector was used to construct single and double mutants by in‐frame deletion technique.

Biofilm assays

Biofilm assay was performed as previously described (O'Toole & Kolter, 1998). Briefly, bacterial cultures grown overnight in LB were diluted 1:100 in Jensen's medium and grown in triplicate in a polyvinylchloride plate (Costar) overnight at 30°C without agitation, followed by staining with 0.1% crystal violet for 30 min at room temperature and then washed twice by dipping into standing water bath. The adherent stain was solubilized in 30% acetic acid and quantified by measuring its optical density at 560 nm.

Motility assays

Motility assay was performed as previously described (Caiazza, Shanks, & O'Toole, 2005). Briefly, bacterial cultures were grown overnight in LB and inoculated culture on swarm agar (nutrient broth‐8 g/L, d‐glucose‐5 g/L, bacto agar‐5 g/L) and swim agar (tryptone‐10 g/L, yeast extract‐5 g/L agar‐1.5 g/L) plates. Plates were incubated upright at 37°C overnight. Swarming and swimming zones were measured accordingly.

Psl immunoblotting

The immunoblotting assay was performed as previously described (Byrd et al., 2009). Briefly, bacterial strains were incubated in Jensen's liquid culture for 24 hr and bacterial surface‐bound polysaccharide were extracted from cultures with OD600nm of 10. Cells were centrifuged, and the pellet was resuspended in 100 µl of 0.5 M EDTA. Cells were boiled at 100°C for 10 min and centrifuged. The supernatant was collected and treated with proteinase K for 60 min at 60°C (final concentration 0.5 mg/ml), followed by proteinase K inactivation for 30 min at 80°C. 3 µl of polysaccharide preparations was spotted onto a nitrocellulose membrane. The blocking step was performed using 10% milk solution in TBST (20 mM Tris, 137 mM NaCl, 0.1% Tween 20, pH 7.6) for 1 hr at room temperature and washed with TBST, and then probed with antibody α‐Psl (1:25,000 in TBST) for 1 hr with agitation then a second antibody (horseradish peroxidase‐conjugated goat anti‐rabbit IgG used at 1:10,000 in TBST) for 1 hr. About 5 ml of Western blue (Promega) was added to stain the dots (Zhu et al., 2016). The analysis was done using Image laboratory 5.1 software.

Quantification of c‐di‐GMP using pCdrA::gfp

Indirect c‐di‐GMP quantified by pCdrA::gfp was performed with some modifications as previously described (Rybtke et al., 2012). Briefly, pCdrA::gfp plasmid was transformed into bacteria by using a standard transformation technique. The strain containing pCdrA::gfp plasmid was grown overnight and was diluted 1:100 in Jensen's medium containing carbenicillin and grown in triplicate in 96‐well plate (NEST) at 37°C for 24 hr. OD600 and fluorescence (excitation 485 nm, emission 520 nm) were read accordingly.

Quantification of c‐di‐GMP using LC‐MS/MS

Quantification of c‐di‐GMP was performed as previously described (Spangler, Bohm, Jenal, Seifert, & Kaever, 2010). Samples of interest were compared to a standard curve derived from measurements of known concentrations of pure c‐di‐GMP to determine the concentration (nM) of c‐di‐GMP in the samples. The data were then normalized to the total protein content of the sample determined by BCA assay (Thermo); for each strain, the experiment was done in biological triplicate, and LC‐MS/MS measurements were repeated in duplicate. Data are presented as pmol of c‐di‐GMP/mg of total protein.

Statistical analysis

The data of this study were analyzed by one‐way ANOVA. Student's t test was used when one‐way ANOVA revealed significant differences (p < .05). All statistical analyses were performed with GraphPad Prism statistical software (GraphPad Software) with the help of Excel (Microsoft).

RESULTS

Quantification of c‐di‐GMP generation potential of predicted DGC‐encoding genes in PAO1

In bacteria, DGC is implicated in the synthesis of intracellular c‐di‐GMP. Imperatively, overexpression of DGC‐encoding genes would produce a higher amount of intracellular c‐di‐GMP and deletion of DGC‐encoding genes would generate less c‐di‐GMP in bacterial cells. We chose 16 predicted DGC‐encoding genes that containing only GGDEF domain to evaluate the methods to measure the intracellular c‐di‐GMP level for a DGC characterization. We generated in‐frame deletion mutants of the 16 genes in P. aeruginosa PAO1. Meanwhile, we cloned each of them into the expression vector pUCP20 and then transferred each plasmid into PAO1 to get overexpression variants. The LC/MS/MS method was used to quantify the intracellular c‐di‐GMP level of strains overexpressed DGC‐encoding genes. As can be seen from Figure 1, among 16 genes, only six genes (PA4332, PA0169, PA4843, PA4929, PA2870, and PA3702) showed a significantly higher amount of c‐di‐GMP than that of wild type (Figure 1a). Among those six genes, four genes (PA4332, PA0169, PA3702, and PA4843) are well‐studied DGC‐encoding genes and confirmed in our study (Jones et al., 2015; Zhu et al., 2016). We then used the pCdrA::gfp as a reporter to monitor the relative c‐di‐GMP level in the 16 deletion mutants. The six mutants out of 16 (∆PA0169, ∆PA0847, ∆PA1120, ∆PA4332, ∆PA3177, and ∆PA3702) exhibited significantly lower c‐di‐GMP level than that of wild type (Figure 1b). PA4332, PA0169, and PA3702 exhibited significant changes in either in‐frame deletion mutants or DGC overexpression strains, while the other genes could not detect a change on the c‐di‐GMP level by one of the methods or both even if the genes have been characterized as DGCs previously, such as PA1107 and PA1120. This revealed that the quantification of c‐di‐GMP is not sufficient for the identification of DGC activities in P. aeruginosa under the tested conditions.

Figure 1

Quantification of c‐di‐GMP level in PAO1‐derived strains with overexpression of a DGC‐encoding gene or a gene deletion. (a) The total intracellular c‐di‐GMP concentration of overexpressed GGDEF domain genes measured by LC/MS/MS. Strains were incubated in Jensen's medium. The total intracellular c‐di‐GMP was extracted from cultures post 24 hr of growth. Shown are the value related to PAO1/pUCP20 (the concentration of c‐di‐GMP was 5.14 pmol/mg total protein); (b) The c‐di‐GMP level of GGDEF domain mutants was indicated by green fluorescent intensity of plasmid pCdrA::gfp measured by OD 560 post 24h of growth. Dotted lines indicate the wild‐type level. Results were normalized to the level of wild type (the green fluorescent intensity of PAO1 is 5,842). *p < .05

C‐di‐GMP‐related phenotypes by overexpression analysis

We then performed four major c‐di‐GMP associated phenotypes (functional output of DGC) of DGC‐overexpressed strains in order to identify a DGC. Firstly, biofilm formation of all overexpression mutants was determined. Quantification of biofilm formation was performed in 96‐well polyvinylchloride plate. There were 10 predicted DGC‐encoding genes that showed significant higher biofilm formation ability than the control strain PAO1/pUCP20 (Figure 2a). Among those genes, PA0169, PA0290, PA0847, and PA4929 produced more than 200% of biofilms as compared to wild‐type PAO1, while another six genes (PA0338, PA1107, PA1120, PA2870, PA3702, and PA4332) produced more than 150% of biofilm compared to control.

Figure 2

C‐di‐GMP‐related phenotypic assays for overexpression of DGC‐encoding genes in PAO1. (a) Biofilm assay; (b) Psl production; (c) swarming motility; and (d) swimming motility. Dotted lines indicate the wild‐type level. Results were normalized to the value of wild‐type control (PAO1/pUC20) that is the biofilm biomass (OD560), 5.82; Psl, 29 µg/OD600; swarm zone, 29 mm; and swim zone, 32 mm. *p < .05

Afterward, we examined the production of Psl exopolysaccharide. Out of 16 genes, six genes (PA0169/siaD, PA0847, PA2870, PA3702, PA4332, PA4929, and PA5487) produced 150% of Psl polysaccharide as compared to wild‐type PAO1, while other overexpression of other genes gave rise to no differential results (Figure 2b). Finally, we detected the swarming and swimming motilities. Strikingly, all 16 genes significantly reduced the swarming and swimming motilities (Figure 2c,d). In particular, PA0847, PA3702, and PA4332 completely lost swimming motility (Figure 2d). Taken together, for overexpression analysis, the motility assay showed the most distinct phenotype and could be a quick way to estimate a predicted DGC.

C‐di‐GMP‐related phenotypes by in‐frame deletion analysis

We then evaluated the phenotypic output of the in‐frame deletion of 16 DGC‐encoding genes. As can be seen from Figure 3a, most of the mutants produced less biofilm; noticeably, five mutants (∆PA0169, ∆PA0338, ∆PA1120, ∆PA3702, and ∆PA4843) produced significantly less biofilm compared to other DGC‐encoding genes and the wild‐type PAO1 control. Psl production was also tested for all mutants. As compared to the wild‐type strain PAO1, seven mutants (∆PA0169, ∆PA0338, ∆PA0847, ∆PA3343, ∆PA3702, ∆PA4332, and ∆PA4929) produced significantly less Psl (Figure 3b). Regard to motilities, seven mutants (∆PA0169, ∆PA0338, ∆PA0847, ∆PA3343, ∆PA3702, ∆PA4332, and ∆PA4929) showed significantly increased in swarming motility (Figure 3c), while 11 mutants (∆PA0169, ∆PA0290, ∆PA0338, ∆PA0847, ∆PA2870, ∆PA3177, ∆PA3343, ∆PA3702, ∆PA4332, and ∆PA4929) showed significant increased swimming motility compared to wild type (Figure 3d), suggesting that the swimming motility assay was also a good way to briefly evaluate a predicted DGC‐encoding gene by deletion analysis.

Figure 3

C‐di‐GMP‐related phenotypic assays for DGC‐encoding gene mutants in PAO1. (a) Biofilm assay; (b) Psl production; (c) swarming motility； (d) swimming motility. Dotted lines indicate the wild‐type level. Results were normalized to the value of PAO1 (OD560, 6.42; swarm zone, 48 mm; Psl, 39 µg/OD600; swim zone, 38 mm). *p < .05

Correlation analysis of c‐di‐GMP‐related phenotypes of DGC‐encoding genes

In order to further evaluate the function of those DGC‐encoding genes, we introduced a quantitative way to classify them into different groups. We first assigned to each phenotype a score based on the relative results as compared to that of wild‐type PAO1 and then calculated the overall contribution of each overexpression or mutation of DGC‐encoding genes to those phenotypes. As can be seen from Table 1, the mutation of 16 DGC‐encoding genes resulted in the overall index above 4, which is the score of wild‐type control. The most pronounced effects were caused by the deletion of PA0169, PA0338, PA3702, and PA4843, which have an overall contribution index higher than 7.5. PA0169 showed the highest score (12.35) among the 16 genes. PA1107 gave the lowest score (6.41 in Table 1) among previously characterized DGC. As for the overexpression of those DGC‐encoding genes (Table 2), there have even higher scores than that of corresponding deletion mutants. Among them, five genes (PA0847, PA2870, PA3177, PA3702, and PA4332) showed a score higher than 10. PA3702/wspR gave the highest value (12.83) among 16 genes. PA1120/yfiN showed a relatively low score (6.74) among previously characterized DGC. Our results also suggested that nine predicted DGC‐encoding genes (PA0290, PA0338, PA0847, PA1851, PA2870, PA3177, PA3343, PA4396, PA4929, and PA5487) were very likely functional DGCs. PA4929 gave the lowest value in both analyses and required further analysis to confirm its function. The overall evaluation also showed that overexpression analysis gave a higher score in general than in‐frame deletion analysis.

Table 1

Contribution index for deletion mutants of predicted DGC‐encoding genes

	Contribution index
	Swarming	Swimming	Biofilm	Psl	Overall
∆PA0169 (siaD)	1.79	1.91	4.98	3.66	12.34
∆PA4843 (adcA)	1.80	1.71	2.50	1.61	7.62
∆PA0338	1.78	1.85	2.60	1.35	7.58
∆PA3702 (wspR)	1.13	1.94	2.75	1.37	7.19
∆PA1120 (yfiN)	1.70	1.59	2.29	1.40	6.98
∆PA2870	1.67	1.93	1.54	1.65	6.78
∆PA4332 (sadC)	1.55	2.00	1.53	1.63	6.71
∆PA3177	1.54	1.99	1.70	1.38	6.61
∆PA3343	1.75	1.93	1.37	1.55	6.60
∆PA0290	1.81	1.87	1.10	1.70	6.48
∆PA1107 (roeA)	1.68	1.90	1.45	1.38	6.40
∆PA4929	1.41	2.21	1.40	1.30	6.33
∆PA0847	1.73	1.93	1.08	1.45	6.18
∆PA5487	1.67	2.00	1.03	1.39	6.09
∆PA1851	1.53	1.78	1.41	1.37	6.09
∆PA4396	1.37	1.90	1.30	1.08	5.65
PAO1	1.00	1.00	1.00	1.00	4.00

Table 2

Contribution index for overexpression of predicted DGC‐encoding genes

	Contribution index
	Swarming	Swimming	Biofilm	Psl	Overall
PA3702 (wspR)	5.93	3.48	1.61	1.80	12.83
PA0847	5.93	2.23	2.07	2.13	12.37
PA4332 (sadC)	5.93	1.85	1.56	1.58	10.93
PA2870	5.84	1.93	1.61	1.51	10.90
PA3177	5.93	1.85	1.20	1.30	10.29
PA4929	4.94	1.18	2.19	1.65	9.95
PA1107 (roeA)	4.94	1.24	1.76	1.34	9.29
PA4843 (adcA)	4.68	1.32	1.43	1.45	8.89
PA0169 (siaD)	3.30	1.21	2.19	1.68	8.38
PA0290	3.07	1.24	2.19	1.44	7.94
PA5487	2.97	1.58	1.38	1.50	7.42
PA1120 (yfiN)	2.28	1.40	1.64	1.42	6.75
PA3343	2.87	1.53	1.23	1.06	6.69
PA1851	1.93	2.02	1.30	1.23	6.49
PA0338	2.23	1.26	1.64	1.25	6.38
PA4396	1.75	1.24	1.43	1.32	5.74
pUCP20	1.00	1.00	1.00	1.00	4.00

DISCUSSION

Opportunistic pathogen P. aeruginosa is making lots of obstacles in human's life by producing diseases such as urinary tract infections, pneumonia, and lung infection, and it is getting resistant to a wide range of antibiotics because of its nature of producing biofilm (Cornforth et al., 2018; Prat & Lacoma, 2016). C‐di‐GMP is an important second messenger involved in the regulation of biofilm formation. Understanding the mechanisms of c‐di‐GMP signaling in bacteria will provide a strategy to solve biofilm‐related problems. In many bacteria, there are multiple genes involved in the biosynthesis and degradation of c‐di‐GMP. In this study, we have used P. aeruginosa PAO1 as a model microorganism to evaluate the methods and functional outputs of 16 putative DGC‐encoding genes in this bacterium. The results will be beneficial for the research on Pseudomona spp. and bacteria using c‐di‐GMP as a signal.

In P. aeruginosa PAO1, 40 predicted genes (16 GGDEF‐only, 16 GGDEF + EAL, five EAL only and three HD‐GYP only) are involved in c‐di‐GMP synthesis and degradation. We have focused on the 16 predicted DGC‐encoding genes that contain only GGDEF domain. We have cloned all 16 genes into shuttle vector pUCP20 and transformed into P. aeruginosa and also have used in‐frame deletion technique to mutate the 16 genes, respectively. In our overexpression study, PA4332 (sadC) and PA0169 (siaD), the two well‐studied DGC in P. aeruginosa, produced remarkably higher c‐di‐GMP than other genes (Valentini & Filloux, 2016; Zhu et al., 2016). Among DGC mutants, ∆PA0169, ∆PA0847, ∆PA1120 (yfiN), ∆PA3702 (wspR), and ∆PA4332 produced significantly less c‐di‐GMP compared to wild type. These are consistent with the previous reports (Malone et al., 2012; Valentini & Filloux, 2016; Zhu et al., 2016). It is interesting to note that the results from our detection of c‐di‐GMP in PAO1 derivatives were comparable to what has been published previously and even can be compared with the PA14, a native Psl‐negative strain (Ha et al., 2014; Ueda & Wood, 2009; Zhu et al., 2016).

Our results in this study have revealed that the intracellular c‐di‐GMP measurement is not enough for evaluation or identification of a functional DGC and several methods have to be used. Overall evaluation is critical because the regulation of c‐di‐GMP level in bacteria is far more complex than what we have imaged. We have also noted that DGC genes still have distinct c‐di‐GMP relevant phenotypes, with some having more of an effect in one assay versus another even though they are overexpressed to the same level in PAO1. This could be due to the localization of DGC that changes the subcellular c‐di‐GMP level, leading the phenotypic difference as reported previously (Zhu et al., 2016). This would also suggest that the DGCs' activity is still regulated at some level even when overexpressed. Another possible explanation is due to the complex protein architecture of those DGCs that would involve more signaling pathways to mediate the downstream physiology (Kulasakara et al., 2006). For example, YfiN (PA1120) has previously been reported to cause a significant increase in the c‐di‐GMP level when overexpressed or mutated, possibly due to a change in the stoichiometry to its regulatory components (YfiB and YfiR) (Malone et al., 2012). In addition, a regulator of swimming motility, FleQ, can function as both a repressor and an activator in response to c‐di‐GMP to control the transcription of pel genes in P. aeruginosa (Baraquet, Murakami, Parsek, & Harwood, 2012).

By comparing the phenotypes of overexpression with deletion of DGC‐encoding genes, our results have shown that the overexpression technique overall has given more distinctive phenotype. Most strikingly, the motility assay has shown relevant phenotype for all 16 predicted DGC‐encoding genes by overexpression. This has suggested that the motility assay could be a quick way to briefly evaluate a predicted DGC for further studies. However, our data have also revealed that it is not easy to evaluate whether putative DGC genes are truly functional DGCs. It is very important to use multiple methods to give an overall evaluation for the characterization of a predicted DGC. By overall evaluation, we have revealed nine functional DGCs in P. aeruginosa, which have not been identified previously. In summary, we have provided a cloning library of 16 DGC‐encoding genes from P. aeruginosa and their corresponding in‐frame deletion mutants. Our work paves ways to briefly evaluate a predicted DGC and would help the researcher to go depth into their interested DGCs in bacteria.

CONFLICT OF INTEREST

None declared.

AUTHOR CONTRIBUTIONS

Pramod Bhasme: Investigation, writing‐original draft, writing‐review & editing; Qing Wei: Investigation, writing‐original draft, writing‐review & editing; Anming Xu: Formal analysis, methodology, writing‐review & editing; Syed Tatheer Alam Naqvi: Formal analysis, methodology, writing‐review & editing; Di Wang: Formal analysis, methodology, writing‐review & editing; and Luyan Zulie Ma: Data curation, funding acquisition, project administration, resources, supervision, writing‐original draft, writing‐review & editing.

ETHICS STATEMENT

None required.